Original Research

Developing a high-efficiency operating room (OR) is both a challenging and rewarding goal for any healthcare system. The OR is traditionally a high-cost/high-revenue environment¹ and operative efficacy has been correlated with low complication rates and surgical success.² An efficient OR is one that maximizes utilization while providing safe, reproducible, cost-effective, high-quality care. Total joint arthroplasty (TJA) has occupied the center stage for OR efficiency research, in part due to increasing demands from our aging population³ and economic pressures related to high implant costs, decreased reimbursement, and competition for market shares when OR time and space are limited.

A PubMed search on OR efficiency in TJA shows a disproportionately high focus on surgical technique, such as use of patient-specific instrumentation (PSI), computer-assisted surgery (CAS), minimally invasive surgery, and closure with barbed suture. In a retrospective review of 352 TKA patients who had PSI vs conventional instrumentation, DeHaan and colleagues⁴ found that PSI was associated with significantly decreased operative and room turnover times (20.4 minutes and 6.4 minutes, respectively). In another prospective multicenter study, Mont and colleagues⁵ showed a reduction in surgical time by 8.90 min for navigated total knee arthroplasty (TKA) performed with single-use instruments, cutting blocks, and trials. Other investigators compared PSI to CAS in TKA and found PSI to be 1.45 times more profitable than CAS, with 3 PSI cases performed in an 8-hour OR day compared to 2 CAS cases.⁶

There is no question that improved surgical technique can enhance OR efficiency. However, this model, while promising, is difficult to implement on a wide scale due to surgeon preferences, vendor limitations, and added costs related to the advanced preoperative imaging studies, manufacturing of the custom guides, and maintenance of navigation equipment. In addition, while interventions such as the use of barbed suture have the potential for speeding closure time, the time saved (4.7 minutes in one randomized trial)⁷ may not be enough to affect major utilization differences per OR per day. These technologies are also frequently employed by high-volume surgeons with high-volume teams and institutions.

Ideally, we need investment in the human capital and a collective change in work cultures to produce high-quality, well-choreographed, easily reproducible routines. An efficient OR requires the synchronous involvement of a large team of individuals, including hospital administrators, surgery schedulers, surgeons, anesthesiologists, preoperative holding area staff, OR nurses, surgical attendants, sterile processing personnel, and recovery room nurses. Case schedulers should match allocated block time with time required for surgery based on the historical performance of the individual surgeon, preferably scheduling similar cases on the same day. Preoperative work-up and medical clearance should be completed prior to scheduling to avoid last-minute cancellations. Patient reminders and accommodations for those traveling from long distances can further minimize late arrivals. Prompt initiation of the perioperative clinical pathway upon a patient’s check-in is important. The surgical site should be marked and the anesthesia plan confirmed upon arrival in the preoperative holding area. Necessary products need to be ready and/or administrated in time for transfer to the OR. These include prophylactic antibiotics, coagulation factors (eg, tranexamic acid), and blood products as indicated. Spinal anesthesia, regional nerve blocks, and intravenous (IV) lines should be completed before transfer to the OR. A “block room” close to the OR can allow concurrent induction of anesthesia and has been shown to increase the number of surgical cases performed during a regular workday.⁸ Hair clipping within the surgical site and pre-scrubbing of the operative extremity should also be performed prior to transfer to the OR in order to minimize micro-organisms and dispersal of loose hair onto the sterile field.

Upon arrival of the patient to the OR, instrument tables based on the surgeon preference cards should be opened, instrument count and implant templating completed, necessary imaging displayed, and OR staff ready with specific responsibilities assigned to each member. Small and colleagues⁹ showed that using dedicated orthopedic staff familiar with the surgical routine decreased operative time by 19 minutes per procedure, or 1.25 hours for a surgeon performing 4 primary TJAs per day. Practices such as routine placement of a urinary catheter should be seriously scrutinized. In a randomized prospective study of patients undergoing total hip arthroplasty under spinal anesthesia, Miller and colleagues¹⁰ found no benefit for indwelling catheters in preventing urinary retention. In another randomized prospective study, Huang and colleagues¹¹ found the prevalence of urinary tract infections was significantly higher in TJA patients who received indwelling urinary catheters.

A scrub nurse familiar with the instruments, their assembly, and the sequence of events can ensure efficient surgical flow. The scrub nurse needs to anticipate missing or defective tools and call for them, ideally before the incision is made. Direct comparison studies are needed to assess the efficacy of routine intraoperative imaging vs commercially available universal cup alignment guides or clinical examinations in determining acceptable component positioning and limb length. Following component implantation and before wound closure, the circulating nurse should initiate the process of acquisition of a recovery room bed, make sure dressing supplies and necessary equipment are available, and call for surgical attendants. Lack of surgical attendants, delayed transfer from the OR table to hospital bed, and prolonged acquisition of a recovery room bed have been identified as major OR inefficiencies in a retrospective study by Attarian and colleagues.¹²

In summary, time is the OR’s most valuable resource.¹³ We believe that a consistent, almost automated attitude to the above procedures decreases variability and improves efficiency. By providing clear communication of the surgical needs with the team, having consistent anesthesia and nursing staff, implementing consistent perioperative protocols, and insuring that all necessary instruments and modalities are available prior to starting the procedure, we were able to sustainably increase OR throughput in a large teaching hospital.^9,14 This process, however, requires constant review to identify and eliminate new gaps, with each member of the team sharing a frank desire to improve. In this regard, hospital administrators share the duty to facilitate the implementation of any necessary changes, allocation of needed resources, and rewarding good effort, which could ultimately increase staff satisfaction and retention. Because efficiency is the ratio of benefits (eg, revenue, safety, etc.) to investment (eg, implant costs, wages, etc.), raises the question: what would be the effect of transitioning from hourly-wage to a salary-based system for key support staff? Unlike hourly-wage personnel, who have no incentive for productivity, a salaried employee assigned to a high-efficiency OR will inherently strive for improvement, employing higher organizational skills to accomplish a common goal. To our knowledge, there is no published data on this topic.

Developing a high-efficiency operating room (OR) is both a challenging and rewarding goal for any healthcare system. The OR is traditionally a high-cost/high-revenue environment¹ and operative efficacy has been correlated with low complication rates and surgical success.² An efficient OR is one that maximizes utilization while providing safe, reproducible, cost-effective, high-quality care. Total joint arthroplasty (TJA) has occupied the center stage for OR efficiency research, in part due to increasing demands from our aging population³ and economic pressures related to high implant costs, decreased reimbursement, and competition for market shares when OR time and space are limited.

A PubMed search on OR efficiency in TJA shows a disproportionately high focus on surgical technique, such as use of patient-specific instrumentation (PSI), computer-assisted surgery (CAS), minimally invasive surgery, and closure with barbed suture. In a retrospective review of 352 TKA patients who had PSI vs conventional instrumentation, DeHaan and colleagues⁴ found that PSI was associated with significantly decreased operative and room turnover times (20.4 minutes and 6.4 minutes, respectively). In another prospective multicenter study, Mont and colleagues⁵ showed a reduction in surgical time by 8.90 min for navigated total knee arthroplasty (TKA) performed with single-use instruments, cutting blocks, and trials. Other investigators compared PSI to CAS in TKA and found PSI to be 1.45 times more profitable than CAS, with 3 PSI cases performed in an 8-hour OR day compared to 2 CAS cases.⁶

There is no question that improved surgical technique can enhance OR efficiency. However, this model, while promising, is difficult to implement on a wide scale due to surgeon preferences, vendor limitations, and added costs related to the advanced preoperative imaging studies, manufacturing of the custom guides, and maintenance of navigation equipment. In addition, while interventions such as the use of barbed suture have the potential for speeding closure time, the time saved (4.7 minutes in one randomized trial)⁷ may not be enough to affect major utilization differences per OR per day. These technologies are also frequently employed by high-volume surgeons with high-volume teams and institutions.

Ideally, we need investment in the human capital and a collective change in work cultures to produce high-quality, well-choreographed, easily reproducible routines. An efficient OR requires the synchronous involvement of a large team of individuals, including hospital administrators, surgery schedulers, surgeons, anesthesiologists, preoperative holding area staff, OR nurses, surgical attendants, sterile processing personnel, and recovery room nurses. Case schedulers should match allocated block time with time required for surgery based on the historical performance of the individual surgeon, preferably scheduling similar cases on the same day. Preoperative work-up and medical clearance should be completed prior to scheduling to avoid last-minute cancellations. Patient reminders and accommodations for those traveling from long distances can further minimize late arrivals. Prompt initiation of the perioperative clinical pathway upon a patient’s check-in is important. The surgical site should be marked and the anesthesia plan confirmed upon arrival in the preoperative holding area. Necessary products need to be ready and/or administrated in time for transfer to the OR. These include prophylactic antibiotics, coagulation factors (eg, tranexamic acid), and blood products as indicated. Spinal anesthesia, regional nerve blocks, and intravenous (IV) lines should be completed before transfer to the OR. A “block room” close to the OR can allow concurrent induction of anesthesia and has been shown to increase the number of surgical cases performed during a regular workday.⁸ Hair clipping within the surgical site and pre-scrubbing of the operative extremity should also be performed prior to transfer to the OR in order to minimize micro-organisms and dispersal of loose hair onto the sterile field.

Upon arrival of the patient to the OR, instrument tables based on the surgeon preference cards should be opened, instrument count and implant templating completed, necessary imaging displayed, and OR staff ready with specific responsibilities assigned to each member. Small and colleagues⁹ showed that using dedicated orthopedic staff familiar with the surgical routine decreased operative time by 19 minutes per procedure, or 1.25 hours for a surgeon performing 4 primary TJAs per day. Practices such as routine placement of a urinary catheter should be seriously scrutinized. In a randomized prospective study of patients undergoing total hip arthroplasty under spinal anesthesia, Miller and colleagues¹⁰ found no benefit for indwelling catheters in preventing urinary retention. In another randomized prospective study, Huang and colleagues¹¹ found the prevalence of urinary tract infections was significantly higher in TJA patients who received indwelling urinary catheters.

A scrub nurse familiar with the instruments, their assembly, and the sequence of events can ensure efficient surgical flow. The scrub nurse needs to anticipate missing or defective tools and call for them, ideally before the incision is made. Direct comparison studies are needed to assess the efficacy of routine intraoperative imaging vs commercially available universal cup alignment guides or clinical examinations in determining acceptable component positioning and limb length. Following component implantation and before wound closure, the circulating nurse should initiate the process of acquisition of a recovery room bed, make sure dressing supplies and necessary equipment are available, and call for surgical attendants. Lack of surgical attendants, delayed transfer from the OR table to hospital bed, and prolonged acquisition of a recovery room bed have been identified as major OR inefficiencies in a retrospective study by Attarian and colleagues.¹²

In summary, time is the OR’s most valuable resource.¹³ We believe that a consistent, almost automated attitude to the above procedures decreases variability and improves efficiency. By providing clear communication of the surgical needs with the team, having consistent anesthesia and nursing staff, implementing consistent perioperative protocols, and insuring that all necessary instruments and modalities are available prior to starting the procedure, we were able to sustainably increase OR throughput in a large teaching hospital.^9,14 This process, however, requires constant review to identify and eliminate new gaps, with each member of the team sharing a frank desire to improve. In this regard, hospital administrators share the duty to facilitate the implementation of any necessary changes, allocation of needed resources, and rewarding good effort, which could ultimately increase staff satisfaction and retention. Because efficiency is the ratio of benefits (eg, revenue, safety, etc.) to investment (eg, implant costs, wages, etc.), raises the question: what would be the effect of transitioning from hourly-wage to a salary-based system for key support staff? Unlike hourly-wage personnel, who have no incentive for productivity, a salaried employee assigned to a high-efficiency OR will inherently strive for improvement, employing higher organizational skills to accomplish a common goal. To our knowledge, there is no published data on this topic.

Developing a high-efficiency operating room (OR) is both a challenging and rewarding goal for any healthcare system. The OR is traditionally a high-cost/high-revenue environment¹ and operative efficacy has been correlated with low complication rates and surgical success.² An efficient OR is one that maximizes utilization while providing safe, reproducible, cost-effective, high-quality care. Total joint arthroplasty (TJA) has occupied the center stage for OR efficiency research, in part due to increasing demands from our aging population³ and economic pressures related to high implant costs, decreased reimbursement, and competition for market shares when OR time and space are limited.

A PubMed search on OR efficiency in TJA shows a disproportionately high focus on surgical technique, such as use of patient-specific instrumentation (PSI), computer-assisted surgery (CAS), minimally invasive surgery, and closure with barbed suture. In a retrospective review of 352 TKA patients who had PSI vs conventional instrumentation, DeHaan and colleagues⁴ found that PSI was associated with significantly decreased operative and room turnover times (20.4 minutes and 6.4 minutes, respectively). In another prospective multicenter study, Mont and colleagues⁵ showed a reduction in surgical time by 8.90 min for navigated total knee arthroplasty (TKA) performed with single-use instruments, cutting blocks, and trials. Other investigators compared PSI to CAS in TKA and found PSI to be 1.45 times more profitable than CAS, with 3 PSI cases performed in an 8-hour OR day compared to 2 CAS cases.⁶

There is no question that improved surgical technique can enhance OR efficiency. However, this model, while promising, is difficult to implement on a wide scale due to surgeon preferences, vendor limitations, and added costs related to the advanced preoperative imaging studies, manufacturing of the custom guides, and maintenance of navigation equipment. In addition, while interventions such as the use of barbed suture have the potential for speeding closure time, the time saved (4.7 minutes in one randomized trial)⁷ may not be enough to affect major utilization differences per OR per day. These technologies are also frequently employed by high-volume surgeons with high-volume teams and institutions.

Ideally, we need investment in the human capital and a collective change in work cultures to produce high-quality, well-choreographed, easily reproducible routines. An efficient OR requires the synchronous involvement of a large team of individuals, including hospital administrators, surgery schedulers, surgeons, anesthesiologists, preoperative holding area staff, OR nurses, surgical attendants, sterile processing personnel, and recovery room nurses. Case schedulers should match allocated block time with time required for surgery based on the historical performance of the individual surgeon, preferably scheduling similar cases on the same day. Preoperative work-up and medical clearance should be completed prior to scheduling to avoid last-minute cancellations. Patient reminders and accommodations for those traveling from long distances can further minimize late arrivals. Prompt initiation of the perioperative clinical pathway upon a patient’s check-in is important. The surgical site should be marked and the anesthesia plan confirmed upon arrival in the preoperative holding area. Necessary products need to be ready and/or administrated in time for transfer to the OR. These include prophylactic antibiotics, coagulation factors (eg, tranexamic acid), and blood products as indicated. Spinal anesthesia, regional nerve blocks, and intravenous (IV) lines should be completed before transfer to the OR. A “block room” close to the OR can allow concurrent induction of anesthesia and has been shown to increase the number of surgical cases performed during a regular workday.⁸ Hair clipping within the surgical site and pre-scrubbing of the operative extremity should also be performed prior to transfer to the OR in order to minimize micro-organisms and dispersal of loose hair onto the sterile field.

Upon arrival of the patient to the OR, instrument tables based on the surgeon preference cards should be opened, instrument count and implant templating completed, necessary imaging displayed, and OR staff ready with specific responsibilities assigned to each member. Small and colleagues⁹ showed that using dedicated orthopedic staff familiar with the surgical routine decreased operative time by 19 minutes per procedure, or 1.25 hours for a surgeon performing 4 primary TJAs per day. Practices such as routine placement of a urinary catheter should be seriously scrutinized. In a randomized prospective study of patients undergoing total hip arthroplasty under spinal anesthesia, Miller and colleagues¹⁰ found no benefit for indwelling catheters in preventing urinary retention. In another randomized prospective study, Huang and colleagues¹¹ found the prevalence of urinary tract infections was significantly higher in TJA patients who received indwelling urinary catheters.

A scrub nurse familiar with the instruments, their assembly, and the sequence of events can ensure efficient surgical flow. The scrub nurse needs to anticipate missing or defective tools and call for them, ideally before the incision is made. Direct comparison studies are needed to assess the efficacy of routine intraoperative imaging vs commercially available universal cup alignment guides or clinical examinations in determining acceptable component positioning and limb length. Following component implantation and before wound closure, the circulating nurse should initiate the process of acquisition of a recovery room bed, make sure dressing supplies and necessary equipment are available, and call for surgical attendants. Lack of surgical attendants, delayed transfer from the OR table to hospital bed, and prolonged acquisition of a recovery room bed have been identified as major OR inefficiencies in a retrospective study by Attarian and colleagues.¹²

In summary, time is the OR’s most valuable resource.¹³ We believe that a consistent, almost automated attitude to the above procedures decreases variability and improves efficiency. By providing clear communication of the surgical needs with the team, having consistent anesthesia and nursing staff, implementing consistent perioperative protocols, and insuring that all necessary instruments and modalities are available prior to starting the procedure, we were able to sustainably increase OR throughput in a large teaching hospital.^9,14 This process, however, requires constant review to identify and eliminate new gaps, with each member of the team sharing a frank desire to improve. In this regard, hospital administrators share the duty to facilitate the implementation of any necessary changes, allocation of needed resources, and rewarding good effort, which could ultimately increase staff satisfaction and retention. Because efficiency is the ratio of benefits (eg, revenue, safety, etc.) to investment (eg, implant costs, wages, etc.), raises the question: what would be the effect of transitioning from hourly-wage to a salary-based system for key support staff? Unlike hourly-wage personnel, who have no incentive for productivity, a salaried employee assigned to a high-efficiency OR will inherently strive for improvement, employing higher organizational skills to accomplish a common goal. To our knowledge, there is no published data on this topic.

Since Neer and colleagues¹ first reported in 1982, glenoid loosening persists as a common cause of anatomic total shoulder arthroplasty (TSA) failure.^1-4 Currently, cemented, all-polyethylene glenoid components are the gold standard, and minimum clinical survival of 10 to 15 years is expected.^3,5 Several clinical studies^5-9 and in vitro biomechanical studies¹⁰ have suggested an advantage of pegged over keeled glenoid components, but glenoid component loosening remains a frequent complication,¹¹ with the cement–implant interface suggested as the weak link of fixation.^10,12 In addition to mechanical loosening, poor cement penetration and heat-induced necrosis have been postulated as contributing to glenoid component loosening.^13,14

Because of these potential complications, there is a growing consideration to minimize or abandon cement fixation and rely on biological fixation to polyethylene for long-term component stability.¹⁵ A newer pegged glenoid component design consists of traditional, peripherally located pegs designed for cement fixation as well as a central, uncemented, fluted, interference-fit peg that allows for bony ingrowth. Short-term clinical studies have shown that bony ingrowth into the space between the flutes can be achieved with a hybrid cementation technique and that, when that occurs, excellent outcomes are likely.^13,16-19 The immediate in vivo stability of this implant design upon initial implantation, before the cement has cured, has prompted some surgeons to consider implanting the device without cement. In a recent series in which this implant design was used without cement, clinical and radiographic results were promising.¹⁵

Despite the widespread clinical use, little biomechanical work has been done to characterize initial fixation of all-polyethylene pegged glenoid implants. We conducted a study to compare glenoid micromotion in an all-polyethylene, centrally fluted pegged glenoid component as a function of 3 fixation techniques: cementless interference-fit fixation, hybrid partial cementation based on manufacturer recommendations, and full cementation to simulate a gold-standard, traditional, cemented, pegged design.

Materials and MethodsBiomechanical Testing

The biomechanical testing methodology used in this study was based on previous studies20-23 and on ASTM standard F2028-1224 using polyurethane bone substitute 0.24 g/cm³ (Pacific Research Laboratories) with ultimate strength of 4.9 MPa and compressive modulus of 123 MPa for component implantation. This material was selected because its mechanical properties are similar to those of cancellous glenoid bone in primary shoulder arthroplasty,25 and it minimizes variability with use of cadaveric specimens. Components were mounted on an MTS 858 Mini-Bionix II materials testing frame (Figure 1). A static compressive load of 756 N (170 lb) was applied via a mass-pulley system simulating the joint compressive force the shoulder is likely to experience during higher load activities.^24,26 The glenoid component was positioned on a linear bearing to allow for joint compression.

Test Groups and Cement Fixation Techniques

All-polyethylene pegged glenoid components (Anchor Peg Glenoid, size 44; DePuy Orthopaedics) were used for biomechanical testing (Figure 1). Polyurethane blocks were reamed with a size 44 reamer until the superior-inferior distance reached 33 mm, ensuring complete seating of implant. Three fixation-technique groups were formed: interference-fit, hybrid cement, and fully cemented. Interference-fit fixation was done without polymethylmethacrylate (PMMA) cement. In hybrid fixation, 2 cm³ of PMMA (SpeedSet Cement, Stryker Orthopaedics) was injected (using a catheter tip syringe) into the peripheral peg holes and manually pressurized; the central peg was press-fit into polyurethane bone substitute. In the fully cemented group, both peripheral and central peg holes received PMMA; the peripheral peg holes were cemented as in hybrid fixation, and the central peg hole was injected with 3 cm³ of PMMA, which was then manually pressurized. The humeral head component (Global Advantage, 44×18 mm; DePuy Synthes) was mounted on the test frame actuator and centrally located within the glenoid at the start of the test.

Determination of Humeral Head Translation via Subluxation Testing

Humeral head subluxation distance, simulating a humeral head rim loading event, was calculated on the basis of preliminary tests outlined in the ASTM standard.²⁴ Three glenoids (1 per fixation technique) were mounted on the test frame with a humeral head positioned centrally within the glenoid. After the joint compressive force was applied, the humeral head was translated along the true superior axis of the glenoid at a rate of 50 mm/min. Testing software was used to record humeral head displacement and load data at a frequency of 100 Hz. Humeral head subluxation displacement was determined at the end of the linear region of the force versus displacement response. This distance, averaged from the 3 subluxation tests, was used as the subluxation distance during cyclic testing.

Determination of Glenoid Component Motion via Cyclic Testing

After subluxation displacement was determined, glenoid components were mounted on the test frame (5 per fixation technique) and subjected to 50,000 cycles of humeral head translation at a frequency of 2 Hz. Amplitude of the humeral head displacement against the glenoid component followed a sinusoidal pattern with maxima and minima represented by the subluxation displacement (positive and negative, respectively). Glenoid edge compression/distraction of the superior edge and glenoid inferior/superior translation were monitored with 2 variable resistance reluctance transducers (Microminiature DVRT; 4.5-µm resolution; MicroStrain) secured to the glenoid component and testing fixture.

Microminiature DVRT measurements of glenoid motion were taken for 5 consecutive cycles at cycles 1, 20, 100, 500, 1000, 5000, 10,000, 15,000, 20,000, 30,000, 40,000, and 50,000. Distraction-compression displacement and superior-inferior translation measurements were recorded relative to the glenoid position with the humeral head at the neutral position at a given cycle. Final glenoid micromotion data were calculated from the mean of consecutive cycles at each cycle time point.

Statistical Analysis

Glenoid motion results are reported as means and standard deviations. Comparisons with 2 factors of fixation technique and number of cycles for glenoid distraction, glenoid compression, and absolute glenoid translation were characterized with 2-way analysis of variance (SigmaPlot Version 11.0; Systat Software), with the Holm-Šídák test used for post hoc determination of significant relationships.

Results

Under subluxation testing, the humeral head translation distance at the end of the linear region was determined to be 0.50 mm. Subsequently for cyclic testing, the humeral head was then translated 0.50 mm from the neutral position of the humeral head along both the superior and inferior axes of the glenoid. All glenoids successfully completed the entire 50,000 cycles of testing. For the glenoid component, Figure 2 depicts distraction and compression, and Figure 3 depicts superior-inferior translation.

Glenoid Component Distraction

Overall, mean (SD) glenoid distraction was significantly higher for interference-fit fixation, 0.21 (0.10) mm, than for hybrid cement fixation, 0.16 (0.05) mm (P < .001), and fully cemented fixation, 0.09 (0.07) mm (P < .001). It was also significantly higher for hybrid fixation than fully cemented fixation (P < .001). From cycle 1000 to cycle 50,000, distraction was significantly higher for interference-fit fixation than for fully cemented fixation at each time point (P < .05).

Glenoid Component Compression

Mean (SD) compression was significantly higher for hybrid cement fixation, 0.31 (0.13) mm, than for interference-fit fixation, 0.17 (0.07) mm (P < .001), and fully cemented fixation, 0.17 (0.08) mm (P < .001). No significant difference was found between interference-fit and fully cemented fixation (P = .793) (Figure 2). At cycles 1, 20, 100, and 500, compression was significantly higher for hybrid fixation than for fully cemented fixation (P < .05). In addition, at cycle 500, it was significantly higher for hybrid fixation than for interference-fit fixation (P < .05).

Glenoid Component Translation

Mean (SD) glenoid translation was significantly lower for fully cemented fixation, 0.10 (0.04) mm, than for interference-fit fixation, 0.13 (0.04) mm (P < .001), and hybrid cement fixation, 0.13 (0.03) mm (P < .001), with all time points considered. There was no significant difference between interference-fit and hybrid fixation (P = .343). Initial translation at cycle 1 was significantly higher for interference-fit and hybid fixation than for fully cemented fixation.

Discussion

Despite advances in glenoid component design, glenoid loosening remains the most common cause of anatomical TSA failure. Recent implants have been designed to take advantage of an all-polyethylene component while providing long-lasting fixation through bony ingrowth into a central peg. In a study of the hybrid cementation technique drescribed here, Groh¹⁷ found no glenoid loosening or radiolucent lines but discovered fingerlike projections of bone between the flanges of the implant in 24 (29%) of 83 cases. Churchill and colleagues¹⁶ also reported bony ingrowth into the central peg in 15 (75%) of 20 patients. Furthermore, Arnold and colleagues13 reported complete bony ingrowth (6/6 inter-fin compartments) in 23 (71%) of 35 shoulders at a mean of 43 months. Wirth and colleagues¹⁹ reported increased radiodensity between the flanges of the central peg in 30 of 44 cases (68%) and osteolysis around the central peg in 3 of 44 cases (7%) at 3 years.

There are also reports of successful bony ingrowth associated with all-polyethylene components implanted without cement. In a canine study using an early ingrowth implant design, Wirth and colleagues²⁷ showed that, though initial fixation was superior with a cemented, keeled implant, pullout strength of the uncemented, pegged implant improved over time and eventually far surpassed that of the cemented, keeled implant owing to both the loosening of the cemented component and the bony ingrowth into the central peg component. Furthermore, Anglin and colleagues¹⁰ confirmed that component micromotion was lower with pegged glenoid components than with keeled components in a biomechanical model. De Wilde and colleagues¹⁵ recently reported on a series of uncemented, central fluted peg glenoids implanted in 34 patients followed clinically and with computed tomography for a minimum of 24 months. The investigators found bony ingrowth into the central peg in 27 (79%) of 34 patients and no signs of loosening in 30 (88%) of 34 patients. Incomplete lucencies around 1 or 2 peripheral pegs were found in 2 (6%) of 34 patients, and complete lucencies around 2 or more peripheral pegs were found in 2 (6%) of 34 patients. However, there was no statistical difference in clinical outcome between patients with and without loosening.

With this type of implant, initial fixation that provides stability while minimizing micromotion under biomechanical loading likely is crucial for attaining bony growth within the central peg flanges. To our knowledge, this is the first biomechanical study to compare micromotion using 3 different fixation methods with a central fluted peg glenoid component design. Of all these fixation methods, fully cemented fixation yielded the most stable glenoid throughout testing with respect to the evaluated parameters. However, this method is not necessarily clinically applicable, as a fully cemented glenoid would inhibit any bony growth within the central flange, which is necessary for long-term biological fixation. Our data showed that, though glenoid distraction was significantly lower with hybrid cement fixation, this fixation method exhibited significantly higher glenoid compression. In addition, there were no significant differences between glenoid components with hybrid fixation and glenoid components with interference-fit fixation with respect to component translation in the superior-inferior direction. These findings may indicate that initial fixation is not significantly improved by the addition of cement to the peripheral pegs in a glenoid component with a central fluted peg design.

The interference fit of the central peg is primarily responsible for conferring long-term implant stability,^13,27 which is ultimately achieved by bony formation between the flutes of the peg. Other authors have reported that, for bony ingrowth to occur, micromotion between the bone–implant interface must not exceed 20 to 150 µm.^28-30 Other than for interference-fit distraction at more than 1000 cycles and hybrid cement fixation compression at each time point throughout testing, our data fall within the reported upper limits of micromotion to support bony ingrowth. Increased micromotion in the interference-fit fixation group is seen at later time points and may be caused by the inability to simulate the potential fixation gained from bony ingrowth allowed with this surgical technique. Research is needed to further explain this increase in distraction.

Results from this study must be interpreted with caution because of limitations of the in vitro testing methodology. This biomechanical model using bone substitute characterizes glenoid fixation at time zero, directly after implantation, followed by repetitive cyclic loading simulating 5 years of implant service. This differs from the clinical scenario in which the shoulder undergoes postoperative immobilization or protected motion during which the early phases of bony remodeling are likely occurring. Furthermore, simulation of 5 years of implant service may not be necessary for an implant that is expected to achieve ultimate fixation by bony ingrowth within the first several months after implantation. Use of this implant without cement is classified off-label, and surgeons should take this into consideration during implantation. Last, this study could not simulate the effect of bony ingrowth on fixation, though our experimental technique of cementing the central peg may be a gross approximation of a fully ingrown central peg and its expected rigid fixation.

Fully cemented fixation of a polyethylene glenoid is superior to hybrid cement fixation and interference-fit fixation with respect to early glenoid micromotion. However, the long-term stability of a fully cemented polyethylene glenoid component remains a clinical concern, as fixation is achieved by bony ingrowth around the central fluted peg of the implant. In this study, interference-fit and hybrid fixation had equivocal component micromotion in biomechanical testing. Our findings suggest that cementation of the peripheral pegs confers no additional initial stability over an uncemented interference-fit technique in a biomechanical model. More research is needed to further evaluate interference-fit fixation as a viable option for implantation of a central fluted, all-polyethylene glenoid component.

Since Neer and colleagues¹ first reported in 1982, glenoid loosening persists as a common cause of anatomic total shoulder arthroplasty (TSA) failure.^1-4 Currently, cemented, all-polyethylene glenoid components are the gold standard, and minimum clinical survival of 10 to 15 years is expected.^3,5 Several clinical studies^5-9 and in vitro biomechanical studies¹⁰ have suggested an advantage of pegged over keeled glenoid components, but glenoid component loosening remains a frequent complication,¹¹ with the cement–implant interface suggested as the weak link of fixation.^10,12 In addition to mechanical loosening, poor cement penetration and heat-induced necrosis have been postulated as contributing to glenoid component loosening.^13,14

Because of these potential complications, there is a growing consideration to minimize or abandon cement fixation and rely on biological fixation to polyethylene for long-term component stability.¹⁵ A newer pegged glenoid component design consists of traditional, peripherally located pegs designed for cement fixation as well as a central, uncemented, fluted, interference-fit peg that allows for bony ingrowth. Short-term clinical studies have shown that bony ingrowth into the space between the flutes can be achieved with a hybrid cementation technique and that, when that occurs, excellent outcomes are likely.^13,16-19 The immediate in vivo stability of this implant design upon initial implantation, before the cement has cured, has prompted some surgeons to consider implanting the device without cement. In a recent series in which this implant design was used without cement, clinical and radiographic results were promising.¹⁵

Despite the widespread clinical use, little biomechanical work has been done to characterize initial fixation of all-polyethylene pegged glenoid implants. We conducted a study to compare glenoid micromotion in an all-polyethylene, centrally fluted pegged glenoid component as a function of 3 fixation techniques: cementless interference-fit fixation, hybrid partial cementation based on manufacturer recommendations, and full cementation to simulate a gold-standard, traditional, cemented, pegged design.

Materials and MethodsBiomechanical Testing

The biomechanical testing methodology used in this study was based on previous studies20-23 and on ASTM standard F2028-1224 using polyurethane bone substitute 0.24 g/cm³ (Pacific Research Laboratories) with ultimate strength of 4.9 MPa and compressive modulus of 123 MPa for component implantation. This material was selected because its mechanical properties are similar to those of cancellous glenoid bone in primary shoulder arthroplasty,25 and it minimizes variability with use of cadaveric specimens. Components were mounted on an MTS 858 Mini-Bionix II materials testing frame (Figure 1). A static compressive load of 756 N (170 lb) was applied via a mass-pulley system simulating the joint compressive force the shoulder is likely to experience during higher load activities.^24,26 The glenoid component was positioned on a linear bearing to allow for joint compression.

Test Groups and Cement Fixation Techniques

All-polyethylene pegged glenoid components (Anchor Peg Glenoid, size 44; DePuy Orthopaedics) were used for biomechanical testing (Figure 1). Polyurethane blocks were reamed with a size 44 reamer until the superior-inferior distance reached 33 mm, ensuring complete seating of implant. Three fixation-technique groups were formed: interference-fit, hybrid cement, and fully cemented. Interference-fit fixation was done without polymethylmethacrylate (PMMA) cement. In hybrid fixation, 2 cm³ of PMMA (SpeedSet Cement, Stryker Orthopaedics) was injected (using a catheter tip syringe) into the peripheral peg holes and manually pressurized; the central peg was press-fit into polyurethane bone substitute. In the fully cemented group, both peripheral and central peg holes received PMMA; the peripheral peg holes were cemented as in hybrid fixation, and the central peg hole was injected with 3 cm³ of PMMA, which was then manually pressurized. The humeral head component (Global Advantage, 44×18 mm; DePuy Synthes) was mounted on the test frame actuator and centrally located within the glenoid at the start of the test.

Determination of Humeral Head Translation via Subluxation Testing

Humeral head subluxation distance, simulating a humeral head rim loading event, was calculated on the basis of preliminary tests outlined in the ASTM standard.²⁴ Three glenoids (1 per fixation technique) were mounted on the test frame with a humeral head positioned centrally within the glenoid. After the joint compressive force was applied, the humeral head was translated along the true superior axis of the glenoid at a rate of 50 mm/min. Testing software was used to record humeral head displacement and load data at a frequency of 100 Hz. Humeral head subluxation displacement was determined at the end of the linear region of the force versus displacement response. This distance, averaged from the 3 subluxation tests, was used as the subluxation distance during cyclic testing.

Determination of Glenoid Component Motion via Cyclic Testing

After subluxation displacement was determined, glenoid components were mounted on the test frame (5 per fixation technique) and subjected to 50,000 cycles of humeral head translation at a frequency of 2 Hz. Amplitude of the humeral head displacement against the glenoid component followed a sinusoidal pattern with maxima and minima represented by the subluxation displacement (positive and negative, respectively). Glenoid edge compression/distraction of the superior edge and glenoid inferior/superior translation were monitored with 2 variable resistance reluctance transducers (Microminiature DVRT; 4.5-µm resolution; MicroStrain) secured to the glenoid component and testing fixture.

Microminiature DVRT measurements of glenoid motion were taken for 5 consecutive cycles at cycles 1, 20, 100, 500, 1000, 5000, 10,000, 15,000, 20,000, 30,000, 40,000, and 50,000. Distraction-compression displacement and superior-inferior translation measurements were recorded relative to the glenoid position with the humeral head at the neutral position at a given cycle. Final glenoid micromotion data were calculated from the mean of consecutive cycles at each cycle time point.

Statistical Analysis

Glenoid motion results are reported as means and standard deviations. Comparisons with 2 factors of fixation technique and number of cycles for glenoid distraction, glenoid compression, and absolute glenoid translation were characterized with 2-way analysis of variance (SigmaPlot Version 11.0; Systat Software), with the Holm-Šídák test used for post hoc determination of significant relationships.

Results

Under subluxation testing, the humeral head translation distance at the end of the linear region was determined to be 0.50 mm. Subsequently for cyclic testing, the humeral head was then translated 0.50 mm from the neutral position of the humeral head along both the superior and inferior axes of the glenoid. All glenoids successfully completed the entire 50,000 cycles of testing. For the glenoid component, Figure 2 depicts distraction and compression, and Figure 3 depicts superior-inferior translation.

Glenoid Component Distraction

Overall, mean (SD) glenoid distraction was significantly higher for interference-fit fixation, 0.21 (0.10) mm, than for hybrid cement fixation, 0.16 (0.05) mm (P < .001), and fully cemented fixation, 0.09 (0.07) mm (P < .001). It was also significantly higher for hybrid fixation than fully cemented fixation (P < .001). From cycle 1000 to cycle 50,000, distraction was significantly higher for interference-fit fixation than for fully cemented fixation at each time point (P < .05).

Glenoid Component Compression

Mean (SD) compression was significantly higher for hybrid cement fixation, 0.31 (0.13) mm, than for interference-fit fixation, 0.17 (0.07) mm (P < .001), and fully cemented fixation, 0.17 (0.08) mm (P < .001). No significant difference was found between interference-fit and fully cemented fixation (P = .793) (Figure 2). At cycles 1, 20, 100, and 500, compression was significantly higher for hybrid fixation than for fully cemented fixation (P < .05). In addition, at cycle 500, it was significantly higher for hybrid fixation than for interference-fit fixation (P < .05).

Glenoid Component Translation

Mean (SD) glenoid translation was significantly lower for fully cemented fixation, 0.10 (0.04) mm, than for interference-fit fixation, 0.13 (0.04) mm (P < .001), and hybrid cement fixation, 0.13 (0.03) mm (P < .001), with all time points considered. There was no significant difference between interference-fit and hybrid fixation (P = .343). Initial translation at cycle 1 was significantly higher for interference-fit and hybid fixation than for fully cemented fixation.

Discussion

Despite advances in glenoid component design, glenoid loosening remains the most common cause of anatomical TSA failure. Recent implants have been designed to take advantage of an all-polyethylene component while providing long-lasting fixation through bony ingrowth into a central peg. In a study of the hybrid cementation technique drescribed here, Groh¹⁷ found no glenoid loosening or radiolucent lines but discovered fingerlike projections of bone between the flanges of the implant in 24 (29%) of 83 cases. Churchill and colleagues¹⁶ also reported bony ingrowth into the central peg in 15 (75%) of 20 patients. Furthermore, Arnold and colleagues13 reported complete bony ingrowth (6/6 inter-fin compartments) in 23 (71%) of 35 shoulders at a mean of 43 months. Wirth and colleagues¹⁹ reported increased radiodensity between the flanges of the central peg in 30 of 44 cases (68%) and osteolysis around the central peg in 3 of 44 cases (7%) at 3 years.

There are also reports of successful bony ingrowth associated with all-polyethylene components implanted without cement. In a canine study using an early ingrowth implant design, Wirth and colleagues²⁷ showed that, though initial fixation was superior with a cemented, keeled implant, pullout strength of the uncemented, pegged implant improved over time and eventually far surpassed that of the cemented, keeled implant owing to both the loosening of the cemented component and the bony ingrowth into the central peg component. Furthermore, Anglin and colleagues¹⁰ confirmed that component micromotion was lower with pegged glenoid components than with keeled components in a biomechanical model. De Wilde and colleagues¹⁵ recently reported on a series of uncemented, central fluted peg glenoids implanted in 34 patients followed clinically and with computed tomography for a minimum of 24 months. The investigators found bony ingrowth into the central peg in 27 (79%) of 34 patients and no signs of loosening in 30 (88%) of 34 patients. Incomplete lucencies around 1 or 2 peripheral pegs were found in 2 (6%) of 34 patients, and complete lucencies around 2 or more peripheral pegs were found in 2 (6%) of 34 patients. However, there was no statistical difference in clinical outcome between patients with and without loosening.

With this type of implant, initial fixation that provides stability while minimizing micromotion under biomechanical loading likely is crucial for attaining bony growth within the central peg flanges. To our knowledge, this is the first biomechanical study to compare micromotion using 3 different fixation methods with a central fluted peg glenoid component design. Of all these fixation methods, fully cemented fixation yielded the most stable glenoid throughout testing with respect to the evaluated parameters. However, this method is not necessarily clinically applicable, as a fully cemented glenoid would inhibit any bony growth within the central flange, which is necessary for long-term biological fixation. Our data showed that, though glenoid distraction was significantly lower with hybrid cement fixation, this fixation method exhibited significantly higher glenoid compression. In addition, there were no significant differences between glenoid components with hybrid fixation and glenoid components with interference-fit fixation with respect to component translation in the superior-inferior direction. These findings may indicate that initial fixation is not significantly improved by the addition of cement to the peripheral pegs in a glenoid component with a central fluted peg design.

The interference fit of the central peg is primarily responsible for conferring long-term implant stability,^13,27 which is ultimately achieved by bony formation between the flutes of the peg. Other authors have reported that, for bony ingrowth to occur, micromotion between the bone–implant interface must not exceed 20 to 150 µm.^28-30 Other than for interference-fit distraction at more than 1000 cycles and hybrid cement fixation compression at each time point throughout testing, our data fall within the reported upper limits of micromotion to support bony ingrowth. Increased micromotion in the interference-fit fixation group is seen at later time points and may be caused by the inability to simulate the potential fixation gained from bony ingrowth allowed with this surgical technique. Research is needed to further explain this increase in distraction.

Results from this study must be interpreted with caution because of limitations of the in vitro testing methodology. This biomechanical model using bone substitute characterizes glenoid fixation at time zero, directly after implantation, followed by repetitive cyclic loading simulating 5 years of implant service. This differs from the clinical scenario in which the shoulder undergoes postoperative immobilization or protected motion during which the early phases of bony remodeling are likely occurring. Furthermore, simulation of 5 years of implant service may not be necessary for an implant that is expected to achieve ultimate fixation by bony ingrowth within the first several months after implantation. Use of this implant without cement is classified off-label, and surgeons should take this into consideration during implantation. Last, this study could not simulate the effect of bony ingrowth on fixation, though our experimental technique of cementing the central peg may be a gross approximation of a fully ingrown central peg and its expected rigid fixation.

Fully cemented fixation of a polyethylene glenoid is superior to hybrid cement fixation and interference-fit fixation with respect to early glenoid micromotion. However, the long-term stability of a fully cemented polyethylene glenoid component remains a clinical concern, as fixation is achieved by bony ingrowth around the central fluted peg of the implant. In this study, interference-fit and hybrid fixation had equivocal component micromotion in biomechanical testing. Our findings suggest that cementation of the peripheral pegs confers no additional initial stability over an uncemented interference-fit technique in a biomechanical model. More research is needed to further evaluate interference-fit fixation as a viable option for implantation of a central fluted, all-polyethylene glenoid component.

Since Neer and colleagues¹ first reported in 1982, glenoid loosening persists as a common cause of anatomic total shoulder arthroplasty (TSA) failure.^1-4 Currently, cemented, all-polyethylene glenoid components are the gold standard, and minimum clinical survival of 10 to 15 years is expected.^3,5 Several clinical studies^5-9 and in vitro biomechanical studies¹⁰ have suggested an advantage of pegged over keeled glenoid components, but glenoid component loosening remains a frequent complication,¹¹ with the cement–implant interface suggested as the weak link of fixation.^10,12 In addition to mechanical loosening, poor cement penetration and heat-induced necrosis have been postulated as contributing to glenoid component loosening.^13,14

Because of these potential complications, there is a growing consideration to minimize or abandon cement fixation and rely on biological fixation to polyethylene for long-term component stability.¹⁵ A newer pegged glenoid component design consists of traditional, peripherally located pegs designed for cement fixation as well as a central, uncemented, fluted, interference-fit peg that allows for bony ingrowth. Short-term clinical studies have shown that bony ingrowth into the space between the flutes can be achieved with a hybrid cementation technique and that, when that occurs, excellent outcomes are likely.^13,16-19 The immediate in vivo stability of this implant design upon initial implantation, before the cement has cured, has prompted some surgeons to consider implanting the device without cement. In a recent series in which this implant design was used without cement, clinical and radiographic results were promising.¹⁵

Despite the widespread clinical use, little biomechanical work has been done to characterize initial fixation of all-polyethylene pegged glenoid implants. We conducted a study to compare glenoid micromotion in an all-polyethylene, centrally fluted pegged glenoid component as a function of 3 fixation techniques: cementless interference-fit fixation, hybrid partial cementation based on manufacturer recommendations, and full cementation to simulate a gold-standard, traditional, cemented, pegged design.

Materials and MethodsBiomechanical Testing

The biomechanical testing methodology used in this study was based on previous studies20-23 and on ASTM standard F2028-1224 using polyurethane bone substitute 0.24 g/cm³ (Pacific Research Laboratories) with ultimate strength of 4.9 MPa and compressive modulus of 123 MPa for component implantation. This material was selected because its mechanical properties are similar to those of cancellous glenoid bone in primary shoulder arthroplasty,25 and it minimizes variability with use of cadaveric specimens. Components were mounted on an MTS 858 Mini-Bionix II materials testing frame (Figure 1). A static compressive load of 756 N (170 lb) was applied via a mass-pulley system simulating the joint compressive force the shoulder is likely to experience during higher load activities.^24,26 The glenoid component was positioned on a linear bearing to allow for joint compression.

Test Groups and Cement Fixation Techniques

All-polyethylene pegged glenoid components (Anchor Peg Glenoid, size 44; DePuy Orthopaedics) were used for biomechanical testing (Figure 1). Polyurethane blocks were reamed with a size 44 reamer until the superior-inferior distance reached 33 mm, ensuring complete seating of implant. Three fixation-technique groups were formed: interference-fit, hybrid cement, and fully cemented. Interference-fit fixation was done without polymethylmethacrylate (PMMA) cement. In hybrid fixation, 2 cm³ of PMMA (SpeedSet Cement, Stryker Orthopaedics) was injected (using a catheter tip syringe) into the peripheral peg holes and manually pressurized; the central peg was press-fit into polyurethane bone substitute. In the fully cemented group, both peripheral and central peg holes received PMMA; the peripheral peg holes were cemented as in hybrid fixation, and the central peg hole was injected with 3 cm³ of PMMA, which was then manually pressurized. The humeral head component (Global Advantage, 44×18 mm; DePuy Synthes) was mounted on the test frame actuator and centrally located within the glenoid at the start of the test.

Determination of Humeral Head Translation via Subluxation Testing

Humeral head subluxation distance, simulating a humeral head rim loading event, was calculated on the basis of preliminary tests outlined in the ASTM standard.²⁴ Three glenoids (1 per fixation technique) were mounted on the test frame with a humeral head positioned centrally within the glenoid. After the joint compressive force was applied, the humeral head was translated along the true superior axis of the glenoid at a rate of 50 mm/min. Testing software was used to record humeral head displacement and load data at a frequency of 100 Hz. Humeral head subluxation displacement was determined at the end of the linear region of the force versus displacement response. This distance, averaged from the 3 subluxation tests, was used as the subluxation distance during cyclic testing.

Determination of Glenoid Component Motion via Cyclic Testing

After subluxation displacement was determined, glenoid components were mounted on the test frame (5 per fixation technique) and subjected to 50,000 cycles of humeral head translation at a frequency of 2 Hz. Amplitude of the humeral head displacement against the glenoid component followed a sinusoidal pattern with maxima and minima represented by the subluxation displacement (positive and negative, respectively). Glenoid edge compression/distraction of the superior edge and glenoid inferior/superior translation were monitored with 2 variable resistance reluctance transducers (Microminiature DVRT; 4.5-µm resolution; MicroStrain) secured to the glenoid component and testing fixture.

Microminiature DVRT measurements of glenoid motion were taken for 5 consecutive cycles at cycles 1, 20, 100, 500, 1000, 5000, 10,000, 15,000, 20,000, 30,000, 40,000, and 50,000. Distraction-compression displacement and superior-inferior translation measurements were recorded relative to the glenoid position with the humeral head at the neutral position at a given cycle. Final glenoid micromotion data were calculated from the mean of consecutive cycles at each cycle time point.

Statistical Analysis

Glenoid motion results are reported as means and standard deviations. Comparisons with 2 factors of fixation technique and number of cycles for glenoid distraction, glenoid compression, and absolute glenoid translation were characterized with 2-way analysis of variance (SigmaPlot Version 11.0; Systat Software), with the Holm-Šídák test used for post hoc determination of significant relationships.

Results

Under subluxation testing, the humeral head translation distance at the end of the linear region was determined to be 0.50 mm. Subsequently for cyclic testing, the humeral head was then translated 0.50 mm from the neutral position of the humeral head along both the superior and inferior axes of the glenoid. All glenoids successfully completed the entire 50,000 cycles of testing. For the glenoid component, Figure 2 depicts distraction and compression, and Figure 3 depicts superior-inferior translation.

Glenoid Component Distraction

Overall, mean (SD) glenoid distraction was significantly higher for interference-fit fixation, 0.21 (0.10) mm, than for hybrid cement fixation, 0.16 (0.05) mm (P < .001), and fully cemented fixation, 0.09 (0.07) mm (P < .001). It was also significantly higher for hybrid fixation than fully cemented fixation (P < .001). From cycle 1000 to cycle 50,000, distraction was significantly higher for interference-fit fixation than for fully cemented fixation at each time point (P < .05).

Glenoid Component Compression

Mean (SD) compression was significantly higher for hybrid cement fixation, 0.31 (0.13) mm, than for interference-fit fixation, 0.17 (0.07) mm (P < .001), and fully cemented fixation, 0.17 (0.08) mm (P < .001). No significant difference was found between interference-fit and fully cemented fixation (P = .793) (Figure 2). At cycles 1, 20, 100, and 500, compression was significantly higher for hybrid fixation than for fully cemented fixation (P < .05). In addition, at cycle 500, it was significantly higher for hybrid fixation than for interference-fit fixation (P < .05).

Glenoid Component Translation

Mean (SD) glenoid translation was significantly lower for fully cemented fixation, 0.10 (0.04) mm, than for interference-fit fixation, 0.13 (0.04) mm (P < .001), and hybrid cement fixation, 0.13 (0.03) mm (P < .001), with all time points considered. There was no significant difference between interference-fit and hybrid fixation (P = .343). Initial translation at cycle 1 was significantly higher for interference-fit and hybid fixation than for fully cemented fixation.

Discussion

Despite advances in glenoid component design, glenoid loosening remains the most common cause of anatomical TSA failure. Recent implants have been designed to take advantage of an all-polyethylene component while providing long-lasting fixation through bony ingrowth into a central peg. In a study of the hybrid cementation technique drescribed here, Groh¹⁷ found no glenoid loosening or radiolucent lines but discovered fingerlike projections of bone between the flanges of the implant in 24 (29%) of 83 cases. Churchill and colleagues¹⁶ also reported bony ingrowth into the central peg in 15 (75%) of 20 patients. Furthermore, Arnold and colleagues13 reported complete bony ingrowth (6/6 inter-fin compartments) in 23 (71%) of 35 shoulders at a mean of 43 months. Wirth and colleagues¹⁹ reported increased radiodensity between the flanges of the central peg in 30 of 44 cases (68%) and osteolysis around the central peg in 3 of 44 cases (7%) at 3 years.

There are also reports of successful bony ingrowth associated with all-polyethylene components implanted without cement. In a canine study using an early ingrowth implant design, Wirth and colleagues²⁷ showed that, though initial fixation was superior with a cemented, keeled implant, pullout strength of the uncemented, pegged implant improved over time and eventually far surpassed that of the cemented, keeled implant owing to both the loosening of the cemented component and the bony ingrowth into the central peg component. Furthermore, Anglin and colleagues¹⁰ confirmed that component micromotion was lower with pegged glenoid components than with keeled components in a biomechanical model. De Wilde and colleagues¹⁵ recently reported on a series of uncemented, central fluted peg glenoids implanted in 34 patients followed clinically and with computed tomography for a minimum of 24 months. The investigators found bony ingrowth into the central peg in 27 (79%) of 34 patients and no signs of loosening in 30 (88%) of 34 patients. Incomplete lucencies around 1 or 2 peripheral pegs were found in 2 (6%) of 34 patients, and complete lucencies around 2 or more peripheral pegs were found in 2 (6%) of 34 patients. However, there was no statistical difference in clinical outcome between patients with and without loosening.

With this type of implant, initial fixation that provides stability while minimizing micromotion under biomechanical loading likely is crucial for attaining bony growth within the central peg flanges. To our knowledge, this is the first biomechanical study to compare micromotion using 3 different fixation methods with a central fluted peg glenoid component design. Of all these fixation methods, fully cemented fixation yielded the most stable glenoid throughout testing with respect to the evaluated parameters. However, this method is not necessarily clinically applicable, as a fully cemented glenoid would inhibit any bony growth within the central flange, which is necessary for long-term biological fixation. Our data showed that, though glenoid distraction was significantly lower with hybrid cement fixation, this fixation method exhibited significantly higher glenoid compression. In addition, there were no significant differences between glenoid components with hybrid fixation and glenoid components with interference-fit fixation with respect to component translation in the superior-inferior direction. These findings may indicate that initial fixation is not significantly improved by the addition of cement to the peripheral pegs in a glenoid component with a central fluted peg design.

The interference fit of the central peg is primarily responsible for conferring long-term implant stability,^13,27 which is ultimately achieved by bony formation between the flutes of the peg. Other authors have reported that, for bony ingrowth to occur, micromotion between the bone–implant interface must not exceed 20 to 150 µm.^28-30 Other than for interference-fit distraction at more than 1000 cycles and hybrid cement fixation compression at each time point throughout testing, our data fall within the reported upper limits of micromotion to support bony ingrowth. Increased micromotion in the interference-fit fixation group is seen at later time points and may be caused by the inability to simulate the potential fixation gained from bony ingrowth allowed with this surgical technique. Research is needed to further explain this increase in distraction.

Results from this study must be interpreted with caution because of limitations of the in vitro testing methodology. This biomechanical model using bone substitute characterizes glenoid fixation at time zero, directly after implantation, followed by repetitive cyclic loading simulating 5 years of implant service. This differs from the clinical scenario in which the shoulder undergoes postoperative immobilization or protected motion during which the early phases of bony remodeling are likely occurring. Furthermore, simulation of 5 years of implant service may not be necessary for an implant that is expected to achieve ultimate fixation by bony ingrowth within the first several months after implantation. Use of this implant without cement is classified off-label, and surgeons should take this into consideration during implantation. Last, this study could not simulate the effect of bony ingrowth on fixation, though our experimental technique of cementing the central peg may be a gross approximation of a fully ingrown central peg and its expected rigid fixation.

Fully cemented fixation of a polyethylene glenoid is superior to hybrid cement fixation and interference-fit fixation with respect to early glenoid micromotion. However, the long-term stability of a fully cemented polyethylene glenoid component remains a clinical concern, as fixation is achieved by bony ingrowth around the central fluted peg of the implant. In this study, interference-fit and hybrid fixation had equivocal component micromotion in biomechanical testing. Our findings suggest that cementation of the peripheral pegs confers no additional initial stability over an uncemented interference-fit technique in a biomechanical model. More research is needed to further evaluate interference-fit fixation as a viable option for implantation of a central fluted, all-polyethylene glenoid component.

From the Departments of Psychology (Ms. Zhao, Drs. Russell, Wesley, and Porter) and Hematology (Mss. Johnson and Pullen, Dr. Hankins), St. Jude Children’s Research Hospital, Memphis, TN.

Abstract

• Background: Children with sickle cell disease (SCD) are surviving into adulthood. Mastery of disease knowledge may facilitate treatment continuity in adult care.
• Objective: To assess the accuracy and extent of medical history knowledge among adolescents with SCD through the use of a personal health record (PHR) form.
• Methods: 68 adolescent patients with SCD (52.9% male; mean age, 16.8 years; 100% African American) completed a PHR listing significant prior medical events (eg, disease complications, diagnostic evaluations, treatments). Responses were compared against participants’ electronic medical record. An agreement percentage was calculated to determine accuracy of knowledge.
• Results: Most adolescents correctly reported their sickle cell genotype (100%), usage of penicillin (97.1%), prior hospitalizations (96.5%), history of prior blood transfusions (93.8%), usage of hydroxyurea (88.2%), and allergies (85.2%). Fewer adolescents accurately reported usage of opioids (52.9%), prior acute chest syndrome events (50.9%), baseline hemoglobin (41.8%), and hepatitis (43.3%), pneumovax (30.2%), and menactra (14.5%) vaccinations.
• Conclusion: Adolescents are aware of most but not all aspects of their medical history. The present findings can inform areas of knowledge deficits. Future targeted interventions for transition education and preparation may be tailored based on individual disease knowledge.

Sickle cell disease (SCD) is a genetic disorder characterized by abnormal sickle hemoglobin resulting in chronic hemolytic anemia and vaso-occlusion [1]. More than 95% of children with SCD in the United States survive into adulthood; however, young adults (YAs) are at risk for mortality shortly after transfer to adult health care [2–5]. Specifically, YAs with SCD (ages 18 to 30) have increased hospital utilization, emergency department visits, and mortality compared to other age-groups [4–7]. During this critical period, transition preparation that includes improving disease literacy and ensuring medical history knowledge may be necessary for optimal outcomes.

In the extant YA literature, significant gaps in medical history knowledge during the transition period were observed in pediatric cancer and inflammatory bowel disease patients [8,9]. YAs often require multidisciplinary management of their chronic disease complications [10]. Therefore, possessing comprehensive knowledge of personal health history may facilitate communication with different adult care providers and promote continuity of care. In the SCD transition literature, transition readiness measures have been developed to assess several aspects of knowledge, including medical and disease knowledge; however, these measures are primarily self-reported perceptions of knowledge and do not evaluate the accuracy of knowledge [11,12]. The current pilot study addresses this gap with the aim of assessing medical history knowledge accuracy in adolescents with SCD.

Methods

Participants

From March 2011 to January 2014, adolescents (aged 15–18 years) with SCD (any genotype) were approached during their regular health maintenance visits by hematology social workers. They were invited to complete the Personal Health Record (PHR) as an implementation effort of transition preparation within our pediatric SCD program.

Personal Health Record

The PHR was developed through literature review and discussions with area adult hematologists. The form was modeled after first visit intake forms used in adult hematology clinics. It was reviewed by the hematology medical team and the institution’s patient education committee. Prior to implementation, the form was piloted to obtain patient feedback on format and content. The PHR consists of 33 questions with 168 possible items/data points covering 12 domains: personal information (eg, contact information, SCD genotype), health provider information, personal health history (ie, health diagnoses), blood transfusion history, sickle cell pain events, hospitalization history in the previous year, diagnostic testing history (eg, laboratory tests), current medications, immunizations, advance directives, resource information (eg, disability benefits), and activities of daily living. Some questions required patients to check “Yes” or “No” (eg, “Have you been hospitalized in the past year? Have you received flu vaccine?”) while some required a written response (eg, “What medicines do you currently take?”).

Adolescents were instructed to complete the PHR independently without the help of their caregivers. After completing the form, the social worker reviewed the answers and/or asked participants’ perspectives about communicating health information to providers. A copy of the completed PHR was provided to the adolescent to promote continued education regarding medical history knowledge. The retrospective review of the PHR answers and participants’ characteristics was approved by the institutional review board with a waiver of consent from participants.

Statistical Methods

PHR answers were compared with each individual’s electronic medical record (EMR) for accuracy of responses. PHR responses were considered accurate only if they matched the information in the EMR. PHR items absent in the EMR were not coded (inability to verify the accuracy of responses) to capture the most accurate depiction of adolescents’ medical history knowledge. Coding was checked by at least 2 coders for response accuracy. Due to lack of EMR information for certain items, we could not verify the accuracy of many PHR items. Therefore, only items with at least 75% of data verified (across all patients who completed the PHR) were included in subsequent analyses.

Using SPSS (version 18), an agreement percentage was calculated for each patient across verifiable items and used as the primary outcome measure of knowledge accuracy. We used t tests to investigate gender or genotype differences in medical history accuracy. To examine genotype differences, we stratified the sample by SCD genotype: HbSS/Sβ⁰ thalassemia and HbSC/Sβ+ thalassemia [13].

Results

Patient Characteristics

Knowledge Accuracy Among Adolescents with SCD

Seventeen items in 6 PHR domains had the highest number of data points (at least 75% verified), and therefore were the only items that could be analyzed. Analyzed items included information about sickle cell genotype, eye doctor care, comorbid health issues (eg, asthma), allergies, hospitalizations, surgeries, transfusions, acute chest syndrome (ACS) episodes, eye problems, baseline hemoglobin level, and vaccination history as well as adolescents’ knowledge of current medications, including hydroxyurea, penicillin, and opioid pain medications.

Gender was not significantly associated with overall accuracy (P = 0.36). A significant difference was found in sickle genotype such that individuals with HbSC/Sβ+ thalassemia genotype (mean number of items, 8.23; SD = 1.70) were more accurate reporters of their medical history than those with HbSS/Sβ⁰ thalassemia genotype (mean number of items, 7.14; SD = 1.75; t(65) = –2.59, P = 0.01). Specifically, those with HbSS/Sβ⁰ thalassemia genotype were significantly less accurate reporters of vaccination history (meningococcus t(60) = 3.55, P = 0.001; pneumococcus t(60) = 2.46,  P = 0.02; hepatitis t(64) = 2.18, P = 0.03, eye problems t(62) = 3.62; P = 0.001, and surgical history t(62) = 2.14, P = 0.04).

Discussion

In the present study, we utilized the PHR to assess the accuracy of medical history knowledge of adolescents with SCD preparing to transition to adult care. Most adolescents were accurate reporters of important disease-relevant information (eg, genotype, transfusion history, hydroxyurea use), which may be a result of these topics being frequently discussed or recently encountered. For example, 97% of adolescents accurately reported penicillin use which may be related to our program’s emphasis on infection prevention education. However, disease knowledge of immunization history, prior ACS events, and opioid medication use might have been more difficult to recall due to the long interval from their occurrence until the completion of the PHR. Further, frequent changes in opioid medication use may have impacted the accuracy of adolescents’ answers with EMR data.

Individuals with HbSC/Sβ+ thalassemia genotype were more accurate reporters of their medical history, but the magnitude of difference was not large. These individuals tend to have fewer health issues and therefore less health information to recall, leading to higher accuracy. Furthermore, evidence demonstrates that individuals with HbSS/Sβ⁰ thalassemia genotype are at greater risk for cerebrovascular events and subsequent cognitive deficits [14], leading to more memory deficits and difficulty understanding and retaining health information [15]. The results suggest that patient health literacy, or an individual’s capacity to understand basic health information [16], may be a mediating factor in assessing for transition readiness. This is especially important given SCD risk for cognitive deficits [17].

Only 17 PHR items were analyzed due to conservative selection of items. Thus the present findings are not representative of the entire medical history. Additionally, the accuracy of medical history knowledge results may be limited by conservatism with abstracting information from the EMR (PHR information was considered accurate if it matched the information found in their EMR). Finally, we did not systematically assess the feasibility and utility of the PHR; ongoing participant feedback would aid in improving the PHR tool and implementation. It would be important to validate the PHR in a larger sample. However, our study is the first to our knowledge to systematically evaluate medical history knowledge among youth with SCD.

Conclusion and Practice Implications

The present study demonstrates that use of the PHR during regular health maintenance visits can help identify gaps in knowledge among adolescents with SCD who are approaching transfer to adult care. Sufficient knowledge of one’s medical history is an important aspect in transition preparation as it can facilitate the communication of medical information, thereby ensuring continuity of care [18,19]. The PHR could be used to teach medical history knowledge, assess a patient’s level of transition readiness at different time points, and identify areas for further targeted intervention. Knowledge tools, such as the PHR, can be investigated prospectively to assess the association of disease literacy and clinical outcomes, serving as a possible predictive instrument for transition health outcomes.

Corresponding author: Jerlym S. Porter, PhD, MPH, St. Jude Children’s Research Hospital, Dept. of Psychology, 262 Danny Thomas Pl., Mail Stop 740, Memphis, TN 38105, jerlym.porter@stjude.org.

Funding/support: This work was supported in part by HRSA grant 6 U1EMC19331-03-02 (PI: Hankins).

Financial disclosures: None.

Author contributions: conception and design, MJ, AP, KMW, JSH, JSP; analysis and interpretation of data, MSZ, KMR, JSP; drafting of article, MSZ, JSP; critical revision of the article, MSZ, MJ, AP, KMW, JSH, JSP; provision of study materials or patients, MJ, AP; statistical expertise, KMR; obtaining of funding, JSH; collection and assembly of data, MSZ, MJ, AP, KMR, KMW.

From the Departments of Psychology (Ms. Zhao, Drs. Russell, Wesley, and Porter) and Hematology (Mss. Johnson and Pullen, Dr. Hankins), St. Jude Children’s Research Hospital, Memphis, TN.

Abstract

• Background: Children with sickle cell disease (SCD) are surviving into adulthood. Mastery of disease knowledge may facilitate treatment continuity in adult care.
• Objective: To assess the accuracy and extent of medical history knowledge among adolescents with SCD through the use of a personal health record (PHR) form.
• Methods: 68 adolescent patients with SCD (52.9% male; mean age, 16.8 years; 100% African American) completed a PHR listing significant prior medical events (eg, disease complications, diagnostic evaluations, treatments). Responses were compared against participants’ electronic medical record. An agreement percentage was calculated to determine accuracy of knowledge.
• Results: Most adolescents correctly reported their sickle cell genotype (100%), usage of penicillin (97.1%), prior hospitalizations (96.5%), history of prior blood transfusions (93.8%), usage of hydroxyurea (88.2%), and allergies (85.2%). Fewer adolescents accurately reported usage of opioids (52.9%), prior acute chest syndrome events (50.9%), baseline hemoglobin (41.8%), and hepatitis (43.3%), pneumovax (30.2%), and menactra (14.5%) vaccinations.
• Conclusion: Adolescents are aware of most but not all aspects of their medical history. The present findings can inform areas of knowledge deficits. Future targeted interventions for transition education and preparation may be tailored based on individual disease knowledge.

Sickle cell disease (SCD) is a genetic disorder characterized by abnormal sickle hemoglobin resulting in chronic hemolytic anemia and vaso-occlusion [1]. More than 95% of children with SCD in the United States survive into adulthood; however, young adults (YAs) are at risk for mortality shortly after transfer to adult health care [2–5]. Specifically, YAs with SCD (ages 18 to 30) have increased hospital utilization, emergency department visits, and mortality compared to other age-groups [4–7]. During this critical period, transition preparation that includes improving disease literacy and ensuring medical history knowledge may be necessary for optimal outcomes.

In the extant YA literature, significant gaps in medical history knowledge during the transition period were observed in pediatric cancer and inflammatory bowel disease patients [8,9]. YAs often require multidisciplinary management of their chronic disease complications [10]. Therefore, possessing comprehensive knowledge of personal health history may facilitate communication with different adult care providers and promote continuity of care. In the SCD transition literature, transition readiness measures have been developed to assess several aspects of knowledge, including medical and disease knowledge; however, these measures are primarily self-reported perceptions of knowledge and do not evaluate the accuracy of knowledge [11,12]. The current pilot study addresses this gap with the aim of assessing medical history knowledge accuracy in adolescents with SCD.

Methods

Participants

From March 2011 to January 2014, adolescents (aged 15–18 years) with SCD (any genotype) were approached during their regular health maintenance visits by hematology social workers. They were invited to complete the Personal Health Record (PHR) as an implementation effort of transition preparation within our pediatric SCD program.

Personal Health Record

The PHR was developed through literature review and discussions with area adult hematologists. The form was modeled after first visit intake forms used in adult hematology clinics. It was reviewed by the hematology medical team and the institution’s patient education committee. Prior to implementation, the form was piloted to obtain patient feedback on format and content. The PHR consists of 33 questions with 168 possible items/data points covering 12 domains: personal information (eg, contact information, SCD genotype), health provider information, personal health history (ie, health diagnoses), blood transfusion history, sickle cell pain events, hospitalization history in the previous year, diagnostic testing history (eg, laboratory tests), current medications, immunizations, advance directives, resource information (eg, disability benefits), and activities of daily living. Some questions required patients to check “Yes” or “No” (eg, “Have you been hospitalized in the past year? Have you received flu vaccine?”) while some required a written response (eg, “What medicines do you currently take?”).

Adolescents were instructed to complete the PHR independently without the help of their caregivers. After completing the form, the social worker reviewed the answers and/or asked participants’ perspectives about communicating health information to providers. A copy of the completed PHR was provided to the adolescent to promote continued education regarding medical history knowledge. The retrospective review of the PHR answers and participants’ characteristics was approved by the institutional review board with a waiver of consent from participants.

Statistical Methods

PHR answers were compared with each individual’s electronic medical record (EMR) for accuracy of responses. PHR responses were considered accurate only if they matched the information in the EMR. PHR items absent in the EMR were not coded (inability to verify the accuracy of responses) to capture the most accurate depiction of adolescents’ medical history knowledge. Coding was checked by at least 2 coders for response accuracy. Due to lack of EMR information for certain items, we could not verify the accuracy of many PHR items. Therefore, only items with at least 75% of data verified (across all patients who completed the PHR) were included in subsequent analyses.

Using SPSS (version 18), an agreement percentage was calculated for each patient across verifiable items and used as the primary outcome measure of knowledge accuracy. We used t tests to investigate gender or genotype differences in medical history accuracy. To examine genotype differences, we stratified the sample by SCD genotype: HbSS/Sβ⁰ thalassemia and HbSC/Sβ+ thalassemia [13].

Results

Patient Characteristics

Knowledge Accuracy Among Adolescents with SCD

Seventeen items in 6 PHR domains had the highest number of data points (at least 75% verified), and therefore were the only items that could be analyzed. Analyzed items included information about sickle cell genotype, eye doctor care, comorbid health issues (eg, asthma), allergies, hospitalizations, surgeries, transfusions, acute chest syndrome (ACS) episodes, eye problems, baseline hemoglobin level, and vaccination history as well as adolescents’ knowledge of current medications, including hydroxyurea, penicillin, and opioid pain medications.

Gender was not significantly associated with overall accuracy (P = 0.36). A significant difference was found in sickle genotype such that individuals with HbSC/Sβ+ thalassemia genotype (mean number of items, 8.23; SD = 1.70) were more accurate reporters of their medical history than those with HbSS/Sβ⁰ thalassemia genotype (mean number of items, 7.14; SD = 1.75; t(65) = –2.59, P = 0.01). Specifically, those with HbSS/Sβ⁰ thalassemia genotype were significantly less accurate reporters of vaccination history (meningococcus t(60) = 3.55, P = 0.001; pneumococcus t(60) = 2.46,  P = 0.02; hepatitis t(64) = 2.18, P = 0.03, eye problems t(62) = 3.62; P = 0.001, and surgical history t(62) = 2.14, P = 0.04).

Discussion

In the present study, we utilized the PHR to assess the accuracy of medical history knowledge of adolescents with SCD preparing to transition to adult care. Most adolescents were accurate reporters of important disease-relevant information (eg, genotype, transfusion history, hydroxyurea use), which may be a result of these topics being frequently discussed or recently encountered. For example, 97% of adolescents accurately reported penicillin use which may be related to our program’s emphasis on infection prevention education. However, disease knowledge of immunization history, prior ACS events, and opioid medication use might have been more difficult to recall due to the long interval from their occurrence until the completion of the PHR. Further, frequent changes in opioid medication use may have impacted the accuracy of adolescents’ answers with EMR data.

Individuals with HbSC/Sβ+ thalassemia genotype were more accurate reporters of their medical history, but the magnitude of difference was not large. These individuals tend to have fewer health issues and therefore less health information to recall, leading to higher accuracy. Furthermore, evidence demonstrates that individuals with HbSS/Sβ⁰ thalassemia genotype are at greater risk for cerebrovascular events and subsequent cognitive deficits [14], leading to more memory deficits and difficulty understanding and retaining health information [15]. The results suggest that patient health literacy, or an individual’s capacity to understand basic health information [16], may be a mediating factor in assessing for transition readiness. This is especially important given SCD risk for cognitive deficits [17].

Only 17 PHR items were analyzed due to conservative selection of items. Thus the present findings are not representative of the entire medical history. Additionally, the accuracy of medical history knowledge results may be limited by conservatism with abstracting information from the EMR (PHR information was considered accurate if it matched the information found in their EMR). Finally, we did not systematically assess the feasibility and utility of the PHR; ongoing participant feedback would aid in improving the PHR tool and implementation. It would be important to validate the PHR in a larger sample. However, our study is the first to our knowledge to systematically evaluate medical history knowledge among youth with SCD.

Conclusion and Practice Implications

The present study demonstrates that use of the PHR during regular health maintenance visits can help identify gaps in knowledge among adolescents with SCD who are approaching transfer to adult care. Sufficient knowledge of one’s medical history is an important aspect in transition preparation as it can facilitate the communication of medical information, thereby ensuring continuity of care [18,19]. The PHR could be used to teach medical history knowledge, assess a patient’s level of transition readiness at different time points, and identify areas for further targeted intervention. Knowledge tools, such as the PHR, can be investigated prospectively to assess the association of disease literacy and clinical outcomes, serving as a possible predictive instrument for transition health outcomes.

Corresponding author: Jerlym S. Porter, PhD, MPH, St. Jude Children’s Research Hospital, Dept. of Psychology, 262 Danny Thomas Pl., Mail Stop 740, Memphis, TN 38105, jerlym.porter@stjude.org.

Funding/support: This work was supported in part by HRSA grant 6 U1EMC19331-03-02 (PI: Hankins).

Financial disclosures: None.

Author contributions: conception and design, MJ, AP, KMW, JSH, JSP; analysis and interpretation of data, MSZ, KMR, JSP; drafting of article, MSZ, JSP; critical revision of the article, MSZ, MJ, AP, KMW, JSH, JSP; provision of study materials or patients, MJ, AP; statistical expertise, KMR; obtaining of funding, JSH; collection and assembly of data, MSZ, MJ, AP, KMR, KMW.

From the Departments of Psychology (Ms. Zhao, Drs. Russell, Wesley, and Porter) and Hematology (Mss. Johnson and Pullen, Dr. Hankins), St. Jude Children’s Research Hospital, Memphis, TN.

Abstract

• Background: Children with sickle cell disease (SCD) are surviving into adulthood. Mastery of disease knowledge may facilitate treatment continuity in adult care.
• Objective: To assess the accuracy and extent of medical history knowledge among adolescents with SCD through the use of a personal health record (PHR) form.
• Methods: 68 adolescent patients with SCD (52.9% male; mean age, 16.8 years; 100% African American) completed a PHR listing significant prior medical events (eg, disease complications, diagnostic evaluations, treatments). Responses were compared against participants’ electronic medical record. An agreement percentage was calculated to determine accuracy of knowledge.
• Results: Most adolescents correctly reported their sickle cell genotype (100%), usage of penicillin (97.1%), prior hospitalizations (96.5%), history of prior blood transfusions (93.8%), usage of hydroxyurea (88.2%), and allergies (85.2%). Fewer adolescents accurately reported usage of opioids (52.9%), prior acute chest syndrome events (50.9%), baseline hemoglobin (41.8%), and hepatitis (43.3%), pneumovax (30.2%), and menactra (14.5%) vaccinations.
• Conclusion: Adolescents are aware of most but not all aspects of their medical history. The present findings can inform areas of knowledge deficits. Future targeted interventions for transition education and preparation may be tailored based on individual disease knowledge.

Sickle cell disease (SCD) is a genetic disorder characterized by abnormal sickle hemoglobin resulting in chronic hemolytic anemia and vaso-occlusion [1]. More than 95% of children with SCD in the United States survive into adulthood; however, young adults (YAs) are at risk for mortality shortly after transfer to adult health care [2–5]. Specifically, YAs with SCD (ages 18 to 30) have increased hospital utilization, emergency department visits, and mortality compared to other age-groups [4–7]. During this critical period, transition preparation that includes improving disease literacy and ensuring medical history knowledge may be necessary for optimal outcomes.

In the extant YA literature, significant gaps in medical history knowledge during the transition period were observed in pediatric cancer and inflammatory bowel disease patients [8,9]. YAs often require multidisciplinary management of their chronic disease complications [10]. Therefore, possessing comprehensive knowledge of personal health history may facilitate communication with different adult care providers and promote continuity of care. In the SCD transition literature, transition readiness measures have been developed to assess several aspects of knowledge, including medical and disease knowledge; however, these measures are primarily self-reported perceptions of knowledge and do not evaluate the accuracy of knowledge [11,12]. The current pilot study addresses this gap with the aim of assessing medical history knowledge accuracy in adolescents with SCD.

Methods

Participants

From March 2011 to January 2014, adolescents (aged 15–18 years) with SCD (any genotype) were approached during their regular health maintenance visits by hematology social workers. They were invited to complete the Personal Health Record (PHR) as an implementation effort of transition preparation within our pediatric SCD program.

Personal Health Record

The PHR was developed through literature review and discussions with area adult hematologists. The form was modeled after first visit intake forms used in adult hematology clinics. It was reviewed by the hematology medical team and the institution’s patient education committee. Prior to implementation, the form was piloted to obtain patient feedback on format and content. The PHR consists of 33 questions with 168 possible items/data points covering 12 domains: personal information (eg, contact information, SCD genotype), health provider information, personal health history (ie, health diagnoses), blood transfusion history, sickle cell pain events, hospitalization history in the previous year, diagnostic testing history (eg, laboratory tests), current medications, immunizations, advance directives, resource information (eg, disability benefits), and activities of daily living. Some questions required patients to check “Yes” or “No” (eg, “Have you been hospitalized in the past year? Have you received flu vaccine?”) while some required a written response (eg, “What medicines do you currently take?”).

Adolescents were instructed to complete the PHR independently without the help of their caregivers. After completing the form, the social worker reviewed the answers and/or asked participants’ perspectives about communicating health information to providers. A copy of the completed PHR was provided to the adolescent to promote continued education regarding medical history knowledge. The retrospective review of the PHR answers and participants’ characteristics was approved by the institutional review board with a waiver of consent from participants.

Statistical Methods

PHR answers were compared with each individual’s electronic medical record (EMR) for accuracy of responses. PHR responses were considered accurate only if they matched the information in the EMR. PHR items absent in the EMR were not coded (inability to verify the accuracy of responses) to capture the most accurate depiction of adolescents’ medical history knowledge. Coding was checked by at least 2 coders for response accuracy. Due to lack of EMR information for certain items, we could not verify the accuracy of many PHR items. Therefore, only items with at least 75% of data verified (across all patients who completed the PHR) were included in subsequent analyses.

Using SPSS (version 18), an agreement percentage was calculated for each patient across verifiable items and used as the primary outcome measure of knowledge accuracy. We used t tests to investigate gender or genotype differences in medical history accuracy. To examine genotype differences, we stratified the sample by SCD genotype: HbSS/Sβ⁰ thalassemia and HbSC/Sβ+ thalassemia [13].

Results

Patient Characteristics

Knowledge Accuracy Among Adolescents with SCD

Seventeen items in 6 PHR domains had the highest number of data points (at least 75% verified), and therefore were the only items that could be analyzed. Analyzed items included information about sickle cell genotype, eye doctor care, comorbid health issues (eg, asthma), allergies, hospitalizations, surgeries, transfusions, acute chest syndrome (ACS) episodes, eye problems, baseline hemoglobin level, and vaccination history as well as adolescents’ knowledge of current medications, including hydroxyurea, penicillin, and opioid pain medications.

Gender was not significantly associated with overall accuracy (P = 0.36). A significant difference was found in sickle genotype such that individuals with HbSC/Sβ+ thalassemia genotype (mean number of items, 8.23; SD = 1.70) were more accurate reporters of their medical history than those with HbSS/Sβ⁰ thalassemia genotype (mean number of items, 7.14; SD = 1.75; t(65) = –2.59, P = 0.01). Specifically, those with HbSS/Sβ⁰ thalassemia genotype were significantly less accurate reporters of vaccination history (meningococcus t(60) = 3.55, P = 0.001; pneumococcus t(60) = 2.46,  P = 0.02; hepatitis t(64) = 2.18, P = 0.03, eye problems t(62) = 3.62; P = 0.001, and surgical history t(62) = 2.14, P = 0.04).

Discussion

In the present study, we utilized the PHR to assess the accuracy of medical history knowledge of adolescents with SCD preparing to transition to adult care. Most adolescents were accurate reporters of important disease-relevant information (eg, genotype, transfusion history, hydroxyurea use), which may be a result of these topics being frequently discussed or recently encountered. For example, 97% of adolescents accurately reported penicillin use which may be related to our program’s emphasis on infection prevention education. However, disease knowledge of immunization history, prior ACS events, and opioid medication use might have been more difficult to recall due to the long interval from their occurrence until the completion of the PHR. Further, frequent changes in opioid medication use may have impacted the accuracy of adolescents’ answers with EMR data.

Individuals with HbSC/Sβ+ thalassemia genotype were more accurate reporters of their medical history, but the magnitude of difference was not large. These individuals tend to have fewer health issues and therefore less health information to recall, leading to higher accuracy. Furthermore, evidence demonstrates that individuals with HbSS/Sβ⁰ thalassemia genotype are at greater risk for cerebrovascular events and subsequent cognitive deficits [14], leading to more memory deficits and difficulty understanding and retaining health information [15]. The results suggest that patient health literacy, or an individual’s capacity to understand basic health information [16], may be a mediating factor in assessing for transition readiness. This is especially important given SCD risk for cognitive deficits [17].

Only 17 PHR items were analyzed due to conservative selection of items. Thus the present findings are not representative of the entire medical history. Additionally, the accuracy of medical history knowledge results may be limited by conservatism with abstracting information from the EMR (PHR information was considered accurate if it matched the information found in their EMR). Finally, we did not systematically assess the feasibility and utility of the PHR; ongoing participant feedback would aid in improving the PHR tool and implementation. It would be important to validate the PHR in a larger sample. However, our study is the first to our knowledge to systematically evaluate medical history knowledge among youth with SCD.

Conclusion and Practice Implications

The present study demonstrates that use of the PHR during regular health maintenance visits can help identify gaps in knowledge among adolescents with SCD who are approaching transfer to adult care. Sufficient knowledge of one’s medical history is an important aspect in transition preparation as it can facilitate the communication of medical information, thereby ensuring continuity of care [18,19]. The PHR could be used to teach medical history knowledge, assess a patient’s level of transition readiness at different time points, and identify areas for further targeted intervention. Knowledge tools, such as the PHR, can be investigated prospectively to assess the association of disease literacy and clinical outcomes, serving as a possible predictive instrument for transition health outcomes.

Corresponding author: Jerlym S. Porter, PhD, MPH, St. Jude Children’s Research Hospital, Dept. of Psychology, 262 Danny Thomas Pl., Mail Stop 740, Memphis, TN 38105, jerlym.porter@stjude.org.

Funding/support: This work was supported in part by HRSA grant 6 U1EMC19331-03-02 (PI: Hankins).

Financial disclosures: None.

Author contributions: conception and design, MJ, AP, KMW, JSH, JSP; analysis and interpretation of data, MSZ, KMR, JSP; drafting of article, MSZ, JSP; critical revision of the article, MSZ, MJ, AP, KMW, JSH, JSP; provision of study materials or patients, MJ, AP; statistical expertise, KMR; obtaining of funding, JSH; collection and assembly of data, MSZ, MJ, AP, KMR, KMW.

Shoulder arthroplasty (SA), including total SA (TSA) and reverse TSA, is an effective surgical treatment for fracture and primary or secondary degenerative disease of the shoulder.¹ Over the past few decades, use of SA has increased dramatically, from about 5000 cases in 1990 to 7000 in 2000 and more than 26,000 in 2008.^1,2

Complications associated with SA generally are classified as perioperative (occurring during the operative index) or long-term (postdischarge).³ Long-term complications include implant loosening, instability, revision, infection, rotator cuff tear, neural injury, and deltoid detachment.^1,4,5 Perioperative complications, which are less commonly reported, include intraoperative fracture, infection, neural injury, venous thromboembolic events (VTEs, including pulmonary embolism [PE] and deep vein thrombosis [DVT]), transfusion, and death.^3,6-10

SA is an attractive treatment option for patients with rheumatoid arthritis (RA), as the effects of pain on these patients are greater in the shoulder joint than in any other joint.¹¹ Patients with RA pose unique orthopedic surgical challenges, including any combination of decreased bone mineralization, poor capsular tissue integrity, and osteonecrosis.^3,12 In addition, RA patients may be taking immunosuppressive medications that have severe side effects, and they may require multiple surgeries.^12,13 These factors predispose patients with RA to complications that include infection and wound dehiscence.^3,5,12-14

The complex nature of RA has prompted investigators to examine outcome measures in this patient group. Hambright and colleagues³ used the Nationwide Inpatient Sample (NIS) to examine perioperative outcomes in RA patients who underwent TSA between 1988 and 2005.³ They found that TSA patients with RA had shorter and less costly hospital stays and were more likely to have a routine discharge.³ Using the same patient population drawn from the period 2006–2011, we conducted a study to determine if this unexpected trend persists as the number of TSAs and quality of postoperative care continue to increase. Given the potential for anemia of chronic disease and the systemic inflammatory nature of RA, we hypothesized that the perioperative complication profile of RA patients would be worse than that of non-RA patients.

Materials and Methods

NIS data were acquired for the period 2006–2011. The NIS is the largest publicly available all-payer inpatient database, with a random 20% sample of about 1000 US hospitals accounting for 7 to 8 million inpatient stays. The database supplies weights used to estimate national totals, at about 35 million inpatient visits per year. NIS inpatient data are limited to the operative index. Postdischarge information is not available. The NIS is managed by the Healthcare Cost and Utilization Project, which is sponsored by the Agency for Healthcare Research and Quality. The quality of NIS data is assessed and validated by an independent contractor. NIS data have been widely used to examine perioperative outcomes.^15-17

NIS data cover patient and hospital demographics, hospital length of stay (LOS), discharge status, payer information, charges, and perioperative outcomes and procedure/diagnosis codes (ICD-9; International Classification of Diseases, Ninth Revision¹⁸).

As our Institutional Review Board (IRB) reviewed the database and determined the project was not human subject research, IRB involvement was not required. This study paralleled successful efforts with similar RA and non-RA patients who had shoulder and elbow surgery.^3,19 SA patients were identified by ICD-9 procedure code 81.80, but this code does not specify whether the prosthesis was unconstrained, semiconstrained, or constrained. ICD-9 coding also does not specify whether the TSA was traditional or reverse. Patients with RA were identified by ICD-9 diagnosis codes 714.0, 714.1, and 714.2. Patients without one of these codes were placed in the non-RA cohort. Patients with codes associated with pathologic fractures secondary to metastatic cancer or bone malignant neoplasm as a secondary or primary diagnosis and patients who had revision surgery indicated by code 81.83 were excluded, as they have a disproportionately higher comorbidity burden.

After each cohort was defined, demographic data (age, sex, race, income quartile based on ZIP postal code) were compared, as were data on primary payer, hospital demographics, LOS (≤5 days, defined as perioperative index), discharge type, inflation-adjusted charges in 2014 dollars based on the Consumer Price Indexes (http://www.bls.gov/cpi/), and mortality. Perioperative complications—respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related (including embolism, fibrosis, hemorrhage, pain, stenosis, or thrombus caused by any device, implant, or graft), cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, postoperative infection complications, and intraoperative transfusions—were considered using ICD-9 codes (996.X-999.X and 99.X, respectively).²⁰ Although commonly used to determine perioperative comorbidity burden using ICD-9 coding, the modified Charlson index was not considered because RA is a component of the index and would therefore bias the variable.^3,21

Statistical analyses, including χ² tests and 2-sample t tests, were performed for categorical and continuous variables, respectively. P < .05 was considered significant. Fisher exact test was used for cohorts with fewer than 5 occurrences. Multivariate logistic regression models were then calculated to determine the effect of RA on different outcomes and complications, with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. Statistical analyses were performed using the R statistical programming language.22

Results

Of the 34,970 patients who underwent SA between 2006 and 2011, 1674 (4.8%) had a diagnosis of RA and 33,296 (95.2%) did not. On average, patients with RA tended to be younger than patients without RA (66.4 vs 69.1 years; P < .001), and a larger percentage of RA patients were female (75.5% vs 54.4%; P < .001). Compared with non-RA patients, RA patients comprised a different ethnic group and had a different expected primary payer (P < .001). SA patients with and without RA did not differ in income quartile based on ZIP code, total number of hospital beds, hospital region, or hospital teaching status (P = .34, .78, .59, and .82, respectively) (Table 1).

LOS was significantly (P < .001) statistically longer for RA patients (2.196 days) than for non-RA patients (2.085 days). RA patients were significantly less likely to be discharged home (63.0% vs 67.6%; P < .001). (Routine discharge was defined as discharge home, whereas nonroutine discharge was defined as discharge to a short-term hospital, skilled nursing facility, intermediate care, another type of facility, home health care, against medical advice, or death.) In addition, inflation-adjusted charges associated with SA were significantly higher (P = .018) for RA patients ($54,284) than for non-RA patients ($52,663) (Table 1).

Regarding the rates of complications that occurred during the perioperative index, there were no significant differences between RA and non-RA cohorts. These complications included respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related, cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, and postoperative infection (Table 2). In addition, there was no significant difference in mortality between the groups (P = .48).

In TSA, blood transfusions were more likely (P < .001) to be given to RA patients (9.00%) than to non-RA patients (6.16%). Multivariate regression analyses were performed with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. These analyses revealed that transfusion (P < .001), discharge type (P = .002), total inflation-adjusted charges (P < .001), and LOS (P = .047) remained significant (Table 3).

Discussion

Large national databases like NIS allow study of uncommon medical occurrences and help delineate risks and trends that otherwise might be indeterminable. Although it has been suggested that patients with RA may have poorer long-term outcomes after SA, the perioperative risk profile indicates that TSA is well tolerated in RA patients during the operative index.^3,23-25

The data on this study’s 34,970 patients, drawn from the period 2006–2011, demonstrated no significant differences in safety profile with respect to the 14 perioperative complications and outcomes examined, except blood transfusion rate. Rates of postoperative infection (RA, 0.24%; non-RA, 0.14%; P = .303), VTE (RA, 0.30%; non-RA, 0.25%; P = .905), and transfusion (RA, 9.00%; non-RA, 6.16%; P < .001) are of particular interest because of the severity of these situations.

Postoperative infection is a potentially serious complication and often occurs secondary to diabetes, RA, lupus erythematosus, prior surgery, or a nosocomial or remote source.¹ The often costly treatment options include antibiotic suppression, irrigation and debridement with implant retention, 1-stage exchange with antibiotic-impregnated cement fixation, staged reimplantation, resection arthroplasty, arthrodesis, and amputation.¹ The overall 0.14% infection rate determined in this study is lower than the 0.7% reported for SA patients in the literature.¹ Given the nature of the NIS database, this rate underestimates the true postoperative infection rate, as any infection that occurred after the perioperative period is not captured.²⁶ The present study’s perioperative infection rates (RA, 0.24%; non-RA, 0.14%) for the period 2006–2011 are comparable to the rates (RA, 0.17%; non-RA, 0.24%) reported by Hambright and colleagues3 for the same patient population over the preceding, 18-year period (1988–2005) and similarly do not significantly differ between groups. Although infection is uncommon in the immediate perioperative period, the ICD-9 codes used refer specifically to infection resulting from surgery and do not represent concomitant infection.

VTEs, which include PEs and DVTs, are rare but potentially life-threatening surgical complications.^27,28 Mechanical prophylaxis and chemical prophylaxis have been recommended for major orthopedic surgery, particularly lower extremity surgery, such as total hip arthroplasty (THA) and total knee arthroplasty (TKA).^28,29 In the present study, VTE rates were low, 0.30% (RA) and 0.25% (non-RA), and not significantly different in bivariate or multivariate analyses. These rates are comparable to those found in other national-database SA studies.²⁸ VTEs that occur outside the index hospital admission are not captured in this database. Therefore, the rates in the present study may be lower than the true incidence after SA. Mortality secondary to VTE usually occurs within 24 hours but may occur up to 90 days after surgery. DVT rates, on the other hand, are difficult to evaluate because of differences in screening practices.^27,28,30,31

That RA patients were more likely than non-RA patients to receive perioperative blood transfusions supports prior findings that SA patients with RA were more likely than SA patients with osteoarthritis (OA) to receive perioperative blood transfusions.⁸ RA patients have been shown to have high rates of anemia of chronic disease, ranging from 22% to 77%.³² During joint replacement, these patients often require transfusions.^32,33 However, these findings differ from prior findings of no differences between RA and non-RA patients in the same patient population during the period 1988–2005.³ This difference may be a product of the constantly changing transfusion guidelines and increased use; transfusion rates increased 140% between 1997 and 2007, making transfusions the fastest growing common procedure in the United States during that time.³⁴ There was no difference between RA and non-RA patients in household income (as determined by ZIP code analysis), number of hospital beds, hospital region, or hospital teaching status. Compared with non-RA patients, RA patients were more likely to be younger, female, and of a difference race and to have a different expected primary payer (P < .001).These findings are consistent with previous findings in the literature.³ In the present SA study, however, RA patients were more likely than non-RA patients to have longer LOS, higher inflation-adjusted hospital charges, and nonroutine discharge. These findings deviate from those of the study covering the preceding 18 years (1988–2005).³ Despite the findings of a changing environment of care for RA patients, by Hambright and colleagues³ and Weiss and colleagues,³⁵ the trend appears to have shifted. Both groups had shorter average LOS than either group from the preceding 18 years.³ Although statistically significant in bivariate analysis, the difference in LOS between the 2 groups differed by an average of 0.11 day (2 hours 24 minutes) and was not clinically relevant.

In addition, the higher charges for patients with RA represent a deviation from the preceding 18 years.³ Other studies have also shown that RA is associated with increased cost in TSA.³⁶ Patients with RA often have rotator cuff pathology, indicating reverse SA may be used more frequently.^37,38 The increased implant cost associated with reverse SA may account for the increased costs in RA patients.³⁹ As mentioned, TSA type is not captured in the NIS database. In addition, that RA patients were less likely than non-RA patients to have routine discharge may indicate RA cases are more complex because of their complications.^1,5,14,40 A recent study of complications in RA patients (1163 who underwent THA, 2692 who underwent TKA) found that THA patients with RA were significantly more likely than THA patients with OA to dislocate, and TKA patients with RA were significantly more likely than TKA patients with OA to develop an infection after surgery.⁴¹ Postoperative dislocation has been shown to increase hospital costs in other orthopedic procedures.⁴² Also, during TSA, patients with RA are more likely than patients with OA to receive intraoperative blood transfusions.⁸ These complications—combined with the fact that RA is a chronic, progressive, systemic inflammatory disease that can affect soft tissue and blood vessel wall healing and is associated with medications having potential side effects—could contribute to the apparent increased hospital charges and LOS.^3,12,13,43 Factors that include surgeon preference, impact of primary payer, and hospital practice may also affect final charges. Total charges in the NIS database include administrative fees, hospital costs, device-related costs, operating room costs, and ancillary staff costs. Total charges do not include professional fees and differ from the total cost that represents the amount reimbursed by the payer. Charges tend to correlate with but overestimate the total costs.⁴⁴

This study had several important limitations. As mentioned, only events that occur during the operative admission are captured in the NIS database, and thus postoperative complications or serious adverse events that lead to readmission cannot be identified. In addition, outpatient TSAs are not captured in the NIS database, and thus inclusion of only inpatient procedures yields higher average LOS and total charges.⁴⁵ Given the limited granularity of ICD-9 coding, this study could not determine RA severity, estimated blood loss, length of surgery, complication severity, type of TSA procedure/prosthesis, or cause of death. Although commonly used to determine comorbidity burden, the modified Charlson index could not be used, and therefore could not be entered as a covariate in multivariate analysis. Furthermore, the NIS database does not include imaging or patient-reported outcomes information, such as improvements in pain or function, which are of crucial importance in considering surgery.

Conclusion

Our findings corroborated findings that the demographics and the perioperative safety profile for TSA were similar for patients with and without RA. The risk for complications or death in the perioperative period was low. Compared with non-RA patients, RA patients had significantly higher charges and longer LOS and were less likely to be discharged home after surgery. The 0.11-day difference in LOS, though statistically significant, was not clinically relevant. These findings differ from those for the preceding, 18-year period (1988–2005). Future research should focus on the causes of these changes.

Shoulder arthroplasty (SA), including total SA (TSA) and reverse TSA, is an effective surgical treatment for fracture and primary or secondary degenerative disease of the shoulder.¹ Over the past few decades, use of SA has increased dramatically, from about 5000 cases in 1990 to 7000 in 2000 and more than 26,000 in 2008.^1,2

Complications associated with SA generally are classified as perioperative (occurring during the operative index) or long-term (postdischarge).³ Long-term complications include implant loosening, instability, revision, infection, rotator cuff tear, neural injury, and deltoid detachment.^1,4,5 Perioperative complications, which are less commonly reported, include intraoperative fracture, infection, neural injury, venous thromboembolic events (VTEs, including pulmonary embolism [PE] and deep vein thrombosis [DVT]), transfusion, and death.^3,6-10

SA is an attractive treatment option for patients with rheumatoid arthritis (RA), as the effects of pain on these patients are greater in the shoulder joint than in any other joint.¹¹ Patients with RA pose unique orthopedic surgical challenges, including any combination of decreased bone mineralization, poor capsular tissue integrity, and osteonecrosis.^3,12 In addition, RA patients may be taking immunosuppressive medications that have severe side effects, and they may require multiple surgeries.^12,13 These factors predispose patients with RA to complications that include infection and wound dehiscence.^3,5,12-14

The complex nature of RA has prompted investigators to examine outcome measures in this patient group. Hambright and colleagues³ used the Nationwide Inpatient Sample (NIS) to examine perioperative outcomes in RA patients who underwent TSA between 1988 and 2005.³ They found that TSA patients with RA had shorter and less costly hospital stays and were more likely to have a routine discharge.³ Using the same patient population drawn from the period 2006–2011, we conducted a study to determine if this unexpected trend persists as the number of TSAs and quality of postoperative care continue to increase. Given the potential for anemia of chronic disease and the systemic inflammatory nature of RA, we hypothesized that the perioperative complication profile of RA patients would be worse than that of non-RA patients.

Materials and Methods

NIS data were acquired for the period 2006–2011. The NIS is the largest publicly available all-payer inpatient database, with a random 20% sample of about 1000 US hospitals accounting for 7 to 8 million inpatient stays. The database supplies weights used to estimate national totals, at about 35 million inpatient visits per year. NIS inpatient data are limited to the operative index. Postdischarge information is not available. The NIS is managed by the Healthcare Cost and Utilization Project, which is sponsored by the Agency for Healthcare Research and Quality. The quality of NIS data is assessed and validated by an independent contractor. NIS data have been widely used to examine perioperative outcomes.^15-17

NIS data cover patient and hospital demographics, hospital length of stay (LOS), discharge status, payer information, charges, and perioperative outcomes and procedure/diagnosis codes (ICD-9; International Classification of Diseases, Ninth Revision¹⁸).

As our Institutional Review Board (IRB) reviewed the database and determined the project was not human subject research, IRB involvement was not required. This study paralleled successful efforts with similar RA and non-RA patients who had shoulder and elbow surgery.^3,19 SA patients were identified by ICD-9 procedure code 81.80, but this code does not specify whether the prosthesis was unconstrained, semiconstrained, or constrained. ICD-9 coding also does not specify whether the TSA was traditional or reverse. Patients with RA were identified by ICD-9 diagnosis codes 714.0, 714.1, and 714.2. Patients without one of these codes were placed in the non-RA cohort. Patients with codes associated with pathologic fractures secondary to metastatic cancer or bone malignant neoplasm as a secondary or primary diagnosis and patients who had revision surgery indicated by code 81.83 were excluded, as they have a disproportionately higher comorbidity burden.

After each cohort was defined, demographic data (age, sex, race, income quartile based on ZIP postal code) were compared, as were data on primary payer, hospital demographics, LOS (≤5 days, defined as perioperative index), discharge type, inflation-adjusted charges in 2014 dollars based on the Consumer Price Indexes (http://www.bls.gov/cpi/), and mortality. Perioperative complications—respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related (including embolism, fibrosis, hemorrhage, pain, stenosis, or thrombus caused by any device, implant, or graft), cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, postoperative infection complications, and intraoperative transfusions—were considered using ICD-9 codes (996.X-999.X and 99.X, respectively).²⁰ Although commonly used to determine perioperative comorbidity burden using ICD-9 coding, the modified Charlson index was not considered because RA is a component of the index and would therefore bias the variable.^3,21

Statistical analyses, including χ² tests and 2-sample t tests, were performed for categorical and continuous variables, respectively. P < .05 was considered significant. Fisher exact test was used for cohorts with fewer than 5 occurrences. Multivariate logistic regression models were then calculated to determine the effect of RA on different outcomes and complications, with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. Statistical analyses were performed using the R statistical programming language.22

Results

Of the 34,970 patients who underwent SA between 2006 and 2011, 1674 (4.8%) had a diagnosis of RA and 33,296 (95.2%) did not. On average, patients with RA tended to be younger than patients without RA (66.4 vs 69.1 years; P < .001), and a larger percentage of RA patients were female (75.5% vs 54.4%; P < .001). Compared with non-RA patients, RA patients comprised a different ethnic group and had a different expected primary payer (P < .001). SA patients with and without RA did not differ in income quartile based on ZIP code, total number of hospital beds, hospital region, or hospital teaching status (P = .34, .78, .59, and .82, respectively) (Table 1).

LOS was significantly (P < .001) statistically longer for RA patients (2.196 days) than for non-RA patients (2.085 days). RA patients were significantly less likely to be discharged home (63.0% vs 67.6%; P < .001). (Routine discharge was defined as discharge home, whereas nonroutine discharge was defined as discharge to a short-term hospital, skilled nursing facility, intermediate care, another type of facility, home health care, against medical advice, or death.) In addition, inflation-adjusted charges associated with SA were significantly higher (P = .018) for RA patients ($54,284) than for non-RA patients ($52,663) (Table 1).

Regarding the rates of complications that occurred during the perioperative index, there were no significant differences between RA and non-RA cohorts. These complications included respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related, cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, and postoperative infection (Table 2). In addition, there was no significant difference in mortality between the groups (P = .48).

In TSA, blood transfusions were more likely (P < .001) to be given to RA patients (9.00%) than to non-RA patients (6.16%). Multivariate regression analyses were performed with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. These analyses revealed that transfusion (P < .001), discharge type (P = .002), total inflation-adjusted charges (P < .001), and LOS (P = .047) remained significant (Table 3).

Discussion

Large national databases like NIS allow study of uncommon medical occurrences and help delineate risks and trends that otherwise might be indeterminable. Although it has been suggested that patients with RA may have poorer long-term outcomes after SA, the perioperative risk profile indicates that TSA is well tolerated in RA patients during the operative index.^3,23-25

The data on this study’s 34,970 patients, drawn from the period 2006–2011, demonstrated no significant differences in safety profile with respect to the 14 perioperative complications and outcomes examined, except blood transfusion rate. Rates of postoperative infection (RA, 0.24%; non-RA, 0.14%; P = .303), VTE (RA, 0.30%; non-RA, 0.25%; P = .905), and transfusion (RA, 9.00%; non-RA, 6.16%; P < .001) are of particular interest because of the severity of these situations.

Postoperative infection is a potentially serious complication and often occurs secondary to diabetes, RA, lupus erythematosus, prior surgery, or a nosocomial or remote source.¹ The often costly treatment options include antibiotic suppression, irrigation and debridement with implant retention, 1-stage exchange with antibiotic-impregnated cement fixation, staged reimplantation, resection arthroplasty, arthrodesis, and amputation.¹ The overall 0.14% infection rate determined in this study is lower than the 0.7% reported for SA patients in the literature.¹ Given the nature of the NIS database, this rate underestimates the true postoperative infection rate, as any infection that occurred after the perioperative period is not captured.²⁶ The present study’s perioperative infection rates (RA, 0.24%; non-RA, 0.14%) for the period 2006–2011 are comparable to the rates (RA, 0.17%; non-RA, 0.24%) reported by Hambright and colleagues3 for the same patient population over the preceding, 18-year period (1988–2005) and similarly do not significantly differ between groups. Although infection is uncommon in the immediate perioperative period, the ICD-9 codes used refer specifically to infection resulting from surgery and do not represent concomitant infection.

VTEs, which include PEs and DVTs, are rare but potentially life-threatening surgical complications.^27,28 Mechanical prophylaxis and chemical prophylaxis have been recommended for major orthopedic surgery, particularly lower extremity surgery, such as total hip arthroplasty (THA) and total knee arthroplasty (TKA).^28,29 In the present study, VTE rates were low, 0.30% (RA) and 0.25% (non-RA), and not significantly different in bivariate or multivariate analyses. These rates are comparable to those found in other national-database SA studies.²⁸ VTEs that occur outside the index hospital admission are not captured in this database. Therefore, the rates in the present study may be lower than the true incidence after SA. Mortality secondary to VTE usually occurs within 24 hours but may occur up to 90 days after surgery. DVT rates, on the other hand, are difficult to evaluate because of differences in screening practices.^27,28,30,31

That RA patients were more likely than non-RA patients to receive perioperative blood transfusions supports prior findings that SA patients with RA were more likely than SA patients with osteoarthritis (OA) to receive perioperative blood transfusions.⁸ RA patients have been shown to have high rates of anemia of chronic disease, ranging from 22% to 77%.³² During joint replacement, these patients often require transfusions.^32,33 However, these findings differ from prior findings of no differences between RA and non-RA patients in the same patient population during the period 1988–2005.³ This difference may be a product of the constantly changing transfusion guidelines and increased use; transfusion rates increased 140% between 1997 and 2007, making transfusions the fastest growing common procedure in the United States during that time.³⁴ There was no difference between RA and non-RA patients in household income (as determined by ZIP code analysis), number of hospital beds, hospital region, or hospital teaching status. Compared with non-RA patients, RA patients were more likely to be younger, female, and of a difference race and to have a different expected primary payer (P < .001).These findings are consistent with previous findings in the literature.³ In the present SA study, however, RA patients were more likely than non-RA patients to have longer LOS, higher inflation-adjusted hospital charges, and nonroutine discharge. These findings deviate from those of the study covering the preceding 18 years (1988–2005).³ Despite the findings of a changing environment of care for RA patients, by Hambright and colleagues³ and Weiss and colleagues,³⁵ the trend appears to have shifted. Both groups had shorter average LOS than either group from the preceding 18 years.³ Although statistically significant in bivariate analysis, the difference in LOS between the 2 groups differed by an average of 0.11 day (2 hours 24 minutes) and was not clinically relevant.

In addition, the higher charges for patients with RA represent a deviation from the preceding 18 years.³ Other studies have also shown that RA is associated with increased cost in TSA.³⁶ Patients with RA often have rotator cuff pathology, indicating reverse SA may be used more frequently.^37,38 The increased implant cost associated with reverse SA may account for the increased costs in RA patients.³⁹ As mentioned, TSA type is not captured in the NIS database. In addition, that RA patients were less likely than non-RA patients to have routine discharge may indicate RA cases are more complex because of their complications.^1,5,14,40 A recent study of complications in RA patients (1163 who underwent THA, 2692 who underwent TKA) found that THA patients with RA were significantly more likely than THA patients with OA to dislocate, and TKA patients with RA were significantly more likely than TKA patients with OA to develop an infection after surgery.⁴¹ Postoperative dislocation has been shown to increase hospital costs in other orthopedic procedures.⁴² Also, during TSA, patients with RA are more likely than patients with OA to receive intraoperative blood transfusions.⁸ These complications—combined with the fact that RA is a chronic, progressive, systemic inflammatory disease that can affect soft tissue and blood vessel wall healing and is associated with medications having potential side effects—could contribute to the apparent increased hospital charges and LOS.^3,12,13,43 Factors that include surgeon preference, impact of primary payer, and hospital practice may also affect final charges. Total charges in the NIS database include administrative fees, hospital costs, device-related costs, operating room costs, and ancillary staff costs. Total charges do not include professional fees and differ from the total cost that represents the amount reimbursed by the payer. Charges tend to correlate with but overestimate the total costs.⁴⁴

This study had several important limitations. As mentioned, only events that occur during the operative admission are captured in the NIS database, and thus postoperative complications or serious adverse events that lead to readmission cannot be identified. In addition, outpatient TSAs are not captured in the NIS database, and thus inclusion of only inpatient procedures yields higher average LOS and total charges.⁴⁵ Given the limited granularity of ICD-9 coding, this study could not determine RA severity, estimated blood loss, length of surgery, complication severity, type of TSA procedure/prosthesis, or cause of death. Although commonly used to determine comorbidity burden, the modified Charlson index could not be used, and therefore could not be entered as a covariate in multivariate analysis. Furthermore, the NIS database does not include imaging or patient-reported outcomes information, such as improvements in pain or function, which are of crucial importance in considering surgery.

Conclusion

Our findings corroborated findings that the demographics and the perioperative safety profile for TSA were similar for patients with and without RA. The risk for complications or death in the perioperative period was low. Compared with non-RA patients, RA patients had significantly higher charges and longer LOS and were less likely to be discharged home after surgery. The 0.11-day difference in LOS, though statistically significant, was not clinically relevant. These findings differ from those for the preceding, 18-year period (1988–2005). Future research should focus on the causes of these changes.

Shoulder arthroplasty (SA), including total SA (TSA) and reverse TSA, is an effective surgical treatment for fracture and primary or secondary degenerative disease of the shoulder.¹ Over the past few decades, use of SA has increased dramatically, from about 5000 cases in 1990 to 7000 in 2000 and more than 26,000 in 2008.^1,2

Complications associated with SA generally are classified as perioperative (occurring during the operative index) or long-term (postdischarge).³ Long-term complications include implant loosening, instability, revision, infection, rotator cuff tear, neural injury, and deltoid detachment.^1,4,5 Perioperative complications, which are less commonly reported, include intraoperative fracture, infection, neural injury, venous thromboembolic events (VTEs, including pulmonary embolism [PE] and deep vein thrombosis [DVT]), transfusion, and death.^3,6-10

SA is an attractive treatment option for patients with rheumatoid arthritis (RA), as the effects of pain on these patients are greater in the shoulder joint than in any other joint.¹¹ Patients with RA pose unique orthopedic surgical challenges, including any combination of decreased bone mineralization, poor capsular tissue integrity, and osteonecrosis.^3,12 In addition, RA patients may be taking immunosuppressive medications that have severe side effects, and they may require multiple surgeries.^12,13 These factors predispose patients with RA to complications that include infection and wound dehiscence.^3,5,12-14

The complex nature of RA has prompted investigators to examine outcome measures in this patient group. Hambright and colleagues³ used the Nationwide Inpatient Sample (NIS) to examine perioperative outcomes in RA patients who underwent TSA between 1988 and 2005.³ They found that TSA patients with RA had shorter and less costly hospital stays and were more likely to have a routine discharge.³ Using the same patient population drawn from the period 2006–2011, we conducted a study to determine if this unexpected trend persists as the number of TSAs and quality of postoperative care continue to increase. Given the potential for anemia of chronic disease and the systemic inflammatory nature of RA, we hypothesized that the perioperative complication profile of RA patients would be worse than that of non-RA patients.

Materials and Methods

NIS data were acquired for the period 2006–2011. The NIS is the largest publicly available all-payer inpatient database, with a random 20% sample of about 1000 US hospitals accounting for 7 to 8 million inpatient stays. The database supplies weights used to estimate national totals, at about 35 million inpatient visits per year. NIS inpatient data are limited to the operative index. Postdischarge information is not available. The NIS is managed by the Healthcare Cost and Utilization Project, which is sponsored by the Agency for Healthcare Research and Quality. The quality of NIS data is assessed and validated by an independent contractor. NIS data have been widely used to examine perioperative outcomes.^15-17

NIS data cover patient and hospital demographics, hospital length of stay (LOS), discharge status, payer information, charges, and perioperative outcomes and procedure/diagnosis codes (ICD-9; International Classification of Diseases, Ninth Revision¹⁸).

As our Institutional Review Board (IRB) reviewed the database and determined the project was not human subject research, IRB involvement was not required. This study paralleled successful efforts with similar RA and non-RA patients who had shoulder and elbow surgery.^3,19 SA patients were identified by ICD-9 procedure code 81.80, but this code does not specify whether the prosthesis was unconstrained, semiconstrained, or constrained. ICD-9 coding also does not specify whether the TSA was traditional or reverse. Patients with RA were identified by ICD-9 diagnosis codes 714.0, 714.1, and 714.2. Patients without one of these codes were placed in the non-RA cohort. Patients with codes associated with pathologic fractures secondary to metastatic cancer or bone malignant neoplasm as a secondary or primary diagnosis and patients who had revision surgery indicated by code 81.83 were excluded, as they have a disproportionately higher comorbidity burden.

After each cohort was defined, demographic data (age, sex, race, income quartile based on ZIP postal code) were compared, as were data on primary payer, hospital demographics, LOS (≤5 days, defined as perioperative index), discharge type, inflation-adjusted charges in 2014 dollars based on the Consumer Price Indexes (http://www.bls.gov/cpi/), and mortality. Perioperative complications—respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related (including embolism, fibrosis, hemorrhage, pain, stenosis, or thrombus caused by any device, implant, or graft), cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, postoperative infection complications, and intraoperative transfusions—were considered using ICD-9 codes (996.X-999.X and 99.X, respectively).²⁰ Although commonly used to determine perioperative comorbidity burden using ICD-9 coding, the modified Charlson index was not considered because RA is a component of the index and would therefore bias the variable.^3,21

Statistical analyses, including χ² tests and 2-sample t tests, were performed for categorical and continuous variables, respectively. P < .05 was considered significant. Fisher exact test was used for cohorts with fewer than 5 occurrences. Multivariate logistic regression models were then calculated to determine the effect of RA on different outcomes and complications, with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. Statistical analyses were performed using the R statistical programming language.22

Results

Of the 34,970 patients who underwent SA between 2006 and 2011, 1674 (4.8%) had a diagnosis of RA and 33,296 (95.2%) did not. On average, patients with RA tended to be younger than patients without RA (66.4 vs 69.1 years; P < .001), and a larger percentage of RA patients were female (75.5% vs 54.4%; P < .001). Compared with non-RA patients, RA patients comprised a different ethnic group and had a different expected primary payer (P < .001). SA patients with and without RA did not differ in income quartile based on ZIP code, total number of hospital beds, hospital region, or hospital teaching status (P = .34, .78, .59, and .82, respectively) (Table 1).

LOS was significantly (P < .001) statistically longer for RA patients (2.196 days) than for non-RA patients (2.085 days). RA patients were significantly less likely to be discharged home (63.0% vs 67.6%; P < .001). (Routine discharge was defined as discharge home, whereas nonroutine discharge was defined as discharge to a short-term hospital, skilled nursing facility, intermediate care, another type of facility, home health care, against medical advice, or death.) In addition, inflation-adjusted charges associated with SA were significantly higher (P = .018) for RA patients ($54,284) than for non-RA patients ($52,663) (Table 1).

Regarding the rates of complications that occurred during the perioperative index, there were no significant differences between RA and non-RA cohorts. These complications included respiratory, gastrointestinal, genitourinary, accidental puncture/laceration, central nervous system, wound dehiscence, device-related, cardiac, hematoma/seroma, acute respiratory distress syndrome, postoperative shock, VTE, and postoperative infection (Table 2). In addition, there was no significant difference in mortality between the groups (P = .48).

In TSA, blood transfusions were more likely (P < .001) to be given to RA patients (9.00%) than to non-RA patients (6.16%). Multivariate regression analyses were performed with age, race, sex, hospital region, hospital type, number of hospital beds, primary payer, and hospital ownership as covariates. These analyses revealed that transfusion (P < .001), discharge type (P = .002), total inflation-adjusted charges (P < .001), and LOS (P = .047) remained significant (Table 3).

Discussion

Large national databases like NIS allow study of uncommon medical occurrences and help delineate risks and trends that otherwise might be indeterminable. Although it has been suggested that patients with RA may have poorer long-term outcomes after SA, the perioperative risk profile indicates that TSA is well tolerated in RA patients during the operative index.^3,23-25

The data on this study’s 34,970 patients, drawn from the period 2006–2011, demonstrated no significant differences in safety profile with respect to the 14 perioperative complications and outcomes examined, except blood transfusion rate. Rates of postoperative infection (RA, 0.24%; non-RA, 0.14%; P = .303), VTE (RA, 0.30%; non-RA, 0.25%; P = .905), and transfusion (RA, 9.00%; non-RA, 6.16%; P < .001) are of particular interest because of the severity of these situations.

Postoperative infection is a potentially serious complication and often occurs secondary to diabetes, RA, lupus erythematosus, prior surgery, or a nosocomial or remote source.¹ The often costly treatment options include antibiotic suppression, irrigation and debridement with implant retention, 1-stage exchange with antibiotic-impregnated cement fixation, staged reimplantation, resection arthroplasty, arthrodesis, and amputation.¹ The overall 0.14% infection rate determined in this study is lower than the 0.7% reported for SA patients in the literature.¹ Given the nature of the NIS database, this rate underestimates the true postoperative infection rate, as any infection that occurred after the perioperative period is not captured.²⁶ The present study’s perioperative infection rates (RA, 0.24%; non-RA, 0.14%) for the period 2006–2011 are comparable to the rates (RA, 0.17%; non-RA, 0.24%) reported by Hambright and colleagues3 for the same patient population over the preceding, 18-year period (1988–2005) and similarly do not significantly differ between groups. Although infection is uncommon in the immediate perioperative period, the ICD-9 codes used refer specifically to infection resulting from surgery and do not represent concomitant infection.

VTEs, which include PEs and DVTs, are rare but potentially life-threatening surgical complications.^27,28 Mechanical prophylaxis and chemical prophylaxis have been recommended for major orthopedic surgery, particularly lower extremity surgery, such as total hip arthroplasty (THA) and total knee arthroplasty (TKA).^28,29 In the present study, VTE rates were low, 0.30% (RA) and 0.25% (non-RA), and not significantly different in bivariate or multivariate analyses. These rates are comparable to those found in other national-database SA studies.²⁸ VTEs that occur outside the index hospital admission are not captured in this database. Therefore, the rates in the present study may be lower than the true incidence after SA. Mortality secondary to VTE usually occurs within 24 hours but may occur up to 90 days after surgery. DVT rates, on the other hand, are difficult to evaluate because of differences in screening practices.^27,28,30,31

That RA patients were more likely than non-RA patients to receive perioperative blood transfusions supports prior findings that SA patients with RA were more likely than SA patients with osteoarthritis (OA) to receive perioperative blood transfusions.⁸ RA patients have been shown to have high rates of anemia of chronic disease, ranging from 22% to 77%.³² During joint replacement, these patients often require transfusions.^32,33 However, these findings differ from prior findings of no differences between RA and non-RA patients in the same patient population during the period 1988–2005.³ This difference may be a product of the constantly changing transfusion guidelines and increased use; transfusion rates increased 140% between 1997 and 2007, making transfusions the fastest growing common procedure in the United States during that time.³⁴ There was no difference between RA and non-RA patients in household income (as determined by ZIP code analysis), number of hospital beds, hospital region, or hospital teaching status. Compared with non-RA patients, RA patients were more likely to be younger, female, and of a difference race and to have a different expected primary payer (P < .001).These findings are consistent with previous findings in the literature.³ In the present SA study, however, RA patients were more likely than non-RA patients to have longer LOS, higher inflation-adjusted hospital charges, and nonroutine discharge. These findings deviate from those of the study covering the preceding 18 years (1988–2005).³ Despite the findings of a changing environment of care for RA patients, by Hambright and colleagues³ and Weiss and colleagues,³⁵ the trend appears to have shifted. Both groups had shorter average LOS than either group from the preceding 18 years.³ Although statistically significant in bivariate analysis, the difference in LOS between the 2 groups differed by an average of 0.11 day (2 hours 24 minutes) and was not clinically relevant.

In addition, the higher charges for patients with RA represent a deviation from the preceding 18 years.³ Other studies have also shown that RA is associated with increased cost in TSA.³⁶ Patients with RA often have rotator cuff pathology, indicating reverse SA may be used more frequently.^37,38 The increased implant cost associated with reverse SA may account for the increased costs in RA patients.³⁹ As mentioned, TSA type is not captured in the NIS database. In addition, that RA patients were less likely than non-RA patients to have routine discharge may indicate RA cases are more complex because of their complications.^1,5,14,40 A recent study of complications in RA patients (1163 who underwent THA, 2692 who underwent TKA) found that THA patients with RA were significantly more likely than THA patients with OA to dislocate, and TKA patients with RA were significantly more likely than TKA patients with OA to develop an infection after surgery.⁴¹ Postoperative dislocation has been shown to increase hospital costs in other orthopedic procedures.⁴² Also, during TSA, patients with RA are more likely than patients with OA to receive intraoperative blood transfusions.⁸ These complications—combined with the fact that RA is a chronic, progressive, systemic inflammatory disease that can affect soft tissue and blood vessel wall healing and is associated with medications having potential side effects—could contribute to the apparent increased hospital charges and LOS.^3,12,13,43 Factors that include surgeon preference, impact of primary payer, and hospital practice may also affect final charges. Total charges in the NIS database include administrative fees, hospital costs, device-related costs, operating room costs, and ancillary staff costs. Total charges do not include professional fees and differ from the total cost that represents the amount reimbursed by the payer. Charges tend to correlate with but overestimate the total costs.⁴⁴

This study had several important limitations. As mentioned, only events that occur during the operative admission are captured in the NIS database, and thus postoperative complications or serious adverse events that lead to readmission cannot be identified. In addition, outpatient TSAs are not captured in the NIS database, and thus inclusion of only inpatient procedures yields higher average LOS and total charges.⁴⁵ Given the limited granularity of ICD-9 coding, this study could not determine RA severity, estimated blood loss, length of surgery, complication severity, type of TSA procedure/prosthesis, or cause of death. Although commonly used to determine comorbidity burden, the modified Charlson index could not be used, and therefore could not be entered as a covariate in multivariate analysis. Furthermore, the NIS database does not include imaging or patient-reported outcomes information, such as improvements in pain or function, which are of crucial importance in considering surgery.

Conclusion

Our findings corroborated findings that the demographics and the perioperative safety profile for TSA were similar for patients with and without RA. The risk for complications or death in the perioperative period was low. Compared with non-RA patients, RA patients had significantly higher charges and longer LOS and were less likely to be discharged home after surgery. The 0.11-day difference in LOS, though statistically significant, was not clinically relevant. These findings differ from those for the preceding, 18-year period (1988–2005). Future research should focus on the causes of these changes.