Original Research

Knee arthroplasty continues to be one of the most common and successful methods for treating severe arthritis and other painful arthropathies. Increasing steadily from 1998 to 2008, and with more than 676,000 procedures performed in 2008, knee arthroplasty remains the most common surgical joint replacement procedure.¹

Although perioperative and long-term complications are uncommon, infection remains one of the most serious complications of total knee arthroplasty (TKA). Some studies have found a post-TKA infection rate of less than 1%.² The solution of 2% chlorhexidine gluconate and 70% isopropyl alcohol (Chloraprep; Medi-Flex, Overland Park, Kansas) is commonly used for antiseptic skin preparation before surgery. Studies have shown significant decreases in post-TKA infection rates with preoperative use.^3,4 Another study evaluated the efficacy of 3 different skin solutions and found Chloraprep to be the most efficient in eradicating bacteria from the foot and ankle before surgery. The investigators noted that, even with preoperative use of Chloraprep, 23% of patients had residual bacteria on the surface of the skin between the toes.⁵ Like the foot, the popliteal fossa is an intertriginous area that may harbor normal flora, including gram-positive cocci, in large numbers, mainly because of the contact between 2 skin surfaces. Although postoperative infection rates decrease with use of Chloraprep, its presurgical efficacy in killing bacteria on another intertriginous area, the popliteal fossa, is largely unknown. Also unknown are susceptible organism species and organism population numbers.

Concerned that our skin preparation might be ineffective, we conducted a study to evaluate the efficacy of Chloraprep skin preparation in eradicating organisms before TKA, to isolate the type and number of organisms, and to evaluate several other contributing factors that could lead to infection.

Materials and Methods

This prospective study included 99 patients who were undergoing primary TKA at John Peter Smith Hospital between July 1, 2011 and August 31, 2012. An attempt was made to enroll consecutive TKA patients, and all patients agreed to participate, but a few were not enrolled because the study team had not asked for their consent before they were taken to the operating room. Patients did not receive monetary compensation for participation. Exclusion criteria were pregnancy, imprisonment, and age under 18 years. The study was approved by the institutional review boards at John Peter Smith Hospital and the University of North Texas Health Science Center.

Each lower extremity was prepared with Chloraprep according to the manufacturer’s instructions. Preparation was done by well-trained operating room staff members who were supervised by the surgeon (Dr. Sanchez or Dr. Wagner) but were not involved in the study. With use of the Chloraprep applicator, the solution was applied in a back-and-forth manner to the entire operative leg for at least 30 seconds, and then discarded. This scrub procedure was repeated with a second applicator before standard drapes were placed. The leg was left to air-dry for at least 30 seconds, and the drapes were placed before postsolution swabbing and before the iodine-impregnated adhesive drape was placed around the knee. During drying, the solution was not blotted, wiped away, or touched with instrumentation. Patients were swabbed with an epidermal sterile swab in the popliteal fossa of the knee undergoing surgery, both before solution application (presolution swab) and after (postsolution swab). Only the operating surgeon participated in swabbing the patients. Aerobic and anaerobic swabs were vigorously rubbed over a 2- to 3-in wide area across the entire posterior flexion crease surface.

The collected pre- and postsolution swabs were sent to John Peter Smith Laboratory for identification of organisms. Anaerobic swabs were cultured in thioglycolate broth and on 4 plates: MacConkey agar, Columbia colistin–nalidixic acid agar, chocolate agar, and sheep blood agar. Aerobic swabs were cultured in thioglycolate broth with hemin and vitamin K and on 4 plates: anaerobic blood agar, bile esculin agar, kanamycin and vancomycin agar, and Columbia colistin–nalidixic acid agar. Anaerobic plates were incubated in an anoxic environment. The plates were then read daily, and final reports were issued after 48 hours (for aerobic bacterial isolates) and 72 hours (for anaerobic bacterial isolates), as was the standard at the time.

Additional patient data were collected for possible correlations: American Society of Anesthesiologists (ASA) classification (physical status),⁶ body mass index (BMI), age, sex, arthroplasty type (unilateral, bilateral), and diabetic status. In addition, patients were asked if they had used Hibiclens antiseptic/antimicrobial skin cleanser daily during the week before surgery—as they had been instructed to do—and the number of times they had used the cleanser.

Study data were analyzed and were used to stratify patients into several groups. Each group had multiple factors evaluated.

 Descriptive statistics were used to characterize the patient demographic information. Chi-square analyses were performed to evaluate the difference between presence of organisms before and after solution application, and the data were also layered with reported Hibiclens cleanser use. In addition, binary logistic regression was used to determine if demographic variables could predict presence of organism isolates before and after solution application. Data analyses were conducted using IBM SPSS Statistics Version 20.

Results

No patient had a postoperative infection. Culture isolates grew in 20 (20%) of the 99 patients before solution application and in 5 (5%) of the 99 after application. Of the 20 patients with presolution culture isolates, 16 (80%) had 1 bacterial isolate, and 4 (20%) had 2 or more species. Presolution isolates included normal flora (10, 50%), coagulase-negative Staphylococcus aureus (6, 30%), rare Bacillus (3, 15%), Micrococcus luteus (1, 5%), rare gram-negative (1, 5%), rare gram-positive (1, 5%), and Staphylococcus hominis (1, 5%) (Figure 1). Postsolution isolates included coagulase-negative S aureus (3, 60%), rare Bacillus (1, 20%), and rare Serratia odorifera (1, 20%) (Figure 2). Two postsolution isolates did not have an associated presolution isolate. Presolution organism isolation was an important predictor of postsolution organism isolation (P < .046).

BMI was recorded for all patients. Mean BMI was 35 (range, 20-63). Distribution was as follows: BMI under 20 (3 patients), under 30 (30 patients), under 40 (47 patients), under 50 (14 patients), under 60 (4 patients), and over 60 (1 patient). Mean presolution BMI was significantly (P < .03) higher for patients with bacterial isolates than for patients without isolates (38 and 34, respectively). Mean postsolution BMI was 40 for patients with bacterial isolates and 35 for patients without isolates (Figure 3). Of the 33 patients with BMI under 30, 3 (9%) had presolution isolates and 1 (3%) had postsolution isolates. Of the 66 patients with BMI over 30, 17 (26%) had presolution isolates and 4 (6%) had postsolution isolates (Table).

Of the 99 patients, 30 (30%) had diabetes. Of these 30 patients, 9 (30%) had presolution isolates (45% of all presolution isolates) and 3 (10%) had postsolution isolates (60% of all postsolution isolates.) Although neither pre- nor postsolution results were statistically significant (P = .172) for increasing organism isolation in patients with diabetes, the odds ratio for these patients was 3.6 when the focus was on the likelihood of postsolution organism isolation.

Mean age was 57 years (range, 29-87 years). Results were not statistically significant for age being a likely factor for organism isolate prediction.

There were 81 women and 18 men in the study. Of the 81 women, 16 (20%) had positive presolution cultures and 5 (6%) had positive postsolution cultures. Of the 18 men, 4 (22%) had positive presolution cultures and none had a positive postsolution culture.

Race was recorded. Forty-nine patients were white, 27 black, 18 Hispanic, and 5 unknown. Presolution, 12 whites (24%), 5 blacks (19%), and 3 Hispanics (17%) had positive cultures. Postsolution, 1 white (2%), 1 black (4%), 3 Hispanics (17%), and 1 patient of unknown race (20%) had positive cultures.

ASA classifications were recorded and analyzed. Of the 99 patients, 38 were classified ASA-2, 60 were ASA-3, and 1 was ASA-4. Presolution, 9 (24%) of the 38 ASA-2 patients and 11 (18%) of the 60 ASA-3 patients had positive cultures; postsolution, 2 (5%) of the 38 ASA-2 patients and 3 (5%) of the 60 ASA-3 patients had positive cultures. The 1 ASA-4 patient had neither presolution nor postsolution positive cultures.

Types of TKA (bilateral, unilateral) were recorded. Of the 99 patients, 89 had unilateral TKAs and 10 had bilateral TKAs. Presolution, 19 (21%) of the 89 unilaterals and 1 (10%) of the 10 bilaterals had positive cultures. Postsolution, 5 (6%) of the 89 unilaterals and none of the 10 bilaterals had positive cultures.

Patients were also verbally asked how many cleanser baths they had taken before surgery. Of the 99 patients, 88 reported having taken 1 or more cleanser baths, and 1 reported no baths; 10 patients’ responses were not available. The 88 patients who had taken at least 1 cleanser bath were divided into 3 groups: 1 bath (35 patients), 2 baths (49 patients), and 3 or more baths (4 patients). Presolution, positive cultures were found for 18 (20%) of the 88 patients; for 7 (20%) of the 35 patients with 1 bath; for 10 (20%) of the 49 patients with 2 baths; and for 1 (25%) of the 4 patients with 3 or more baths. Postsolution, positive cultures were found for 5 (6%) of the 88 patients; for 2 (6%) of the 35 patients with 1 bath; for 3 (6%) of the 49 patients with 2 baths; and for 0 (0%) of the 4 patients with 3 or more baths. The 1 patient with no baths did not have a positive culture. Of the 10 patients whose responses were unavailable, 2 patients had positive presolution cultures and no patients had a positive postsolution culture.

Discussion

The efficacy of using Chloraprep before TKA has not been well assessed in orthopedic practice. However, compared with other preoperative solutions, chlorhexidine has been shown to be significantly better in preventing post-TKA infections.⁴ Other studies have found it far more effective than other commonly used surgical preparations in eliminating microorganisms in hip arthroplasty and foot surgery.^5,7 Our study, focused on the efficacy of Chloraprep in killing bacteria, found the solution effective in removing 85% (17/20) of cultured presolution organisms.

Of the bacterial isolates cultured, normal flora were effectively removed from all associated postsolution cultures. Although most of the bacterial isolates were eliminated after solution application, both coagulase-negative S aureus and rare Bacillus species were found both pre- and postsolution, suggesting either inadequate skin preparation or resistant bacteria.

With respect to the secondary variables, our study data showed that BMI was an important predictor for bacterial isolates, significantly so presolution (P < .03). Mean BMI for the overall study was 35, firmly in the obese category. Only when BMI increased to 38 did it become significant as a predictor for postsolution organisms. Mean postsolution BMI was even higher, 40, which is in the morbidly obese category. Interestingly, the percentage of nonobese patients (BMI, <30) with positive presolution cultures was only 9%, versus the 20% with positive presolution cultures overall. In addition, 1 nonobese patient had positive postsolution cultures.

Other studies have linked higher BMI to higher rates of surgical site infection and other complications, but it is unknown if the infections are due to higher bacterial counts in the patients with high BMI or to other factors, such as reduced wound healing or decreased immune response. More research is needed to determine if the number of organisms in patients with high BMI correlates to a higher risk for surgical site infection.⁸ As expected, along with BMI (>38), presolution organism isolation was an important predictor for postsolution organism isolation. Patients with presolution organism isolation were 24 times more likely to have postsolution isolates.

Even though diabetic status was not significant for predicting bacterial isolation, patients with diabetes were 3.6 times more likely than patients without diabetes to have a positive culture. Other studies have shown that, compared with patients without diabetes, patients with diabetes had a higher chance of postoperative infection.^9,10

In this study, 18 of 20 patients with presolution organism isolates reported they had been compliant in taking the recommended preoperative cleanser baths. This finding may indicate that preoperative cleanser baths are ineffective. However, only 20% of our patients had positive presolution cultures, whereas Ostrander and colleagues⁵ reported 30% positive pre-preparation cultures from the anterior knee. A recent Cochrane Database System Review did not provide clear evidence of benefit for preoperative showering or bathing with chlorhexidine over other wash products.¹¹ Although their benefit may be questionable, we will continue to recommend preoperative cleanser baths.

One limitation of this study is sample size. Although size was sufficient for determining the efficacy of Chloraprep in the intertriginous area of the back of the knee, the lack of statistical significance (eg, effect of diabetes) may not be accurate. In addition, because the nurse who prepared patients’ skin was aware of the study and was supervised in every case, it is possible that the preparation was done more carefully than usual, resulting in more negative cultures than average. Also, compliance in taking preoperative cleanser baths was subjectively determined. Patients may have reported more baths than were actually taken. Still another study limitation is that 2 postsolution isolates did not have an associated presolution isolate. Although we think this may have resulted from laboratory contamination, it is possible the presolution swabs did not accurately determine true bacterial counts in these cases.

Conclusion

A study that showed significant residual bacteria between patients’ toes after chlorhexidine skin preparation⁵ left us concerned that Chloraprep skin preparation for TKA might not be adequate. The present study showed that this solution was effective in eliminating bacteria from the intertriginous area of the back of the knee in 95% of patients. Skin preparation appears to be less effective in patients with higher BMI.

Knee arthroplasty continues to be one of the most common and successful methods for treating severe arthritis and other painful arthropathies. Increasing steadily from 1998 to 2008, and with more than 676,000 procedures performed in 2008, knee arthroplasty remains the most common surgical joint replacement procedure.¹

Although perioperative and long-term complications are uncommon, infection remains one of the most serious complications of total knee arthroplasty (TKA). Some studies have found a post-TKA infection rate of less than 1%.² The solution of 2% chlorhexidine gluconate and 70% isopropyl alcohol (Chloraprep; Medi-Flex, Overland Park, Kansas) is commonly used for antiseptic skin preparation before surgery. Studies have shown significant decreases in post-TKA infection rates with preoperative use.^3,4 Another study evaluated the efficacy of 3 different skin solutions and found Chloraprep to be the most efficient in eradicating bacteria from the foot and ankle before surgery. The investigators noted that, even with preoperative use of Chloraprep, 23% of patients had residual bacteria on the surface of the skin between the toes.⁵ Like the foot, the popliteal fossa is an intertriginous area that may harbor normal flora, including gram-positive cocci, in large numbers, mainly because of the contact between 2 skin surfaces. Although postoperative infection rates decrease with use of Chloraprep, its presurgical efficacy in killing bacteria on another intertriginous area, the popliteal fossa, is largely unknown. Also unknown are susceptible organism species and organism population numbers.

Concerned that our skin preparation might be ineffective, we conducted a study to evaluate the efficacy of Chloraprep skin preparation in eradicating organisms before TKA, to isolate the type and number of organisms, and to evaluate several other contributing factors that could lead to infection.

Materials and Methods

This prospective study included 99 patients who were undergoing primary TKA at John Peter Smith Hospital between July 1, 2011 and August 31, 2012. An attempt was made to enroll consecutive TKA patients, and all patients agreed to participate, but a few were not enrolled because the study team had not asked for their consent before they were taken to the operating room. Patients did not receive monetary compensation for participation. Exclusion criteria were pregnancy, imprisonment, and age under 18 years. The study was approved by the institutional review boards at John Peter Smith Hospital and the University of North Texas Health Science Center.

Each lower extremity was prepared with Chloraprep according to the manufacturer’s instructions. Preparation was done by well-trained operating room staff members who were supervised by the surgeon (Dr. Sanchez or Dr. Wagner) but were not involved in the study. With use of the Chloraprep applicator, the solution was applied in a back-and-forth manner to the entire operative leg for at least 30 seconds, and then discarded. This scrub procedure was repeated with a second applicator before standard drapes were placed. The leg was left to air-dry for at least 30 seconds, and the drapes were placed before postsolution swabbing and before the iodine-impregnated adhesive drape was placed around the knee. During drying, the solution was not blotted, wiped away, or touched with instrumentation. Patients were swabbed with an epidermal sterile swab in the popliteal fossa of the knee undergoing surgery, both before solution application (presolution swab) and after (postsolution swab). Only the operating surgeon participated in swabbing the patients. Aerobic and anaerobic swabs were vigorously rubbed over a 2- to 3-in wide area across the entire posterior flexion crease surface.

The collected pre- and postsolution swabs were sent to John Peter Smith Laboratory for identification of organisms. Anaerobic swabs were cultured in thioglycolate broth and on 4 plates: MacConkey agar, Columbia colistin–nalidixic acid agar, chocolate agar, and sheep blood agar. Aerobic swabs were cultured in thioglycolate broth with hemin and vitamin K and on 4 plates: anaerobic blood agar, bile esculin agar, kanamycin and vancomycin agar, and Columbia colistin–nalidixic acid agar. Anaerobic plates were incubated in an anoxic environment. The plates were then read daily, and final reports were issued after 48 hours (for aerobic bacterial isolates) and 72 hours (for anaerobic bacterial isolates), as was the standard at the time.

Additional patient data were collected for possible correlations: American Society of Anesthesiologists (ASA) classification (physical status),⁶ body mass index (BMI), age, sex, arthroplasty type (unilateral, bilateral), and diabetic status. In addition, patients were asked if they had used Hibiclens antiseptic/antimicrobial skin cleanser daily during the week before surgery—as they had been instructed to do—and the number of times they had used the cleanser.

Study data were analyzed and were used to stratify patients into several groups. Each group had multiple factors evaluated.

 Descriptive statistics were used to characterize the patient demographic information. Chi-square analyses were performed to evaluate the difference between presence of organisms before and after solution application, and the data were also layered with reported Hibiclens cleanser use. In addition, binary logistic regression was used to determine if demographic variables could predict presence of organism isolates before and after solution application. Data analyses were conducted using IBM SPSS Statistics Version 20.

Results

No patient had a postoperative infection. Culture isolates grew in 20 (20%) of the 99 patients before solution application and in 5 (5%) of the 99 after application. Of the 20 patients with presolution culture isolates, 16 (80%) had 1 bacterial isolate, and 4 (20%) had 2 or more species. Presolution isolates included normal flora (10, 50%), coagulase-negative Staphylococcus aureus (6, 30%), rare Bacillus (3, 15%), Micrococcus luteus (1, 5%), rare gram-negative (1, 5%), rare gram-positive (1, 5%), and Staphylococcus hominis (1, 5%) (Figure 1). Postsolution isolates included coagulase-negative S aureus (3, 60%), rare Bacillus (1, 20%), and rare Serratia odorifera (1, 20%) (Figure 2). Two postsolution isolates did not have an associated presolution isolate. Presolution organism isolation was an important predictor of postsolution organism isolation (P < .046).

BMI was recorded for all patients. Mean BMI was 35 (range, 20-63). Distribution was as follows: BMI under 20 (3 patients), under 30 (30 patients), under 40 (47 patients), under 50 (14 patients), under 60 (4 patients), and over 60 (1 patient). Mean presolution BMI was significantly (P < .03) higher for patients with bacterial isolates than for patients without isolates (38 and 34, respectively). Mean postsolution BMI was 40 for patients with bacterial isolates and 35 for patients without isolates (Figure 3). Of the 33 patients with BMI under 30, 3 (9%) had presolution isolates and 1 (3%) had postsolution isolates. Of the 66 patients with BMI over 30, 17 (26%) had presolution isolates and 4 (6%) had postsolution isolates (Table).

Of the 99 patients, 30 (30%) had diabetes. Of these 30 patients, 9 (30%) had presolution isolates (45% of all presolution isolates) and 3 (10%) had postsolution isolates (60% of all postsolution isolates.) Although neither pre- nor postsolution results were statistically significant (P = .172) for increasing organism isolation in patients with diabetes, the odds ratio for these patients was 3.6 when the focus was on the likelihood of postsolution organism isolation.

Mean age was 57 years (range, 29-87 years). Results were not statistically significant for age being a likely factor for organism isolate prediction.

There were 81 women and 18 men in the study. Of the 81 women, 16 (20%) had positive presolution cultures and 5 (6%) had positive postsolution cultures. Of the 18 men, 4 (22%) had positive presolution cultures and none had a positive postsolution culture.

Race was recorded. Forty-nine patients were white, 27 black, 18 Hispanic, and 5 unknown. Presolution, 12 whites (24%), 5 blacks (19%), and 3 Hispanics (17%) had positive cultures. Postsolution, 1 white (2%), 1 black (4%), 3 Hispanics (17%), and 1 patient of unknown race (20%) had positive cultures.

ASA classifications were recorded and analyzed. Of the 99 patients, 38 were classified ASA-2, 60 were ASA-3, and 1 was ASA-4. Presolution, 9 (24%) of the 38 ASA-2 patients and 11 (18%) of the 60 ASA-3 patients had positive cultures; postsolution, 2 (5%) of the 38 ASA-2 patients and 3 (5%) of the 60 ASA-3 patients had positive cultures. The 1 ASA-4 patient had neither presolution nor postsolution positive cultures.

Types of TKA (bilateral, unilateral) were recorded. Of the 99 patients, 89 had unilateral TKAs and 10 had bilateral TKAs. Presolution, 19 (21%) of the 89 unilaterals and 1 (10%) of the 10 bilaterals had positive cultures. Postsolution, 5 (6%) of the 89 unilaterals and none of the 10 bilaterals had positive cultures.

Patients were also verbally asked how many cleanser baths they had taken before surgery. Of the 99 patients, 88 reported having taken 1 or more cleanser baths, and 1 reported no baths; 10 patients’ responses were not available. The 88 patients who had taken at least 1 cleanser bath were divided into 3 groups: 1 bath (35 patients), 2 baths (49 patients), and 3 or more baths (4 patients). Presolution, positive cultures were found for 18 (20%) of the 88 patients; for 7 (20%) of the 35 patients with 1 bath; for 10 (20%) of the 49 patients with 2 baths; and for 1 (25%) of the 4 patients with 3 or more baths. Postsolution, positive cultures were found for 5 (6%) of the 88 patients; for 2 (6%) of the 35 patients with 1 bath; for 3 (6%) of the 49 patients with 2 baths; and for 0 (0%) of the 4 patients with 3 or more baths. The 1 patient with no baths did not have a positive culture. Of the 10 patients whose responses were unavailable, 2 patients had positive presolution cultures and no patients had a positive postsolution culture.

Discussion

The efficacy of using Chloraprep before TKA has not been well assessed in orthopedic practice. However, compared with other preoperative solutions, chlorhexidine has been shown to be significantly better in preventing post-TKA infections.⁴ Other studies have found it far more effective than other commonly used surgical preparations in eliminating microorganisms in hip arthroplasty and foot surgery.^5,7 Our study, focused on the efficacy of Chloraprep in killing bacteria, found the solution effective in removing 85% (17/20) of cultured presolution organisms.

Of the bacterial isolates cultured, normal flora were effectively removed from all associated postsolution cultures. Although most of the bacterial isolates were eliminated after solution application, both coagulase-negative S aureus and rare Bacillus species were found both pre- and postsolution, suggesting either inadequate skin preparation or resistant bacteria.

With respect to the secondary variables, our study data showed that BMI was an important predictor for bacterial isolates, significantly so presolution (P < .03). Mean BMI for the overall study was 35, firmly in the obese category. Only when BMI increased to 38 did it become significant as a predictor for postsolution organisms. Mean postsolution BMI was even higher, 40, which is in the morbidly obese category. Interestingly, the percentage of nonobese patients (BMI, <30) with positive presolution cultures was only 9%, versus the 20% with positive presolution cultures overall. In addition, 1 nonobese patient had positive postsolution cultures.

Other studies have linked higher BMI to higher rates of surgical site infection and other complications, but it is unknown if the infections are due to higher bacterial counts in the patients with high BMI or to other factors, such as reduced wound healing or decreased immune response. More research is needed to determine if the number of organisms in patients with high BMI correlates to a higher risk for surgical site infection.⁸ As expected, along with BMI (>38), presolution organism isolation was an important predictor for postsolution organism isolation. Patients with presolution organism isolation were 24 times more likely to have postsolution isolates.

Even though diabetic status was not significant for predicting bacterial isolation, patients with diabetes were 3.6 times more likely than patients without diabetes to have a positive culture. Other studies have shown that, compared with patients without diabetes, patients with diabetes had a higher chance of postoperative infection.^9,10

In this study, 18 of 20 patients with presolution organism isolates reported they had been compliant in taking the recommended preoperative cleanser baths. This finding may indicate that preoperative cleanser baths are ineffective. However, only 20% of our patients had positive presolution cultures, whereas Ostrander and colleagues⁵ reported 30% positive pre-preparation cultures from the anterior knee. A recent Cochrane Database System Review did not provide clear evidence of benefit for preoperative showering or bathing with chlorhexidine over other wash products.¹¹ Although their benefit may be questionable, we will continue to recommend preoperative cleanser baths.

One limitation of this study is sample size. Although size was sufficient for determining the efficacy of Chloraprep in the intertriginous area of the back of the knee, the lack of statistical significance (eg, effect of diabetes) may not be accurate. In addition, because the nurse who prepared patients’ skin was aware of the study and was supervised in every case, it is possible that the preparation was done more carefully than usual, resulting in more negative cultures than average. Also, compliance in taking preoperative cleanser baths was subjectively determined. Patients may have reported more baths than were actually taken. Still another study limitation is that 2 postsolution isolates did not have an associated presolution isolate. Although we think this may have resulted from laboratory contamination, it is possible the presolution swabs did not accurately determine true bacterial counts in these cases.

Conclusion

A study that showed significant residual bacteria between patients’ toes after chlorhexidine skin preparation⁵ left us concerned that Chloraprep skin preparation for TKA might not be adequate. The present study showed that this solution was effective in eliminating bacteria from the intertriginous area of the back of the knee in 95% of patients. Skin preparation appears to be less effective in patients with higher BMI.

Knee arthroplasty continues to be one of the most common and successful methods for treating severe arthritis and other painful arthropathies. Increasing steadily from 1998 to 2008, and with more than 676,000 procedures performed in 2008, knee arthroplasty remains the most common surgical joint replacement procedure.¹

Although perioperative and long-term complications are uncommon, infection remains one of the most serious complications of total knee arthroplasty (TKA). Some studies have found a post-TKA infection rate of less than 1%.² The solution of 2% chlorhexidine gluconate and 70% isopropyl alcohol (Chloraprep; Medi-Flex, Overland Park, Kansas) is commonly used for antiseptic skin preparation before surgery. Studies have shown significant decreases in post-TKA infection rates with preoperative use.^3,4 Another study evaluated the efficacy of 3 different skin solutions and found Chloraprep to be the most efficient in eradicating bacteria from the foot and ankle before surgery. The investigators noted that, even with preoperative use of Chloraprep, 23% of patients had residual bacteria on the surface of the skin between the toes.⁵ Like the foot, the popliteal fossa is an intertriginous area that may harbor normal flora, including gram-positive cocci, in large numbers, mainly because of the contact between 2 skin surfaces. Although postoperative infection rates decrease with use of Chloraprep, its presurgical efficacy in killing bacteria on another intertriginous area, the popliteal fossa, is largely unknown. Also unknown are susceptible organism species and organism population numbers.

Concerned that our skin preparation might be ineffective, we conducted a study to evaluate the efficacy of Chloraprep skin preparation in eradicating organisms before TKA, to isolate the type and number of organisms, and to evaluate several other contributing factors that could lead to infection.

Materials and Methods

This prospective study included 99 patients who were undergoing primary TKA at John Peter Smith Hospital between July 1, 2011 and August 31, 2012. An attempt was made to enroll consecutive TKA patients, and all patients agreed to participate, but a few were not enrolled because the study team had not asked for their consent before they were taken to the operating room. Patients did not receive monetary compensation for participation. Exclusion criteria were pregnancy, imprisonment, and age under 18 years. The study was approved by the institutional review boards at John Peter Smith Hospital and the University of North Texas Health Science Center.

Each lower extremity was prepared with Chloraprep according to the manufacturer’s instructions. Preparation was done by well-trained operating room staff members who were supervised by the surgeon (Dr. Sanchez or Dr. Wagner) but were not involved in the study. With use of the Chloraprep applicator, the solution was applied in a back-and-forth manner to the entire operative leg for at least 30 seconds, and then discarded. This scrub procedure was repeated with a second applicator before standard drapes were placed. The leg was left to air-dry for at least 30 seconds, and the drapes were placed before postsolution swabbing and before the iodine-impregnated adhesive drape was placed around the knee. During drying, the solution was not blotted, wiped away, or touched with instrumentation. Patients were swabbed with an epidermal sterile swab in the popliteal fossa of the knee undergoing surgery, both before solution application (presolution swab) and after (postsolution swab). Only the operating surgeon participated in swabbing the patients. Aerobic and anaerobic swabs were vigorously rubbed over a 2- to 3-in wide area across the entire posterior flexion crease surface.

The collected pre- and postsolution swabs were sent to John Peter Smith Laboratory for identification of organisms. Anaerobic swabs were cultured in thioglycolate broth and on 4 plates: MacConkey agar, Columbia colistin–nalidixic acid agar, chocolate agar, and sheep blood agar. Aerobic swabs were cultured in thioglycolate broth with hemin and vitamin K and on 4 plates: anaerobic blood agar, bile esculin agar, kanamycin and vancomycin agar, and Columbia colistin–nalidixic acid agar. Anaerobic plates were incubated in an anoxic environment. The plates were then read daily, and final reports were issued after 48 hours (for aerobic bacterial isolates) and 72 hours (for anaerobic bacterial isolates), as was the standard at the time.

Additional patient data were collected for possible correlations: American Society of Anesthesiologists (ASA) classification (physical status),⁶ body mass index (BMI), age, sex, arthroplasty type (unilateral, bilateral), and diabetic status. In addition, patients were asked if they had used Hibiclens antiseptic/antimicrobial skin cleanser daily during the week before surgery—as they had been instructed to do—and the number of times they had used the cleanser.

Study data were analyzed and were used to stratify patients into several groups. Each group had multiple factors evaluated.

 Descriptive statistics were used to characterize the patient demographic information. Chi-square analyses were performed to evaluate the difference between presence of organisms before and after solution application, and the data were also layered with reported Hibiclens cleanser use. In addition, binary logistic regression was used to determine if demographic variables could predict presence of organism isolates before and after solution application. Data analyses were conducted using IBM SPSS Statistics Version 20.

Results

No patient had a postoperative infection. Culture isolates grew in 20 (20%) of the 99 patients before solution application and in 5 (5%) of the 99 after application. Of the 20 patients with presolution culture isolates, 16 (80%) had 1 bacterial isolate, and 4 (20%) had 2 or more species. Presolution isolates included normal flora (10, 50%), coagulase-negative Staphylococcus aureus (6, 30%), rare Bacillus (3, 15%), Micrococcus luteus (1, 5%), rare gram-negative (1, 5%), rare gram-positive (1, 5%), and Staphylococcus hominis (1, 5%) (Figure 1). Postsolution isolates included coagulase-negative S aureus (3, 60%), rare Bacillus (1, 20%), and rare Serratia odorifera (1, 20%) (Figure 2). Two postsolution isolates did not have an associated presolution isolate. Presolution organism isolation was an important predictor of postsolution organism isolation (P < .046).

BMI was recorded for all patients. Mean BMI was 35 (range, 20-63). Distribution was as follows: BMI under 20 (3 patients), under 30 (30 patients), under 40 (47 patients), under 50 (14 patients), under 60 (4 patients), and over 60 (1 patient). Mean presolution BMI was significantly (P < .03) higher for patients with bacterial isolates than for patients without isolates (38 and 34, respectively). Mean postsolution BMI was 40 for patients with bacterial isolates and 35 for patients without isolates (Figure 3). Of the 33 patients with BMI under 30, 3 (9%) had presolution isolates and 1 (3%) had postsolution isolates. Of the 66 patients with BMI over 30, 17 (26%) had presolution isolates and 4 (6%) had postsolution isolates (Table).

Of the 99 patients, 30 (30%) had diabetes. Of these 30 patients, 9 (30%) had presolution isolates (45% of all presolution isolates) and 3 (10%) had postsolution isolates (60% of all postsolution isolates.) Although neither pre- nor postsolution results were statistically significant (P = .172) for increasing organism isolation in patients with diabetes, the odds ratio for these patients was 3.6 when the focus was on the likelihood of postsolution organism isolation.

Mean age was 57 years (range, 29-87 years). Results were not statistically significant for age being a likely factor for organism isolate prediction.

There were 81 women and 18 men in the study. Of the 81 women, 16 (20%) had positive presolution cultures and 5 (6%) had positive postsolution cultures. Of the 18 men, 4 (22%) had positive presolution cultures and none had a positive postsolution culture.

Race was recorded. Forty-nine patients were white, 27 black, 18 Hispanic, and 5 unknown. Presolution, 12 whites (24%), 5 blacks (19%), and 3 Hispanics (17%) had positive cultures. Postsolution, 1 white (2%), 1 black (4%), 3 Hispanics (17%), and 1 patient of unknown race (20%) had positive cultures.

ASA classifications were recorded and analyzed. Of the 99 patients, 38 were classified ASA-2, 60 were ASA-3, and 1 was ASA-4. Presolution, 9 (24%) of the 38 ASA-2 patients and 11 (18%) of the 60 ASA-3 patients had positive cultures; postsolution, 2 (5%) of the 38 ASA-2 patients and 3 (5%) of the 60 ASA-3 patients had positive cultures. The 1 ASA-4 patient had neither presolution nor postsolution positive cultures.

Types of TKA (bilateral, unilateral) were recorded. Of the 99 patients, 89 had unilateral TKAs and 10 had bilateral TKAs. Presolution, 19 (21%) of the 89 unilaterals and 1 (10%) of the 10 bilaterals had positive cultures. Postsolution, 5 (6%) of the 89 unilaterals and none of the 10 bilaterals had positive cultures.

Patients were also verbally asked how many cleanser baths they had taken before surgery. Of the 99 patients, 88 reported having taken 1 or more cleanser baths, and 1 reported no baths; 10 patients’ responses were not available. The 88 patients who had taken at least 1 cleanser bath were divided into 3 groups: 1 bath (35 patients), 2 baths (49 patients), and 3 or more baths (4 patients). Presolution, positive cultures were found for 18 (20%) of the 88 patients; for 7 (20%) of the 35 patients with 1 bath; for 10 (20%) of the 49 patients with 2 baths; and for 1 (25%) of the 4 patients with 3 or more baths. Postsolution, positive cultures were found for 5 (6%) of the 88 patients; for 2 (6%) of the 35 patients with 1 bath; for 3 (6%) of the 49 patients with 2 baths; and for 0 (0%) of the 4 patients with 3 or more baths. The 1 patient with no baths did not have a positive culture. Of the 10 patients whose responses were unavailable, 2 patients had positive presolution cultures and no patients had a positive postsolution culture.

Discussion

The efficacy of using Chloraprep before TKA has not been well assessed in orthopedic practice. However, compared with other preoperative solutions, chlorhexidine has been shown to be significantly better in preventing post-TKA infections.⁴ Other studies have found it far more effective than other commonly used surgical preparations in eliminating microorganisms in hip arthroplasty and foot surgery.^5,7 Our study, focused on the efficacy of Chloraprep in killing bacteria, found the solution effective in removing 85% (17/20) of cultured presolution organisms.

Of the bacterial isolates cultured, normal flora were effectively removed from all associated postsolution cultures. Although most of the bacterial isolates were eliminated after solution application, both coagulase-negative S aureus and rare Bacillus species were found both pre- and postsolution, suggesting either inadequate skin preparation or resistant bacteria.

With respect to the secondary variables, our study data showed that BMI was an important predictor for bacterial isolates, significantly so presolution (P < .03). Mean BMI for the overall study was 35, firmly in the obese category. Only when BMI increased to 38 did it become significant as a predictor for postsolution organisms. Mean postsolution BMI was even higher, 40, which is in the morbidly obese category. Interestingly, the percentage of nonobese patients (BMI, <30) with positive presolution cultures was only 9%, versus the 20% with positive presolution cultures overall. In addition, 1 nonobese patient had positive postsolution cultures.

Other studies have linked higher BMI to higher rates of surgical site infection and other complications, but it is unknown if the infections are due to higher bacterial counts in the patients with high BMI or to other factors, such as reduced wound healing or decreased immune response. More research is needed to determine if the number of organisms in patients with high BMI correlates to a higher risk for surgical site infection.⁸ As expected, along with BMI (>38), presolution organism isolation was an important predictor for postsolution organism isolation. Patients with presolution organism isolation were 24 times more likely to have postsolution isolates.

Even though diabetic status was not significant for predicting bacterial isolation, patients with diabetes were 3.6 times more likely than patients without diabetes to have a positive culture. Other studies have shown that, compared with patients without diabetes, patients with diabetes had a higher chance of postoperative infection.^9,10

In this study, 18 of 20 patients with presolution organism isolates reported they had been compliant in taking the recommended preoperative cleanser baths. This finding may indicate that preoperative cleanser baths are ineffective. However, only 20% of our patients had positive presolution cultures, whereas Ostrander and colleagues⁵ reported 30% positive pre-preparation cultures from the anterior knee. A recent Cochrane Database System Review did not provide clear evidence of benefit for preoperative showering or bathing with chlorhexidine over other wash products.¹¹ Although their benefit may be questionable, we will continue to recommend preoperative cleanser baths.

One limitation of this study is sample size. Although size was sufficient for determining the efficacy of Chloraprep in the intertriginous area of the back of the knee, the lack of statistical significance (eg, effect of diabetes) may not be accurate. In addition, because the nurse who prepared patients’ skin was aware of the study and was supervised in every case, it is possible that the preparation was done more carefully than usual, resulting in more negative cultures than average. Also, compliance in taking preoperative cleanser baths was subjectively determined. Patients may have reported more baths than were actually taken. Still another study limitation is that 2 postsolution isolates did not have an associated presolution isolate. Although we think this may have resulted from laboratory contamination, it is possible the presolution swabs did not accurately determine true bacterial counts in these cases.

Conclusion

A study that showed significant residual bacteria between patients’ toes after chlorhexidine skin preparation⁵ left us concerned that Chloraprep skin preparation for TKA might not be adequate. The present study showed that this solution was effective in eliminating bacteria from the intertriginous area of the back of the knee in 95% of patients. Skin preparation appears to be less effective in patients with higher BMI.

With health care costs increasing and economic resources diminishing, substantial efforts have been directed toward improving the quality of care delivered in a cost-effective manner. For a total hip arthroplasty (THA) performed in the United States between 1997 and 2001, total hospital cost, including direct and indirect costs, was estimated as averaging $13,339.¹ In 2012, this cost was estimated to be between $43,000 and $100,000.² This overall cost estimate, along with the rate at which the procedure is performed, may present an opportunity for cost savings.

Length of hospital stay (LHS) is an important outcome measure that has been assessed for optimal health care delivery. Prolonged LHS implies increased resource expenditure. Therefore, it is crucial to identify factors associated with prolonged LHS in order to reduce costs. Investigations have identified factors shown to affect LHS after THA. These factors include advanced age, medical comorbidities, obesity, intraoperative time, anesthesia technique, surgical site infection, and incision length.^3-7

We conducted a study to identify the patient and clinical factors that affect LHS and to determine whether the specific day of the week when primary THA is performed affects LHS at a large tertiary-care university-based medical center. This information may prove valuable to hospital planning committees allotting operating room time and floor staffing for elective surgical cases with the goal of delivering cost-efficient care.

Materials and Methods

After obtaining institutional review board approval for this study, we retrospectively analyzed all primary unilateral THAs (273 patients) performed at our institution, a tertiary-care teaching hospital, between January 2010 and May 2011. The majority of the surgeries were performed through a posterior approach, and a majority of the implants were uncemented. All patients followed the same postoperative clinical pathway; no fast-track pathway was used.

The combined effects of day of surgery, American Society of Anesthesiologists (ASA) grade, anesthesia type, intraoperative time, estimated blood loss (EBL), incision length, presence of complications, age, sex, body mass index (BMI), disposition (skilled nursing facility vs home), transfusion, hematocrit, and hemoglobin on LHS were analyzed using a multiple quasi-Poisson regression model that included a random effect for surgeon. A Poisson regression model (typically used for count data) was deemed appropriate, as LHS was reported in whole days; a quasi-Poisson model relaxes the Poisson model assumption that the variance in the data equals the mean. The random effect for surgeon adjusts for any correlation among data from surgeries conducted by the same surgeon.

All complications were recorded. Complications included excess wound drainage,⁸ wound hematoma (a case of excess wound drainage necessitated surgical irrigation and débridement), new-onset atrial fibrillation, non-ST-elevation myocardial infarction, atrial flutter, urinary tract infection, pulmonary embolism, disseminated intravascular coagulation, hepatic decompensation as manifested by elevated liver enzymes, pneumonia, gastroesophageal reflux disease, gastric ulcer, sepsis, delirium, hypotension, and dysphagia.

The parameter estimates reported from the quasi-Poisson regression model are incident rate ratios (IRRs). IRR represents the change in expected LHS for a 1-unit change in a continuous variable (eg, age) or between categories of a categorical variable (eg, sex). IRR higher than 1 indicates higher risk as the continuous variable increases or a higher risk relative to the comparator group for a categorical variable. IRR lower than 1 indicates lower risk.

Results

Table 1 summarizes patient characteristics by surgical day. Mean LHS ranged from a minimum of 3.7 days for patients who had surgery on a Monday to a maximum of 4.2 days for patients who had surgery on a Thursday.

Table 2 summarizes results of the multivariate quasi-Poisson regression analysis of LHS by surgical day, ASA grade, anesthesia type, intraoperative time, EBL, incision length, presence of complications, age, sex, and BMI. With all other variables included in the model adjusted for, each additional point in ASA grade was associated with a 12% increase in LHS (P = .019). In addition, with all other variables included in the model adjusted for, LHS was 33% longer for patients with complications than for patients without complications (P < .001) and 12% longer for patients who received transfusions than for patients who did not (P = .046). LHS did not differ significantly by the day of the week when the surgery was performed (P = .496). Disposition status (skilled nursing facility vs home) as a variable to determine LHS did approach statistical significance (P = .061). As the effect size we were interested in detecting was an approximate 1-day increase in LHS for patients who had surgery later in the week relative to patients who had surgery earlier in the week, our sample size was adequate (range of required sample size, 200-300 patients). This study had 99% power to detect a 27% increase in LHS (equivalent to 1 day or more).

Discussion

This retrospective analysis explored how day of the week of primary THA affected LHS. Various confounders, such as surgery and patient factors, were also examined so that the multivariate analysis would be able to isolate the effects of surgical day of the week on LHS.

Effect of day of the week of primary THA on LHS was not investigated in the United States before. In Denmark, in a study similar to ours, Husted and colleagues⁴ found a 400% increase in the probability of LHS of more than 3 days when patients operated on a Thursday were compared with patients operated on a Friday. The authors reasoned that the Thursday patients most likely had a compromised physical therapy protocol owing to the inclusion of weekend days in the crucial postoperative period. LHS was consequently increased so that these patients would achieve their therapy goals before being discharged. Our investigation showed that LHS did not differ significantly by surgical day of the week. Although patients who had THA on a Thursday had 15% longer LHS than patients who had THA on a Monday, this difference was not statistically significant (P = .496), even though the study was adequately powered to detect a change in LHS of a whole day.

Table 3 summarizes the difference in quantum of workforce on weekdays and weekends at our center. The physiotherapy sessions were reduced to 1 per day. Nurse practitioners and discharge planners were not available on weekends, and some skilled nursing facilities and rehabilitation centers refused to accept patients on weekends. At our center, a teaching institute, the clinical duties of discharge planners and nurse practitioners were assumed by licensed physicians (orthopedic residents covering the arthroplasty team on weekends). This could be one of several possible reasons our study failed to detect statistically significant difference between the 2 groups. This kind of alternative arrangement may not be possible at many other centers. However, our study results provide a reasonably accurate logistical aim with regard to workforce availability on weekends to keep LHS in check.

The importance of giving patients an inpatient physical therapy regimen in timely fashion has been demonstrated in other studies. Munin and colleagues,⁹ in a randomized controlled trial, evaluated 71 patients who underwent elective hip and knee arthroplasty and received 2 different physical therapy regimens. Patients started their in-treatment physical therapy on postoperative day 3 or 7. Mean total LHS was shorter in the 3-day group (11.7 days) than in the 7-day group (14.5 days) (P < .001). Brusco and colleagues¹⁰ also showed that introducing weekend physical therapy services significantly reduced LHS in patients who underwent THA (10.6 vs 12.5 days; P < .05). Rapoport and Judd-Van Eerd¹¹ retrospectively analyzed orthopedic surgery LHS, comparing patients treated in a community hospital during a period of 5-days-a-week physical therapy coverage and patients treated during a period of 7-days-a-week physical therapy coverage. The 7-days-a-week group had significantly statistically shorter mean LHS.

Another rationale for analyzing the impact of surgical day of the week stems from the expectation that patients who undergo THA on Wednesday or Thursday and are scheduled to have physical therapy or be discharged on the weekend may be affected not only by reduced inpatient weekend physical therapy coverage but also by difficulties in being transferred to a skilled nursing facility or rehabilitation center if not discharged home. In our study, the patients who were to be discharged to a rehabilitation center were delayed by 12.5%, and this statistic trended toward significance (P = .061). Our literature search did not turn up any studies, US or European, specifically linking LHS to discharge disposition (whether patient is discharged home or to a skilled nursing facility or rehabilitation center).

Reduced medical staffing on weekends may not only affect the quality of in-hospital patient care but may also result in unnecessary delays in discharge. Chow and Szeto¹² retrospectively analyzed the medical records of all acute medical wards in a university hospital and compared weekend discharge rates before and after implementation of a work ordinance, which decreased the physician workforce by half on Saturday and Sunday. Results showed a 2.7% decrease in the weekend discharge rate after the work ordinance was established. The number of weekday discharges between the 2 time periods did not differ. Increasing the workforce availability presents a challenge in academic medical centers where graduate medical education enforces a strict cap on resident duty hours. Under these circumstances, a more feasible approach to decreasing LHS for THA patients is for surgical planning committees to provide the joint replacement services with operative block times early in the workweek.

Even though the organizational structure at our center is strong enough to provide for an adequate weekend workforce to discharge these patients, this study had a few limitations. We could not study readmission rates and whether the transition to home health and home physical therapy for the patients who went home was seamless.

We found that only 3 patient characteristics had a significant effect on LHS: higher ASA grade (a surrogate for medical comorbidities), requirement for blood transfusion, and presence of complications. In Denmark, blood transfusion increased the likelihood of longer LHS by 400%.⁴ In that study, patients who were ASA grades 1 and 2 had 60% and 20% decreased likelihood of LHS of more than 3 days compared with patients who were ASA grade 3. Similarly, in 2009, Mears and colleagues⁵ found 4 factors related to increased LHS: female sex (P < .001), older age (P < .001), higher ASA grade (3, P < .01; 4, P  < .001), and increased blood loss (P < .001).⁵

Conclusion

Over the past decade, there has been a significant reduction in LHS after THA, from a mean of 3 weeks to 4 days. Advances in implant technology, delivery of in-home physical therapy, and improved prevention and management of postoperative complications have contributed to this decline. Early identification of patients with transfusion requirements may be helpful in expediting their care. Although guidelines are in place for transfusion, further study in this regard may be needed. It is important to continue to identify surgery and patient factors that affect LHS, but the importance of organizational and planning issues in optimizing hospital health care expenditures cannot be ignored. Further study of providing a specific discharge planning service to identify patients’ discharge needs (home vs extended care facility) may help reduce LHS.

With health care costs increasing and economic resources diminishing, substantial efforts have been directed toward improving the quality of care delivered in a cost-effective manner. For a total hip arthroplasty (THA) performed in the United States between 1997 and 2001, total hospital cost, including direct and indirect costs, was estimated as averaging $13,339.¹ In 2012, this cost was estimated to be between $43,000 and $100,000.² This overall cost estimate, along with the rate at which the procedure is performed, may present an opportunity for cost savings.

Length of hospital stay (LHS) is an important outcome measure that has been assessed for optimal health care delivery. Prolonged LHS implies increased resource expenditure. Therefore, it is crucial to identify factors associated with prolonged LHS in order to reduce costs. Investigations have identified factors shown to affect LHS after THA. These factors include advanced age, medical comorbidities, obesity, intraoperative time, anesthesia technique, surgical site infection, and incision length.^3-7

We conducted a study to identify the patient and clinical factors that affect LHS and to determine whether the specific day of the week when primary THA is performed affects LHS at a large tertiary-care university-based medical center. This information may prove valuable to hospital planning committees allotting operating room time and floor staffing for elective surgical cases with the goal of delivering cost-efficient care.

Materials and Methods

After obtaining institutional review board approval for this study, we retrospectively analyzed all primary unilateral THAs (273 patients) performed at our institution, a tertiary-care teaching hospital, between January 2010 and May 2011. The majority of the surgeries were performed through a posterior approach, and a majority of the implants were uncemented. All patients followed the same postoperative clinical pathway; no fast-track pathway was used.

The combined effects of day of surgery, American Society of Anesthesiologists (ASA) grade, anesthesia type, intraoperative time, estimated blood loss (EBL), incision length, presence of complications, age, sex, body mass index (BMI), disposition (skilled nursing facility vs home), transfusion, hematocrit, and hemoglobin on LHS were analyzed using a multiple quasi-Poisson regression model that included a random effect for surgeon. A Poisson regression model (typically used for count data) was deemed appropriate, as LHS was reported in whole days; a quasi-Poisson model relaxes the Poisson model assumption that the variance in the data equals the mean. The random effect for surgeon adjusts for any correlation among data from surgeries conducted by the same surgeon.

All complications were recorded. Complications included excess wound drainage,⁸ wound hematoma (a case of excess wound drainage necessitated surgical irrigation and débridement), new-onset atrial fibrillation, non-ST-elevation myocardial infarction, atrial flutter, urinary tract infection, pulmonary embolism, disseminated intravascular coagulation, hepatic decompensation as manifested by elevated liver enzymes, pneumonia, gastroesophageal reflux disease, gastric ulcer, sepsis, delirium, hypotension, and dysphagia.

The parameter estimates reported from the quasi-Poisson regression model are incident rate ratios (IRRs). IRR represents the change in expected LHS for a 1-unit change in a continuous variable (eg, age) or between categories of a categorical variable (eg, sex). IRR higher than 1 indicates higher risk as the continuous variable increases or a higher risk relative to the comparator group for a categorical variable. IRR lower than 1 indicates lower risk.

Results

Table 1 summarizes patient characteristics by surgical day. Mean LHS ranged from a minimum of 3.7 days for patients who had surgery on a Monday to a maximum of 4.2 days for patients who had surgery on a Thursday.

Table 2 summarizes results of the multivariate quasi-Poisson regression analysis of LHS by surgical day, ASA grade, anesthesia type, intraoperative time, EBL, incision length, presence of complications, age, sex, and BMI. With all other variables included in the model adjusted for, each additional point in ASA grade was associated with a 12% increase in LHS (P = .019). In addition, with all other variables included in the model adjusted for, LHS was 33% longer for patients with complications than for patients without complications (P < .001) and 12% longer for patients who received transfusions than for patients who did not (P = .046). LHS did not differ significantly by the day of the week when the surgery was performed (P = .496). Disposition status (skilled nursing facility vs home) as a variable to determine LHS did approach statistical significance (P = .061). As the effect size we were interested in detecting was an approximate 1-day increase in LHS for patients who had surgery later in the week relative to patients who had surgery earlier in the week, our sample size was adequate (range of required sample size, 200-300 patients). This study had 99% power to detect a 27% increase in LHS (equivalent to 1 day or more).

Discussion

This retrospective analysis explored how day of the week of primary THA affected LHS. Various confounders, such as surgery and patient factors, were also examined so that the multivariate analysis would be able to isolate the effects of surgical day of the week on LHS.

Effect of day of the week of primary THA on LHS was not investigated in the United States before. In Denmark, in a study similar to ours, Husted and colleagues⁴ found a 400% increase in the probability of LHS of more than 3 days when patients operated on a Thursday were compared with patients operated on a Friday. The authors reasoned that the Thursday patients most likely had a compromised physical therapy protocol owing to the inclusion of weekend days in the crucial postoperative period. LHS was consequently increased so that these patients would achieve their therapy goals before being discharged. Our investigation showed that LHS did not differ significantly by surgical day of the week. Although patients who had THA on a Thursday had 15% longer LHS than patients who had THA on a Monday, this difference was not statistically significant (P = .496), even though the study was adequately powered to detect a change in LHS of a whole day.

Table 3 summarizes the difference in quantum of workforce on weekdays and weekends at our center. The physiotherapy sessions were reduced to 1 per day. Nurse practitioners and discharge planners were not available on weekends, and some skilled nursing facilities and rehabilitation centers refused to accept patients on weekends. At our center, a teaching institute, the clinical duties of discharge planners and nurse practitioners were assumed by licensed physicians (orthopedic residents covering the arthroplasty team on weekends). This could be one of several possible reasons our study failed to detect statistically significant difference between the 2 groups. This kind of alternative arrangement may not be possible at many other centers. However, our study results provide a reasonably accurate logistical aim with regard to workforce availability on weekends to keep LHS in check.

The importance of giving patients an inpatient physical therapy regimen in timely fashion has been demonstrated in other studies. Munin and colleagues,⁹ in a randomized controlled trial, evaluated 71 patients who underwent elective hip and knee arthroplasty and received 2 different physical therapy regimens. Patients started their in-treatment physical therapy on postoperative day 3 or 7. Mean total LHS was shorter in the 3-day group (11.7 days) than in the 7-day group (14.5 days) (P < .001). Brusco and colleagues¹⁰ also showed that introducing weekend physical therapy services significantly reduced LHS in patients who underwent THA (10.6 vs 12.5 days; P < .05). Rapoport and Judd-Van Eerd¹¹ retrospectively analyzed orthopedic surgery LHS, comparing patients treated in a community hospital during a period of 5-days-a-week physical therapy coverage and patients treated during a period of 7-days-a-week physical therapy coverage. The 7-days-a-week group had significantly statistically shorter mean LHS.

Another rationale for analyzing the impact of surgical day of the week stems from the expectation that patients who undergo THA on Wednesday or Thursday and are scheduled to have physical therapy or be discharged on the weekend may be affected not only by reduced inpatient weekend physical therapy coverage but also by difficulties in being transferred to a skilled nursing facility or rehabilitation center if not discharged home. In our study, the patients who were to be discharged to a rehabilitation center were delayed by 12.5%, and this statistic trended toward significance (P = .061). Our literature search did not turn up any studies, US or European, specifically linking LHS to discharge disposition (whether patient is discharged home or to a skilled nursing facility or rehabilitation center).

Reduced medical staffing on weekends may not only affect the quality of in-hospital patient care but may also result in unnecessary delays in discharge. Chow and Szeto¹² retrospectively analyzed the medical records of all acute medical wards in a university hospital and compared weekend discharge rates before and after implementation of a work ordinance, which decreased the physician workforce by half on Saturday and Sunday. Results showed a 2.7% decrease in the weekend discharge rate after the work ordinance was established. The number of weekday discharges between the 2 time periods did not differ. Increasing the workforce availability presents a challenge in academic medical centers where graduate medical education enforces a strict cap on resident duty hours. Under these circumstances, a more feasible approach to decreasing LHS for THA patients is for surgical planning committees to provide the joint replacement services with operative block times early in the workweek.

Even though the organizational structure at our center is strong enough to provide for an adequate weekend workforce to discharge these patients, this study had a few limitations. We could not study readmission rates and whether the transition to home health and home physical therapy for the patients who went home was seamless.

We found that only 3 patient characteristics had a significant effect on LHS: higher ASA grade (a surrogate for medical comorbidities), requirement for blood transfusion, and presence of complications. In Denmark, blood transfusion increased the likelihood of longer LHS by 400%.⁴ In that study, patients who were ASA grades 1 and 2 had 60% and 20% decreased likelihood of LHS of more than 3 days compared with patients who were ASA grade 3. Similarly, in 2009, Mears and colleagues⁵ found 4 factors related to increased LHS: female sex (P < .001), older age (P < .001), higher ASA grade (3, P < .01; 4, P  < .001), and increased blood loss (P < .001).⁵

Conclusion

Over the past decade, there has been a significant reduction in LHS after THA, from a mean of 3 weeks to 4 days. Advances in implant technology, delivery of in-home physical therapy, and improved prevention and management of postoperative complications have contributed to this decline. Early identification of patients with transfusion requirements may be helpful in expediting their care. Although guidelines are in place for transfusion, further study in this regard may be needed. It is important to continue to identify surgery and patient factors that affect LHS, but the importance of organizational and planning issues in optimizing hospital health care expenditures cannot be ignored. Further study of providing a specific discharge planning service to identify patients’ discharge needs (home vs extended care facility) may help reduce LHS.

With health care costs increasing and economic resources diminishing, substantial efforts have been directed toward improving the quality of care delivered in a cost-effective manner. For a total hip arthroplasty (THA) performed in the United States between 1997 and 2001, total hospital cost, including direct and indirect costs, was estimated as averaging $13,339.¹ In 2012, this cost was estimated to be between $43,000 and $100,000.² This overall cost estimate, along with the rate at which the procedure is performed, may present an opportunity for cost savings.

Length of hospital stay (LHS) is an important outcome measure that has been assessed for optimal health care delivery. Prolonged LHS implies increased resource expenditure. Therefore, it is crucial to identify factors associated with prolonged LHS in order to reduce costs. Investigations have identified factors shown to affect LHS after THA. These factors include advanced age, medical comorbidities, obesity, intraoperative time, anesthesia technique, surgical site infection, and incision length.^3-7

We conducted a study to identify the patient and clinical factors that affect LHS and to determine whether the specific day of the week when primary THA is performed affects LHS at a large tertiary-care university-based medical center. This information may prove valuable to hospital planning committees allotting operating room time and floor staffing for elective surgical cases with the goal of delivering cost-efficient care.

Materials and Methods

After obtaining institutional review board approval for this study, we retrospectively analyzed all primary unilateral THAs (273 patients) performed at our institution, a tertiary-care teaching hospital, between January 2010 and May 2011. The majority of the surgeries were performed through a posterior approach, and a majority of the implants were uncemented. All patients followed the same postoperative clinical pathway; no fast-track pathway was used.

The combined effects of day of surgery, American Society of Anesthesiologists (ASA) grade, anesthesia type, intraoperative time, estimated blood loss (EBL), incision length, presence of complications, age, sex, body mass index (BMI), disposition (skilled nursing facility vs home), transfusion, hematocrit, and hemoglobin on LHS were analyzed using a multiple quasi-Poisson regression model that included a random effect for surgeon. A Poisson regression model (typically used for count data) was deemed appropriate, as LHS was reported in whole days; a quasi-Poisson model relaxes the Poisson model assumption that the variance in the data equals the mean. The random effect for surgeon adjusts for any correlation among data from surgeries conducted by the same surgeon.

All complications were recorded. Complications included excess wound drainage,⁸ wound hematoma (a case of excess wound drainage necessitated surgical irrigation and débridement), new-onset atrial fibrillation, non-ST-elevation myocardial infarction, atrial flutter, urinary tract infection, pulmonary embolism, disseminated intravascular coagulation, hepatic decompensation as manifested by elevated liver enzymes, pneumonia, gastroesophageal reflux disease, gastric ulcer, sepsis, delirium, hypotension, and dysphagia.

The parameter estimates reported from the quasi-Poisson regression model are incident rate ratios (IRRs). IRR represents the change in expected LHS for a 1-unit change in a continuous variable (eg, age) or between categories of a categorical variable (eg, sex). IRR higher than 1 indicates higher risk as the continuous variable increases or a higher risk relative to the comparator group for a categorical variable. IRR lower than 1 indicates lower risk.

Results

Table 1 summarizes patient characteristics by surgical day. Mean LHS ranged from a minimum of 3.7 days for patients who had surgery on a Monday to a maximum of 4.2 days for patients who had surgery on a Thursday.

Table 2 summarizes results of the multivariate quasi-Poisson regression analysis of LHS by surgical day, ASA grade, anesthesia type, intraoperative time, EBL, incision length, presence of complications, age, sex, and BMI. With all other variables included in the model adjusted for, each additional point in ASA grade was associated with a 12% increase in LHS (P = .019). In addition, with all other variables included in the model adjusted for, LHS was 33% longer for patients with complications than for patients without complications (P < .001) and 12% longer for patients who received transfusions than for patients who did not (P = .046). LHS did not differ significantly by the day of the week when the surgery was performed (P = .496). Disposition status (skilled nursing facility vs home) as a variable to determine LHS did approach statistical significance (P = .061). As the effect size we were interested in detecting was an approximate 1-day increase in LHS for patients who had surgery later in the week relative to patients who had surgery earlier in the week, our sample size was adequate (range of required sample size, 200-300 patients). This study had 99% power to detect a 27% increase in LHS (equivalent to 1 day or more).

Discussion

This retrospective analysis explored how day of the week of primary THA affected LHS. Various confounders, such as surgery and patient factors, were also examined so that the multivariate analysis would be able to isolate the effects of surgical day of the week on LHS.

Effect of day of the week of primary THA on LHS was not investigated in the United States before. In Denmark, in a study similar to ours, Husted and colleagues⁴ found a 400% increase in the probability of LHS of more than 3 days when patients operated on a Thursday were compared with patients operated on a Friday. The authors reasoned that the Thursday patients most likely had a compromised physical therapy protocol owing to the inclusion of weekend days in the crucial postoperative period. LHS was consequently increased so that these patients would achieve their therapy goals before being discharged. Our investigation showed that LHS did not differ significantly by surgical day of the week. Although patients who had THA on a Thursday had 15% longer LHS than patients who had THA on a Monday, this difference was not statistically significant (P = .496), even though the study was adequately powered to detect a change in LHS of a whole day.

Table 3 summarizes the difference in quantum of workforce on weekdays and weekends at our center. The physiotherapy sessions were reduced to 1 per day. Nurse practitioners and discharge planners were not available on weekends, and some skilled nursing facilities and rehabilitation centers refused to accept patients on weekends. At our center, a teaching institute, the clinical duties of discharge planners and nurse practitioners were assumed by licensed physicians (orthopedic residents covering the arthroplasty team on weekends). This could be one of several possible reasons our study failed to detect statistically significant difference between the 2 groups. This kind of alternative arrangement may not be possible at many other centers. However, our study results provide a reasonably accurate logistical aim with regard to workforce availability on weekends to keep LHS in check.

The importance of giving patients an inpatient physical therapy regimen in timely fashion has been demonstrated in other studies. Munin and colleagues,⁹ in a randomized controlled trial, evaluated 71 patients who underwent elective hip and knee arthroplasty and received 2 different physical therapy regimens. Patients started their in-treatment physical therapy on postoperative day 3 or 7. Mean total LHS was shorter in the 3-day group (11.7 days) than in the 7-day group (14.5 days) (P < .001). Brusco and colleagues¹⁰ also showed that introducing weekend physical therapy services significantly reduced LHS in patients who underwent THA (10.6 vs 12.5 days; P < .05). Rapoport and Judd-Van Eerd¹¹ retrospectively analyzed orthopedic surgery LHS, comparing patients treated in a community hospital during a period of 5-days-a-week physical therapy coverage and patients treated during a period of 7-days-a-week physical therapy coverage. The 7-days-a-week group had significantly statistically shorter mean LHS.

Another rationale for analyzing the impact of surgical day of the week stems from the expectation that patients who undergo THA on Wednesday or Thursday and are scheduled to have physical therapy or be discharged on the weekend may be affected not only by reduced inpatient weekend physical therapy coverage but also by difficulties in being transferred to a skilled nursing facility or rehabilitation center if not discharged home. In our study, the patients who were to be discharged to a rehabilitation center were delayed by 12.5%, and this statistic trended toward significance (P = .061). Our literature search did not turn up any studies, US or European, specifically linking LHS to discharge disposition (whether patient is discharged home or to a skilled nursing facility or rehabilitation center).

Reduced medical staffing on weekends may not only affect the quality of in-hospital patient care but may also result in unnecessary delays in discharge. Chow and Szeto¹² retrospectively analyzed the medical records of all acute medical wards in a university hospital and compared weekend discharge rates before and after implementation of a work ordinance, which decreased the physician workforce by half on Saturday and Sunday. Results showed a 2.7% decrease in the weekend discharge rate after the work ordinance was established. The number of weekday discharges between the 2 time periods did not differ. Increasing the workforce availability presents a challenge in academic medical centers where graduate medical education enforces a strict cap on resident duty hours. Under these circumstances, a more feasible approach to decreasing LHS for THA patients is for surgical planning committees to provide the joint replacement services with operative block times early in the workweek.

Even though the organizational structure at our center is strong enough to provide for an adequate weekend workforce to discharge these patients, this study had a few limitations. We could not study readmission rates and whether the transition to home health and home physical therapy for the patients who went home was seamless.

We found that only 3 patient characteristics had a significant effect on LHS: higher ASA grade (a surrogate for medical comorbidities), requirement for blood transfusion, and presence of complications. In Denmark, blood transfusion increased the likelihood of longer LHS by 400%.⁴ In that study, patients who were ASA grades 1 and 2 had 60% and 20% decreased likelihood of LHS of more than 3 days compared with patients who were ASA grade 3. Similarly, in 2009, Mears and colleagues⁵ found 4 factors related to increased LHS: female sex (P < .001), older age (P < .001), higher ASA grade (3, P < .01; 4, P  < .001), and increased blood loss (P < .001).⁵

Conclusion

Over the past decade, there has been a significant reduction in LHS after THA, from a mean of 3 weeks to 4 days. Advances in implant technology, delivery of in-home physical therapy, and improved prevention and management of postoperative complications have contributed to this decline. Early identification of patients with transfusion requirements may be helpful in expediting their care. Although guidelines are in place for transfusion, further study in this regard may be needed. It is important to continue to identify surgery and patient factors that affect LHS, but the importance of organizational and planning issues in optimizing hospital health care expenditures cannot be ignored. Further study of providing a specific discharge planning service to identify patients’ discharge needs (home vs extended care facility) may help reduce LHS.

Serum metal ion levels are increased after primary total hip arthroplasty (THA) with all types of head-liner bearing surfaces.^1-4 In recent years, there has been increasing concern regarding elevated metal ion levels and adverse local and systemic effects, in particular, from metal-on-metal (MOM) implants.⁵ There have been reports of adverse local tissue reactions (ALTR) and systemic effects associated with elevated metal ion levels from MOM THA.^6-10 MOM hip resurfacings have been advocated in the literature for use in select patient populations, such as young, active individuals.^11,12 Purported benefits of MOM-bearing surfaces in this patient population include reduced wear and the prevention of osteolysis.¹³ While the incidence of ALTR has been reported to be approximately 1% within 5 years after MOM hip resurfacing, the prevalence of ALTR at long-term follow-up is unknown.¹⁴ Adverse local tissue reactions in hip resurfacing are related in part to femoral head coverage by the acetabular cup, as well as component design and orientation. The risk for ALTR with MOM THA has been reported to correlate with the level of serum metal ion levels because of the bearing surface, along with taper corrosion and corrosion secondary to the large surface area of the femoral head.^13-15 The overall clinical and prognostic value of metal ion levels in cases of MOM bearings remains controversial and without clear cut-off values.¹⁵

The long-term physiologic response to elevated serum metal ion levels from THA implants remains unknown and is of particular concern in pregnant women because of potential transplacental transfer. Several studies have shown that DNA (deoxyribonucleic acid) and chromosomal changes can occur in patients with both MOM and non-MOM hip implants.^16-18 Techniques to accurately measure the levels of metal ions, including cobalt, chromium, and titanium in the serum, have improved substantially in recent years and have been used successfully in clinical applications with low detection limits.^2,4,19,20 Evidence shows that pregnancy-related complications in women with well-functioning cemented, hybrid, or uncemented metal-on-polyethylene THA are not different from those in healthy women.²¹ However, it is unclear if metal ions can cross the placenta and negatively affect the development of a fetus in mothers with MOM-bearing THAs.²² We conducted this study to determine if metal ions can traverse the placenta by measuring serum metal ion concentrations in mothers with and without MOM THA and the corresponding levels in umbilical cord blood samples. 

Materials and Methods

Ten patients were prospectively enrolled in this study, which was approved by the institutional review boards at all 3 participating hospitals prior to initiation. All patients provided informed consent and agreed to maternal and umbilical cord blood sampling. Seven of the 10 patients did not have any metallic implants and served as controls. The remaining 3 patients had either a MOM unilateral primary THA (1 patient) or unilateral MOM hip-resurfacing arthroplasty (2 patients) with no other metal implants in the body. For all subjects, maternal and umbilical cord blood was obtained at the time of delivery.

Group Demographics

The 3 women in the implant group had a mean age of 32.3 years (range, 23-39 years) at time of delivery. The first patient had posttraumatic osteoarthritis and underwent right THA using a modular MOM 60-mm acetabular component, a 36-mm cobalt chrome head, and a cementless, titanium proximally porous-coated femoral component (Pinnacle Total Hip System and Summit Total Hip System; DePuy Synthes, Warsaw, Indiana). Her infant was born 2 years after index THA. The second patient had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent a hybrid MOM hip resurfacing with a monoblock 50-mm cup and 40-mm cobalt chrome head (Conserve Plus; Wright Medical Technology, Arlington, Tennessee). She gave birth to her infant 6 years after her hip resurfacing. The third patient also had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent hip resurfacing with a 54-mm monoblock cup and 44-mm cobalt chrome head (Conserve Plus). Her infant was born 4 years after her resurfacing arthroplasty. All of the infants were born healthy, and the deliveries were uneventful and without complications. Seven women with a mean age of 32.1 years (range, 24-37 years) and their infants served as controls at the time of delivery. None of the women in the control group had a history of renal impairment, inherited genetic disorders, or metal implants.

Serum Metal Ion Analysis

Blood samples were collected using S-Monovette polypropylene syringes (Sarstedt, Princeton, New Jersey), a multi-adapter, and infusion set following an established technique.¹ All vessels and utensils used for specimen collection were verified to be free of metal contamination. Three 10-mL syringes were drawn, and each syringe was labeled to indicate the sequence of collection. The first 10 mL were drawn to rinse the needle and adapter. Blood was then allowed to clot naturally and centrifuged at 1850 rpm for 30 minutes, separating samples into cell and serum fractions that were stored in labeled vials at -80ºC. All specimen manipulations after collection were carried out in a class-100 environment using a SterilGARD Biological Safety Cabinet (Baker, Sanford, Maine) and class-100 gloves (Oak Technical, Ravenna, Ohio) to minimize atmospheric and manual contamination. Maternal and umbilical cord blood metal ion levels were tested using high-resolution sector-field inductively-coupled plasma-mass spectrometry (HR-SF-ICPMS) (Element 2; Thermo Fisher Scientific, Bremen, Germany) following the method of additions as described previously.²³ The HR-SF-ICPMS machine was equipped with an SC-E2 autosampler, Teflon nebulizer and spray chamber, sapphire injector (Elemental Scientific, Omaha, Nebraska) and platinum cones. All calibration and internal standard solutions were prepared by gradual dilutions of single-element standard solutions (1000 μg/mL from High Purity Standards, Charleston, South Carolina). The certified reference material, Seronorm Trace Elements Serum (SERO, Billingstad, Norway), were routinely analyzed with samples. The serum concentrations of cobalt, chromium, titanium, and nickel were measured with detection limits in ng/mL (parts per billion) of 0.04 for cobalt, 0.015 for chromium, 0.2 for titanium, and 0.17 for nickel.⁴ Concentrations below the detection limit were approximated as one-half of the detection limit by convention to calculate means.

Statistical Analysis

The data reported are the means for each group for each of the metal ion levels analyzed.  Intergroup comparisons were made with the Mann-Whitney-Wilcoxon test using SPSS statistics software (SPSS Science Inc, Chicago, Illinois) to compare implant and control groups in regards to serum metal ion levels. Intragroup comparisons were made using the Friedman test with significance set at P < .05. Spearman rank-order correlation tests were used to investigate relationships between maternal and infant serum metal ion levels. 

Results

The Table shows the mean serum metal ion levels of chromium, cobalt, titanium, and nickel for both groups. The implant-group mothers had significantly higher chromium and cobalt levels than the control-group mothers, with mean chromium levels of 1.87 ng/mL vs 0.16 ng/mL (P = .01) and mean cobalt levels of 0.97 ng/mL vs 0.20 ng/mL (P = .01), respectively. All control-group maternal chromium and cobalt levels were lower than the implant group. There were no significant differences detected between the implant-group and control-group mothers with respect to serum titanium or nickel levels.

The implant-group cord blood samples also had significantly higher chromium and cobalt levels than did the control-group infants with mean chromium levels of 0.29 ng/mL vs 0.10 ng/mL (P = .03) and mean cobalt levels of 0.49 ng/mL vs 0.16 ng/mL (P = .01), respectively. All but 1 of the control-group infants had chromium levels that were lower than the implant-group infant chromium levels. All of the control-group infant cobalt levels were lower than the implant-group infant cobalt levels. In the mother (I-1) who had a titanium-containing implant, her titanium level was 1.77 ng/mL and her cord blood level was 0.78 ng/mL. In contrast, the other 2 patients did not have titanium-containing implants and had corresponding metal levels of 0.10 ng/mL and cord blood levels either below the detection limit or just slightly above it. No statistically significant differences were found between the implant- and control-group infants with respect to serum titanium or nickel levels.

Considering the implant and control groups separately, we found no statistically significant differences between the maternal and infant titanium levels or the maternal and infant nickel levels. In the implant group, the mother’s chromium level was always higher than her infant’s chromium level, and the mother’s cobalt level was always higher than her infant’s cobalt level (P= .08). In the control group, there was no correlation between the maternal and infant chromium levels, or between the maternal and infant cobalt levels. In the implant group, the maternal and infant chromium levels were highly correlated (r = 1), as were the maternal and infant cobalt levels (r = 1). 

When infants’ chromium levels were expressed as a percentage of their mothers’ chromium levels, the mean was 15.4% (range, 12.3%-18%) for the implant group and 58.9% (range, 13.5%-165.7%) for the control group (P = .05). The mean infant cobalt level, expressed as percentage of maternal cobalt level, was 50% (range, 45.8%-53.1%) for the implant group and 76.6% (range, 59.1%-150.8%) for the control group (P = .01).

Discussion 

Cobalt and chromium serum metal ion levels obtained from mothers with MOM implants and umbilical cord samples were significantly elevated in comparison with controls. There was also a strong correlation between implant-group maternal cobalt and chromium levels and implant-group infant cobalt and chromium levels; however, no such correlation existed in the control group. Our results suggest that cobalt and chromium cross the placental barrier. Implant mothers had higher chromium and cobalt levels compared with implant babies, and these babies had approximately 15% of the level of chromium and 50% of the level of cobalt when compared to implant maternal levels. This finding suggests that the placenta modulates the transfer of chromium and cobalt to the fetus. 

Two studies have reported chromium and cobalt levels in maternal and umbilical cord sera after MOM THA.^5,24 Brodner and colleagues⁵ determined the maternal serum levels of cobalt and chromium in 3 women 3.8 years after MOM THA and compared those to cobalt and chromium levels obtained from umbilical cord blood. At the time of delivery, the maternal chromium concentrations in the 3 patients were 1.6 ng/mL, 0.5 ng/mL, and 0.9 ng/mL, and the cobalt concentrations were 1 ng/mL in 1 patient and below the detection limit in the other 2 patients. Cobalt and chromium concentrations of the 3 umbilical cord sera were below the detection limit. The authors concluded that cobalt and chromium did not cross the placenta based on their laboratory detection limits. Metal ion levels were measured using atomic absorption spectrometry with relatively high detection limits of 0.3 ng/mL, which is not as sensitive as the HR-SF-ICPMS technique used in the present study that has detection limits of 0.04 ng/mL for cobalt and 0.015 ng/mL for chromium. The relatively high detection limits of atomic absorption spectrometry were likely responsible for the authors’ inability to detect elevated chromium and cobalt levels in umbilical cord sera.

Ziaee and colleagues²⁴ used HR-SF-ICPMS, as we did in this study, to measure the mean concentrations of cobalt and chromium ion levels in 10 maternal and umbilical cord blood samples in women with Birmingham MOM hip-resurfacing prostheses (Smith & Nephew, Warwick, United Kingdom). Nine of those patients had a unilateral resurfacing and 1 patient had bilateral-resurfacing prostheses. The mean maternal age was 31 years, and mean duration between hip resurfacing and delivery was 53 months. Ten normal controls were also tested with a mean maternal age of 30.9 years. The authors found that the mean cord blood level of cobalt in the study patients was 0.83 ng/mL, significantly higher (P < .01) than cobalt levels in the control group, which measured 0.33 ng/mL. The mean cord blood levels of chromium in the study and control groups were 0.37 ng/mL and 0.19 ng/mL, respectively. No children were reported to have evidence of congenital anomalies. Similar to our findings, they noted a modulatory effect on the transfer of metal ions across the placenta in patients with MOM prostheses.²⁴ They reported the relative levels of cobalt and chromium in offspring to be 60.4% and 29.4% of the maternal ion levels, respectively. Control-group infants had mean cobalt and chromium levels that were 98.5% and 97.2% of mean maternal levels, respectively. 

The transfer of metal ions across the placenta in control subjects is an expected finding because cobalt and chromium are essential trace elements required by the developing fetus. Rudge and colleagues²⁵ estimated a transplacental transfer rate of 45% for cobalt in a series of 62 paired samples of maternal and cord blood. DeSouza and coauthors²⁶ reported a series of 3 patients with MOM hip resurfacings who had the prosthesis in situ during pregnancy and found no teratogenic effects of metal ion transfer across the placenta. Umbilical cord blood chromium levels were less than 25% of the maternal serum levels and cord blood cobalt levels were approximately 50% that of maternal blood.²⁶ In an animal experiment, Wallach and Verch¹⁸ also reported that maternal chromium levels can be decreased because of placental uptake.

Ziaee and colleagues²⁴ tested metal ion levels using whole blood in contrast to serum, as we did in the present study. Daniel and coauthors¹⁶, who reported on the validity of serum levels as a surrogate measure of systemic exposure to metal ions in hip replacement, suggested that serum and whole blood metal ion levels cannot be interconverted because metal within cells are not in dynamic equilibrium with extracellular levels. They concluded that serum metal ion concentrations are not a useful surrogate measure of systemic metal ion exposure based on the wide variability seen in normalized and Bland-Altman scatterplots.¹⁶ However, it is important to note that Bland-Altman plots are user-dependent in determining significance, and results can vary based on the parameters tested. A high correlation does not automatically imply that there is a good agreement between 2 methods because a widespread sample could influence results. Whole blood analysis requires more processing steps, thus providing an increased chance of contamination and variability compared with serum metal level analysis. In our experiences, serum metal ion analysis has been shown to have accurate and reproducible results in clinical situations.^2,4

While there is insufficient literature that specifically studies the effects of elevated metal ion levels on maternal and fetal subjects, there have been no reported negative effects in human babies even when maternal ion levels are elevated enough to be associated with ALTR. A case report by Fritzsche and colleagues²⁷ reported a mother with bilateral MOM THA, a recurrent pseudotumor, and high blood levels of chromium (39 ng/mL) and cobalt (138 ng/mL) at 12 weeks gestation. The child was born at 38 weeks gestation with cord blood chromium and cobalt levels of 2.1 ng/mL and 75 ng/mL, respectively. The infant’s metal ion levels remained elevated at age 8 weeks with a chromium level of 2.5 ng/mL and cobalt level of 13 ng/mL and no signs of toxicity by age 14 weeks. In an animal model, Saxena and colleagues²⁸ found that chromium in the hexavalent form passed through the placenta in mice and rats that were fed high doses of potassium dichromate. Trivalent chromium was not found to cross the placenta. In a follow-up study, Junaid and coauthors²⁹ investigated the effects of elevated chromium levels in female mice given potassium dichromate in drinking water on days 14 to 19 of pregnancy. Animals receiving high-dose chromium had significantly higher incidences of postimplantation loss along with subdermal hemorrhagic patches and reduced ossification.²⁹ Cobalt has not been shown to be teratogenic or cause fetotoxicity in a rat animal model given daily doses of as much as 100 mg/kg cobalt (II) chloride on days 6 to 15 of gestation.³⁰

It is important to recognize that rodent data are limited and may not provide accurate translational insight into the effects of metal ions in human maternal and fetal subjects. Mammalian species have significant heterogeneity in the structure and function of their placentas. Rurak³¹ has shown that rodents have an additional persisting yolk sac placenta that allows the transfer of maternal immunoglobulins to the fetus. Humans, on the other hand, have a yolk sac placenta that regresses early in pregnancy. Differing placental biologic function makes it difficult to extrapolate the effects of metal ions in rodents to human subjects.

It is also important to note that serum levels of cobalt, chromium, and titanium can remain persistently elevated in well-functioning metal-on-polyethylene THA for several years and that elevated metal ion levels are not confined to MOM bearings.² Levine and colleagues⁴ reported that serum levels of cobalt, chromium, and titanium remain persistently elevated after 10 years in a cohort of 27 well-functioning primary metal-on-polyethylene THA (hybrid, cobalt-chrome, titanium). Cobalt concentrations were elevated in all implants compared with controls at all follow-up periods through 10 years with absolute values less than 1 ng/mL. The authors noted that metal release at the modular femoral head-neck junctions was likely the dominant source of serum cobalt and chromium rather than passive dissolution. Hsu and colleagues³² have also shown that patients undergoing a second metal-on-polyethylene THA after primary THA have elevated serum metal ion levels (cobalt, chromium, titanium) up to 6 years after second surgery. Reported cobalt concentrations in patients with unilateral THA reached a maximum of 0.5 ng/mL during the follow-up course compared with 1.5 ng/mL for patients with bilateral THA. It is unknown what the potential metal ion transfer load would be in mothers with metal-on-polyethylene THA and associated taper corrosion to their infants. 

Conclusion

Mothers with MOM-bearing implants and their children have higher cobalt and chromium levels than control subjects, demonstrating that the placenta is not a complete barrier to metal ion transport, although it seems to have a modulating effect. Physicians and women of child-bearing age should be aware of these findings when considering the use of MOM-bearing couples for THA. The effects of metal ions on long-term maternal and fetal health require research through serial clinical exams and metal ion level testing in prospective studies of different THA-bearing surfaces.

Serum metal ion levels are increased after primary total hip arthroplasty (THA) with all types of head-liner bearing surfaces.^1-4 In recent years, there has been increasing concern regarding elevated metal ion levels and adverse local and systemic effects, in particular, from metal-on-metal (MOM) implants.⁵ There have been reports of adverse local tissue reactions (ALTR) and systemic effects associated with elevated metal ion levels from MOM THA.^6-10 MOM hip resurfacings have been advocated in the literature for use in select patient populations, such as young, active individuals.^11,12 Purported benefits of MOM-bearing surfaces in this patient population include reduced wear and the prevention of osteolysis.¹³ While the incidence of ALTR has been reported to be approximately 1% within 5 years after MOM hip resurfacing, the prevalence of ALTR at long-term follow-up is unknown.¹⁴ Adverse local tissue reactions in hip resurfacing are related in part to femoral head coverage by the acetabular cup, as well as component design and orientation. The risk for ALTR with MOM THA has been reported to correlate with the level of serum metal ion levels because of the bearing surface, along with taper corrosion and corrosion secondary to the large surface area of the femoral head.^13-15 The overall clinical and prognostic value of metal ion levels in cases of MOM bearings remains controversial and without clear cut-off values.¹⁵

The long-term physiologic response to elevated serum metal ion levels from THA implants remains unknown and is of particular concern in pregnant women because of potential transplacental transfer. Several studies have shown that DNA (deoxyribonucleic acid) and chromosomal changes can occur in patients with both MOM and non-MOM hip implants.^16-18 Techniques to accurately measure the levels of metal ions, including cobalt, chromium, and titanium in the serum, have improved substantially in recent years and have been used successfully in clinical applications with low detection limits.^2,4,19,20 Evidence shows that pregnancy-related complications in women with well-functioning cemented, hybrid, or uncemented metal-on-polyethylene THA are not different from those in healthy women.²¹ However, it is unclear if metal ions can cross the placenta and negatively affect the development of a fetus in mothers with MOM-bearing THAs.²² We conducted this study to determine if metal ions can traverse the placenta by measuring serum metal ion concentrations in mothers with and without MOM THA and the corresponding levels in umbilical cord blood samples. 

Materials and Methods

Ten patients were prospectively enrolled in this study, which was approved by the institutional review boards at all 3 participating hospitals prior to initiation. All patients provided informed consent and agreed to maternal and umbilical cord blood sampling. Seven of the 10 patients did not have any metallic implants and served as controls. The remaining 3 patients had either a MOM unilateral primary THA (1 patient) or unilateral MOM hip-resurfacing arthroplasty (2 patients) with no other metal implants in the body. For all subjects, maternal and umbilical cord blood was obtained at the time of delivery.

Group Demographics

The 3 women in the implant group had a mean age of 32.3 years (range, 23-39 years) at time of delivery. The first patient had posttraumatic osteoarthritis and underwent right THA using a modular MOM 60-mm acetabular component, a 36-mm cobalt chrome head, and a cementless, titanium proximally porous-coated femoral component (Pinnacle Total Hip System and Summit Total Hip System; DePuy Synthes, Warsaw, Indiana). Her infant was born 2 years after index THA. The second patient had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent a hybrid MOM hip resurfacing with a monoblock 50-mm cup and 40-mm cobalt chrome head (Conserve Plus; Wright Medical Technology, Arlington, Tennessee). She gave birth to her infant 6 years after her hip resurfacing. The third patient also had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent hip resurfacing with a 54-mm monoblock cup and 44-mm cobalt chrome head (Conserve Plus). Her infant was born 4 years after her resurfacing arthroplasty. All of the infants were born healthy, and the deliveries were uneventful and without complications. Seven women with a mean age of 32.1 years (range, 24-37 years) and their infants served as controls at the time of delivery. None of the women in the control group had a history of renal impairment, inherited genetic disorders, or metal implants.

Serum Metal Ion Analysis

Blood samples were collected using S-Monovette polypropylene syringes (Sarstedt, Princeton, New Jersey), a multi-adapter, and infusion set following an established technique.¹ All vessels and utensils used for specimen collection were verified to be free of metal contamination. Three 10-mL syringes were drawn, and each syringe was labeled to indicate the sequence of collection. The first 10 mL were drawn to rinse the needle and adapter. Blood was then allowed to clot naturally and centrifuged at 1850 rpm for 30 minutes, separating samples into cell and serum fractions that were stored in labeled vials at -80ºC. All specimen manipulations after collection were carried out in a class-100 environment using a SterilGARD Biological Safety Cabinet (Baker, Sanford, Maine) and class-100 gloves (Oak Technical, Ravenna, Ohio) to minimize atmospheric and manual contamination. Maternal and umbilical cord blood metal ion levels were tested using high-resolution sector-field inductively-coupled plasma-mass spectrometry (HR-SF-ICPMS) (Element 2; Thermo Fisher Scientific, Bremen, Germany) following the method of additions as described previously.²³ The HR-SF-ICPMS machine was equipped with an SC-E2 autosampler, Teflon nebulizer and spray chamber, sapphire injector (Elemental Scientific, Omaha, Nebraska) and platinum cones. All calibration and internal standard solutions were prepared by gradual dilutions of single-element standard solutions (1000 μg/mL from High Purity Standards, Charleston, South Carolina). The certified reference material, Seronorm Trace Elements Serum (SERO, Billingstad, Norway), were routinely analyzed with samples. The serum concentrations of cobalt, chromium, titanium, and nickel were measured with detection limits in ng/mL (parts per billion) of 0.04 for cobalt, 0.015 for chromium, 0.2 for titanium, and 0.17 for nickel.⁴ Concentrations below the detection limit were approximated as one-half of the detection limit by convention to calculate means.

Statistical Analysis

The data reported are the means for each group for each of the metal ion levels analyzed.  Intergroup comparisons were made with the Mann-Whitney-Wilcoxon test using SPSS statistics software (SPSS Science Inc, Chicago, Illinois) to compare implant and control groups in regards to serum metal ion levels. Intragroup comparisons were made using the Friedman test with significance set at P < .05. Spearman rank-order correlation tests were used to investigate relationships between maternal and infant serum metal ion levels. 

Results

The Table shows the mean serum metal ion levels of chromium, cobalt, titanium, and nickel for both groups. The implant-group mothers had significantly higher chromium and cobalt levels than the control-group mothers, with mean chromium levels of 1.87 ng/mL vs 0.16 ng/mL (P = .01) and mean cobalt levels of 0.97 ng/mL vs 0.20 ng/mL (P = .01), respectively. All control-group maternal chromium and cobalt levels were lower than the implant group. There were no significant differences detected between the implant-group and control-group mothers with respect to serum titanium or nickel levels.

The implant-group cord blood samples also had significantly higher chromium and cobalt levels than did the control-group infants with mean chromium levels of 0.29 ng/mL vs 0.10 ng/mL (P = .03) and mean cobalt levels of 0.49 ng/mL vs 0.16 ng/mL (P = .01), respectively. All but 1 of the control-group infants had chromium levels that were lower than the implant-group infant chromium levels. All of the control-group infant cobalt levels were lower than the implant-group infant cobalt levels. In the mother (I-1) who had a titanium-containing implant, her titanium level was 1.77 ng/mL and her cord blood level was 0.78 ng/mL. In contrast, the other 2 patients did not have titanium-containing implants and had corresponding metal levels of 0.10 ng/mL and cord blood levels either below the detection limit or just slightly above it. No statistically significant differences were found between the implant- and control-group infants with respect to serum titanium or nickel levels.

Considering the implant and control groups separately, we found no statistically significant differences between the maternal and infant titanium levels or the maternal and infant nickel levels. In the implant group, the mother’s chromium level was always higher than her infant’s chromium level, and the mother’s cobalt level was always higher than her infant’s cobalt level (P= .08). In the control group, there was no correlation between the maternal and infant chromium levels, or between the maternal and infant cobalt levels. In the implant group, the maternal and infant chromium levels were highly correlated (r = 1), as were the maternal and infant cobalt levels (r = 1). 

When infants’ chromium levels were expressed as a percentage of their mothers’ chromium levels, the mean was 15.4% (range, 12.3%-18%) for the implant group and 58.9% (range, 13.5%-165.7%) for the control group (P = .05). The mean infant cobalt level, expressed as percentage of maternal cobalt level, was 50% (range, 45.8%-53.1%) for the implant group and 76.6% (range, 59.1%-150.8%) for the control group (P = .01).

Discussion 

Cobalt and chromium serum metal ion levels obtained from mothers with MOM implants and umbilical cord samples were significantly elevated in comparison with controls. There was also a strong correlation between implant-group maternal cobalt and chromium levels and implant-group infant cobalt and chromium levels; however, no such correlation existed in the control group. Our results suggest that cobalt and chromium cross the placental barrier. Implant mothers had higher chromium and cobalt levels compared with implant babies, and these babies had approximately 15% of the level of chromium and 50% of the level of cobalt when compared to implant maternal levels. This finding suggests that the placenta modulates the transfer of chromium and cobalt to the fetus. 

Two studies have reported chromium and cobalt levels in maternal and umbilical cord sera after MOM THA.^5,24 Brodner and colleagues⁵ determined the maternal serum levels of cobalt and chromium in 3 women 3.8 years after MOM THA and compared those to cobalt and chromium levels obtained from umbilical cord blood. At the time of delivery, the maternal chromium concentrations in the 3 patients were 1.6 ng/mL, 0.5 ng/mL, and 0.9 ng/mL, and the cobalt concentrations were 1 ng/mL in 1 patient and below the detection limit in the other 2 patients. Cobalt and chromium concentrations of the 3 umbilical cord sera were below the detection limit. The authors concluded that cobalt and chromium did not cross the placenta based on their laboratory detection limits. Metal ion levels were measured using atomic absorption spectrometry with relatively high detection limits of 0.3 ng/mL, which is not as sensitive as the HR-SF-ICPMS technique used in the present study that has detection limits of 0.04 ng/mL for cobalt and 0.015 ng/mL for chromium. The relatively high detection limits of atomic absorption spectrometry were likely responsible for the authors’ inability to detect elevated chromium and cobalt levels in umbilical cord sera.

Ziaee and colleagues²⁴ used HR-SF-ICPMS, as we did in this study, to measure the mean concentrations of cobalt and chromium ion levels in 10 maternal and umbilical cord blood samples in women with Birmingham MOM hip-resurfacing prostheses (Smith & Nephew, Warwick, United Kingdom). Nine of those patients had a unilateral resurfacing and 1 patient had bilateral-resurfacing prostheses. The mean maternal age was 31 years, and mean duration between hip resurfacing and delivery was 53 months. Ten normal controls were also tested with a mean maternal age of 30.9 years. The authors found that the mean cord blood level of cobalt in the study patients was 0.83 ng/mL, significantly higher (P < .01) than cobalt levels in the control group, which measured 0.33 ng/mL. The mean cord blood levels of chromium in the study and control groups were 0.37 ng/mL and 0.19 ng/mL, respectively. No children were reported to have evidence of congenital anomalies. Similar to our findings, they noted a modulatory effect on the transfer of metal ions across the placenta in patients with MOM prostheses.²⁴ They reported the relative levels of cobalt and chromium in offspring to be 60.4% and 29.4% of the maternal ion levels, respectively. Control-group infants had mean cobalt and chromium levels that were 98.5% and 97.2% of mean maternal levels, respectively. 

The transfer of metal ions across the placenta in control subjects is an expected finding because cobalt and chromium are essential trace elements required by the developing fetus. Rudge and colleagues²⁵ estimated a transplacental transfer rate of 45% for cobalt in a series of 62 paired samples of maternal and cord blood. DeSouza and coauthors²⁶ reported a series of 3 patients with MOM hip resurfacings who had the prosthesis in situ during pregnancy and found no teratogenic effects of metal ion transfer across the placenta. Umbilical cord blood chromium levels were less than 25% of the maternal serum levels and cord blood cobalt levels were approximately 50% that of maternal blood.²⁶ In an animal experiment, Wallach and Verch¹⁸ also reported that maternal chromium levels can be decreased because of placental uptake.

Ziaee and colleagues²⁴ tested metal ion levels using whole blood in contrast to serum, as we did in the present study. Daniel and coauthors¹⁶, who reported on the validity of serum levels as a surrogate measure of systemic exposure to metal ions in hip replacement, suggested that serum and whole blood metal ion levels cannot be interconverted because metal within cells are not in dynamic equilibrium with extracellular levels. They concluded that serum metal ion concentrations are not a useful surrogate measure of systemic metal ion exposure based on the wide variability seen in normalized and Bland-Altman scatterplots.¹⁶ However, it is important to note that Bland-Altman plots are user-dependent in determining significance, and results can vary based on the parameters tested. A high correlation does not automatically imply that there is a good agreement between 2 methods because a widespread sample could influence results. Whole blood analysis requires more processing steps, thus providing an increased chance of contamination and variability compared with serum metal level analysis. In our experiences, serum metal ion analysis has been shown to have accurate and reproducible results in clinical situations.^2,4

While there is insufficient literature that specifically studies the effects of elevated metal ion levels on maternal and fetal subjects, there have been no reported negative effects in human babies even when maternal ion levels are elevated enough to be associated with ALTR. A case report by Fritzsche and colleagues²⁷ reported a mother with bilateral MOM THA, a recurrent pseudotumor, and high blood levels of chromium (39 ng/mL) and cobalt (138 ng/mL) at 12 weeks gestation. The child was born at 38 weeks gestation with cord blood chromium and cobalt levels of 2.1 ng/mL and 75 ng/mL, respectively. The infant’s metal ion levels remained elevated at age 8 weeks with a chromium level of 2.5 ng/mL and cobalt level of 13 ng/mL and no signs of toxicity by age 14 weeks. In an animal model, Saxena and colleagues²⁸ found that chromium in the hexavalent form passed through the placenta in mice and rats that were fed high doses of potassium dichromate. Trivalent chromium was not found to cross the placenta. In a follow-up study, Junaid and coauthors²⁹ investigated the effects of elevated chromium levels in female mice given potassium dichromate in drinking water on days 14 to 19 of pregnancy. Animals receiving high-dose chromium had significantly higher incidences of postimplantation loss along with subdermal hemorrhagic patches and reduced ossification.²⁹ Cobalt has not been shown to be teratogenic or cause fetotoxicity in a rat animal model given daily doses of as much as 100 mg/kg cobalt (II) chloride on days 6 to 15 of gestation.³⁰

It is important to recognize that rodent data are limited and may not provide accurate translational insight into the effects of metal ions in human maternal and fetal subjects. Mammalian species have significant heterogeneity in the structure and function of their placentas. Rurak³¹ has shown that rodents have an additional persisting yolk sac placenta that allows the transfer of maternal immunoglobulins to the fetus. Humans, on the other hand, have a yolk sac placenta that regresses early in pregnancy. Differing placental biologic function makes it difficult to extrapolate the effects of metal ions in rodents to human subjects.

It is also important to note that serum levels of cobalt, chromium, and titanium can remain persistently elevated in well-functioning metal-on-polyethylene THA for several years and that elevated metal ion levels are not confined to MOM bearings.² Levine and colleagues⁴ reported that serum levels of cobalt, chromium, and titanium remain persistently elevated after 10 years in a cohort of 27 well-functioning primary metal-on-polyethylene THA (hybrid, cobalt-chrome, titanium). Cobalt concentrations were elevated in all implants compared with controls at all follow-up periods through 10 years with absolute values less than 1 ng/mL. The authors noted that metal release at the modular femoral head-neck junctions was likely the dominant source of serum cobalt and chromium rather than passive dissolution. Hsu and colleagues³² have also shown that patients undergoing a second metal-on-polyethylene THA after primary THA have elevated serum metal ion levels (cobalt, chromium, titanium) up to 6 years after second surgery. Reported cobalt concentrations in patients with unilateral THA reached a maximum of 0.5 ng/mL during the follow-up course compared with 1.5 ng/mL for patients with bilateral THA. It is unknown what the potential metal ion transfer load would be in mothers with metal-on-polyethylene THA and associated taper corrosion to their infants. 

Conclusion

Mothers with MOM-bearing implants and their children have higher cobalt and chromium levels than control subjects, demonstrating that the placenta is not a complete barrier to metal ion transport, although it seems to have a modulating effect. Physicians and women of child-bearing age should be aware of these findings when considering the use of MOM-bearing couples for THA. The effects of metal ions on long-term maternal and fetal health require research through serial clinical exams and metal ion level testing in prospective studies of different THA-bearing surfaces.

Serum metal ion levels are increased after primary total hip arthroplasty (THA) with all types of head-liner bearing surfaces.^1-4 In recent years, there has been increasing concern regarding elevated metal ion levels and adverse local and systemic effects, in particular, from metal-on-metal (MOM) implants.⁵ There have been reports of adverse local tissue reactions (ALTR) and systemic effects associated with elevated metal ion levels from MOM THA.^6-10 MOM hip resurfacings have been advocated in the literature for use in select patient populations, such as young, active individuals.^11,12 Purported benefits of MOM-bearing surfaces in this patient population include reduced wear and the prevention of osteolysis.¹³ While the incidence of ALTR has been reported to be approximately 1% within 5 years after MOM hip resurfacing, the prevalence of ALTR at long-term follow-up is unknown.¹⁴ Adverse local tissue reactions in hip resurfacing are related in part to femoral head coverage by the acetabular cup, as well as component design and orientation. The risk for ALTR with MOM THA has been reported to correlate with the level of serum metal ion levels because of the bearing surface, along with taper corrosion and corrosion secondary to the large surface area of the femoral head.^13-15 The overall clinical and prognostic value of metal ion levels in cases of MOM bearings remains controversial and without clear cut-off values.¹⁵

The long-term physiologic response to elevated serum metal ion levels from THA implants remains unknown and is of particular concern in pregnant women because of potential transplacental transfer. Several studies have shown that DNA (deoxyribonucleic acid) and chromosomal changes can occur in patients with both MOM and non-MOM hip implants.^16-18 Techniques to accurately measure the levels of metal ions, including cobalt, chromium, and titanium in the serum, have improved substantially in recent years and have been used successfully in clinical applications with low detection limits.^2,4,19,20 Evidence shows that pregnancy-related complications in women with well-functioning cemented, hybrid, or uncemented metal-on-polyethylene THA are not different from those in healthy women.²¹ However, it is unclear if metal ions can cross the placenta and negatively affect the development of a fetus in mothers with MOM-bearing THAs.²² We conducted this study to determine if metal ions can traverse the placenta by measuring serum metal ion concentrations in mothers with and without MOM THA and the corresponding levels in umbilical cord blood samples. 

Materials and Methods

Ten patients were prospectively enrolled in this study, which was approved by the institutional review boards at all 3 participating hospitals prior to initiation. All patients provided informed consent and agreed to maternal and umbilical cord blood sampling. Seven of the 10 patients did not have any metallic implants and served as controls. The remaining 3 patients had either a MOM unilateral primary THA (1 patient) or unilateral MOM hip-resurfacing arthroplasty (2 patients) with no other metal implants in the body. For all subjects, maternal and umbilical cord blood was obtained at the time of delivery.

Group Demographics

The 3 women in the implant group had a mean age of 32.3 years (range, 23-39 years) at time of delivery. The first patient had posttraumatic osteoarthritis and underwent right THA using a modular MOM 60-mm acetabular component, a 36-mm cobalt chrome head, and a cementless, titanium proximally porous-coated femoral component (Pinnacle Total Hip System and Summit Total Hip System; DePuy Synthes, Warsaw, Indiana). Her infant was born 2 years after index THA. The second patient had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent a hybrid MOM hip resurfacing with a monoblock 50-mm cup and 40-mm cobalt chrome head (Conserve Plus; Wright Medical Technology, Arlington, Tennessee). She gave birth to her infant 6 years after her hip resurfacing. The third patient also had a diagnosis of degenerative osteoarthritis secondary to developmental dysplasia of the hip and underwent hip resurfacing with a 54-mm monoblock cup and 44-mm cobalt chrome head (Conserve Plus). Her infant was born 4 years after her resurfacing arthroplasty. All of the infants were born healthy, and the deliveries were uneventful and without complications. Seven women with a mean age of 32.1 years (range, 24-37 years) and their infants served as controls at the time of delivery. None of the women in the control group had a history of renal impairment, inherited genetic disorders, or metal implants.

Serum Metal Ion Analysis

Blood samples were collected using S-Monovette polypropylene syringes (Sarstedt, Princeton, New Jersey), a multi-adapter, and infusion set following an established technique.¹ All vessels and utensils used for specimen collection were verified to be free of metal contamination. Three 10-mL syringes were drawn, and each syringe was labeled to indicate the sequence of collection. The first 10 mL were drawn to rinse the needle and adapter. Blood was then allowed to clot naturally and centrifuged at 1850 rpm for 30 minutes, separating samples into cell and serum fractions that were stored in labeled vials at -80ºC. All specimen manipulations after collection were carried out in a class-100 environment using a SterilGARD Biological Safety Cabinet (Baker, Sanford, Maine) and class-100 gloves (Oak Technical, Ravenna, Ohio) to minimize atmospheric and manual contamination. Maternal and umbilical cord blood metal ion levels were tested using high-resolution sector-field inductively-coupled plasma-mass spectrometry (HR-SF-ICPMS) (Element 2; Thermo Fisher Scientific, Bremen, Germany) following the method of additions as described previously.²³ The HR-SF-ICPMS machine was equipped with an SC-E2 autosampler, Teflon nebulizer and spray chamber, sapphire injector (Elemental Scientific, Omaha, Nebraska) and platinum cones. All calibration and internal standard solutions were prepared by gradual dilutions of single-element standard solutions (1000 μg/mL from High Purity Standards, Charleston, South Carolina). The certified reference material, Seronorm Trace Elements Serum (SERO, Billingstad, Norway), were routinely analyzed with samples. The serum concentrations of cobalt, chromium, titanium, and nickel were measured with detection limits in ng/mL (parts per billion) of 0.04 for cobalt, 0.015 for chromium, 0.2 for titanium, and 0.17 for nickel.⁴ Concentrations below the detection limit were approximated as one-half of the detection limit by convention to calculate means.

Statistical Analysis

The data reported are the means for each group for each of the metal ion levels analyzed.  Intergroup comparisons were made with the Mann-Whitney-Wilcoxon test using SPSS statistics software (SPSS Science Inc, Chicago, Illinois) to compare implant and control groups in regards to serum metal ion levels. Intragroup comparisons were made using the Friedman test with significance set at P < .05. Spearman rank-order correlation tests were used to investigate relationships between maternal and infant serum metal ion levels. 

Results

The Table shows the mean serum metal ion levels of chromium, cobalt, titanium, and nickel for both groups. The implant-group mothers had significantly higher chromium and cobalt levels than the control-group mothers, with mean chromium levels of 1.87 ng/mL vs 0.16 ng/mL (P = .01) and mean cobalt levels of 0.97 ng/mL vs 0.20 ng/mL (P = .01), respectively. All control-group maternal chromium and cobalt levels were lower than the implant group. There were no significant differences detected between the implant-group and control-group mothers with respect to serum titanium or nickel levels.

The implant-group cord blood samples also had significantly higher chromium and cobalt levels than did the control-group infants with mean chromium levels of 0.29 ng/mL vs 0.10 ng/mL (P = .03) and mean cobalt levels of 0.49 ng/mL vs 0.16 ng/mL (P = .01), respectively. All but 1 of the control-group infants had chromium levels that were lower than the implant-group infant chromium levels. All of the control-group infant cobalt levels were lower than the implant-group infant cobalt levels. In the mother (I-1) who had a titanium-containing implant, her titanium level was 1.77 ng/mL and her cord blood level was 0.78 ng/mL. In contrast, the other 2 patients did not have titanium-containing implants and had corresponding metal levels of 0.10 ng/mL and cord blood levels either below the detection limit or just slightly above it. No statistically significant differences were found between the implant- and control-group infants with respect to serum titanium or nickel levels.

Considering the implant and control groups separately, we found no statistically significant differences between the maternal and infant titanium levels or the maternal and infant nickel levels. In the implant group, the mother’s chromium level was always higher than her infant’s chromium level, and the mother’s cobalt level was always higher than her infant’s cobalt level (P= .08). In the control group, there was no correlation between the maternal and infant chromium levels, or between the maternal and infant cobalt levels. In the implant group, the maternal and infant chromium levels were highly correlated (r = 1), as were the maternal and infant cobalt levels (r = 1). 

When infants’ chromium levels were expressed as a percentage of their mothers’ chromium levels, the mean was 15.4% (range, 12.3%-18%) for the implant group and 58.9% (range, 13.5%-165.7%) for the control group (P = .05). The mean infant cobalt level, expressed as percentage of maternal cobalt level, was 50% (range, 45.8%-53.1%) for the implant group and 76.6% (range, 59.1%-150.8%) for the control group (P = .01).

Discussion 

Cobalt and chromium serum metal ion levels obtained from mothers with MOM implants and umbilical cord samples were significantly elevated in comparison with controls. There was also a strong correlation between implant-group maternal cobalt and chromium levels and implant-group infant cobalt and chromium levels; however, no such correlation existed in the control group. Our results suggest that cobalt and chromium cross the placental barrier. Implant mothers had higher chromium and cobalt levels compared with implant babies, and these babies had approximately 15% of the level of chromium and 50% of the level of cobalt when compared to implant maternal levels. This finding suggests that the placenta modulates the transfer of chromium and cobalt to the fetus. 

Two studies have reported chromium and cobalt levels in maternal and umbilical cord sera after MOM THA.^5,24 Brodner and colleagues⁵ determined the maternal serum levels of cobalt and chromium in 3 women 3.8 years after MOM THA and compared those to cobalt and chromium levels obtained from umbilical cord blood. At the time of delivery, the maternal chromium concentrations in the 3 patients were 1.6 ng/mL, 0.5 ng/mL, and 0.9 ng/mL, and the cobalt concentrations were 1 ng/mL in 1 patient and below the detection limit in the other 2 patients. Cobalt and chromium concentrations of the 3 umbilical cord sera were below the detection limit. The authors concluded that cobalt and chromium did not cross the placenta based on their laboratory detection limits. Metal ion levels were measured using atomic absorption spectrometry with relatively high detection limits of 0.3 ng/mL, which is not as sensitive as the HR-SF-ICPMS technique used in the present study that has detection limits of 0.04 ng/mL for cobalt and 0.015 ng/mL for chromium. The relatively high detection limits of atomic absorption spectrometry were likely responsible for the authors’ inability to detect elevated chromium and cobalt levels in umbilical cord sera.

Ziaee and colleagues²⁴ used HR-SF-ICPMS, as we did in this study, to measure the mean concentrations of cobalt and chromium ion levels in 10 maternal and umbilical cord blood samples in women with Birmingham MOM hip-resurfacing prostheses (Smith & Nephew, Warwick, United Kingdom). Nine of those patients had a unilateral resurfacing and 1 patient had bilateral-resurfacing prostheses. The mean maternal age was 31 years, and mean duration between hip resurfacing and delivery was 53 months. Ten normal controls were also tested with a mean maternal age of 30.9 years. The authors found that the mean cord blood level of cobalt in the study patients was 0.83 ng/mL, significantly higher (P < .01) than cobalt levels in the control group, which measured 0.33 ng/mL. The mean cord blood levels of chromium in the study and control groups were 0.37 ng/mL and 0.19 ng/mL, respectively. No children were reported to have evidence of congenital anomalies. Similar to our findings, they noted a modulatory effect on the transfer of metal ions across the placenta in patients with MOM prostheses.²⁴ They reported the relative levels of cobalt and chromium in offspring to be 60.4% and 29.4% of the maternal ion levels, respectively. Control-group infants had mean cobalt and chromium levels that were 98.5% and 97.2% of mean maternal levels, respectively. 

The transfer of metal ions across the placenta in control subjects is an expected finding because cobalt and chromium are essential trace elements required by the developing fetus. Rudge and colleagues²⁵ estimated a transplacental transfer rate of 45% for cobalt in a series of 62 paired samples of maternal and cord blood. DeSouza and coauthors²⁶ reported a series of 3 patients with MOM hip resurfacings who had the prosthesis in situ during pregnancy and found no teratogenic effects of metal ion transfer across the placenta. Umbilical cord blood chromium levels were less than 25% of the maternal serum levels and cord blood cobalt levels were approximately 50% that of maternal blood.²⁶ In an animal experiment, Wallach and Verch¹⁸ also reported that maternal chromium levels can be decreased because of placental uptake.

Ziaee and colleagues²⁴ tested metal ion levels using whole blood in contrast to serum, as we did in the present study. Daniel and coauthors¹⁶, who reported on the validity of serum levels as a surrogate measure of systemic exposure to metal ions in hip replacement, suggested that serum and whole blood metal ion levels cannot be interconverted because metal within cells are not in dynamic equilibrium with extracellular levels. They concluded that serum metal ion concentrations are not a useful surrogate measure of systemic metal ion exposure based on the wide variability seen in normalized and Bland-Altman scatterplots.¹⁶ However, it is important to note that Bland-Altman plots are user-dependent in determining significance, and results can vary based on the parameters tested. A high correlation does not automatically imply that there is a good agreement between 2 methods because a widespread sample could influence results. Whole blood analysis requires more processing steps, thus providing an increased chance of contamination and variability compared with serum metal level analysis. In our experiences, serum metal ion analysis has been shown to have accurate and reproducible results in clinical situations.^2,4

While there is insufficient literature that specifically studies the effects of elevated metal ion levels on maternal and fetal subjects, there have been no reported negative effects in human babies even when maternal ion levels are elevated enough to be associated with ALTR. A case report by Fritzsche and colleagues²⁷ reported a mother with bilateral MOM THA, a recurrent pseudotumor, and high blood levels of chromium (39 ng/mL) and cobalt (138 ng/mL) at 12 weeks gestation. The child was born at 38 weeks gestation with cord blood chromium and cobalt levels of 2.1 ng/mL and 75 ng/mL, respectively. The infant’s metal ion levels remained elevated at age 8 weeks with a chromium level of 2.5 ng/mL and cobalt level of 13 ng/mL and no signs of toxicity by age 14 weeks. In an animal model, Saxena and colleagues²⁸ found that chromium in the hexavalent form passed through the placenta in mice and rats that were fed high doses of potassium dichromate. Trivalent chromium was not found to cross the placenta. In a follow-up study, Junaid and coauthors²⁹ investigated the effects of elevated chromium levels in female mice given potassium dichromate in drinking water on days 14 to 19 of pregnancy. Animals receiving high-dose chromium had significantly higher incidences of postimplantation loss along with subdermal hemorrhagic patches and reduced ossification.²⁹ Cobalt has not been shown to be teratogenic or cause fetotoxicity in a rat animal model given daily doses of as much as 100 mg/kg cobalt (II) chloride on days 6 to 15 of gestation.³⁰

It is important to recognize that rodent data are limited and may not provide accurate translational insight into the effects of metal ions in human maternal and fetal subjects. Mammalian species have significant heterogeneity in the structure and function of their placentas. Rurak³¹ has shown that rodents have an additional persisting yolk sac placenta that allows the transfer of maternal immunoglobulins to the fetus. Humans, on the other hand, have a yolk sac placenta that regresses early in pregnancy. Differing placental biologic function makes it difficult to extrapolate the effects of metal ions in rodents to human subjects.

It is also important to note that serum levels of cobalt, chromium, and titanium can remain persistently elevated in well-functioning metal-on-polyethylene THA for several years and that elevated metal ion levels are not confined to MOM bearings.² Levine and colleagues⁴ reported that serum levels of cobalt, chromium, and titanium remain persistently elevated after 10 years in a cohort of 27 well-functioning primary metal-on-polyethylene THA (hybrid, cobalt-chrome, titanium). Cobalt concentrations were elevated in all implants compared with controls at all follow-up periods through 10 years with absolute values less than 1 ng/mL. The authors noted that metal release at the modular femoral head-neck junctions was likely the dominant source of serum cobalt and chromium rather than passive dissolution. Hsu and colleagues³² have also shown that patients undergoing a second metal-on-polyethylene THA after primary THA have elevated serum metal ion levels (cobalt, chromium, titanium) up to 6 years after second surgery. Reported cobalt concentrations in patients with unilateral THA reached a maximum of 0.5 ng/mL during the follow-up course compared with 1.5 ng/mL for patients with bilateral THA. It is unknown what the potential metal ion transfer load would be in mothers with metal-on-polyethylene THA and associated taper corrosion to their infants. 

Conclusion

Mothers with MOM-bearing implants and their children have higher cobalt and chromium levels than control subjects, demonstrating that the placenta is not a complete barrier to metal ion transport, although it seems to have a modulating effect. Physicians and women of child-bearing age should be aware of these findings when considering the use of MOM-bearing couples for THA. The effects of metal ions on long-term maternal and fetal health require research through serial clinical exams and metal ion level testing in prospective studies of different THA-bearing surfaces.

Partial meniscectomy of tears in the avascular zone remains one of the most common orthopedic procedures. While results of partial meniscectomy in younger patients have excellent short- to medium-term results, the long-term clinical outcomes are often less favorable.^1-3 Repair in the avascular “white-white” zone has resulted in lower patient satisfaction scores and higher revision surgery rates.^4-7 Consequently, most tears in this region have been treated with partial meniscectomy.

The inability to repair rather than resect menisci with avascular tears has led to extensive research. Techniques such as trephination and rasping to initiate an angiogenic response have had inconsistent and unreliable results when applied to the white-white zone.^8-13 In contrast, Tasto and colleagues¹⁴ have shown that radiofrequency (RF) applied to hypovascular tissue can not only stimulate tissue vascularity, but also increase organization of fibroblastic cells. In addition, Tasto and colleagues^15,16 have shown that RF application can significantly improve histologic healing and clinical outcomes in refractory cases of Achilles tendinopathy and lateral epicondylitis. In Japanese white rabbit menisci, Higuchi and coauthors¹⁷  applied monopolar RF at 60°C and 40W to avascular zone tears to fuse the tissue. They found a significant increase in fibroblast proliferation and fusion of collagen fibers at 2, 4, and 12 weeks after surgery. They also found significant acellular zones of meniscus tissue and attributed these findings to fibrochondrocyte death because of thermal treatment. 

This body of research led to the present study, which evaluates the effect of low-temperature, bipolar RF stimulation, in conjunction with suture repair, on the healing of tears in the avascular white-white zone of the meniscus both in vivo and ex vivo. We performed gross and histologic analyses of the treatment groups for the in vivo aspect of the study and biochemical analyses to study the ex vivo effects of RF treatment. ³H-thymidine incorporation has been shown to be a reliable indicator of cellular proliferation in several studies, and this was measured in our treatment groups.^18-21 In addition, the response of mitogenic growth factors (IGF-1, bFGF) and angiogenic markers (VEGF, αV) to RF treatment was studied.²² We hypothesized that bipolar RF application would show increased gross, histologic, and biochemical healing when combined with suture repair of longitudinal avascular zone meniscus tears.

Materials and Methods

Creation of Meniscal Tears

Fifty-four healthy, skeletally mature male and female adult New Zealand white rabbits aged 7 to 18 months were used for the study. All procedures conformed to the guidelines of our university’s institutional animal care and use committee and the American Association for Accreditation of Laboratory Animal Care. All rabbits underwent a surgical procedure in which a longitudinal tear was created in the avascular white-white zone of the medial meniscus. Using sterile technique and instrumentation, a medial parapatellar incision was made on the left knee of each rabbit. The patella was retracted laterally, exposing the medial meniscus. The tibia was then externally rotated to sublux the anterior horn of the medial meniscus anteriorly. A longitudinal full-thickness meniscal tear (3-4 mm in length) was created in the avascular zone (inner third) of the anterior horn of the medial meniscus using an 11-blade scalpel (Figures 1A, 1B). The location of the tear was grossly performed in the same location in each meniscus. The rabbits were randomly divided into 3 treatment groups: 1, 2, and 3 (Table 1). Group 1 (n = 6) served as a control with no repair or RF treatment applied. Group 2 (n = 15) underwent suture repair only of the meniscal tear using 5-0 nylon suture in a horizontal mattress pattern (Figure 2A). Group 3 (n = 33) underwent suture repair after RF stimulation was applied to both sides of the meniscus tear (Figures 2B, 2C). RF was applied using a 0.8-mm TOPAZ MicroDebrider (ArthroCare, Sunnyvale, California) set at level 4 (175 V-RMS) for 500 milliseconds. Lactated ringer’s solution was continuously infused through the probe via sterile tubing to prevent overheating.

After meniscal treatment, hemostasis of the surrounding surgical dissection was achieved to prevent hematoma formation, and the wounds were irrigated. The patellae were relocated and the arthrotomies were closed with a running 2-0 vicryl suture. Fascial and subcutaneous layers were closed with a running 3-0 vicryl suture, and skin was closed with subcuticular 4‑0 vicryl sutures. The rabbit limbs were allowed weight-bearing with unrestricted range of motion within 2x2x2-ft cages. 

For all groups, menisci were explanted at 28 and 84 days for gross and histologic analysis. For biochemical assessments, menisci were explanted at 9, 28, and 84 days (Table 1).  

Gross Analysis

Immediately after specimen removal, all medial menisci were evaluated for gross morphology. A grading system was used for organization and classification of data (Table 2). Three blinded orthopedic surgeon-observers performed all grading. Grade A was considered complete healing of the meniscus. Grade B involved complete healing with a trace of injury remaining on the surface of the meniscus. Grade C represented incomplete healing with a full-thickness injury that was stable to stress of the repair site with an arthroscopic probe. Grade D had no healing with the injured region unstable to stress of the repair site with an arthroscopic probe.

Histologic Analysis and Microscopic Grading of Meniscal Healing

After gross evaluation by the 3 blinded observers, each meniscus was fixed for 24 hours in 10% buffered formalin. Each specimen was then embedded in paraffin and cut into 6-µm slices along the radial plane. The tissue samples were stained with hematoxylin-eosin, and microscopic grading was assigned. The grading system was the same as that used for gross morphologic analysis. 

Biochemical Analysis

To determine whether RF treatment stimulated a healing response in the avascular zone of the meniscus, measurements of specific biochemical markers were analyzed at 9, 28, and 84 days after treatment. As a control, unrepaired meniscal tissue from the contralateral knee was also analyzed. ³H-thymidine incorporation into the meniscus was measured to assess cell proliferation.²³ At sacrifice, control and treated menisci were dissected and immediately placed into sterile culture media (Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, antibiotic, and fungicide). ³H-thymidine was added at a concentration of 5µCi/mL of media to each tube. After incubation for 48 hours at 37°C under 5% CO₂, the menisci were removed and dialyzed against water for 24 hours to remove unincorporated thymidine. After washing, the menisci were lyophilized, aliquots weighed, and radioactivity determined by liquid scintillation spectrometry. Results are expressed as counts per minute per mg dry tissue weight.

Semiquantitative reverse transcription polymerase chain reaction (RT-PCR) was used to determine mRNA expression of mitogenic growth factors, IGF-1 and bFGF, and angiogenic markers, αV and VEFG.²⁴ National Institutes of Health (NIH) image-analysis software (version 1.61; NIH, Bethesda, Maryland) was used to quantitatively scan RT-PCR profiles after agarose gel electrophoresis and ethidium bromide visualization. Values were normalized to the housekeeping gene, GAPDH.

Statistical Analysis

Data are expressed as mean (SD) and evaluated using an unpaired Student t test between groups. Statistical significance was established at P < .05.

Results

Gross Morphology

Analysis of gross morphology showed signs of healing only in the group treated with suture repair combined with RF treatment (Table 3). In group 1 (meniscal injury only) and group 2 (suture repair only), no healing occurred at 28 and 84 days (Figure 3A). A meniscal grading system was developed to better describe the varying levels of healing shown in the suture-plus-RF-treatment group (Table 2). Of the specimens that showed healing in group 3, 1 had complete healing (grade A) within the avascular zone of the meniscus at 84 days (Figure 3B). In addition, 4 specimens subjected to suture repair and RF treatment had complete healing with only a trace of injured tissue remaining (grade B). Fourteen specimens in group 3 had incomplete healing with lesions stable to stress suggesting early signs of healing (grade C). In total, 58% of menisci treated with RF showed signs of healing while the remaining 14 specimens in group 3 showed none (grade D).

Histologic Examination  

The histology correlated well with gross analysis. No microscopic evidence of healing was seen in groups 1 and 2 (Figure 3C). Of the specimens treated with suture repair combined with RF, 19 (58%) showed varying degrees of histologic healing. While gross morphologic examination showed that only 1 specimen had complete healing, microscopic analysis showed that 1 specimen from group 3 had grade B healing on gross analysis but grade A healing on histologic analysis. Thus, upon histologic examination, 2 specimens showed complete healing of injuries in the avascular zone of the meniscus when treated with suture repair combined with RF treatment rather than the 1 specimen seen on gross morphology (Figure 3D).

Biochemical Analysis

Biochemical assessments were performed at 9, 28, and 84 days after surgery. ³H-thymidine incorporation was studied as a marker for cellular proliferation, and its levels were significantly higher in meniscus explants treated with RF (Figure 4). The mean (SD) rate of incorporation for meniscal tears treated with suture repair plus RF was 590 (80) cpm/mg dry tissue at 9 days. This value was approximately 40% greater than the menisci treated with suture repair only, which had a mean (SD) value of 380 (30) cpm/mg (P < .05). Normal, unrepaired meniscal tissue had a mean (SD) ³H-thymidine incorporation rate of 250 (35) cpm/mg. By 84 days, thymidine levels returned to uninjured levels in both suture-only and RF-treated menisci. Semiquantitative RT-PCR analysis showed that, 9 days after repair, the RF-treated menisci had increased mRNA expression of IGF-1, bFGF, VEGF, and αV relative to untreated repairs (Figure 5). There was a statistically significant acute phase response in IGF-1, bFGF, VEGF and αV in groups treated with RF at 9 days (P > .05).

Adverse Outcomes

There were no surgical complications. During the histologic evaluation, there were no incidences of fibrochondrocyte cell death or damage from the application of RF treatment. 

Discussion

RF treatments have been used for many years in various medical and surgical applications. Presently, the most common implementation of RF is for cutting and coagulating tissue during surgery. More recently, however, several publications have shown that when used properly and safely, RF can be an effective surgical adjunct for tendinosis recalcitrant to conservative therapy.^15-17,25-32

Many have suggested that RF coblation is successful in these clinical scenarios because of its ability to promote an increased angiogenic and fibroblastic response in hypovascular tissue.^29,33,34

This body of literature led to the evaluation of RF coblation in treating meniscal tears in the avascular zone. Studies have shown poor success of meniscus repairs done in the avascular zone; however, our data demonstrate that supplementing suture repair with RF treatment may improve the acute-phase healing response. Although the control and suture-repair groups showed no signs of healing, the suture-repair-combined-with-RF-treatment group had 2 specimens in which complete gross and histologic healing occurred. In addition, 19 (58%) specimens in the RF group showed gross or histologic signs of healing.

Biochemically, ³H-thymidine incorporation was examined to assess cellular proliferation. Mitogenic (IGF, bFGF) and angiogenic (VEGF, αV) growth factors were measured as markers of an increased healing response. Compared with noninjured meniscal tissue, ³H-thymidine incorporation was significantly increased in both the suture and suture-combined-with-RF-treatment groups at 9 and 28 days after surgery. Between the suture and suture-RF groups, RF treatment led to a 40% greater increase in ³H-thymidine incorporation suggesting greater cellular proliferation in the immediate postoperative period. With respect to mitogenic and angiogenic factors, IGF, bFGF, VEGF, and αV were only significantly increased when RF was combined with suture repair. All 4 factors are important regulators of vasculogenesis, angiogenesis, wound healing, bone remodeling, and neurogenesis. The suture repair–only group showed no upregulation of these factors compared with uninjured controls.

Our study has several strengths. Using an animal model with menisci grossly similar to that of humans, we performed a controlled study comparing 2 treatment options, suture repair only and suture repair combined with RF treatment.^35,36 The animal model also enabled second-look examinations at designated intervals. We analyzed the effect of RF treatment on concrete measures, such as gross, histologic, and biochemical healing. In particular, the biochemical analysis may indicate that RF treatment can increase the proliferative, mitogenic, and angiogenic capabilities of surrounding progenitor cells. This was evidenced by the statistically significant increase we saw in IGF-1, bFGF, VEGF, and αV at 9 and 28 days compared with controls.

Meniscal tears in the avascular zone represent a significant treatment dilemma for the physiologically young patient population. While partial meniscectomy provides excellent short-term relief, the long-term outcome of this intervention is degenerative joint disease. Meniscal repair in the central two-thirds of the meniscus has shown poor results. Our study presents data that show supplementing suture repair of avascular meniscal tears with RF can lead to increased gross, histologic, and biochemical healing in the New Zealand white rabbit. While these results are encouraging, studies with longer follow-up and specimens that represent the human menisci are necessary to determine whether these preliminary results would translate to human meniscal tears in the avascular zone.

Weaknesses of our study include the use of an animal model and the location of the tear created in the menisci. While using an animal model had many strengths, the results of our study are probably not strong enough to immediately extrapolate the use of RF in human meniscal repairs. However, the data we obtained are very encouraging and perhaps suggest that RF warrants human trials. Our open surgical technique made it difficult to create and repair a tear on the posterior horn of the medical meniscus without completely dislocating the knee anteriorly. As a result, the knees were subluxed anteriorly, and the meniscal tears and repairs were performed more anteriorly. The more anterior aspects of the menisci do not undergo the same rotational and axial loads as the posterior horn, and it is unclear whether this difference would contribute to the results we obtained from RF treatment. In addition, the tears were surgically created and the repair was done during the same procedure. Patients do not present in this manner, and this further underscores the need for a clinical trial to determine the effectiveness of this treatment option in humans.

Conclusion

RF-based supplementation of meniscal repairs in the avascular zone showed acute signs of biochemical healing in 58% of New Zealand white rabbit specimens. In addition, gross and histologic evaluations showed an increase in healing compared with controls. Two specimens treated with RF in combination with suture repair had complete healing. These results illustrate the effectiveness of RF in stimulating a healing response in hypovascular tissue. Clinical trials are necessary to determine the effectiveness of this treatment in humans.

Partial meniscectomy of tears in the avascular zone remains one of the most common orthopedic procedures. While results of partial meniscectomy in younger patients have excellent short- to medium-term results, the long-term clinical outcomes are often less favorable.^1-3 Repair in the avascular “white-white” zone has resulted in lower patient satisfaction scores and higher revision surgery rates.^4-7 Consequently, most tears in this region have been treated with partial meniscectomy.

The inability to repair rather than resect menisci with avascular tears has led to extensive research. Techniques such as trephination and rasping to initiate an angiogenic response have had inconsistent and unreliable results when applied to the white-white zone.^8-13 In contrast, Tasto and colleagues¹⁴ have shown that radiofrequency (RF) applied to hypovascular tissue can not only stimulate tissue vascularity, but also increase organization of fibroblastic cells. In addition, Tasto and colleagues^15,16 have shown that RF application can significantly improve histologic healing and clinical outcomes in refractory cases of Achilles tendinopathy and lateral epicondylitis. In Japanese white rabbit menisci, Higuchi and coauthors¹⁷  applied monopolar RF at 60°C and 40W to avascular zone tears to fuse the tissue. They found a significant increase in fibroblast proliferation and fusion of collagen fibers at 2, 4, and 12 weeks after surgery. They also found significant acellular zones of meniscus tissue and attributed these findings to fibrochondrocyte death because of thermal treatment. 

This body of research led to the present study, which evaluates the effect of low-temperature, bipolar RF stimulation, in conjunction with suture repair, on the healing of tears in the avascular white-white zone of the meniscus both in vivo and ex vivo. We performed gross and histologic analyses of the treatment groups for the in vivo aspect of the study and biochemical analyses to study the ex vivo effects of RF treatment. ³H-thymidine incorporation has been shown to be a reliable indicator of cellular proliferation in several studies, and this was measured in our treatment groups.^18-21 In addition, the response of mitogenic growth factors (IGF-1, bFGF) and angiogenic markers (VEGF, αV) to RF treatment was studied.²² We hypothesized that bipolar RF application would show increased gross, histologic, and biochemical healing when combined with suture repair of longitudinal avascular zone meniscus tears.

Materials and Methods

Creation of Meniscal Tears

Fifty-four healthy, skeletally mature male and female adult New Zealand white rabbits aged 7 to 18 months were used for the study. All procedures conformed to the guidelines of our university’s institutional animal care and use committee and the American Association for Accreditation of Laboratory Animal Care. All rabbits underwent a surgical procedure in which a longitudinal tear was created in the avascular white-white zone of the medial meniscus. Using sterile technique and instrumentation, a medial parapatellar incision was made on the left knee of each rabbit. The patella was retracted laterally, exposing the medial meniscus. The tibia was then externally rotated to sublux the anterior horn of the medial meniscus anteriorly. A longitudinal full-thickness meniscal tear (3-4 mm in length) was created in the avascular zone (inner third) of the anterior horn of the medial meniscus using an 11-blade scalpel (Figures 1A, 1B). The location of the tear was grossly performed in the same location in each meniscus. The rabbits were randomly divided into 3 treatment groups: 1, 2, and 3 (Table 1). Group 1 (n = 6) served as a control with no repair or RF treatment applied. Group 2 (n = 15) underwent suture repair only of the meniscal tear using 5-0 nylon suture in a horizontal mattress pattern (Figure 2A). Group 3 (n = 33) underwent suture repair after RF stimulation was applied to both sides of the meniscus tear (Figures 2B, 2C). RF was applied using a 0.8-mm TOPAZ MicroDebrider (ArthroCare, Sunnyvale, California) set at level 4 (175 V-RMS) for 500 milliseconds. Lactated ringer’s solution was continuously infused through the probe via sterile tubing to prevent overheating.

After meniscal treatment, hemostasis of the surrounding surgical dissection was achieved to prevent hematoma formation, and the wounds were irrigated. The patellae were relocated and the arthrotomies were closed with a running 2-0 vicryl suture. Fascial and subcutaneous layers were closed with a running 3-0 vicryl suture, and skin was closed with subcuticular 4‑0 vicryl sutures. The rabbit limbs were allowed weight-bearing with unrestricted range of motion within 2x2x2-ft cages. 

For all groups, menisci were explanted at 28 and 84 days for gross and histologic analysis. For biochemical assessments, menisci were explanted at 9, 28, and 84 days (Table 1).  

Gross Analysis

Immediately after specimen removal, all medial menisci were evaluated for gross morphology. A grading system was used for organization and classification of data (Table 2). Three blinded orthopedic surgeon-observers performed all grading. Grade A was considered complete healing of the meniscus. Grade B involved complete healing with a trace of injury remaining on the surface of the meniscus. Grade C represented incomplete healing with a full-thickness injury that was stable to stress of the repair site with an arthroscopic probe. Grade D had no healing with the injured region unstable to stress of the repair site with an arthroscopic probe.

Histologic Analysis and Microscopic Grading of Meniscal Healing

After gross evaluation by the 3 blinded observers, each meniscus was fixed for 24 hours in 10% buffered formalin. Each specimen was then embedded in paraffin and cut into 6-µm slices along the radial plane. The tissue samples were stained with hematoxylin-eosin, and microscopic grading was assigned. The grading system was the same as that used for gross morphologic analysis. 

Biochemical Analysis

To determine whether RF treatment stimulated a healing response in the avascular zone of the meniscus, measurements of specific biochemical markers were analyzed at 9, 28, and 84 days after treatment. As a control, unrepaired meniscal tissue from the contralateral knee was also analyzed. ³H-thymidine incorporation into the meniscus was measured to assess cell proliferation.²³ At sacrifice, control and treated menisci were dissected and immediately placed into sterile culture media (Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum, antibiotic, and fungicide). ³H-thymidine was added at a concentration of 5µCi/mL of media to each tube. After incubation for 48 hours at 37°C under 5% CO₂, the menisci were removed and dialyzed against water for 24 hours to remove unincorporated thymidine. After washing, the menisci were lyophilized, aliquots weighed, and radioactivity determined by liquid scintillation spectrometry. Results are expressed as counts per minute per mg dry tissue weight.

Semiquantitative reverse transcription polymerase chain reaction (RT-PCR) was used to determine mRNA expression of mitogenic growth factors, IGF-1 and bFGF, and angiogenic markers, αV and VEFG.²⁴ National Institutes of Health (NIH) image-analysis software (version 1.61; NIH, Bethesda, Maryland) was used to quantitatively scan RT-PCR profiles after agarose gel electrophoresis and ethidium bromide visualization. Values were normalized to the housekeeping gene, GAPDH.

Statistical Analysis

Data are expressed as mean (SD) and evaluated using an unpaired Student t test between groups. Statistical significance was established at P < .05.