Case Reports

Reverse total shoulder arthroplasty (RTSA) performed in carefully selected patients often leads to satisfactory outcomes.^1,2 In recent years, its indications and the number performed per year have expanded. Subsequently, there has been a concomitant rise in reported complications,^2,3 with a rate ranging from 19% to 68%.^2,3 Some common complications include scapular notching,^2-4 fracture,^2,3,5-7 dislocation,^2,3,7 and infection.^2,3,7

In this series, we describe 2 cases of coracoid fracture after RTSA. The patients provided written informed consent for print and electronic publication of these case reports.

Case Series

Case 1

An independently functioning 81-year-old right hand–dominant woman (BMI, 22.1 [height, 160 cm; weight, 56.7 kg]) presented with increasing left shoulder pain and dysfunction after a motor vehicle accident 2 months earlier. She had reported vague chronic left shoulder pain in the past, but after the accident her pain was significantly worse. A subacromial corticosteroid injection by her primary care physician provided temporary symptomatic relief, but her symptoms recurred.

On presentation, there was obvious anterior superior escape of the humeral head, which was accentuated by shoulder shrug. Her deltoid motor function was found to be intact, and her active shoulder range of motion was severely limited (pseudoparesis). There was notable crepitation as well as significant weakness and pain with abduction and external rotation strength testing.

Radiographic imaging showed anterior superior escape of the humeral head with some early degenerative changes (Seebauer type IIB⁸ [Figure 1A]). Magnetic resonance imaging confirmed a full-thickness retracted massive rotator cuff tear with complete involvement of the supraspinatus, infraspinatus, and most of the subscapularis muscles. Significant glenohumeral degenerative changes consistent with cuff tear arthropathy were also seen without any evidence of fracture.

After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. The patient underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 12, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 3 screws, and the stem was placed in neutral version. The patient’s shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiograph (Figure 1B) showed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful, and rehabilitation consisted of 6 weeks of sling protection, with advancing passive and active range of motion. Strengthening exercises were initiated 6 weeks after surgery.

At the patient’s 6-week postoperative visit, she demonstrated pain-free passive elevation to 80° and active forward elevation to 70°. At her 3-month postoperative visit, she reported a 1-week onset of anterior shoulder pain accompanied by a strange noise at the anterior aspect of the operative shoulder. She denied any recent trauma. She continued to have minimal shoulder pain with passive forward flexion of 80°; however, her active forward elevation was very limited because of pain in the anterior aspect of her shoulder. Active external rotation was noted to be 20° and internal rotation was to her buttock. She had pain to palpation of the coracoid process. Radiographs were unchanged from immediate postoperative radiographs. Computed tomography (CT), which was ordered to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis and no loosening. However, CT was notable for a nondisplaced fracture through the base of the coracoid (Figures 2A-2D). The patient stopped formal physical therapy, and sling immobilization was initiated. After 3 weeks, the sling was discontinued and physical therapy was begun again. She responded satisfactorily to this treatment approach, and, at her 6-month postoperative follow-up, she was without pain, instability, or crepitation. Her range of motion had improved with pain-free active forward flexion, external rotation, and abduction of 100°, 15°, and 90°, respectively. At 28-month postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 3, 73, and 67, respectively. 

Case 2

A 68-year-old, right-handed woman (BMI, 22.5 [height, 160 cm; weight, 57.6 kg]) presented with right shoulder pain and dysfunction of 3 years’ duration. She had undergone an open rotator cuff repair at an outside facility 4 years ago that was unsuccessful. At the time of her presentation to our institution, she had already undergone a failed course of physical therapy. A trial of corticosteroid subacromial injections did not adequately manage her symptoms.

On presentation, her active forward flexion, abduction, and external rotation were 40°, 30°, and 10°, respectively. She had full passive range of motion and pain with active and passive shoulder motion. Radiographic imaging showed superior migration of the humeral head with evidence of glenohumeral arthropathy suggestive of rotator cuff arthropathy (Seebauer type IIA⁸). After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. She underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 8, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 4 screws, and the stem was placed in neutral version. Her shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiographs revealed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful. She was taken out of her shoulder immobilizer 4 weeks after surgery and began home-based physical therapy.

At 1 year after surgery, the patient had minimal shoulder pain with active forward flexion, external rotation, and abduction of 135°, 20°, and 85°, respectively. She presented to our clinic 15 months after RTSA with acute onset of pain about her anterior shoulder. She denied any recent trauma or infectious exposures. On examination, her motion was unchanged from prior examinations. However, she was tender on palpation of the coracoid. Radiographs at that time were unchanged (Figures 3A, 3B). Laboratory tests (erythrocyte sedimentation rate, C-reactive protein, and complete blood count with differential) that were subsequently ordered to rule out an occult infection were within normal limits. Computed tomography, which was ordered for further assessment and to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis without loosening. However, a lucency was noted in the midportion of the coracoid that was suggestive of a fracture (Figures 4A, 4B). A conservative plan of treatment was advised with sling immobilization for 3 weeks and follow-up visits. The patient responded satisfactorily to this treatment approach, and, at her latest follow-up, 8 months after presenting with a coracoid fracture, she was pain-free. At the 5-year postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 1-2, 78, and 75, respectively. 

Discussion

The reverse prosthesis, a semi-constrained ball-and-socket device, provides satisfactory functional outcomes when used in carefully selected patients with rotator cuff arthropathy and pseudoparalysis, failed shoulder arthroplasty, and fracture sequelae.^1,9-11 By the traditional Grammont principles of medializing the center of rotation and lowering the humerus, shear forces about the glenoid are reduced and the deltoid muscle is tensioned, allowing for adequate torque generation, required to facilitate shoulder motion.^12,13 While long-term outcomes  concerning durability and survivorship are pending, some studies have attempted to improve our understanding of implant and functional longevity. Guery and colleagues¹⁴ noted an implant survival of 91% at 120 months. However, increased pain and decreased function were seen at the 6-year mark.¹⁴ A more recent study by Cuff and colleagues¹⁵ revealed 94% implant survivorship and sustained improvement in range of motion and pain at 5 years. 

Despite considerable success, RTSA can be associated with a myriad of complications. The most common complications of RTSA include scapular notching (44%-96%), glenoid side failure (5%-40%), instability (2.4%-31%), and infection (1%-15.3%).^2,3 In the setting of inflammatory arthropathy, there is an increased risk for intraoperative and postoperative fractures.^16,17 To date, there are only 2 reported cases of coracoid process fractures after RTSA.^18,19 In the case by Nolan and colleagues,¹⁸ conservative management with a sling for 6 weeks led to successful resolution of symptoms. Although little information is provided on the management of these rare fractures, literature on the slightly more common scapular (0.9%-7.2%) and acromial (0.9%-4.9%) fractures suggest that periscapular fractures are on the rise, may increase the risk for revision surgery, and can lead to inferior outcomes when compared with patients without fractures.^5,20,21

Acromial fractures after RTSA have been reported to occur at a rate of 0.9% to 4.9%.^5,21 This is a concern because of RTSA reliance on a functional deltoid.^5,6 The cause of these fractures remains to be fully elucidated. Wahlquist and colleagues⁶ in 2011 reported the cases of 5 patients that sustained acromial base fractures after RTSA. All 5 patients were noted to have unsatisfactory functional results despite achieving union (3 were treated with open reduction and internal fixation, and 2 were treated nonoperatively). Acromial fractures tend to present with pain within 6 months of surgery, which may indicate excessive constraint about the scapula, eventually leading to fracture. Furthermore, disruption of this bony structure can lead to devastating results because the acromial base serves as a fulcrum for the deltoid.

Despite a well-placed reverse prosthesis, there is increased reliance on surrounding glenohumeral musculature, resulting from poor rotator cuff function and biomechanical differences compared with a native shoulder. Both our patients were found to have relatively small body habitus. It is possible that, by nature of their smaller statures, they were more susceptible to consequences of excessive joint and soft-tissue tension after RTSA. One explanation for acromial fractures after RTSA is that, by excessively lengthening and/or lateralizing the deltoid, the tension on the acromion in these elderly patients may be sufficient to cause a fracture. A similar mechanism may explain their coracoid fractures. As the arm is lengthened and the prosthesis is tightened, the conjoint tendon is significantly tensioned. We routinely check the tension of these muscles as an extra confirmation of joint stability. However, excessive tension for a significant duration may provide too much stress for bone turnover to match with the inherent repair process, potentially causing a fracture. Recent evidence has also found that bone mineral density of the coracoid diminishes with age, suggesting some predisposition to fracture with lower-energy mechanisms.²²

Another possible cause for coracoid fractures may be the orientation of the implants. While we did not have mechanistic evidence, it is possible that, with adduction and internal rotation, prosthetic impingement against the coracoid is feasible, particularly in patients of small stature. Although a glenoid implant placed high can increase the chance for coracoid–implant impingement, the fact that the patients improved without revision makes chronic mechanical impingement less likely. Drill holes, especially multiple ones, placed throughout the base of the coracoid may also predispose to coracoid fractures.

Patients with periscapular fractures (acromion, scapular spine, or coracoid) after RTSA often present with pain and occasional deficits in function. Both patients in this series noted pain out of proportion to examination. The onset of this pain differed, with 1 patient noting pain within the first 3 months and 1 noting discomfort later. Neither patient had any trauma. In the presence of significant symptoms, negative radiographs, and a poor response to conservative treatment, we recommend advanced imaging to rule out fracture. However, prior to obtaining advanced imaging, proper radiographic techniques should be utilized. Eyres and colleagues,²³ in a series of 12 fractures of the coracoid process, relied primarily on coracoid views directed 45° in a cephalic direction and thin-slice CT. An isotope bone scan identified 1 case not initially found on radiographs.²³

Conservative management with use of a sling until resolution of symptoms was successful in our series. If symptoms persist, a bone stimulator can be used prior to implementing a surgical solution; however, current evidence does not expound on timing and utility of such modalities. Perhaps as important as treatment is education of the patient and the rehabilitation team about the importance of identifying increasing pain as a potential sign of impending fracture in this population. Subsequent activity modification until the pain resolves can help avoid the setback in postoperative recovery that this complication may cause.

Conclusion

We present 2 patients with coracoid fractures encountered at 3 months and 15 months after RTSA. Nonoperative management proved adequate in treating both cases. We suggest a high level of suspicion for possible fracture in the patient who comes in with new-onset pain in a localized region with or without functional deficits.

Reverse total shoulder arthroplasty (RTSA) performed in carefully selected patients often leads to satisfactory outcomes.^1,2 In recent years, its indications and the number performed per year have expanded. Subsequently, there has been a concomitant rise in reported complications,^2,3 with a rate ranging from 19% to 68%.^2,3 Some common complications include scapular notching,^2-4 fracture,^2,3,5-7 dislocation,^2,3,7 and infection.^2,3,7

In this series, we describe 2 cases of coracoid fracture after RTSA. The patients provided written informed consent for print and electronic publication of these case reports.

Case Series

Case 1

An independently functioning 81-year-old right hand–dominant woman (BMI, 22.1 [height, 160 cm; weight, 56.7 kg]) presented with increasing left shoulder pain and dysfunction after a motor vehicle accident 2 months earlier. She had reported vague chronic left shoulder pain in the past, but after the accident her pain was significantly worse. A subacromial corticosteroid injection by her primary care physician provided temporary symptomatic relief, but her symptoms recurred.

On presentation, there was obvious anterior superior escape of the humeral head, which was accentuated by shoulder shrug. Her deltoid motor function was found to be intact, and her active shoulder range of motion was severely limited (pseudoparesis). There was notable crepitation as well as significant weakness and pain with abduction and external rotation strength testing.

Radiographic imaging showed anterior superior escape of the humeral head with some early degenerative changes (Seebauer type IIB⁸ [Figure 1A]). Magnetic resonance imaging confirmed a full-thickness retracted massive rotator cuff tear with complete involvement of the supraspinatus, infraspinatus, and most of the subscapularis muscles. Significant glenohumeral degenerative changes consistent with cuff tear arthropathy were also seen without any evidence of fracture.

After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. The patient underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 12, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 3 screws, and the stem was placed in neutral version. The patient’s shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiograph (Figure 1B) showed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful, and rehabilitation consisted of 6 weeks of sling protection, with advancing passive and active range of motion. Strengthening exercises were initiated 6 weeks after surgery.

At the patient’s 6-week postoperative visit, she demonstrated pain-free passive elevation to 80° and active forward elevation to 70°. At her 3-month postoperative visit, she reported a 1-week onset of anterior shoulder pain accompanied by a strange noise at the anterior aspect of the operative shoulder. She denied any recent trauma. She continued to have minimal shoulder pain with passive forward flexion of 80°; however, her active forward elevation was very limited because of pain in the anterior aspect of her shoulder. Active external rotation was noted to be 20° and internal rotation was to her buttock. She had pain to palpation of the coracoid process. Radiographs were unchanged from immediate postoperative radiographs. Computed tomography (CT), which was ordered to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis and no loosening. However, CT was notable for a nondisplaced fracture through the base of the coracoid (Figures 2A-2D). The patient stopped formal physical therapy, and sling immobilization was initiated. After 3 weeks, the sling was discontinued and physical therapy was begun again. She responded satisfactorily to this treatment approach, and, at her 6-month postoperative follow-up, she was without pain, instability, or crepitation. Her range of motion had improved with pain-free active forward flexion, external rotation, and abduction of 100°, 15°, and 90°, respectively. At 28-month postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 3, 73, and 67, respectively. 

Case 2

A 68-year-old, right-handed woman (BMI, 22.5 [height, 160 cm; weight, 57.6 kg]) presented with right shoulder pain and dysfunction of 3 years’ duration. She had undergone an open rotator cuff repair at an outside facility 4 years ago that was unsuccessful. At the time of her presentation to our institution, she had already undergone a failed course of physical therapy. A trial of corticosteroid subacromial injections did not adequately manage her symptoms.

On presentation, her active forward flexion, abduction, and external rotation were 40°, 30°, and 10°, respectively. She had full passive range of motion and pain with active and passive shoulder motion. Radiographic imaging showed superior migration of the humeral head with evidence of glenohumeral arthropathy suggestive of rotator cuff arthropathy (Seebauer type IIA⁸). After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. She underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 8, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 4 screws, and the stem was placed in neutral version. Her shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiographs revealed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful. She was taken out of her shoulder immobilizer 4 weeks after surgery and began home-based physical therapy.

At 1 year after surgery, the patient had minimal shoulder pain with active forward flexion, external rotation, and abduction of 135°, 20°, and 85°, respectively. She presented to our clinic 15 months after RTSA with acute onset of pain about her anterior shoulder. She denied any recent trauma or infectious exposures. On examination, her motion was unchanged from prior examinations. However, she was tender on palpation of the coracoid. Radiographs at that time were unchanged (Figures 3A, 3B). Laboratory tests (erythrocyte sedimentation rate, C-reactive protein, and complete blood count with differential) that were subsequently ordered to rule out an occult infection were within normal limits. Computed tomography, which was ordered for further assessment and to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis without loosening. However, a lucency was noted in the midportion of the coracoid that was suggestive of a fracture (Figures 4A, 4B). A conservative plan of treatment was advised with sling immobilization for 3 weeks and follow-up visits. The patient responded satisfactorily to this treatment approach, and, at her latest follow-up, 8 months after presenting with a coracoid fracture, she was pain-free. At the 5-year postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 1-2, 78, and 75, respectively. 

Discussion

The reverse prosthesis, a semi-constrained ball-and-socket device, provides satisfactory functional outcomes when used in carefully selected patients with rotator cuff arthropathy and pseudoparalysis, failed shoulder arthroplasty, and fracture sequelae.^1,9-11 By the traditional Grammont principles of medializing the center of rotation and lowering the humerus, shear forces about the glenoid are reduced and the deltoid muscle is tensioned, allowing for adequate torque generation, required to facilitate shoulder motion.^12,13 While long-term outcomes  concerning durability and survivorship are pending, some studies have attempted to improve our understanding of implant and functional longevity. Guery and colleagues¹⁴ noted an implant survival of 91% at 120 months. However, increased pain and decreased function were seen at the 6-year mark.¹⁴ A more recent study by Cuff and colleagues¹⁵ revealed 94% implant survivorship and sustained improvement in range of motion and pain at 5 years. 

Despite considerable success, RTSA can be associated with a myriad of complications. The most common complications of RTSA include scapular notching (44%-96%), glenoid side failure (5%-40%), instability (2.4%-31%), and infection (1%-15.3%).^2,3 In the setting of inflammatory arthropathy, there is an increased risk for intraoperative and postoperative fractures.^16,17 To date, there are only 2 reported cases of coracoid process fractures after RTSA.^18,19 In the case by Nolan and colleagues,¹⁸ conservative management with a sling for 6 weeks led to successful resolution of symptoms. Although little information is provided on the management of these rare fractures, literature on the slightly more common scapular (0.9%-7.2%) and acromial (0.9%-4.9%) fractures suggest that periscapular fractures are on the rise, may increase the risk for revision surgery, and can lead to inferior outcomes when compared with patients without fractures.^5,20,21

Acromial fractures after RTSA have been reported to occur at a rate of 0.9% to 4.9%.^5,21 This is a concern because of RTSA reliance on a functional deltoid.^5,6 The cause of these fractures remains to be fully elucidated. Wahlquist and colleagues⁶ in 2011 reported the cases of 5 patients that sustained acromial base fractures after RTSA. All 5 patients were noted to have unsatisfactory functional results despite achieving union (3 were treated with open reduction and internal fixation, and 2 were treated nonoperatively). Acromial fractures tend to present with pain within 6 months of surgery, which may indicate excessive constraint about the scapula, eventually leading to fracture. Furthermore, disruption of this bony structure can lead to devastating results because the acromial base serves as a fulcrum for the deltoid.

Despite a well-placed reverse prosthesis, there is increased reliance on surrounding glenohumeral musculature, resulting from poor rotator cuff function and biomechanical differences compared with a native shoulder. Both our patients were found to have relatively small body habitus. It is possible that, by nature of their smaller statures, they were more susceptible to consequences of excessive joint and soft-tissue tension after RTSA. One explanation for acromial fractures after RTSA is that, by excessively lengthening and/or lateralizing the deltoid, the tension on the acromion in these elderly patients may be sufficient to cause a fracture. A similar mechanism may explain their coracoid fractures. As the arm is lengthened and the prosthesis is tightened, the conjoint tendon is significantly tensioned. We routinely check the tension of these muscles as an extra confirmation of joint stability. However, excessive tension for a significant duration may provide too much stress for bone turnover to match with the inherent repair process, potentially causing a fracture. Recent evidence has also found that bone mineral density of the coracoid diminishes with age, suggesting some predisposition to fracture with lower-energy mechanisms.²²

Another possible cause for coracoid fractures may be the orientation of the implants. While we did not have mechanistic evidence, it is possible that, with adduction and internal rotation, prosthetic impingement against the coracoid is feasible, particularly in patients of small stature. Although a glenoid implant placed high can increase the chance for coracoid–implant impingement, the fact that the patients improved without revision makes chronic mechanical impingement less likely. Drill holes, especially multiple ones, placed throughout the base of the coracoid may also predispose to coracoid fractures.

Patients with periscapular fractures (acromion, scapular spine, or coracoid) after RTSA often present with pain and occasional deficits in function. Both patients in this series noted pain out of proportion to examination. The onset of this pain differed, with 1 patient noting pain within the first 3 months and 1 noting discomfort later. Neither patient had any trauma. In the presence of significant symptoms, negative radiographs, and a poor response to conservative treatment, we recommend advanced imaging to rule out fracture. However, prior to obtaining advanced imaging, proper radiographic techniques should be utilized. Eyres and colleagues,²³ in a series of 12 fractures of the coracoid process, relied primarily on coracoid views directed 45° in a cephalic direction and thin-slice CT. An isotope bone scan identified 1 case not initially found on radiographs.²³

Conservative management with use of a sling until resolution of symptoms was successful in our series. If symptoms persist, a bone stimulator can be used prior to implementing a surgical solution; however, current evidence does not expound on timing and utility of such modalities. Perhaps as important as treatment is education of the patient and the rehabilitation team about the importance of identifying increasing pain as a potential sign of impending fracture in this population. Subsequent activity modification until the pain resolves can help avoid the setback in postoperative recovery that this complication may cause.

Conclusion

We present 2 patients with coracoid fractures encountered at 3 months and 15 months after RTSA. Nonoperative management proved adequate in treating both cases. We suggest a high level of suspicion for possible fracture in the patient who comes in with new-onset pain in a localized region with or without functional deficits.

Reverse total shoulder arthroplasty (RTSA) performed in carefully selected patients often leads to satisfactory outcomes.^1,2 In recent years, its indications and the number performed per year have expanded. Subsequently, there has been a concomitant rise in reported complications,^2,3 with a rate ranging from 19% to 68%.^2,3 Some common complications include scapular notching,^2-4 fracture,^2,3,5-7 dislocation,^2,3,7 and infection.^2,3,7

In this series, we describe 2 cases of coracoid fracture after RTSA. The patients provided written informed consent for print and electronic publication of these case reports.

Case Series

Case 1

An independently functioning 81-year-old right hand–dominant woman (BMI, 22.1 [height, 160 cm; weight, 56.7 kg]) presented with increasing left shoulder pain and dysfunction after a motor vehicle accident 2 months earlier. She had reported vague chronic left shoulder pain in the past, but after the accident her pain was significantly worse. A subacromial corticosteroid injection by her primary care physician provided temporary symptomatic relief, but her symptoms recurred.

On presentation, there was obvious anterior superior escape of the humeral head, which was accentuated by shoulder shrug. Her deltoid motor function was found to be intact, and her active shoulder range of motion was severely limited (pseudoparesis). There was notable crepitation as well as significant weakness and pain with abduction and external rotation strength testing.

Radiographic imaging showed anterior superior escape of the humeral head with some early degenerative changes (Seebauer type IIB⁸ [Figure 1A]). Magnetic resonance imaging confirmed a full-thickness retracted massive rotator cuff tear with complete involvement of the supraspinatus, infraspinatus, and most of the subscapularis muscles. Significant glenohumeral degenerative changes consistent with cuff tear arthropathy were also seen without any evidence of fracture.

After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. The patient underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 12, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 3 screws, and the stem was placed in neutral version. The patient’s shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiograph (Figure 1B) showed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful, and rehabilitation consisted of 6 weeks of sling protection, with advancing passive and active range of motion. Strengthening exercises were initiated 6 weeks after surgery.

At the patient’s 6-week postoperative visit, she demonstrated pain-free passive elevation to 80° and active forward elevation to 70°. At her 3-month postoperative visit, she reported a 1-week onset of anterior shoulder pain accompanied by a strange noise at the anterior aspect of the operative shoulder. She denied any recent trauma. She continued to have minimal shoulder pain with passive forward flexion of 80°; however, her active forward elevation was very limited because of pain in the anterior aspect of her shoulder. Active external rotation was noted to be 20° and internal rotation was to her buttock. She had pain to palpation of the coracoid process. Radiographs were unchanged from immediate postoperative radiographs. Computed tomography (CT), which was ordered to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis and no loosening. However, CT was notable for a nondisplaced fracture through the base of the coracoid (Figures 2A-2D). The patient stopped formal physical therapy, and sling immobilization was initiated. After 3 weeks, the sling was discontinued and physical therapy was begun again. She responded satisfactorily to this treatment approach, and, at her 6-month postoperative follow-up, she was without pain, instability, or crepitation. Her range of motion had improved with pain-free active forward flexion, external rotation, and abduction of 100°, 15°, and 90°, respectively. At 28-month postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 3, 73, and 67, respectively. 

Case 2

A 68-year-old, right-handed woman (BMI, 22.5 [height, 160 cm; weight, 57.6 kg]) presented with right shoulder pain and dysfunction of 3 years’ duration. She had undergone an open rotator cuff repair at an outside facility 4 years ago that was unsuccessful. At the time of her presentation to our institution, she had already undergone a failed course of physical therapy. A trial of corticosteroid subacromial injections did not adequately manage her symptoms.

On presentation, her active forward flexion, abduction, and external rotation were 40°, 30°, and 10°, respectively. She had full passive range of motion and pain with active and passive shoulder motion. Radiographic imaging showed superior migration of the humeral head with evidence of glenohumeral arthropathy suggestive of rotator cuff arthropathy (Seebauer type IIA⁸). After thorough discussion of options, risks, and benefits, the decision was made to proceed with RTSA. She underwent the procedure without complications. A DePuy Delta Xtend prosthesis was used with a cemented humeral stem, polyethylene, and glenosphere, sizes of 8, +3, and 38, respectively. The glenosphere component, positioned inferiorly to avoid scapular notching, was secured with 4 screws, and the stem was placed in neutral version. Her shoulder was reduced, ranged, and noted to be stable, allowing for supple passive range of motion without evidence of excessive tightness. She was placed in a sling with the shoulder positioned in neutral alignment. Her postoperative radiographs revealed satisfactory implantation of the reverse total shoulder prosthesis. Her postoperative course was uneventful. She was taken out of her shoulder immobilizer 4 weeks after surgery and began home-based physical therapy.

At 1 year after surgery, the patient had minimal shoulder pain with active forward flexion, external rotation, and abduction of 135°, 20°, and 85°, respectively. She presented to our clinic 15 months after RTSA with acute onset of pain about her anterior shoulder. She denied any recent trauma or infectious exposures. On examination, her motion was unchanged from prior examinations. However, she was tender on palpation of the coracoid. Radiographs at that time were unchanged (Figures 3A, 3B). Laboratory tests (erythrocyte sedimentation rate, C-reactive protein, and complete blood count with differential) that were subsequently ordered to rule out an occult infection were within normal limits. Computed tomography, which was ordered for further assessment and to ensure that the implant was stable with no loosening, showed satisfactory alignment of the prosthesis without loosening. However, a lucency was noted in the midportion of the coracoid that was suggestive of a fracture (Figures 4A, 4B). A conservative plan of treatment was advised with sling immobilization for 3 weeks and follow-up visits. The patient responded satisfactorily to this treatment approach, and, at her latest follow-up, 8 months after presenting with a coracoid fracture, she was pain-free. At the 5-year postoperative follow-up, her visual analog scale, American Shoulder and Elbow Surgeons score, and Simple Shoulder Test score were 1-2, 78, and 75, respectively. 

Discussion

The reverse prosthesis, a semi-constrained ball-and-socket device, provides satisfactory functional outcomes when used in carefully selected patients with rotator cuff arthropathy and pseudoparalysis, failed shoulder arthroplasty, and fracture sequelae.^1,9-11 By the traditional Grammont principles of medializing the center of rotation and lowering the humerus, shear forces about the glenoid are reduced and the deltoid muscle is tensioned, allowing for adequate torque generation, required to facilitate shoulder motion.^12,13 While long-term outcomes  concerning durability and survivorship are pending, some studies have attempted to improve our understanding of implant and functional longevity. Guery and colleagues¹⁴ noted an implant survival of 91% at 120 months. However, increased pain and decreased function were seen at the 6-year mark.¹⁴ A more recent study by Cuff and colleagues¹⁵ revealed 94% implant survivorship and sustained improvement in range of motion and pain at 5 years. 

Despite considerable success, RTSA can be associated with a myriad of complications. The most common complications of RTSA include scapular notching (44%-96%), glenoid side failure (5%-40%), instability (2.4%-31%), and infection (1%-15.3%).^2,3 In the setting of inflammatory arthropathy, there is an increased risk for intraoperative and postoperative fractures.^16,17 To date, there are only 2 reported cases of coracoid process fractures after RTSA.^18,19 In the case by Nolan and colleagues,¹⁸ conservative management with a sling for 6 weeks led to successful resolution of symptoms. Although little information is provided on the management of these rare fractures, literature on the slightly more common scapular (0.9%-7.2%) and acromial (0.9%-4.9%) fractures suggest that periscapular fractures are on the rise, may increase the risk for revision surgery, and can lead to inferior outcomes when compared with patients without fractures.^5,20,21

Acromial fractures after RTSA have been reported to occur at a rate of 0.9% to 4.9%.^5,21 This is a concern because of RTSA reliance on a functional deltoid.^5,6 The cause of these fractures remains to be fully elucidated. Wahlquist and colleagues⁶ in 2011 reported the cases of 5 patients that sustained acromial base fractures after RTSA. All 5 patients were noted to have unsatisfactory functional results despite achieving union (3 were treated with open reduction and internal fixation, and 2 were treated nonoperatively). Acromial fractures tend to present with pain within 6 months of surgery, which may indicate excessive constraint about the scapula, eventually leading to fracture. Furthermore, disruption of this bony structure can lead to devastating results because the acromial base serves as a fulcrum for the deltoid.

Despite a well-placed reverse prosthesis, there is increased reliance on surrounding glenohumeral musculature, resulting from poor rotator cuff function and biomechanical differences compared with a native shoulder. Both our patients were found to have relatively small body habitus. It is possible that, by nature of their smaller statures, they were more susceptible to consequences of excessive joint and soft-tissue tension after RTSA. One explanation for acromial fractures after RTSA is that, by excessively lengthening and/or lateralizing the deltoid, the tension on the acromion in these elderly patients may be sufficient to cause a fracture. A similar mechanism may explain their coracoid fractures. As the arm is lengthened and the prosthesis is tightened, the conjoint tendon is significantly tensioned. We routinely check the tension of these muscles as an extra confirmation of joint stability. However, excessive tension for a significant duration may provide too much stress for bone turnover to match with the inherent repair process, potentially causing a fracture. Recent evidence has also found that bone mineral density of the coracoid diminishes with age, suggesting some predisposition to fracture with lower-energy mechanisms.²²

Another possible cause for coracoid fractures may be the orientation of the implants. While we did not have mechanistic evidence, it is possible that, with adduction and internal rotation, prosthetic impingement against the coracoid is feasible, particularly in patients of small stature. Although a glenoid implant placed high can increase the chance for coracoid–implant impingement, the fact that the patients improved without revision makes chronic mechanical impingement less likely. Drill holes, especially multiple ones, placed throughout the base of the coracoid may also predispose to coracoid fractures.

Patients with periscapular fractures (acromion, scapular spine, or coracoid) after RTSA often present with pain and occasional deficits in function. Both patients in this series noted pain out of proportion to examination. The onset of this pain differed, with 1 patient noting pain within the first 3 months and 1 noting discomfort later. Neither patient had any trauma. In the presence of significant symptoms, negative radiographs, and a poor response to conservative treatment, we recommend advanced imaging to rule out fracture. However, prior to obtaining advanced imaging, proper radiographic techniques should be utilized. Eyres and colleagues,²³ in a series of 12 fractures of the coracoid process, relied primarily on coracoid views directed 45° in a cephalic direction and thin-slice CT. An isotope bone scan identified 1 case not initially found on radiographs.²³

Conservative management with use of a sling until resolution of symptoms was successful in our series. If symptoms persist, a bone stimulator can be used prior to implementing a surgical solution; however, current evidence does not expound on timing and utility of such modalities. Perhaps as important as treatment is education of the patient and the rehabilitation team about the importance of identifying increasing pain as a potential sign of impending fracture in this population. Subsequent activity modification until the pain resolves can help avoid the setback in postoperative recovery that this complication may cause.

Conclusion

We present 2 patients with coracoid fractures encountered at 3 months and 15 months after RTSA. Nonoperative management proved adequate in treating both cases. We suggest a high level of suspicion for possible fracture in the patient who comes in with new-onset pain in a localized region with or without functional deficits.

Neuralgia and neuroma secondary to iatrogenic saphenous nerve injury have been described in the setting of orthopedic surgical interventions using a medial parapatellar approach, and in vascular surgery procedures for harvest of the saphenous vein.^1-3 However, postoperative neuropathic pain caused by saphenous neuroma in the setting of orthopedic trauma has not been described. 

We present a case of symptomatic posttraumatic saphenous neuroma after a displaced and laterally angulated open distal one-third tibial fracture. This unreported cause of postinjury neuralgia is an important complication to address as other similar and more common conditions, such as peripheral neuropathy and complex regional pain syndrome (CRPS), can present in a similar manner. Reaching the correct diagnosis can be challenging for clinicians unfamiliar with this condition or its clinical presentation. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 43-year-old woman presented to our practice 2 years after an open distal one-third metadiaphyseal fracture of the tibia with associated segmental fibular fracture (Gustilo-Anderson type II)⁴ after an automobile/bicycle accident. At the time of injury, she was noted to have a complex medial wound in the region of an open fracture at the junction of the middle and distal thirds of her tibial shaft. She underwent definitive treatment at an outside facility with initial irrigation and débridement and primary wound closure, followed by staged intramedullary nail fixation. Both soft-tissue and bony injuries healed within the expected time frame, and the patient was discharged from orthopedic care. 

Approximately 1 year after her initial injury, the patient began to complain of progressive and persistent anteromedial knee pain as well as gradual-onset, medial-sided leg pain. The leg pain began at the level of her previous fracture site, at the distal one-third metadiaphyseal tibial junction, and radiated from the site of her previous medial open wound distally to the medial aspect of her foot. The pain was burning and tingling in nature, with associated hyperesthesia of the affected area. A diagnosis of CRPS was made, and the patient was prescribed a course of desensitization therapy, oral neuromodulating agents, and physical therapy. After 3 months’ therapy, she remained symptomatic and underwent removal of her proximal tibial interlocking screw fixation (Figure 1). When these measures failed to provide symptomatic relief, and having seen several therapists and physicians, including physiatrists, pain management specialists, and orthopedic surgeons, she presented to our clinic for consultation.

Diagnostic Assessment

On presentation, the patient’s surgical incisions were well healed. At the junction of the middle and distal thirds of the tibia, a 4-cm oblique scar was noted over the anteromedial border of the tibia, the site of her previous open fracture. She demonstrated decreased sensation along the length of this oblique scar, as well as in the distribution of the saphenous nerve distally. Further examination of the previously injured region revealed a positive Tinel sign over the course of the saphenous nerve, with radiating pain down the medial aspect of her leg, recreating her symptoms. She otherwise had full range of motion at the knee with mild tenderness to palpation at the medial joint line and patellar tendon. Her lower extremity motor examination, reflexes, and the remainder of her sensory examination were benign.

These findings were consistent with isolated saphenous neuralgia, and selective injection of the saphenous distribution over the injury site was performed. This injection provided immediate symptomatic relief, with the patient reporting preinjection and postinjection pain scores of 7/10 and 2/10, respectively. Because of the clinical improvement demonstrated with selective injection, surgical intervention with exploration and neurolysis of the saphenous nerve was recommended. 

Therapeutic Intervention

The patient underwent surgical exploration of her saphenous nerve at the level of her original open fracture. This was done concurrently with a left-knee diagnostic arthroscopy and removal of her intramedullary tibial implant both to exclude intra-articular pathology (given her medial joint-line tenderness and the limitation of magnetic resonance imaging to diagnose meniscus tear in the presence of her tibial hardware) and to remove any potential hardware irritation in the setting of her anterior knee pain.^5,6 

Preoperatively, the path of the saphenous nerve was marked using the saphenous vein as a guide. An incision overlying the presumed saphenous nerve course was made at the site of her previous open wound and clinical Tinel sign. The saphenous nerve was carefully dissected with loupe magnification, and the distal divisions of the anterior and posterior branches were identified. The anterior branch was found to be in continuity but encased in fibrotic neuroma. Selective neurolysis of this anterior branch was performed. The posterior branch was found to have been traumatically severed, with both the proximal and distal ends encased in neuromatous scar (Figure 2). Neurectomy of the posterior branch was performed and the severed proximal end of the nerve was buried into the adjacent medial gastrocnemius muscle beneath the fascia of the superficial posterior compartment.^7-9

Postoperative Course and Outcome

Upon transport to the postsurgical care unit and emergence from sedation, the patient experienced immediate resolution of her neuralgic symptoms. Pathology of the operative specimen showed a benign, disorganized arrangement of axons, Schwann cells, and perineural fibroblasts amidst a fibrous stroma, consistent with traumatic neuroma. At 1-month and 6-month follow-up visits, the patient remained symptom-free, aside from some continued anterior left knee pain near the site of intramedullary nail entry. Her positive Tinel sign had completely resolved, as did her neuralgic symptoms down the medial aspect of her leg. This proved consistent with a diagnosis of neuroma as the cause of the majority of her symptoms. Subjectively, she reported excellent overall pain relief and satisfaction with her treatment and postoperative course. 

Discussion

Postoperative pain after intramedullary fixation of tibial shaft fractures is common and can be caused by several clinical entities.^5,10 Anterior knee pain is a well-known complication present in up to 73% of patients treated with tibial nailing.¹⁰ Osteoarthritis of the knee or ankle as well as nonarthritic ipsilateral ankle pain are also common complaints, often resulting from tibial malunion or malrotation, leading to altered joint kinematics.¹¹ Additionally, superficial peroneal nerve and tibial neurovascular bundle injuries have been reported as potential complications of distal interlocking screw placement, and should be considered in such patients.¹² 

Another consideration for the development of postoperative pain is CRPS, which is thought to be caused by postinjury sympathetic activation that produces pain out of proportion to clinical examination findings.¹³ Although no postoperative incidence of CRPS in the setting of tibial nailing has been reported, it is a known contributor to poor functional outcomes after fractures or crush injuries to the lower extremity.⁹ When attempting to diagnose and treat chronic postoperative pain after tibial nailing, the clinician must keep these common etiologies in mind as well as an understanding of the adjacent anatomy. 

The saphenous nerve originates from the third and fourth lumbar nerve roots, coursing beneath the inguinal ligament as part of the femoral nerve. As the terminal branch of the femoral nerve, the saphenous nerve runs in the Hunter canal beneath the fascia of the sartorius muscle. It is bordered laterally by the vastus medialis muscle, and posteriorly and laterally by the adductor longus and magnus muscles. The saphenous nerve then crosses the femoral artery superficially from medial to lateral as it courses distally in the canal. As it emerges from the adductor hiatus, the saphenous nerve runs superficial to the gracilis muscle around the posterior border of the sartorius muscle with the descending genicular artery, and becomes a subcutaneous structure at the level of the knee joint. The infrapatellar branch of the saphenous nerve provides sensation to the medial knee, and continues in a subcutaneous course just medial to the posterior aspect of the tibial shaft with the great saphenous vein. The nerve distally supplies sensory input from the medial foot and ankle.^1,3,14

There are several causes of saphenous neuralgia related to surgical and nonsurgical trauma.^2,3,15,16 The most common cause of nerve injury is iatrogenic traction or transection causing neuralgic sequelae from subsequent neuroma formation. The anatomy of the saphenous nerve puts it at particular risk when performing saphenectomy for vascular procedures, and its infrapatellar branch is at particular risk when performing a medial parapatellar approach for total knee arthroplasty.^2,3 In the case of the surgically naïve patient, saphenous nerve entrapment syndromes have also been described, and occur most frequently at the level of the adductor hiatus or as the saphenous nerve courses between the sartorius and gracilis muscles proximal to the knee joint.¹⁶ 

As is illustrated in the present case, orthopedic trauma may be an additional cause of saphenous neuroma formation, leading to symptomatic neuralgia. This case suggests that symptomatic neuroma should be included in the differential diagnosis of posttraumatic pain in the orthopedic trauma patient. It is important to note that, although this case occurred after a severe injury, the intimate association of the saphenous nerve with the tibia places it in a vulnerable position, and traumatic transection is possible after closed injuries to the tibial metadiaphyseal junction or tibial shaft. 

Neuroma formation occurs in response to damage to the endoneurium and axon. For an axon to repair properly, the damaged proximal segment must join with, and reenter, the distal stump. As axons attempt to regenerate, occasionally the proximal stump can escape into the surrounding tissue and form a painful neuroma consisting of a disorganized mass of Schwann cells, fibroblasts, blood vessels, and axons with various degrees of myelination. The subsequent neuralgia associated with neuroma formation is caused by chemical or mechanical stimulation of the damaged axons or by spontaneously evoked potentials in the damaged axons. These signals can manifest as a variety of symptoms, including paresthesia and allodynia.^17,18

Making the diagnosis of neuroma-related neuralgia can be challenging and nebulous. A characteristic history and positive Tinel sign over the affected area are helpful clinical indicators. However, the clinical finding most predictive of favorable surgical outcome is symptomatic relief after local injection of 1% lidocaine to the affected area. This is an important diagnostic test, especially when attempting to differentiate painful neuroma from other causes of posttraumatic lower extremity pain (eg, CRPS). Such an injection should be performed in the diagnosis and treatment of symptomatic neuroma, and some authors would suggest that insufficient relief of symptoms with diagnostic nerve block is a contraindication to surgical treatment.¹⁹

Several treatments for painful neuromas have been described, with variable results.^19,20 The most widely accepted treatment of a complete nerve transection with associated neuroma is neurectomy with reimplantation of the proximal end into adjacent bone, muscle, or vein.^14,15 Balcin and colleagues²¹ suggest that vein transposition produces the most favorable outcomes. Simple neurolysis of in-continuity neuromas has also been described with favorable results.

Conclusion

Neuralgia-producing neuromas of the saphenous nerve are relatively uncommon but can lead to persistent pain and frustrating symptoms for the patient. As noted, the diagnosis may elude clinicians, especially in patients with less obvious clinical presentations. We suggest the following algorithm to help distinguish between painful neuroma and other causes of posttraumatic leg pain: (1) physical examination (including testing for instability, joint line tenderness, patellofemoral pain, Tinel sign, and Semmes-Weinstein testing) should be performed, and plain radiographs taken of the involved bones and joints; (2) if all of the above reveal no abnormality, and there is a positive Tinel sign directly over the course of a nerve, an injection of lidocaine over the region of the potential neuroma can be diagnostic; (3) should several abnormalities be present, further investigation using magnetic resonance imaging, bone scan, and/or electromyography may provide additional information that leads to a diagnosis. 

Neuralgia and neuroma secondary to iatrogenic saphenous nerve injury have been described in the setting of orthopedic surgical interventions using a medial parapatellar approach, and in vascular surgery procedures for harvest of the saphenous vein.^1-3 However, postoperative neuropathic pain caused by saphenous neuroma in the setting of orthopedic trauma has not been described. 

We present a case of symptomatic posttraumatic saphenous neuroma after a displaced and laterally angulated open distal one-third tibial fracture. This unreported cause of postinjury neuralgia is an important complication to address as other similar and more common conditions, such as peripheral neuropathy and complex regional pain syndrome (CRPS), can present in a similar manner. Reaching the correct diagnosis can be challenging for clinicians unfamiliar with this condition or its clinical presentation. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 43-year-old woman presented to our practice 2 years after an open distal one-third metadiaphyseal fracture of the tibia with associated segmental fibular fracture (Gustilo-Anderson type II)⁴ after an automobile/bicycle accident. At the time of injury, she was noted to have a complex medial wound in the region of an open fracture at the junction of the middle and distal thirds of her tibial shaft. She underwent definitive treatment at an outside facility with initial irrigation and débridement and primary wound closure, followed by staged intramedullary nail fixation. Both soft-tissue and bony injuries healed within the expected time frame, and the patient was discharged from orthopedic care. 

Approximately 1 year after her initial injury, the patient began to complain of progressive and persistent anteromedial knee pain as well as gradual-onset, medial-sided leg pain. The leg pain began at the level of her previous fracture site, at the distal one-third metadiaphyseal tibial junction, and radiated from the site of her previous medial open wound distally to the medial aspect of her foot. The pain was burning and tingling in nature, with associated hyperesthesia of the affected area. A diagnosis of CRPS was made, and the patient was prescribed a course of desensitization therapy, oral neuromodulating agents, and physical therapy. After 3 months’ therapy, she remained symptomatic and underwent removal of her proximal tibial interlocking screw fixation (Figure 1). When these measures failed to provide symptomatic relief, and having seen several therapists and physicians, including physiatrists, pain management specialists, and orthopedic surgeons, she presented to our clinic for consultation.

Diagnostic Assessment

On presentation, the patient’s surgical incisions were well healed. At the junction of the middle and distal thirds of the tibia, a 4-cm oblique scar was noted over the anteromedial border of the tibia, the site of her previous open fracture. She demonstrated decreased sensation along the length of this oblique scar, as well as in the distribution of the saphenous nerve distally. Further examination of the previously injured region revealed a positive Tinel sign over the course of the saphenous nerve, with radiating pain down the medial aspect of her leg, recreating her symptoms. She otherwise had full range of motion at the knee with mild tenderness to palpation at the medial joint line and patellar tendon. Her lower extremity motor examination, reflexes, and the remainder of her sensory examination were benign.

These findings were consistent with isolated saphenous neuralgia, and selective injection of the saphenous distribution over the injury site was performed. This injection provided immediate symptomatic relief, with the patient reporting preinjection and postinjection pain scores of 7/10 and 2/10, respectively. Because of the clinical improvement demonstrated with selective injection, surgical intervention with exploration and neurolysis of the saphenous nerve was recommended. 

Therapeutic Intervention

The patient underwent surgical exploration of her saphenous nerve at the level of her original open fracture. This was done concurrently with a left-knee diagnostic arthroscopy and removal of her intramedullary tibial implant both to exclude intra-articular pathology (given her medial joint-line tenderness and the limitation of magnetic resonance imaging to diagnose meniscus tear in the presence of her tibial hardware) and to remove any potential hardware irritation in the setting of her anterior knee pain.^5,6 

Preoperatively, the path of the saphenous nerve was marked using the saphenous vein as a guide. An incision overlying the presumed saphenous nerve course was made at the site of her previous open wound and clinical Tinel sign. The saphenous nerve was carefully dissected with loupe magnification, and the distal divisions of the anterior and posterior branches were identified. The anterior branch was found to be in continuity but encased in fibrotic neuroma. Selective neurolysis of this anterior branch was performed. The posterior branch was found to have been traumatically severed, with both the proximal and distal ends encased in neuromatous scar (Figure 2). Neurectomy of the posterior branch was performed and the severed proximal end of the nerve was buried into the adjacent medial gastrocnemius muscle beneath the fascia of the superficial posterior compartment.^7-9

Postoperative Course and Outcome

Upon transport to the postsurgical care unit and emergence from sedation, the patient experienced immediate resolution of her neuralgic symptoms. Pathology of the operative specimen showed a benign, disorganized arrangement of axons, Schwann cells, and perineural fibroblasts amidst a fibrous stroma, consistent with traumatic neuroma. At 1-month and 6-month follow-up visits, the patient remained symptom-free, aside from some continued anterior left knee pain near the site of intramedullary nail entry. Her positive Tinel sign had completely resolved, as did her neuralgic symptoms down the medial aspect of her leg. This proved consistent with a diagnosis of neuroma as the cause of the majority of her symptoms. Subjectively, she reported excellent overall pain relief and satisfaction with her treatment and postoperative course. 

Discussion

Postoperative pain after intramedullary fixation of tibial shaft fractures is common and can be caused by several clinical entities.^5,10 Anterior knee pain is a well-known complication present in up to 73% of patients treated with tibial nailing.¹⁰ Osteoarthritis of the knee or ankle as well as nonarthritic ipsilateral ankle pain are also common complaints, often resulting from tibial malunion or malrotation, leading to altered joint kinematics.¹¹ Additionally, superficial peroneal nerve and tibial neurovascular bundle injuries have been reported as potential complications of distal interlocking screw placement, and should be considered in such patients.¹² 

Another consideration for the development of postoperative pain is CRPS, which is thought to be caused by postinjury sympathetic activation that produces pain out of proportion to clinical examination findings.¹³ Although no postoperative incidence of CRPS in the setting of tibial nailing has been reported, it is a known contributor to poor functional outcomes after fractures or crush injuries to the lower extremity.⁹ When attempting to diagnose and treat chronic postoperative pain after tibial nailing, the clinician must keep these common etiologies in mind as well as an understanding of the adjacent anatomy. 

The saphenous nerve originates from the third and fourth lumbar nerve roots, coursing beneath the inguinal ligament as part of the femoral nerve. As the terminal branch of the femoral nerve, the saphenous nerve runs in the Hunter canal beneath the fascia of the sartorius muscle. It is bordered laterally by the vastus medialis muscle, and posteriorly and laterally by the adductor longus and magnus muscles. The saphenous nerve then crosses the femoral artery superficially from medial to lateral as it courses distally in the canal. As it emerges from the adductor hiatus, the saphenous nerve runs superficial to the gracilis muscle around the posterior border of the sartorius muscle with the descending genicular artery, and becomes a subcutaneous structure at the level of the knee joint. The infrapatellar branch of the saphenous nerve provides sensation to the medial knee, and continues in a subcutaneous course just medial to the posterior aspect of the tibial shaft with the great saphenous vein. The nerve distally supplies sensory input from the medial foot and ankle.^1,3,14

There are several causes of saphenous neuralgia related to surgical and nonsurgical trauma.^2,3,15,16 The most common cause of nerve injury is iatrogenic traction or transection causing neuralgic sequelae from subsequent neuroma formation. The anatomy of the saphenous nerve puts it at particular risk when performing saphenectomy for vascular procedures, and its infrapatellar branch is at particular risk when performing a medial parapatellar approach for total knee arthroplasty.^2,3 In the case of the surgically naïve patient, saphenous nerve entrapment syndromes have also been described, and occur most frequently at the level of the adductor hiatus or as the saphenous nerve courses between the sartorius and gracilis muscles proximal to the knee joint.¹⁶ 

As is illustrated in the present case, orthopedic trauma may be an additional cause of saphenous neuroma formation, leading to symptomatic neuralgia. This case suggests that symptomatic neuroma should be included in the differential diagnosis of posttraumatic pain in the orthopedic trauma patient. It is important to note that, although this case occurred after a severe injury, the intimate association of the saphenous nerve with the tibia places it in a vulnerable position, and traumatic transection is possible after closed injuries to the tibial metadiaphyseal junction or tibial shaft. 

Neuroma formation occurs in response to damage to the endoneurium and axon. For an axon to repair properly, the damaged proximal segment must join with, and reenter, the distal stump. As axons attempt to regenerate, occasionally the proximal stump can escape into the surrounding tissue and form a painful neuroma consisting of a disorganized mass of Schwann cells, fibroblasts, blood vessels, and axons with various degrees of myelination. The subsequent neuralgia associated with neuroma formation is caused by chemical or mechanical stimulation of the damaged axons or by spontaneously evoked potentials in the damaged axons. These signals can manifest as a variety of symptoms, including paresthesia and allodynia.^17,18

Making the diagnosis of neuroma-related neuralgia can be challenging and nebulous. A characteristic history and positive Tinel sign over the affected area are helpful clinical indicators. However, the clinical finding most predictive of favorable surgical outcome is symptomatic relief after local injection of 1% lidocaine to the affected area. This is an important diagnostic test, especially when attempting to differentiate painful neuroma from other causes of posttraumatic lower extremity pain (eg, CRPS). Such an injection should be performed in the diagnosis and treatment of symptomatic neuroma, and some authors would suggest that insufficient relief of symptoms with diagnostic nerve block is a contraindication to surgical treatment.¹⁹

Several treatments for painful neuromas have been described, with variable results.^19,20 The most widely accepted treatment of a complete nerve transection with associated neuroma is neurectomy with reimplantation of the proximal end into adjacent bone, muscle, or vein.^14,15 Balcin and colleagues²¹ suggest that vein transposition produces the most favorable outcomes. Simple neurolysis of in-continuity neuromas has also been described with favorable results.

Conclusion

Neuralgia-producing neuromas of the saphenous nerve are relatively uncommon but can lead to persistent pain and frustrating symptoms for the patient. As noted, the diagnosis may elude clinicians, especially in patients with less obvious clinical presentations. We suggest the following algorithm to help distinguish between painful neuroma and other causes of posttraumatic leg pain: (1) physical examination (including testing for instability, joint line tenderness, patellofemoral pain, Tinel sign, and Semmes-Weinstein testing) should be performed, and plain radiographs taken of the involved bones and joints; (2) if all of the above reveal no abnormality, and there is a positive Tinel sign directly over the course of a nerve, an injection of lidocaine over the region of the potential neuroma can be diagnostic; (3) should several abnormalities be present, further investigation using magnetic resonance imaging, bone scan, and/or electromyography may provide additional information that leads to a diagnosis. 

Neuralgia and neuroma secondary to iatrogenic saphenous nerve injury have been described in the setting of orthopedic surgical interventions using a medial parapatellar approach, and in vascular surgery procedures for harvest of the saphenous vein.^1-3 However, postoperative neuropathic pain caused by saphenous neuroma in the setting of orthopedic trauma has not been described. 

We present a case of symptomatic posttraumatic saphenous neuroma after a displaced and laterally angulated open distal one-third tibial fracture. This unreported cause of postinjury neuralgia is an important complication to address as other similar and more common conditions, such as peripheral neuropathy and complex regional pain syndrome (CRPS), can present in a similar manner. Reaching the correct diagnosis can be challenging for clinicians unfamiliar with this condition or its clinical presentation. The patient provided written informed consent for print and electronic publication of this case report.

Case Report

A 43-year-old woman presented to our practice 2 years after an open distal one-third metadiaphyseal fracture of the tibia with associated segmental fibular fracture (Gustilo-Anderson type II)⁴ after an automobile/bicycle accident. At the time of injury, she was noted to have a complex medial wound in the region of an open fracture at the junction of the middle and distal thirds of her tibial shaft. She underwent definitive treatment at an outside facility with initial irrigation and débridement and primary wound closure, followed by staged intramedullary nail fixation. Both soft-tissue and bony injuries healed within the expected time frame, and the patient was discharged from orthopedic care. 

Approximately 1 year after her initial injury, the patient began to complain of progressive and persistent anteromedial knee pain as well as gradual-onset, medial-sided leg pain. The leg pain began at the level of her previous fracture site, at the distal one-third metadiaphyseal tibial junction, and radiated from the site of her previous medial open wound distally to the medial aspect of her foot. The pain was burning and tingling in nature, with associated hyperesthesia of the affected area. A diagnosis of CRPS was made, and the patient was prescribed a course of desensitization therapy, oral neuromodulating agents, and physical therapy. After 3 months’ therapy, she remained symptomatic and underwent removal of her proximal tibial interlocking screw fixation (Figure 1). When these measures failed to provide symptomatic relief, and having seen several therapists and physicians, including physiatrists, pain management specialists, and orthopedic surgeons, she presented to our clinic for consultation.

Diagnostic Assessment

On presentation, the patient’s surgical incisions were well healed. At the junction of the middle and distal thirds of the tibia, a 4-cm oblique scar was noted over the anteromedial border of the tibia, the site of her previous open fracture. She demonstrated decreased sensation along the length of this oblique scar, as well as in the distribution of the saphenous nerve distally. Further examination of the previously injured region revealed a positive Tinel sign over the course of the saphenous nerve, with radiating pain down the medial aspect of her leg, recreating her symptoms. She otherwise had full range of motion at the knee with mild tenderness to palpation at the medial joint line and patellar tendon. Her lower extremity motor examination, reflexes, and the remainder of her sensory examination were benign.

These findings were consistent with isolated saphenous neuralgia, and selective injection of the saphenous distribution over the injury site was performed. This injection provided immediate symptomatic relief, with the patient reporting preinjection and postinjection pain scores of 7/10 and 2/10, respectively. Because of the clinical improvement demonstrated with selective injection, surgical intervention with exploration and neurolysis of the saphenous nerve was recommended. 

Therapeutic Intervention

The patient underwent surgical exploration of her saphenous nerve at the level of her original open fracture. This was done concurrently with a left-knee diagnostic arthroscopy and removal of her intramedullary tibial implant both to exclude intra-articular pathology (given her medial joint-line tenderness and the limitation of magnetic resonance imaging to diagnose meniscus tear in the presence of her tibial hardware) and to remove any potential hardware irritation in the setting of her anterior knee pain.^5,6 

Preoperatively, the path of the saphenous nerve was marked using the saphenous vein as a guide. An incision overlying the presumed saphenous nerve course was made at the site of her previous open wound and clinical Tinel sign. The saphenous nerve was carefully dissected with loupe magnification, and the distal divisions of the anterior and posterior branches were identified. The anterior branch was found to be in continuity but encased in fibrotic neuroma. Selective neurolysis of this anterior branch was performed. The posterior branch was found to have been traumatically severed, with both the proximal and distal ends encased in neuromatous scar (Figure 2). Neurectomy of the posterior branch was performed and the severed proximal end of the nerve was buried into the adjacent medial gastrocnemius muscle beneath the fascia of the superficial posterior compartment.^7-9

Postoperative Course and Outcome

Upon transport to the postsurgical care unit and emergence from sedation, the patient experienced immediate resolution of her neuralgic symptoms. Pathology of the operative specimen showed a benign, disorganized arrangement of axons, Schwann cells, and perineural fibroblasts amidst a fibrous stroma, consistent with traumatic neuroma. At 1-month and 6-month follow-up visits, the patient remained symptom-free, aside from some continued anterior left knee pain near the site of intramedullary nail entry. Her positive Tinel sign had completely resolved, as did her neuralgic symptoms down the medial aspect of her leg. This proved consistent with a diagnosis of neuroma as the cause of the majority of her symptoms. Subjectively, she reported excellent overall pain relief and satisfaction with her treatment and postoperative course. 

Discussion

Postoperative pain after intramedullary fixation of tibial shaft fractures is common and can be caused by several clinical entities.^5,10 Anterior knee pain is a well-known complication present in up to 73% of patients treated with tibial nailing.¹⁰ Osteoarthritis of the knee or ankle as well as nonarthritic ipsilateral ankle pain are also common complaints, often resulting from tibial malunion or malrotation, leading to altered joint kinematics.¹¹ Additionally, superficial peroneal nerve and tibial neurovascular bundle injuries have been reported as potential complications of distal interlocking screw placement, and should be considered in such patients.¹² 

Another consideration for the development of postoperative pain is CRPS, which is thought to be caused by postinjury sympathetic activation that produces pain out of proportion to clinical examination findings.¹³ Although no postoperative incidence of CRPS in the setting of tibial nailing has been reported, it is a known contributor to poor functional outcomes after fractures or crush injuries to the lower extremity.⁹ When attempting to diagnose and treat chronic postoperative pain after tibial nailing, the clinician must keep these common etiologies in mind as well as an understanding of the adjacent anatomy. 

The saphenous nerve originates from the third and fourth lumbar nerve roots, coursing beneath the inguinal ligament as part of the femoral nerve. As the terminal branch of the femoral nerve, the saphenous nerve runs in the Hunter canal beneath the fascia of the sartorius muscle. It is bordered laterally by the vastus medialis muscle, and posteriorly and laterally by the adductor longus and magnus muscles. The saphenous nerve then crosses the femoral artery superficially from medial to lateral as it courses distally in the canal. As it emerges from the adductor hiatus, the saphenous nerve runs superficial to the gracilis muscle around the posterior border of the sartorius muscle with the descending genicular artery, and becomes a subcutaneous structure at the level of the knee joint. The infrapatellar branch of the saphenous nerve provides sensation to the medial knee, and continues in a subcutaneous course just medial to the posterior aspect of the tibial shaft with the great saphenous vein. The nerve distally supplies sensory input from the medial foot and ankle.^1,3,14

There are several causes of saphenous neuralgia related to surgical and nonsurgical trauma.^2,3,15,16 The most common cause of nerve injury is iatrogenic traction or transection causing neuralgic sequelae from subsequent neuroma formation. The anatomy of the saphenous nerve puts it at particular risk when performing saphenectomy for vascular procedures, and its infrapatellar branch is at particular risk when performing a medial parapatellar approach for total knee arthroplasty.^2,3 In the case of the surgically naïve patient, saphenous nerve entrapment syndromes have also been described, and occur most frequently at the level of the adductor hiatus or as the saphenous nerve courses between the sartorius and gracilis muscles proximal to the knee joint.¹⁶ 

As is illustrated in the present case, orthopedic trauma may be an additional cause of saphenous neuroma formation, leading to symptomatic neuralgia. This case suggests that symptomatic neuroma should be included in the differential diagnosis of posttraumatic pain in the orthopedic trauma patient. It is important to note that, although this case occurred after a severe injury, the intimate association of the saphenous nerve with the tibia places it in a vulnerable position, and traumatic transection is possible after closed injuries to the tibial metadiaphyseal junction or tibial shaft. 

Neuroma formation occurs in response to damage to the endoneurium and axon. For an axon to repair properly, the damaged proximal segment must join with, and reenter, the distal stump. As axons attempt to regenerate, occasionally the proximal stump can escape into the surrounding tissue and form a painful neuroma consisting of a disorganized mass of Schwann cells, fibroblasts, blood vessels, and axons with various degrees of myelination. The subsequent neuralgia associated with neuroma formation is caused by chemical or mechanical stimulation of the damaged axons or by spontaneously evoked potentials in the damaged axons. These signals can manifest as a variety of symptoms, including paresthesia and allodynia.^17,18

Making the diagnosis of neuroma-related neuralgia can be challenging and nebulous. A characteristic history and positive Tinel sign over the affected area are helpful clinical indicators. However, the clinical finding most predictive of favorable surgical outcome is symptomatic relief after local injection of 1% lidocaine to the affected area. This is an important diagnostic test, especially when attempting to differentiate painful neuroma from other causes of posttraumatic lower extremity pain (eg, CRPS). Such an injection should be performed in the diagnosis and treatment of symptomatic neuroma, and some authors would suggest that insufficient relief of symptoms with diagnostic nerve block is a contraindication to surgical treatment.¹⁹

Several treatments for painful neuromas have been described, with variable results.^19,20 The most widely accepted treatment of a complete nerve transection with associated neuroma is neurectomy with reimplantation of the proximal end into adjacent bone, muscle, or vein.^14,15 Balcin and colleagues²¹ suggest that vein transposition produces the most favorable outcomes. Simple neurolysis of in-continuity neuromas has also been described with favorable results.

Conclusion

Neuralgia-producing neuromas of the saphenous nerve are relatively uncommon but can lead to persistent pain and frustrating symptoms for the patient. As noted, the diagnosis may elude clinicians, especially in patients with less obvious clinical presentations. We suggest the following algorithm to help distinguish between painful neuroma and other causes of posttraumatic leg pain: (1) physical examination (including testing for instability, joint line tenderness, patellofemoral pain, Tinel sign, and Semmes-Weinstein testing) should be performed, and plain radiographs taken of the involved bones and joints; (2) if all of the above reveal no abnormality, and there is a positive Tinel sign directly over the course of a nerve, an injection of lidocaine over the region of the potential neuroma can be diagnostic; (3) should several abnormalities be present, further investigation using magnetic resonance imaging, bone scan, and/or electromyography may provide additional information that leads to a diagnosis. 

According to the CDC, the medical use of opioid painkillers has increased at least 10-fold during the past 
20 years, “because of a movement toward more aggressive management of pain.”¹ Although opioid therapy is generally considered effective for the treatment of pain, long-term use (both orally and intrathecally) is associated with adverse effects (AEs) such as constipation, fatigue, nausea, sleep disturbances, depression, sexual dysfunction, and hypogonadism.^2,3Opioid-induced androgen deficiency (OPIAD), as defined by Smith and Elliot, is a clinical syndrome characterized by inappropriately low concentrations of gonadotropins (specifically, follicle-stimulating hormone [FSH] and luteinizing hormone [LH]), which leads to inadequate production of sex hormones, including estradiol and testosterone.⁴

Related: Testosterone Replacement Therapy: Playing Catch-up With Patients

The mechanism behind this phenomenon is initiated by either endogenous or exogenous opioids acting on opioid receptors in the hypothalamus, which causes a decrease in the release of gonadotropin-
releasing hormone (GnRH). This decrease in GnRH causes a reduction in the release of LH and FSH from the pituitary gland as well as testosterone or estradiol from the gonads.^4,5 Various guidelines report different cutoffs for the lower limit of normal total testosterone: The Endocrine Society recommends 300 ng/dL, the American Association of Clinical Endocrinologists suggests 200 ng/dL, and various other organizations suggest 230 ng/dL.^6-8 Hypotestosteronism can result in patients presenting with a broad spectrum of clinical symptoms, including reduced libido, erectile dysfunction (ED), fatigue, hot flashes, depression, anemia, decreased muscle mass, weight gain, and osteopenia or osteoporosis.⁴ Women with low testosterone levels can experience irregular menstrual periods, oligomenorrhea, or amenorrhea.⁹ Opioid-induced androgen deficiency often goes unrecognized and untreated. The reported prevalence of opioid-induced hypogonadism ranges from 21% to 86%.^4,9 Given the growing number of patients on chronic opioid therapy, OPIAD warrants further investigation to identify the prevalence in the veteran population to appropriately monitor and manage this deficiency.

The objective of this retrospective review was to identify the presence of secondary hypogonadism in chronic opioid users among a cohort of veterans receiving chronic opioids for nonmalignant pain. In addition to identifying the presence of secondary hypogonadism, the relationship between testosterone concentrations and total daily morphine equivalent doses (MEDs) was reviewed. These data along with the new information recently published on testosterone replacement therapy (TRT) and cardiovascular (CV) risk were then used to evaluate current practices at the West Palm Beach VAMC for OPIAD monitoring and management and to modify and update the local Criteria for Use (CFU) for TRT.

Methods

Patient data from the West Palm Beach VAMC in Florida from January 2013 to December 2013 were reviewed to identify patients who had a total testosterone (TT) level measured. All patient appointments for evaluation and treatment by the clinical pharmacy specialist in pain management were reviewed for data collection. This retrospective review was approved by the scientific advisory committee as part of the facility’s ongoing performance improvement efforts as defined by VHA Handbook 1058.05 and did not require written patient consent.¹⁰

Several distinct TT level data were collected. The descriptive data included patient age; gender; type of treated pain; testosterone level(s) drawn, including TT level before opioid therapy, TT level before/during/after TRT, and current total testosterone level; total daily MED of opioid therapy; duration of chronic opioid therapy; symptoms of exhibited hypogonadism; TRT formulation, dose, and duration; TRT prescriber; symptom change (if any); and laboratory tests ordered for TRT monitoring (lipid profile, liver profile, complete blood count, LH/FSH, and prostate specific antigen [PSA] panel).^5,11,12

Related: Combination Treatment Relieves Opioid-Induced Constipation

Daily MED of opioid therapy was calculated using the VA/DoD opioid conversion table for patients on oxycodone, hydromorphone, or hydrocodone.¹³ For those on the fentanyl patch or methadone, conversion factors of 1:2 (fentanyl [µg/h]:morphine [mg/d]) and 1:3 (methadone:morphine) were used to convert to the MED.¹⁴ For patients on the buprenorphine patch, the package insert was used to convert to the corresponding MED.¹⁵ Combination therapies used the applicable conversions to calculate the total daily MED.

Once the data were collected, descriptive statistics were used to analyze the data. In addition, 4 graphs were generated to review potential relationships. The correlation coefficient was calculated using the Alcula Online Statistics Calculator (http://www.alcula.com; Correlation Coefficient Calculator).

Results

A total of 316 unique veteran patients were seen by the clinical pharmacy specialist in pain management from January 1, 2013, through December 31, 2013. Of these, 73 patients (23.1%) had at least 1 TT level drawn in 2013. Three patients with testosterone levels drawn (4.1%) were excluded from the data analysis for the following reasons: 1 patient did not have testosterone levels on file before receiving testosterone replacement from a non-VA source, 
1 patient received opioids from a non-VA source (MED and duration of opioid therapy could not be calculated), and 1 patient inconsistently received opioids and MED used at the time of testosterone level draw. Per the local TRT CFU, a TT level > 350 ng/dL
does not require treatment, whereas levels < 230 ng/dL (with symptoms) may require TRT, and 
< 200 ng/dL should be treated as 
hypogonadal (interpretation based on local laboratory’s reference range for TT).16 Of the 70 patients included in the analysis, 34 (48.6%) had a TT level < 230 ng/dL and would be considered eligible for TRT if they presented with symptoms 
of low testosterone. Of these 34 patients with a low testosterone level, 28 (40%) were being treated or had been treated with TRT (Figure 1).

The average age of the male patients with a testosterone level drawn was 58.3 years, which was not significantly different from the calculated median age of 60 years. No female patients had a testosterone level drawn. On average, the TT level was normal before starting opioids (reference range per local laboratory: 175-781 ng/dL). Once opioids were initiated, patients were treated for an average duration of 52.5 months (calculated through December 2013) with an average daily dose of 126.8 MED (Table). Fifty of the 70 patients (71.4%) with testosterone levels drawn in 2013 received TRT. The most common symptoms reported by patients related to low testosterone included ED, decreased libido, depression, chronic fatigue, generalized weakness, and hot flashes or night sweats.

The average TT level prior to TRT was 145.3, and the average 
testosterone level after initiation of or 
during treatment with TRT was 292.4, which is within the normal TT level range. Most patients receiving TRT were treated with testosterone cypionate injections, and this was also the formulation used for the longest periods, likely due to the local CFU. In addition to testosterone cypionate injections, patients were also treated with testosterone enanthate injections, testosterone patches, and testosterone gel.

Figure 1 compares current testosterone level and testosterone level before TRT with total daily MEDs. Figure 2 compares current testosterone level and testosterone level before TRT with length of opioid therapy. The 2 figures use data from all patients included in the analysis and indicate a potential inverse relationship between the total daily MED and duration of therapy with the testosterone level, although none of the calculated correlation coefficients indicate that a strong relationship was present.

Figures 3 and 4 include data only for patients who had both a testosterone level collected before opioids (baseline testosterone level) and a current testosterone level. Figure 3 trends the data using total daily MED, and Figure 4 uses the duration of opioid therapy. The correlation for Figure 4 is slightly stronger; the strongest negative correlations were identified between total daily MED and testosterone level before opioid therapy (r = -0.273) and duration of opioid therapy and testosterone level prior to opioid therapy (r = -0.396). The trends indicate that most patients had a normal TT level before opioid treatment and that patients treated with higher MEDs and for longer durations of time were more likely to have lower total testosterone levels.

Discussion

Low testosterone levels can adversely affect patients’ quality of life (QOL) and add to patients’ medication burden with the initiation of TRT. Given new data analyzing the potential effects of TRT on CV event risk, the use of TRT should be carefully considered, as it may carry significant risks and may not be suitable for all patients.

In November 2013, a study was published regarding TRT and 
increased CV risk.¹⁷ This was a retrospective cohort study of men with low testosterone levels (< 300 ng/dL) who had undergone coronary angiography in the VA system between 2005 and 2011 (average age in testosterone group was 60.6 years). The results were significant for an absolute rate of events (all-cause mortality, myocardial infarction [MI], and ischemic stroke) of 19.9% in the no testosterone group and 25.7% in the TRT group, an absolute risk difference of 5.8% at 3 years after coronary angiography. Kaplan-Meier survival curves demonstrated that testosterone use was associated with increased risk of death, MI, and stroke. This result was unchanged when adjusted for the presence of coronary artery disease (CAD). In 
addition, no significant difference was found between the groups in terms of systolic blood pressure, low-
density lipoprotein cholesterol level, or in the use of beta-blocker and statin medications. What is important to note is that in this cohort, 20% had a prior history of MI and heart failure, and more than 50% had confirmed obstructive CAD on angiography. In addition, as this was an observational study, confounding or bias may exist, and given the study population, 
generalizability may be limited to a veteran population.

Related: A Multidisciplinary Chronic Pain Management Clinic in an Indian Health Service Facility

Another retrospective cohort study assessed the risk of acute nonfatal MI following an initial TRT prescription in a large health care database (average age based on TRT prescription was 54.4 years).¹⁸ In men aged ≥ 65 years, a 2-fold increase in the risk of MI in the immediate 90 days after filling an initial TRT prescription declined to baseline after 91 to 180 days among those who did not refill their prescription. Younger men with a history of heart disease had a 2- to 3-fold increased risk of MI in the 90 days following initial TRT prescription. No excess risk was observed in the younger men without such a history. Again, this study has its limitations related to the retrospective design and use of a health care database as opposed to a randomized controlled trial.

In February 2014, a VA National Pharmacy Benefits Management (PBM) bulletin addressed 2 recent studies that had identified a possible risk of increased CV events in men receiving TRT. The bulletin noted that these studies had prompted the FDA to reassess the CV safety of TRT.¹⁹ The TRT CFU was updated by VISN 8 to ensure that the patients receive appropriate treatment and are monitored accordingly.

One of the major changes to the CFU was defining the reference ranges for TRT (interpretation based on a local laboratory’s reference range for total testosterone): serum TT < 200 ng/dL be “treated as 
hypogonadal, those with TT 
> 400 ng/dL be considered normal and those with TT 200-400 ng/dL be treated based on their clinical presentation if symptomatic; TT levels 
> 350 ng/dL do not require treatment, and levels below 230 ng/dL (with symptoms) may require testosterone replacement therapy.”¹⁶ Other important updates included revision of the exclusion criteria as well as 
highlighting special considerations related to TRT, including the use of free testosterone levels rather than TT levels in patients with suspected protein-binding issues, role in fertility treatments, limited use in patients on spironolactone therapy (due to spironolactone’s anti-androgen effects), and potential association with mood and behavior.¹⁶

As chronic opioid therapy is associated with OPIAD, the renewed interest in TRT and its potential AEs provides yet another reason to reconsider opioid therapy. This is especially valid when opioids are the potential cause of hypogonadism and the reaction is treating the AEs of opioids (as opposed to considering elimination of the causative agent) with a therapy that can potentially increase the risk for CV events so that opioids can be continued. Outside the potential CV risk with TRT, opioids carry the innate risk for substance abuse and addiction.

The Opioid Safety Initiative Requirements was released as a memorandum in April 2014 and is the VHA’s effort to “reduce harm from unsafe medications and/or excessive doses while adequately controlling pain in Veterans.”²⁰ Although it does not discuss the risk of OPIAD, it does highlight the need to identify and mitigate high-risk patients as well as high-risk opioid regimens. All these factors, including the possibility of hypogonadism, should be considered before starting opioid therapy and at the time of opioid renewal, as it is known that opioid therapy is not without risks.

At the West Palm Beach VAMC, the primary care providers (PCPs) are responsible for the management of TRT, including the workup, renewal, and monitoring. The Chronic Nonmalignant Pain Management Clinic (CNMPMC) orders testosterone levels on patients who report symptoms of low testosterone, such as hot flashes, depression, and low energy level and refers them to their PCP as indicated. The authors believe that this is most appropriate for a number of reasons: (1) the CNMPMC is a consult service, and patients are not followed indefinitely; (2) patients should be fully evaluated for appropriateness of TRT (including assessment of CV risk) before starting therapy; and (3) the necessary monitoring parameters (laboratory testing, digital rectal exam, and osteoporosis screening) are not typically within the VA pain clinic provider’s scope of practice or expertise. A consideration for future practice would be to incorporate the use of a standardized questionnaire for OPIAD monitoring in patients receiving ≥ 100 mg of morphine daily (eg, the Aging Males’ Symptoms scale).²¹ It should, however, be at the forefront of the pain specialist’s and PCP’s minds that all patients on chronic opioid therapy or considering chronic opioid therapy should be counseled on the risk for OPIAD. If OPIAD is identified, the patient should be carefully considered for an opioid dose reduction as an initial management strategy.

Limitations

A limitation of this review is the lack of consistency or adequacy of serum testosterone sampling, noting that valid testosterone levels need to be drawn in the morning and not obtained during a time of acute illness. In addition, testosterone levels need to be drawn at an appropriate interval while on TRT (eg, at the midpoint between testosterone injections).¹⁶ Although the time of the sample collection is documented in the Computerized Patient Record System (CPRS), it is unknown whether the patient was acutely ill on the day of the sampling unless a progress note is entered, and it is difficult to determine whether the level timing was accurate based on the testosterone replacement formulation. Another limitation is that the average decline in serum testosterone levels with aging in men is 1% to 2% per year. A significant fraction of older men have levels below the lower limit of the normal range for healthy young men, so in older men it can be more difficult to determine whether low testosterone is related to chronic opioid use or to older age.^5,16

As this was a retrospective review, additional limitations included the inability to measure subclinical OPIAD, and the data collection related to symptoms of hypogonadism was restricted by documentation in the CPRS progress notes. The lack of data for females does not contribute to the literature on OPIAD in women. Finally, as the total daily MED does not distinguish between short-acting and long-acting opioid therapy, no differences between the impacts of short-acting vs long-
acting opioid therapy on risk for 
hypogonadism can be inferred. There is evidence to suggest that long-acting opioids are associated with a significantly higher risk for OPIAD compared with short-acting opioids, although the mechanism behind this is not well established.^22,23

Conclusions

The average age of the patients 
on chronic opioid therapy with a 
testosterone level drawn in this cohort was 58.3 years, which is younger than originally anticipated. The median age of 60 years is not significantly different from the average age, indicating that outliers did not impact this calculation. On average, the TT level was normal before starting opioids. Once opioids were started, patients were treated for an average duration of 52.5 months with an average daily dose of 126.8 mg MED. In this veteran cohort, 48.6% of patients met the criteria for TRT based on TT level alone, which is within the reported prevalence range of opioid-induced hypogonadism already published.^4,9 These results are in line with the original hypothesis that chronic opioid use can adversely impact testosterone levels and can have a poor effect on a patient’s QOL due to symptoms of low testosterone. In addition to TRT, possible and suggested (but not proven) treatment options for OPIAD include discontinuation of opioid therapy, opioid rotation, or conversion to buprenorphine.²¹ The approach used should account for multiple patient-specific factors and should be individualized.

Based on the data, there is a trend toward lower testosterone levels in veterans treated with higher MED and for longer periods with chronic opioids. Given recent data that infer that TRT carries increased CV risk as well as the VHA’s Opioid Safety Initiative, it is imperative that providers closely evaluate the appropriateness of starting TRT and/or continuing chronic opioid therapy. All patients generally should have failed non-
opioid management prior to opioid therapy for chronic nonmalignant pain, and this should be documented accordingly. It is also crucial to have the “opioid talk” with patients from time to time and discuss the risks vs benefits, the potential for addiction, overdose, dependence, tolerance, constipation, and OPIAD so patients can continue to be an active and informed participants in their care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

According to the CDC, the medical use of opioid painkillers has increased at least 10-fold during the past 
20 years, “because of a movement toward more aggressive management of pain.”¹ Although opioid therapy is generally considered effective for the treatment of pain, long-term use (both orally and intrathecally) is associated with adverse effects (AEs) such as constipation, fatigue, nausea, sleep disturbances, depression, sexual dysfunction, and hypogonadism.^2,3Opioid-induced androgen deficiency (OPIAD), as defined by Smith and Elliot, is a clinical syndrome characterized by inappropriately low concentrations of gonadotropins (specifically, follicle-stimulating hormone [FSH] and luteinizing hormone [LH]), which leads to inadequate production of sex hormones, including estradiol and testosterone.⁴

Related: Testosterone Replacement Therapy: Playing Catch-up With Patients

The mechanism behind this phenomenon is initiated by either endogenous or exogenous opioids acting on opioid receptors in the hypothalamus, which causes a decrease in the release of gonadotropin-
releasing hormone (GnRH). This decrease in GnRH causes a reduction in the release of LH and FSH from the pituitary gland as well as testosterone or estradiol from the gonads.^4,5 Various guidelines report different cutoffs for the lower limit of normal total testosterone: The Endocrine Society recommends 300 ng/dL, the American Association of Clinical Endocrinologists suggests 200 ng/dL, and various other organizations suggest 230 ng/dL.^6-8 Hypotestosteronism can result in patients presenting with a broad spectrum of clinical symptoms, including reduced libido, erectile dysfunction (ED), fatigue, hot flashes, depression, anemia, decreased muscle mass, weight gain, and osteopenia or osteoporosis.⁴ Women with low testosterone levels can experience irregular menstrual periods, oligomenorrhea, or amenorrhea.⁹ Opioid-induced androgen deficiency often goes unrecognized and untreated. The reported prevalence of opioid-induced hypogonadism ranges from 21% to 86%.^4,9 Given the growing number of patients on chronic opioid therapy, OPIAD warrants further investigation to identify the prevalence in the veteran population to appropriately monitor and manage this deficiency.

The objective of this retrospective review was to identify the presence of secondary hypogonadism in chronic opioid users among a cohort of veterans receiving chronic opioids for nonmalignant pain. In addition to identifying the presence of secondary hypogonadism, the relationship between testosterone concentrations and total daily morphine equivalent doses (MEDs) was reviewed. These data along with the new information recently published on testosterone replacement therapy (TRT) and cardiovascular (CV) risk were then used to evaluate current practices at the West Palm Beach VAMC for OPIAD monitoring and management and to modify and update the local Criteria for Use (CFU) for TRT.

Methods

Patient data from the West Palm Beach VAMC in Florida from January 2013 to December 2013 were reviewed to identify patients who had a total testosterone (TT) level measured. All patient appointments for evaluation and treatment by the clinical pharmacy specialist in pain management were reviewed for data collection. This retrospective review was approved by the scientific advisory committee as part of the facility’s ongoing performance improvement efforts as defined by VHA Handbook 1058.05 and did not require written patient consent.¹⁰

Several distinct TT level data were collected. The descriptive data included patient age; gender; type of treated pain; testosterone level(s) drawn, including TT level before opioid therapy, TT level before/during/after TRT, and current total testosterone level; total daily MED of opioid therapy; duration of chronic opioid therapy; symptoms of exhibited hypogonadism; TRT formulation, dose, and duration; TRT prescriber; symptom change (if any); and laboratory tests ordered for TRT monitoring (lipid profile, liver profile, complete blood count, LH/FSH, and prostate specific antigen [PSA] panel).^5,11,12

Related: Combination Treatment Relieves Opioid-Induced Constipation

Daily MED of opioid therapy was calculated using the VA/DoD opioid conversion table for patients on oxycodone, hydromorphone, or hydrocodone.¹³ For those on the fentanyl patch or methadone, conversion factors of 1:2 (fentanyl [µg/h]:morphine [mg/d]) and 1:3 (methadone:morphine) were used to convert to the MED.¹⁴ For patients on the buprenorphine patch, the package insert was used to convert to the corresponding MED.¹⁵ Combination therapies used the applicable conversions to calculate the total daily MED.

Once the data were collected, descriptive statistics were used to analyze the data. In addition, 4 graphs were generated to review potential relationships. The correlation coefficient was calculated using the Alcula Online Statistics Calculator (http://www.alcula.com; Correlation Coefficient Calculator).

Results

A total of 316 unique veteran patients were seen by the clinical pharmacy specialist in pain management from January 1, 2013, through December 31, 2013. Of these, 73 patients (23.1%) had at least 1 TT level drawn in 2013. Three patients with testosterone levels drawn (4.1%) were excluded from the data analysis for the following reasons: 1 patient did not have testosterone levels on file before receiving testosterone replacement from a non-VA source, 
1 patient received opioids from a non-VA source (MED and duration of opioid therapy could not be calculated), and 1 patient inconsistently received opioids and MED used at the time of testosterone level draw. Per the local TRT CFU, a TT level > 350 ng/dL
does not require treatment, whereas levels < 230 ng/dL (with symptoms) may require TRT, and 
< 200 ng/dL should be treated as 
hypogonadal (interpretation based on local laboratory’s reference range for TT).16 Of the 70 patients included in the analysis, 34 (48.6%) had a TT level < 230 ng/dL and would be considered eligible for TRT if they presented with symptoms 
of low testosterone. Of these 34 patients with a low testosterone level, 28 (40%) were being treated or had been treated with TRT (Figure 1).

The average age of the male patients with a testosterone level drawn was 58.3 years, which was not significantly different from the calculated median age of 60 years. No female patients had a testosterone level drawn. On average, the TT level was normal before starting opioids (reference range per local laboratory: 175-781 ng/dL). Once opioids were initiated, patients were treated for an average duration of 52.5 months (calculated through December 2013) with an average daily dose of 126.8 MED (Table). Fifty of the 70 patients (71.4%) with testosterone levels drawn in 2013 received TRT. The most common symptoms reported by patients related to low testosterone included ED, decreased libido, depression, chronic fatigue, generalized weakness, and hot flashes or night sweats.

The average TT level prior to TRT was 145.3, and the average 
testosterone level after initiation of or 
during treatment with TRT was 292.4, which is within the normal TT level range. Most patients receiving TRT were treated with testosterone cypionate injections, and this was also the formulation used for the longest periods, likely due to the local CFU. In addition to testosterone cypionate injections, patients were also treated with testosterone enanthate injections, testosterone patches, and testosterone gel.

Figure 1 compares current testosterone level and testosterone level before TRT with total daily MEDs. Figure 2 compares current testosterone level and testosterone level before TRT with length of opioid therapy. The 2 figures use data from all patients included in the analysis and indicate a potential inverse relationship between the total daily MED and duration of therapy with the testosterone level, although none of the calculated correlation coefficients indicate that a strong relationship was present.

Figures 3 and 4 include data only for patients who had both a testosterone level collected before opioids (baseline testosterone level) and a current testosterone level. Figure 3 trends the data using total daily MED, and Figure 4 uses the duration of opioid therapy. The correlation for Figure 4 is slightly stronger; the strongest negative correlations were identified between total daily MED and testosterone level before opioid therapy (r = -0.273) and duration of opioid therapy and testosterone level prior to opioid therapy (r = -0.396). The trends indicate that most patients had a normal TT level before opioid treatment and that patients treated with higher MEDs and for longer durations of time were more likely to have lower total testosterone levels.

Discussion

Low testosterone levels can adversely affect patients’ quality of life (QOL) and add to patients’ medication burden with the initiation of TRT. Given new data analyzing the potential effects of TRT on CV event risk, the use of TRT should be carefully considered, as it may carry significant risks and may not be suitable for all patients.

In November 2013, a study was published regarding TRT and 
increased CV risk.¹⁷ This was a retrospective cohort study of men with low testosterone levels (< 300 ng/dL) who had undergone coronary angiography in the VA system between 2005 and 2011 (average age in testosterone group was 60.6 years). The results were significant for an absolute rate of events (all-cause mortality, myocardial infarction [MI], and ischemic stroke) of 19.9% in the no testosterone group and 25.7% in the TRT group, an absolute risk difference of 5.8% at 3 years after coronary angiography. Kaplan-Meier survival curves demonstrated that testosterone use was associated with increased risk of death, MI, and stroke. This result was unchanged when adjusted for the presence of coronary artery disease (CAD). In 
addition, no significant difference was found between the groups in terms of systolic blood pressure, low-
density lipoprotein cholesterol level, or in the use of beta-blocker and statin medications. What is important to note is that in this cohort, 20% had a prior history of MI and heart failure, and more than 50% had confirmed obstructive CAD on angiography. In addition, as this was an observational study, confounding or bias may exist, and given the study population, 
generalizability may be limited to a veteran population.

Related: A Multidisciplinary Chronic Pain Management Clinic in an Indian Health Service Facility

Another retrospective cohort study assessed the risk of acute nonfatal MI following an initial TRT prescription in a large health care database (average age based on TRT prescription was 54.4 years).¹⁸ In men aged ≥ 65 years, a 2-fold increase in the risk of MI in the immediate 90 days after filling an initial TRT prescription declined to baseline after 91 to 180 days among those who did not refill their prescription. Younger men with a history of heart disease had a 2- to 3-fold increased risk of MI in the 90 days following initial TRT prescription. No excess risk was observed in the younger men without such a history. Again, this study has its limitations related to the retrospective design and use of a health care database as opposed to a randomized controlled trial.

In February 2014, a VA National Pharmacy Benefits Management (PBM) bulletin addressed 2 recent studies that had identified a possible risk of increased CV events in men receiving TRT. The bulletin noted that these studies had prompted the FDA to reassess the CV safety of TRT.¹⁹ The TRT CFU was updated by VISN 8 to ensure that the patients receive appropriate treatment and are monitored accordingly.

One of the major changes to the CFU was defining the reference ranges for TRT (interpretation based on a local laboratory’s reference range for total testosterone): serum TT < 200 ng/dL be “treated as 
hypogonadal, those with TT 
> 400 ng/dL be considered normal and those with TT 200-400 ng/dL be treated based on their clinical presentation if symptomatic; TT levels 
> 350 ng/dL do not require treatment, and levels below 230 ng/dL (with symptoms) may require testosterone replacement therapy.”¹⁶ Other important updates included revision of the exclusion criteria as well as 
highlighting special considerations related to TRT, including the use of free testosterone levels rather than TT levels in patients with suspected protein-binding issues, role in fertility treatments, limited use in patients on spironolactone therapy (due to spironolactone’s anti-androgen effects), and potential association with mood and behavior.¹⁶

As chronic opioid therapy is associated with OPIAD, the renewed interest in TRT and its potential AEs provides yet another reason to reconsider opioid therapy. This is especially valid when opioids are the potential cause of hypogonadism and the reaction is treating the AEs of opioids (as opposed to considering elimination of the causative agent) with a therapy that can potentially increase the risk for CV events so that opioids can be continued. Outside the potential CV risk with TRT, opioids carry the innate risk for substance abuse and addiction.

The Opioid Safety Initiative Requirements was released as a memorandum in April 2014 and is the VHA’s effort to “reduce harm from unsafe medications and/or excessive doses while adequately controlling pain in Veterans.”²⁰ Although it does not discuss the risk of OPIAD, it does highlight the need to identify and mitigate high-risk patients as well as high-risk opioid regimens. All these factors, including the possibility of hypogonadism, should be considered before starting opioid therapy and at the time of opioid renewal, as it is known that opioid therapy is not without risks.

At the West Palm Beach VAMC, the primary care providers (PCPs) are responsible for the management of TRT, including the workup, renewal, and monitoring. The Chronic Nonmalignant Pain Management Clinic (CNMPMC) orders testosterone levels on patients who report symptoms of low testosterone, such as hot flashes, depression, and low energy level and refers them to their PCP as indicated. The authors believe that this is most appropriate for a number of reasons: (1) the CNMPMC is a consult service, and patients are not followed indefinitely; (2) patients should be fully evaluated for appropriateness of TRT (including assessment of CV risk) before starting therapy; and (3) the necessary monitoring parameters (laboratory testing, digital rectal exam, and osteoporosis screening) are not typically within the VA pain clinic provider’s scope of practice or expertise. A consideration for future practice would be to incorporate the use of a standardized questionnaire for OPIAD monitoring in patients receiving ≥ 100 mg of morphine daily (eg, the Aging Males’ Symptoms scale).²¹ It should, however, be at the forefront of the pain specialist’s and PCP’s minds that all patients on chronic opioid therapy or considering chronic opioid therapy should be counseled on the risk for OPIAD. If OPIAD is identified, the patient should be carefully considered for an opioid dose reduction as an initial management strategy.

Limitations

A limitation of this review is the lack of consistency or adequacy of serum testosterone sampling, noting that valid testosterone levels need to be drawn in the morning and not obtained during a time of acute illness. In addition, testosterone levels need to be drawn at an appropriate interval while on TRT (eg, at the midpoint between testosterone injections).¹⁶ Although the time of the sample collection is documented in the Computerized Patient Record System (CPRS), it is unknown whether the patient was acutely ill on the day of the sampling unless a progress note is entered, and it is difficult to determine whether the level timing was accurate based on the testosterone replacement formulation. Another limitation is that the average decline in serum testosterone levels with aging in men is 1% to 2% per year. A significant fraction of older men have levels below the lower limit of the normal range for healthy young men, so in older men it can be more difficult to determine whether low testosterone is related to chronic opioid use or to older age.^5,16

As this was a retrospective review, additional limitations included the inability to measure subclinical OPIAD, and the data collection related to symptoms of hypogonadism was restricted by documentation in the CPRS progress notes. The lack of data for females does not contribute to the literature on OPIAD in women. Finally, as the total daily MED does not distinguish between short-acting and long-acting opioid therapy, no differences between the impacts of short-acting vs long-
acting opioid therapy on risk for 
hypogonadism can be inferred. There is evidence to suggest that long-acting opioids are associated with a significantly higher risk for OPIAD compared with short-acting opioids, although the mechanism behind this is not well established.^22,23

Conclusions

The average age of the patients 
on chronic opioid therapy with a 
testosterone level drawn in this cohort was 58.3 years, which is younger than originally anticipated. The median age of 60 years is not significantly different from the average age, indicating that outliers did not impact this calculation. On average, the TT level was normal before starting opioids. Once opioids were started, patients were treated for an average duration of 52.5 months with an average daily dose of 126.8 mg MED. In this veteran cohort, 48.6% of patients met the criteria for TRT based on TT level alone, which is within the reported prevalence range of opioid-induced hypogonadism already published.^4,9 These results are in line with the original hypothesis that chronic opioid use can adversely impact testosterone levels and can have a poor effect on a patient’s QOL due to symptoms of low testosterone. In addition to TRT, possible and suggested (but not proven) treatment options for OPIAD include discontinuation of opioid therapy, opioid rotation, or conversion to buprenorphine.²¹ The approach used should account for multiple patient-specific factors and should be individualized.

Based on the data, there is a trend toward lower testosterone levels in veterans treated with higher MED and for longer periods with chronic opioids. Given recent data that infer that TRT carries increased CV risk as well as the VHA’s Opioid Safety Initiative, it is imperative that providers closely evaluate the appropriateness of starting TRT and/or continuing chronic opioid therapy. All patients generally should have failed non-
opioid management prior to opioid therapy for chronic nonmalignant pain, and this should be documented accordingly. It is also crucial to have the “opioid talk” with patients from time to time and discuss the risks vs benefits, the potential for addiction, overdose, dependence, tolerance, constipation, and OPIAD so patients can continue to be an active and informed participants in their care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

According to the CDC, the medical use of opioid painkillers has increased at least 10-fold during the past 
20 years, “because of a movement toward more aggressive management of pain.”¹ Although opioid therapy is generally considered effective for the treatment of pain, long-term use (both orally and intrathecally) is associated with adverse effects (AEs) such as constipation, fatigue, nausea, sleep disturbances, depression, sexual dysfunction, and hypogonadism.^2,3Opioid-induced androgen deficiency (OPIAD), as defined by Smith and Elliot, is a clinical syndrome characterized by inappropriately low concentrations of gonadotropins (specifically, follicle-stimulating hormone [FSH] and luteinizing hormone [LH]), which leads to inadequate production of sex hormones, including estradiol and testosterone.⁴

Related: Testosterone Replacement Therapy: Playing Catch-up With Patients

The mechanism behind this phenomenon is initiated by either endogenous or exogenous opioids acting on opioid receptors in the hypothalamus, which causes a decrease in the release of gonadotropin-
releasing hormone (GnRH). This decrease in GnRH causes a reduction in the release of LH and FSH from the pituitary gland as well as testosterone or estradiol from the gonads.^4,5 Various guidelines report different cutoffs for the lower limit of normal total testosterone: The Endocrine Society recommends 300 ng/dL, the American Association of Clinical Endocrinologists suggests 200 ng/dL, and various other organizations suggest 230 ng/dL.^6-8 Hypotestosteronism can result in patients presenting with a broad spectrum of clinical symptoms, including reduced libido, erectile dysfunction (ED), fatigue, hot flashes, depression, anemia, decreased muscle mass, weight gain, and osteopenia or osteoporosis.⁴ Women with low testosterone levels can experience irregular menstrual periods, oligomenorrhea, or amenorrhea.⁹ Opioid-induced androgen deficiency often goes unrecognized and untreated. The reported prevalence of opioid-induced hypogonadism ranges from 21% to 86%.^4,9 Given the growing number of patients on chronic opioid therapy, OPIAD warrants further investigation to identify the prevalence in the veteran population to appropriately monitor and manage this deficiency.

The objective of this retrospective review was to identify the presence of secondary hypogonadism in chronic opioid users among a cohort of veterans receiving chronic opioids for nonmalignant pain. In addition to identifying the presence of secondary hypogonadism, the relationship between testosterone concentrations and total daily morphine equivalent doses (MEDs) was reviewed. These data along with the new information recently published on testosterone replacement therapy (TRT) and cardiovascular (CV) risk were then used to evaluate current practices at the West Palm Beach VAMC for OPIAD monitoring and management and to modify and update the local Criteria for Use (CFU) for TRT.

Methods

Patient data from the West Palm Beach VAMC in Florida from January 2013 to December 2013 were reviewed to identify patients who had a total testosterone (TT) level measured. All patient appointments for evaluation and treatment by the clinical pharmacy specialist in pain management were reviewed for data collection. This retrospective review was approved by the scientific advisory committee as part of the facility’s ongoing performance improvement efforts as defined by VHA Handbook 1058.05 and did not require written patient consent.¹⁰

Several distinct TT level data were collected. The descriptive data included patient age; gender; type of treated pain; testosterone level(s) drawn, including TT level before opioid therapy, TT level before/during/after TRT, and current total testosterone level; total daily MED of opioid therapy; duration of chronic opioid therapy; symptoms of exhibited hypogonadism; TRT formulation, dose, and duration; TRT prescriber; symptom change (if any); and laboratory tests ordered for TRT monitoring (lipid profile, liver profile, complete blood count, LH/FSH, and prostate specific antigen [PSA] panel).^5,11,12

Related: Combination Treatment Relieves Opioid-Induced Constipation

Daily MED of opioid therapy was calculated using the VA/DoD opioid conversion table for patients on oxycodone, hydromorphone, or hydrocodone.¹³ For those on the fentanyl patch or methadone, conversion factors of 1:2 (fentanyl [µg/h]:morphine [mg/d]) and 1:3 (methadone:morphine) were used to convert to the MED.¹⁴ For patients on the buprenorphine patch, the package insert was used to convert to the corresponding MED.¹⁵ Combination therapies used the applicable conversions to calculate the total daily MED.

Once the data were collected, descriptive statistics were used to analyze the data. In addition, 4 graphs were generated to review potential relationships. The correlation coefficient was calculated using the Alcula Online Statistics Calculator (http://www.alcula.com; Correlation Coefficient Calculator).

Results

A total of 316 unique veteran patients were seen by the clinical pharmacy specialist in pain management from January 1, 2013, through December 31, 2013. Of these, 73 patients (23.1%) had at least 1 TT level drawn in 2013. Three patients with testosterone levels drawn (4.1%) were excluded from the data analysis for the following reasons: 1 patient did not have testosterone levels on file before receiving testosterone replacement from a non-VA source, 
1 patient received opioids from a non-VA source (MED and duration of opioid therapy could not be calculated), and 1 patient inconsistently received opioids and MED used at the time of testosterone level draw. Per the local TRT CFU, a TT level > 350 ng/dL
does not require treatment, whereas levels < 230 ng/dL (with symptoms) may require TRT, and 
< 200 ng/dL should be treated as 
hypogonadal (interpretation based on local laboratory’s reference range for TT).16 Of the 70 patients included in the analysis, 34 (48.6%) had a TT level < 230 ng/dL and would be considered eligible for TRT if they presented with symptoms 
of low testosterone. Of these 34 patients with a low testosterone level, 28 (40%) were being treated or had been treated with TRT (Figure 1).

The average age of the male patients with a testosterone level drawn was 58.3 years, which was not significantly different from the calculated median age of 60 years. No female patients had a testosterone level drawn. On average, the TT level was normal before starting opioids (reference range per local laboratory: 175-781 ng/dL). Once opioids were initiated, patients were treated for an average duration of 52.5 months (calculated through December 2013) with an average daily dose of 126.8 MED (Table). Fifty of the 70 patients (71.4%) with testosterone levels drawn in 2013 received TRT. The most common symptoms reported by patients related to low testosterone included ED, decreased libido, depression, chronic fatigue, generalized weakness, and hot flashes or night sweats.

The average TT level prior to TRT was 145.3, and the average 
testosterone level after initiation of or 
during treatment with TRT was 292.4, which is within the normal TT level range. Most patients receiving TRT were treated with testosterone cypionate injections, and this was also the formulation used for the longest periods, likely due to the local CFU. In addition to testosterone cypionate injections, patients were also treated with testosterone enanthate injections, testosterone patches, and testosterone gel.

Figure 1 compares current testosterone level and testosterone level before TRT with total daily MEDs. Figure 2 compares current testosterone level and testosterone level before TRT with length of opioid therapy. The 2 figures use data from all patients included in the analysis and indicate a potential inverse relationship between the total daily MED and duration of therapy with the testosterone level, although none of the calculated correlation coefficients indicate that a strong relationship was present.

Figures 3 and 4 include data only for patients who had both a testosterone level collected before opioids (baseline testosterone level) and a current testosterone level. Figure 3 trends the data using total daily MED, and Figure 4 uses the duration of opioid therapy. The correlation for Figure 4 is slightly stronger; the strongest negative correlations were identified between total daily MED and testosterone level before opioid therapy (r = -0.273) and duration of opioid therapy and testosterone level prior to opioid therapy (r = -0.396). The trends indicate that most patients had a normal TT level before opioid treatment and that patients treated with higher MEDs and for longer durations of time were more likely to have lower total testosterone levels.

Discussion

Low testosterone levels can adversely affect patients’ quality of life (QOL) and add to patients’ medication burden with the initiation of TRT. Given new data analyzing the potential effects of TRT on CV event risk, the use of TRT should be carefully considered, as it may carry significant risks and may not be suitable for all patients.

In November 2013, a study was published regarding TRT and 
increased CV risk.¹⁷ This was a retrospective cohort study of men with low testosterone levels (< 300 ng/dL) who had undergone coronary angiography in the VA system between 2005 and 2011 (average age in testosterone group was 60.6 years). The results were significant for an absolute rate of events (all-cause mortality, myocardial infarction [MI], and ischemic stroke) of 19.9% in the no testosterone group and 25.7% in the TRT group, an absolute risk difference of 5.8% at 3 years after coronary angiography. Kaplan-Meier survival curves demonstrated that testosterone use was associated with increased risk of death, MI, and stroke. This result was unchanged when adjusted for the presence of coronary artery disease (CAD). In 
addition, no significant difference was found between the groups in terms of systolic blood pressure, low-
density lipoprotein cholesterol level, or in the use of beta-blocker and statin medications. What is important to note is that in this cohort, 20% had a prior history of MI and heart failure, and more than 50% had confirmed obstructive CAD on angiography. In addition, as this was an observational study, confounding or bias may exist, and given the study population, 
generalizability may be limited to a veteran population.

Related: A Multidisciplinary Chronic Pain Management Clinic in an Indian Health Service Facility

Another retrospective cohort study assessed the risk of acute nonfatal MI following an initial TRT prescription in a large health care database (average age based on TRT prescription was 54.4 years).¹⁸ In men aged ≥ 65 years, a 2-fold increase in the risk of MI in the immediate 90 days after filling an initial TRT prescription declined to baseline after 91 to 180 days among those who did not refill their prescription. Younger men with a history of heart disease had a 2- to 3-fold increased risk of MI in the 90 days following initial TRT prescription. No excess risk was observed in the younger men without such a history. Again, this study has its limitations related to the retrospective design and use of a health care database as opposed to a randomized controlled trial.

In February 2014, a VA National Pharmacy Benefits Management (PBM) bulletin addressed 2 recent studies that had identified a possible risk of increased CV events in men receiving TRT. The bulletin noted that these studies had prompted the FDA to reassess the CV safety of TRT.¹⁹ The TRT CFU was updated by VISN 8 to ensure that the patients receive appropriate treatment and are monitored accordingly.

One of the major changes to the CFU was defining the reference ranges for TRT (interpretation based on a local laboratory’s reference range for total testosterone): serum TT < 200 ng/dL be “treated as 
hypogonadal, those with TT 
> 400 ng/dL be considered normal and those with TT 200-400 ng/dL be treated based on their clinical presentation if symptomatic; TT levels 
> 350 ng/dL do not require treatment, and levels below 230 ng/dL (with symptoms) may require testosterone replacement therapy.”¹⁶ Other important updates included revision of the exclusion criteria as well as 
highlighting special considerations related to TRT, including the use of free testosterone levels rather than TT levels in patients with suspected protein-binding issues, role in fertility treatments, limited use in patients on spironolactone therapy (due to spironolactone’s anti-androgen effects), and potential association with mood and behavior.¹⁶

As chronic opioid therapy is associated with OPIAD, the renewed interest in TRT and its potential AEs provides yet another reason to reconsider opioid therapy. This is especially valid when opioids are the potential cause of hypogonadism and the reaction is treating the AEs of opioids (as opposed to considering elimination of the causative agent) with a therapy that can potentially increase the risk for CV events so that opioids can be continued. Outside the potential CV risk with TRT, opioids carry the innate risk for substance abuse and addiction.

The Opioid Safety Initiative Requirements was released as a memorandum in April 2014 and is the VHA’s effort to “reduce harm from unsafe medications and/or excessive doses while adequately controlling pain in Veterans.”²⁰ Although it does not discuss the risk of OPIAD, it does highlight the need to identify and mitigate high-risk patients as well as high-risk opioid regimens. All these factors, including the possibility of hypogonadism, should be considered before starting opioid therapy and at the time of opioid renewal, as it is known that opioid therapy is not without risks.

At the West Palm Beach VAMC, the primary care providers (PCPs) are responsible for the management of TRT, including the workup, renewal, and monitoring. The Chronic Nonmalignant Pain Management Clinic (CNMPMC) orders testosterone levels on patients who report symptoms of low testosterone, such as hot flashes, depression, and low energy level and refers them to their PCP as indicated. The authors believe that this is most appropriate for a number of reasons: (1) the CNMPMC is a consult service, and patients are not followed indefinitely; (2) patients should be fully evaluated for appropriateness of TRT (including assessment of CV risk) before starting therapy; and (3) the necessary monitoring parameters (laboratory testing, digital rectal exam, and osteoporosis screening) are not typically within the VA pain clinic provider’s scope of practice or expertise. A consideration for future practice would be to incorporate the use of a standardized questionnaire for OPIAD monitoring in patients receiving ≥ 100 mg of morphine daily (eg, the Aging Males’ Symptoms scale).²¹ It should, however, be at the forefront of the pain specialist’s and PCP’s minds that all patients on chronic opioid therapy or considering chronic opioid therapy should be counseled on the risk for OPIAD. If OPIAD is identified, the patient should be carefully considered for an opioid dose reduction as an initial management strategy.

Limitations

A limitation of this review is the lack of consistency or adequacy of serum testosterone sampling, noting that valid testosterone levels need to be drawn in the morning and not obtained during a time of acute illness. In addition, testosterone levels need to be drawn at an appropriate interval while on TRT (eg, at the midpoint between testosterone injections).¹⁶ Although the time of the sample collection is documented in the Computerized Patient Record System (CPRS), it is unknown whether the patient was acutely ill on the day of the sampling unless a progress note is entered, and it is difficult to determine whether the level timing was accurate based on the testosterone replacement formulation. Another limitation is that the average decline in serum testosterone levels with aging in men is 1% to 2% per year. A significant fraction of older men have levels below the lower limit of the normal range for healthy young men, so in older men it can be more difficult to determine whether low testosterone is related to chronic opioid use or to older age.^5,16

As this was a retrospective review, additional limitations included the inability to measure subclinical OPIAD, and the data collection related to symptoms of hypogonadism was restricted by documentation in the CPRS progress notes. The lack of data for females does not contribute to the literature on OPIAD in women. Finally, as the total daily MED does not distinguish between short-acting and long-acting opioid therapy, no differences between the impacts of short-acting vs long-
acting opioid therapy on risk for 
hypogonadism can be inferred. There is evidence to suggest that long-acting opioids are associated with a significantly higher risk for OPIAD compared with short-acting opioids, although the mechanism behind this is not well established.^22,23

Conclusions

The average age of the patients 
on chronic opioid therapy with a 
testosterone level drawn in this cohort was 58.3 years, which is younger than originally anticipated. The median age of 60 years is not significantly different from the average age, indicating that outliers did not impact this calculation. On average, the TT level was normal before starting opioids. Once opioids were started, patients were treated for an average duration of 52.5 months with an average daily dose of 126.8 mg MED. In this veteran cohort, 48.6% of patients met the criteria for TRT based on TT level alone, which is within the reported prevalence range of opioid-induced hypogonadism already published.^4,9 These results are in line with the original hypothesis that chronic opioid use can adversely impact testosterone levels and can have a poor effect on a patient’s QOL due to symptoms of low testosterone. In addition to TRT, possible and suggested (but not proven) treatment options for OPIAD include discontinuation of opioid therapy, opioid rotation, or conversion to buprenorphine.²¹ The approach used should account for multiple patient-specific factors and should be individualized.

Based on the data, there is a trend toward lower testosterone levels in veterans treated with higher MED and for longer periods with chronic opioids. Given recent data that infer that TRT carries increased CV risk as well as the VHA’s Opioid Safety Initiative, it is imperative that providers closely evaluate the appropriateness of starting TRT and/or continuing chronic opioid therapy. All patients generally should have failed non-
opioid management prior to opioid therapy for chronic nonmalignant pain, and this should be documented accordingly. It is also crucial to have the “opioid talk” with patients from time to time and discuss the risks vs benefits, the potential for addiction, overdose, dependence, tolerance, constipation, and OPIAD so patients can continue to be an active and informed participants in their care.

Author disclosures
The authors report no actual or potential conflicts of interest with regard to this article.

Disclaimer
The opinions expressed herein are those of the authors and do not necessarily reflect those of Federal Practitioner, Frontline Medical Communications Inc., the U.S. Government, or any of its agencies. This article may discuss unlabeled or investigational use of certain drugs. Please review the complete prescribing information for specific drugs or drug combinations—including indications, contraindications, warnings, and adverse effects—before administering pharmacologic therapy to patients.

Case

A 64-year-old man presented to the ED seeking assistance in withdrawing from his prescription of oxycodone, which he had been taking to manage chronic pain due to rheumatoid arthritis. He stated that he had discontinued the oxycodone approximately 4 days prior to presentation and over the past 3 days, had been experiencing abdominal pain, nausea, vomiting, diarrhea, and diaphoresis. He noted that his symptoms were identical to those he had during previous unsuccessful attempts to wean himself from the narcotic medication. He denied any fever, dysuria, penile discharge, or any skin changes. Further evaluation revealed a remote history of a cholecystectomy and an appendectomy.

During the initial examination, the patient appeared uncomfortable but in no acute distress. His vital signs were: heart rate, 110 beats/minute; blood pressure, 107/76 mm Hg, respiratory rate, 22 breaths/minute; and temperature, 99˚F. Oxygen saturation was 97% on room air. The abdominal examination revealed moderate diffuse tenderness, mild distension without guarding or rebound, and some well-healed surgical scars.

In light of the abnormal sonographic findings, a computed tomography (CT) scan with intravenous (IV) contrast was performed, which confirmed the diagnosis of a distal small bowel obstruction (SBO). Surgery services were consulted. As the patient’s current symptoms were believed to be the result of an SBO and not from narcotic withdrawal, surgery services instructed nothing by mouth and elected nonsurgical management. They placed a nasogastric tube and administered fluids and analgesics via IV. The patient was discharged 4 days after presentation to the ED, with complete resolution of his symptoms.

Discussion

Annually in the United States, less than 1% of all patients presenting to EDs are subsequently diagnosed with SBO.¹ However, this disease comprises 15% of all surgical hospital admissions, costing upward of $1 billion in annual hospital charges.² Moreover, patients with SBO suffer from a disproportionately high morbidity (eg, bowel strangulation, necrosis) and mortality than the general population,^3-5 and delayed diagnosis is associated with a higher risk of bowel resection. One study by Bickell et al⁶ showed that only 4% of patients appropriately managed less than 24 hours after symptom onset required resection compared with 10% to 14% of patients managed more than 24 hours after symptom onset.

As most patients diagnosed with SBO are first seen in the ED, emergency physicians (EPs) have a distinctive role in lowering the likelihood of a poor outcome by making this diagnosis early.⁷ Multiple methods of diagnosing SBO are at the disposal of the clinician, including the history and physical examination, abdominal X-ray, ultrasound, CT, and magnetic resonance imaging (MRI).

The history and physical examination can be rapidly performed at the bedside in patients with suspected SBO. The factors typically associated with SBO include constipation, a previous history of abdominal surgeries, abnormal bowel sounds, and abdominal distension.³ However, these findings are not sufficient to accurately and adequately rule in or rule out disease.^3,8,9

Diagnostic Imaging

While patient history and physical examination may be helpful in diagnosing SBO, imaging plays a critical role in the definitive diagnosis. The imaging modality that is the de facto gold standard for diagnosis is the CT scan.¹⁰  A meta-analysis by Taylor et al,³ which included 64-slice multidetector CT imaging studies (using both oral and IV contrast), demonstrated sensitivities of 93% to 96% and specificities of 93% to 100% in diagnosing SBO.

In patients in whom CT is contraindicated, MRI can be a useful alternative, with studies showing a similar diagnostic accuracy to 64-slice CT.^13,14 Both CT and MRI are highly accurate in diagnosing SBO; however, there are disadvantages to their use. Such disadvantages include the inability to perform these studies at bedside; the length of time to perform these studies; the higher cost compared to other modalities; and, in CT, the adverse side effects of radiation and possible contrast reactions.

Bedside Ultrasound

Abdominal X-ray traditionally has been the initial choice in bedside imaging for SBO; however, a recent meta-analysis found this modality to have a summary sensitivity of 75%, specificity of 66%, positive likelihood ratio of 1.6, and negative likelihood ratio of 0.43 in diagnosing SBO.³ Based on these statistics, bedside ultrasound has recently ascended as a viable alternative to abdominal X-ray.

Although there is limited research regarding the accuracy of ultrasound to evaluate SBO, initial study results are encouraging. The previously cited meta-analysis by Taylor et al³ identified six ultrasound studies, two of which were performed in the ED. In one of these two studies, Unlüer et al¹⁰ performed a prospective study that enrolled 174 patients in the ED, 90 of whom were subsequently found to have an SBO. In addition, Unlüer et al’s study found that relatively inexperienced emergency medicine (EM) residents were able to use bedside ultrasound in the diagnosis of SBO with a sensitivity of 97.7% and a specificity of 92.7%.

Another ED study by Jang et al¹⁵ enrolled 76 patients, 33 of whom were diagnosed with SBO using CT. In this study, the authors found ultrasound to have a 91% sensitivity and 84% specificity for dilated bowel, and a specificity of 98% and sensitivity of 27% for decreased peristalsis.¹⁵ Imaging in this study was performed by EM residents, who received only 10 minutes of didactic lecture.  

The criteria used in the abovementioned studies varied slightly. The study by Jang et al¹⁵ used either fluid-filled dilated bowel >2.5 cm or decreased/absent forward bowel peristalsis,  while the study by Unlüer et al¹⁰ defined sonographic SBO as two of the three following criteria: greater than 3 dilated loops of either jejunum (>25 mm), or of ileum (>15 mm), increased peristalsis or a collapsed colonic lumen. In cases of higher-grade obstruction, the Tanga sign, fluid seen outside of the dilated loops of bowel, has also been reported (Figure 2).¹⁶

Conclusion

There are several distinct advantages to using bedside ultrasound in cases of suspected SBO, including its lack of ionizing radiation, the ability to perform the scan rapidly, and the high accuracy rate in detecting this condition—even in the hands of providers with minimal training. In addition to its cost-effectiveness, ultrasound may be preferred in patients with relative contraindications to CT, such as pregnant patients and patients with contrast allergies. Even in patients in whom there is no contraindication to CT, ultrasound may be used to safely and quickly identify and risk-stratify those who require further imaging versus those who can be safely discharged home—or possibly even finding alternative diagnoses of acute abdominal pain (eg, acute cholecystitis, ureterolithiasis, abdominal aortic aneurysm).

Dr Avila is an attending physician and ultrasound fellow in the department of emergency medicine at the University of Kentucky, Lexington. Dr Smith is the director of emergency ultrasound in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga. Dr Whittle is the director of research in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga.

Case

A 64-year-old man presented to the ED seeking assistance in withdrawing from his prescription of oxycodone, which he had been taking to manage chronic pain due to rheumatoid arthritis. He stated that he had discontinued the oxycodone approximately 4 days prior to presentation and over the past 3 days, had been experiencing abdominal pain, nausea, vomiting, diarrhea, and diaphoresis. He noted that his symptoms were identical to those he had during previous unsuccessful attempts to wean himself from the narcotic medication. He denied any fever, dysuria, penile discharge, or any skin changes. Further evaluation revealed a remote history of a cholecystectomy and an appendectomy.

During the initial examination, the patient appeared uncomfortable but in no acute distress. His vital signs were: heart rate, 110 beats/minute; blood pressure, 107/76 mm Hg, respiratory rate, 22 breaths/minute; and temperature, 99˚F. Oxygen saturation was 97% on room air. The abdominal examination revealed moderate diffuse tenderness, mild distension without guarding or rebound, and some well-healed surgical scars.

In light of the abnormal sonographic findings, a computed tomography (CT) scan with intravenous (IV) contrast was performed, which confirmed the diagnosis of a distal small bowel obstruction (SBO). Surgery services were consulted. As the patient’s current symptoms were believed to be the result of an SBO and not from narcotic withdrawal, surgery services instructed nothing by mouth and elected nonsurgical management. They placed a nasogastric tube and administered fluids and analgesics via IV. The patient was discharged 4 days after presentation to the ED, with complete resolution of his symptoms.

Discussion

Annually in the United States, less than 1% of all patients presenting to EDs are subsequently diagnosed with SBO.¹ However, this disease comprises 15% of all surgical hospital admissions, costing upward of $1 billion in annual hospital charges.² Moreover, patients with SBO suffer from a disproportionately high morbidity (eg, bowel strangulation, necrosis) and mortality than the general population,^3-5 and delayed diagnosis is associated with a higher risk of bowel resection. One study by Bickell et al⁶ showed that only 4% of patients appropriately managed less than 24 hours after symptom onset required resection compared with 10% to 14% of patients managed more than 24 hours after symptom onset.

As most patients diagnosed with SBO are first seen in the ED, emergency physicians (EPs) have a distinctive role in lowering the likelihood of a poor outcome by making this diagnosis early.⁷ Multiple methods of diagnosing SBO are at the disposal of the clinician, including the history and physical examination, abdominal X-ray, ultrasound, CT, and magnetic resonance imaging (MRI).

The history and physical examination can be rapidly performed at the bedside in patients with suspected SBO. The factors typically associated with SBO include constipation, a previous history of abdominal surgeries, abnormal bowel sounds, and abdominal distension.³ However, these findings are not sufficient to accurately and adequately rule in or rule out disease.^3,8,9

Diagnostic Imaging

While patient history and physical examination may be helpful in diagnosing SBO, imaging plays a critical role in the definitive diagnosis. The imaging modality that is the de facto gold standard for diagnosis is the CT scan.¹⁰  A meta-analysis by Taylor et al,³ which included 64-slice multidetector CT imaging studies (using both oral and IV contrast), demonstrated sensitivities of 93% to 96% and specificities of 93% to 100% in diagnosing SBO.

In patients in whom CT is contraindicated, MRI can be a useful alternative, with studies showing a similar diagnostic accuracy to 64-slice CT.^13,14 Both CT and MRI are highly accurate in diagnosing SBO; however, there are disadvantages to their use. Such disadvantages include the inability to perform these studies at bedside; the length of time to perform these studies; the higher cost compared to other modalities; and, in CT, the adverse side effects of radiation and possible contrast reactions.

Bedside Ultrasound

Abdominal X-ray traditionally has been the initial choice in bedside imaging for SBO; however, a recent meta-analysis found this modality to have a summary sensitivity of 75%, specificity of 66%, positive likelihood ratio of 1.6, and negative likelihood ratio of 0.43 in diagnosing SBO.³ Based on these statistics, bedside ultrasound has recently ascended as a viable alternative to abdominal X-ray.

Although there is limited research regarding the accuracy of ultrasound to evaluate SBO, initial study results are encouraging. The previously cited meta-analysis by Taylor et al³ identified six ultrasound studies, two of which were performed in the ED. In one of these two studies, Unlüer et al¹⁰ performed a prospective study that enrolled 174 patients in the ED, 90 of whom were subsequently found to have an SBO. In addition, Unlüer et al’s study found that relatively inexperienced emergency medicine (EM) residents were able to use bedside ultrasound in the diagnosis of SBO with a sensitivity of 97.7% and a specificity of 92.7%.

Another ED study by Jang et al¹⁵ enrolled 76 patients, 33 of whom were diagnosed with SBO using CT. In this study, the authors found ultrasound to have a 91% sensitivity and 84% specificity for dilated bowel, and a specificity of 98% and sensitivity of 27% for decreased peristalsis.¹⁵ Imaging in this study was performed by EM residents, who received only 10 minutes of didactic lecture.  

The criteria used in the abovementioned studies varied slightly. The study by Jang et al¹⁵ used either fluid-filled dilated bowel >2.5 cm or decreased/absent forward bowel peristalsis,  while the study by Unlüer et al¹⁰ defined sonographic SBO as two of the three following criteria: greater than 3 dilated loops of either jejunum (>25 mm), or of ileum (>15 mm), increased peristalsis or a collapsed colonic lumen. In cases of higher-grade obstruction, the Tanga sign, fluid seen outside of the dilated loops of bowel, has also been reported (Figure 2).¹⁶

Conclusion

There are several distinct advantages to using bedside ultrasound in cases of suspected SBO, including its lack of ionizing radiation, the ability to perform the scan rapidly, and the high accuracy rate in detecting this condition—even in the hands of providers with minimal training. In addition to its cost-effectiveness, ultrasound may be preferred in patients with relative contraindications to CT, such as pregnant patients and patients with contrast allergies. Even in patients in whom there is no contraindication to CT, ultrasound may be used to safely and quickly identify and risk-stratify those who require further imaging versus those who can be safely discharged home—or possibly even finding alternative diagnoses of acute abdominal pain (eg, acute cholecystitis, ureterolithiasis, abdominal aortic aneurysm).

Dr Avila is an attending physician and ultrasound fellow in the department of emergency medicine at the University of Kentucky, Lexington. Dr Smith is the director of emergency ultrasound in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga. Dr Whittle is the director of research in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga.

Case

A 64-year-old man presented to the ED seeking assistance in withdrawing from his prescription of oxycodone, which he had been taking to manage chronic pain due to rheumatoid arthritis. He stated that he had discontinued the oxycodone approximately 4 days prior to presentation and over the past 3 days, had been experiencing abdominal pain, nausea, vomiting, diarrhea, and diaphoresis. He noted that his symptoms were identical to those he had during previous unsuccessful attempts to wean himself from the narcotic medication. He denied any fever, dysuria, penile discharge, or any skin changes. Further evaluation revealed a remote history of a cholecystectomy and an appendectomy.

During the initial examination, the patient appeared uncomfortable but in no acute distress. His vital signs were: heart rate, 110 beats/minute; blood pressure, 107/76 mm Hg, respiratory rate, 22 breaths/minute; and temperature, 99˚F. Oxygen saturation was 97% on room air. The abdominal examination revealed moderate diffuse tenderness, mild distension without guarding or rebound, and some well-healed surgical scars.

In light of the abnormal sonographic findings, a computed tomography (CT) scan with intravenous (IV) contrast was performed, which confirmed the diagnosis of a distal small bowel obstruction (SBO). Surgery services were consulted. As the patient’s current symptoms were believed to be the result of an SBO and not from narcotic withdrawal, surgery services instructed nothing by mouth and elected nonsurgical management. They placed a nasogastric tube and administered fluids and analgesics via IV. The patient was discharged 4 days after presentation to the ED, with complete resolution of his symptoms.

Discussion

Annually in the United States, less than 1% of all patients presenting to EDs are subsequently diagnosed with SBO.¹ However, this disease comprises 15% of all surgical hospital admissions, costing upward of $1 billion in annual hospital charges.² Moreover, patients with SBO suffer from a disproportionately high morbidity (eg, bowel strangulation, necrosis) and mortality than the general population,^3-5 and delayed diagnosis is associated with a higher risk of bowel resection. One study by Bickell et al⁶ showed that only 4% of patients appropriately managed less than 24 hours after symptom onset required resection compared with 10% to 14% of patients managed more than 24 hours after symptom onset.

As most patients diagnosed with SBO are first seen in the ED, emergency physicians (EPs) have a distinctive role in lowering the likelihood of a poor outcome by making this diagnosis early.⁷ Multiple methods of diagnosing SBO are at the disposal of the clinician, including the history and physical examination, abdominal X-ray, ultrasound, CT, and magnetic resonance imaging (MRI).

The history and physical examination can be rapidly performed at the bedside in patients with suspected SBO. The factors typically associated with SBO include constipation, a previous history of abdominal surgeries, abnormal bowel sounds, and abdominal distension.³ However, these findings are not sufficient to accurately and adequately rule in or rule out disease.^3,8,9

Diagnostic Imaging

While patient history and physical examination may be helpful in diagnosing SBO, imaging plays a critical role in the definitive diagnosis. The imaging modality that is the de facto gold standard for diagnosis is the CT scan.¹⁰  A meta-analysis by Taylor et al,³ which included 64-slice multidetector CT imaging studies (using both oral and IV contrast), demonstrated sensitivities of 93% to 96% and specificities of 93% to 100% in diagnosing SBO.

In patients in whom CT is contraindicated, MRI can be a useful alternative, with studies showing a similar diagnostic accuracy to 64-slice CT.^13,14 Both CT and MRI are highly accurate in diagnosing SBO; however, there are disadvantages to their use. Such disadvantages include the inability to perform these studies at bedside; the length of time to perform these studies; the higher cost compared to other modalities; and, in CT, the adverse side effects of radiation and possible contrast reactions.

Bedside Ultrasound

Abdominal X-ray traditionally has been the initial choice in bedside imaging for SBO; however, a recent meta-analysis found this modality to have a summary sensitivity of 75%, specificity of 66%, positive likelihood ratio of 1.6, and negative likelihood ratio of 0.43 in diagnosing SBO.³ Based on these statistics, bedside ultrasound has recently ascended as a viable alternative to abdominal X-ray.

Although there is limited research regarding the accuracy of ultrasound to evaluate SBO, initial study results are encouraging. The previously cited meta-analysis by Taylor et al³ identified six ultrasound studies, two of which were performed in the ED. In one of these two studies, Unlüer et al¹⁰ performed a prospective study that enrolled 174 patients in the ED, 90 of whom were subsequently found to have an SBO. In addition, Unlüer et al’s study found that relatively inexperienced emergency medicine (EM) residents were able to use bedside ultrasound in the diagnosis of SBO with a sensitivity of 97.7% and a specificity of 92.7%.

Another ED study by Jang et al¹⁵ enrolled 76 patients, 33 of whom were diagnosed with SBO using CT. In this study, the authors found ultrasound to have a 91% sensitivity and 84% specificity for dilated bowel, and a specificity of 98% and sensitivity of 27% for decreased peristalsis.¹⁵ Imaging in this study was performed by EM residents, who received only 10 minutes of didactic lecture.  

The criteria used in the abovementioned studies varied slightly. The study by Jang et al¹⁵ used either fluid-filled dilated bowel >2.5 cm or decreased/absent forward bowel peristalsis,  while the study by Unlüer et al¹⁰ defined sonographic SBO as two of the three following criteria: greater than 3 dilated loops of either jejunum (>25 mm), or of ileum (>15 mm), increased peristalsis or a collapsed colonic lumen. In cases of higher-grade obstruction, the Tanga sign, fluid seen outside of the dilated loops of bowel, has also been reported (Figure 2).¹⁶

Conclusion

There are several distinct advantages to using bedside ultrasound in cases of suspected SBO, including its lack of ionizing radiation, the ability to perform the scan rapidly, and the high accuracy rate in detecting this condition—even in the hands of providers with minimal training. In addition to its cost-effectiveness, ultrasound may be preferred in patients with relative contraindications to CT, such as pregnant patients and patients with contrast allergies. Even in patients in whom there is no contraindication to CT, ultrasound may be used to safely and quickly identify and risk-stratify those who require further imaging versus those who can be safely discharged home—or possibly even finding alternative diagnoses of acute abdominal pain (eg, acute cholecystitis, ureterolithiasis, abdominal aortic aneurysm).

Dr Avila is an attending physician and ultrasound fellow in the department of emergency medicine at the University of Kentucky, Lexington. Dr Smith is the director of emergency ultrasound in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga. Dr Whittle is the director of research in the department of emergency medicine at the University of Tennessee College of Medicine Chattanooga.