Meniscal injuries are among the most common intra-articular knee pathologies, particularly in young and physically active individuals. The menisci play essential roles in load transmission, shock absorption, joint stability, and preservation of articular cartilage[1]. Meniscal deficiency alters knee biomechanics, increases tibiofemoral contact stresses, and accelerates cartilage degeneration, potentially contributing to early osteoarthritis. Consequently, contemporary orthopaedic management increasingly emphasizes meniscal preservation over resection whenever feasible. Arthroscopic meniscal repair has emerged as an important surgical intervention aimed at restoring meniscal integrity and maintaining long-term knee function [2,3].Advances in arthroscopic instrumentation, fixation devices, and surgical techniques have expanded the indications for meniscal repair[4]. Inside-out, outside-in, and all-inside techniques are commonly employed according to tear morphology, anatomical location, and surgeon preference. Despite these developments, successful healing remains challenging because of the limited vascularity of meniscal tissue, particularly within the central avascular region. Healing potential is influenced by tear configuration, vascular zone, chronicity, tissue quality, and associated ligamentous injuries [5,6].Although arthroscopic meniscal repair offers substantial benefits, postoperative failure remains an important clinical concern[7]. Failure may present as persistent pain, recurrent mechanical symptoms, re-tearing, or the requirement for revision arthroscopy or partial meniscectomy. Reported failure rates vary across studies because of differences in patient characteristics, surgical techniques, rehabilitation protocols, follow-up duration, and definitions of failure. Long-term studies have demonstrated that a considerable proportion of patients experience unsuccessful outcomes despite initially satisfactory postoperative recovery [8,9].Several patient-related, injury-related, and surgical factors have been investigated as potential predictors of meniscal repair failure[10]. Patient characteristics, including age, body mass index, smoking status, and activity level, may influence tissue healing. Similarly, meniscal side, tear length, configuration, vascularity, and interval between injury and surgery may affect repair integrity. Concomitant anterior cruciate ligament (ACL) injury, simultaneous ACL reconstruction, fixation technique, and postoperative rehabilitation may further influence clinical outcomes. However, findings regarding the independent contribution of these factors remain inconsistent [11,12].Rönnblad et al. (2020), in a retrospective analysis of 918 meniscal repairs, reported an overall failure rate of 22.5%, with medial meniscal repair and bioabsorbable arrow fixation associated with increased failure risk, whereas concomitant ACL reconstruction demonstrated a protective association [10]. Similarly, Schweizer et al. (2024), in a meta-analysis involving 3,829 patients, reported an overall failure rate of approximately 14.8%, with lower failure rates following lateral meniscal repair and concomitant ACL reconstruction [11]. These findings highlight the multifactorial nature of meniscal healing and the importance of identifying clinically relevant predictors. Identifying predictors of repair failure is essential for appropriate patient selection, preoperative counselling, surgical planning, and individualized rehabilitation. Nevertheless, variations in study populations, operative techniques, and outcome definitions limit the generalizability of existing evidence. Further evaluation of patient-related and surgical characteristics in routine clinical practice may improve risk stratification and treatment outcomes [9–12].Therefore, the present retrospective cohort study aimed to determine the incidence of failure following arthroscopic meniscal repair and identify patient-related, injury-related, and surgical factors associated with unsuccessful outcomes.

This retrospective cohort study was conducted in the Department of Orthopaedics at a tertiary care hospital to evaluate the incidence and predictors of failure following arthroscopic meniscal repair. Medical records, operative notes, imaging findings, and postoperative follow-up data of patients who underwent arthroscopic meniscal repair were retrospectively reviewed. The study assessed patient-related, injury-related, and surgical factors associated with meniscal repair failure.A total of 100 patients who underwent arthroscopic meniscal repair and fulfilled the eligibility criteria were included in the study. Consecutive sampling was used, and all eligible patients with adequate postoperative follow-up records during the study period were considered for inclusion. Inclusion Criteria 1. Patients aged 18 years and above who underwent arthroscopic meniscal repair. 2. Patients with isolated meniscal tears or meniscal injuries associated with anterior cruciate ligament (ACL) tears. 3. Patients who underwent all-inside, inside-out, or outside-in arthroscopic meniscal repair. 4. Patients with a minimum postoperative follow-up of 12 months. 5. Patients with complete clinical, operative, and follow-up records. Exclusion Criteria 1. Patients who underwent partial or total meniscectomy without meniscal repair. 2. Patients with previous meniscal surgery involving the same knee. 3. Patients with advanced knee osteoarthritis (Kellgren–Lawrence grade III or IV). 4. Patients with associated fractures or complex multiligament knee injuries requiring additional major surgical procedures. 5. Patients with incomplete medical records or postoperative follow-up of less than 12 months. Grouping of Patients Based on postoperative outcomes, the 100 patients were categorized into two groups: • Group A (Successful Repair): Patients without documented repair failure during the postoperative follow-up period. • Group B (Failed Repair): Patients with documented recurrent meniscal tearing at the repaired site requiring revision arthroscopy, repeat meniscal repair, or partial meniscectomy. The distribution of patients between the groups was determined by the observed postoperative outcomes. Both groups were compared to identify potential predictors of meniscal repair failure. Data Collection Data were retrospectively collected from hospital records, operative registers, imaging reports, and follow-up documentation using a structured data extraction proforma. Demographic characteristics, including age, sex, body mass index (BMI), smoking status, and activity level, were recorded.Injury-related variables included the side of the affected knee, medial or lateral meniscal involvement, tear pattern, tear length, vascular zone, and duration between injury and surgery. Surgical variables included the repair technique, number of sutures or fixation devices, concomitant ACL injury, and simultaneous ACL reconstruction. Postoperative complications, recurrent symptoms, revision procedures, and duration of follow-up were also documented. Surgical Procedure All patients underwent arthroscopic meniscal repair using standard arthroscopic techniques. The choice of repair technique, including all-inside, inside-out, or outside-in repair, was determined by tear location, morphology, tissue quality, and surgeon preference. Meniscal tears were assessed arthroscopically, and repair was performed using appropriate sutures or fixation devices. Concomitant ACL reconstruction was performed when clinically indicated. Intraoperative findings and procedural details were obtained from operative records. Postoperative Rehabilitation and Follow-up Postoperative rehabilitation was undertaken according to institutional protocols and the type of meniscal repair performed. Rehabilitation included protected weight-bearing, controlled knee range-of-motion exercises, progressive muscle strengthening, and gradual return to routine activities.Available postoperative records were reviewed at approximately 6 weeks, 3 months, 6 months, and 12 months. Clinical symptoms, functional recovery, recurrent mechanical complaints, and subsequent surgical interventions were documented. Follow-up duration was calculated from the date of primary surgery to the last available clinical assessment. Outcome Measures Primary Outcome: The primary outcome was failure following arthroscopic meniscal repair, defined as documented recurrent meniscal tearing at the repaired site requiring revision arthroscopy, repeat meniscal repair, or partial meniscectomy during follow-up. Secondary Outcomes: Secondary outcomes included the association of demographic characteristics, tear-related variables, surgical techniques, and concomitant ACL reconstruction with repair failure. The interval between primary repair and revision surgery and the occurrence of postoperative complications were also assessed. Statistical Analysis Collected data were entered into Microsoft Excel and analyzed using IBM SPSS Statistics.26 Continuous variables were expressed as mean ± standard deviation or median with interquartile range, depending on their distribution, while categorical variables were presented as frequencies and percentages. Normality was assessed using the Shapiro–Wilk test. Comparisons between successful and failed repair groups were performed using the independent-samples t-test or Mann–Whitney U test for continuous variables and the Chi-square test or Fisher's exact test for categorical variables, as appropriate.Univariate logistic regression analysis was used to examine associations between potential predictors and meniscal repair failure. Clinically relevant variables and those showing associations in univariate analysis were considered for multivariable logistic regression, depending on the number of observed failure events. Results were expressed as odds ratios (OR) with 95% confidence intervals (CI). A p-value of less than 0.05 was considered statistically significant.

A total of 100 patients who underwent arthroscopic meniscal repair were included in the study, of whom 80 (80.0%) had successful repair and 20 (20.0%) experienced repair failure.The mean age was significantly higher in the failed repair group compared with the successful repair group (35.6 ± 9.2 vs. 29.8 ± 8.4 years; p = 0.008). Patients aged above 30 years constituted a significantly greater proportion of the failed repair group (70.0% vs. 43.8%; p = 0.036). The mean BMI was also significantly higher among patients with failed repair (27.1 ± 3.6 vs. 24.3 ± 3.1 kg/m²; p = 0.001). Smoking history was more prevalent in the failed repair group (40.0% vs. 15.0%; p = 0.013). However, no statistically significant differences were observed between the groups regarding sex (p = 0.822) or high physical activity level (p = 0.549) (Table 1).Medial meniscal involvement was significantly more frequent among patients with failed repair compared with those with successful repair (80.0% vs. 55.0%; p = 0.040). The distribution of tears according to vascular zone differed significantly between the groups (p = 0.015). Red-red zone tears were more common in the successful repair group (43.8% vs. 15.0%), whereas red-white (60.0% vs. 45.0%) and white-white zone tears (25.0% vs. 11.3%) were more frequent in the failed repair group. The mean tear length was significantly greater in the failed repair group (23.4 ± 6.1 vs. 18.6 ± 5.2 mm; p = 0.001). Similarly, the mean injury-to-surgery interval was significantly longer among patients with failed repair (5.1 ± 3.0 vs. 3.2 ± 2.1 months; p = 0.002). However, tear pattern was not significantly associated with repair outcome (p = 0.119) (Table 2).The distribution of surgical repair techniques was comparable between the successful and failed repair groups (p = 0.730). The all-inside technique was the most frequently employed approach in both groups (62.5% vs. 75.0%). Concomitant anterior cruciate ligament (ACL) reconstruction was significantly more frequent in the successful repair group compared with the failed repair group (47.5% vs. 20.0%; p = 0.025). The mean number of sutures used was comparable between the groups (3.4 ± 1.1 vs. 3.1 ± 1.0; p = 0.270). An injury-to-surgery interval exceeding three months was significantly more frequent among patients with failed repair (70.0% vs. 37.5%; p = 0.013) (Table 3, Figure 1). The mean postoperative follow-up duration was 18.6 ± 5.2 months. Successful meniscal repair was observed in 80 (80.0%) patients, while repair failure occurred in 20 (20.0%) patients. Among patients with failed repair, 7 (35.0%) underwent revision meniscal repair, whereas 13 (65.0%) underwent partial meniscectomy. The mean time to repair failure was 11.4 ± 4.6 months. Most failures occurred between 6 and 12 months (50.0%), followed by failures after 12 months (35.0%) and within six months (15.0%). Postoperative infection and stiffness were observed in 2.0% and 5.0% of patients, respectively (Table 4).Spearman correlation analysis demonstrated significant positive correlations between meniscal repair failure and age (r = 0.271; p = 0.006), BMI (r = 0.318; p = 0.001), tear length (r = 0.336; p = 0.001), and injury-to-surgery interval (r = 0.294; p = 0.003). Tear length demonstrated the strongest positive correlation with repair failure, followed by BMI, injury-to-surgery interval, and age. In contrast, the number of sutures showed a weak negative correlation with repair failure, which was not statistically significant (r = −0.108; p = 0.285) (Table 5, Figure 2).Univariate binary logistic regression analysis demonstrated that age above 30 years (OR = 3.00; 95% CI: 1.04–8.66; p = 0.042), smoking history (OR = 4.00; 95% CI: 1.35–11.87; p = 0.012), medial meniscal involvement (OR = 3.27; 95% CI: 1.01–10.64; p = 0.049), and an injury-to-surgery interval exceeding three months (OR = 3.89; 95% CI: 1.35–11.22; p = 0.012) were significantly associated with increased odds of meniscal repair failure. Conversely, concomitant ACL reconstruction was associated with significantly reduced odds of repair failure (OR = 0.28; 95% CI: 0.09–0.90; p = 0.032). Male sex was not significantly associated with repair failure (OR = 1.14; 95% CI: 0.37–3.51; p = 0.822) (Table 6, Figure 3). Table 1. Comparison of Demographic and Baseline Characteristics Between Successful and Failed Meniscal Repair Groups (N = 100) Parameter Successful Repair (n = 80) Failed Repair (n = 20) p-value Statistical Test Age (years), mean ± SD 29.8 ± 8.4 35.6 ± 9.2 0.008 Independent t-test Age ≤30 years 45 (56.3%) 6 (30.0%) 0.036 Chi-square test Age >30 years 35 (43.8%) 14 (70.0%) Male 58 (72.5%) 15 (75.0%) 0.822 Chi-square test Female 22 (27.5%) 5 (25.0%) BMI (kg/m²), mean ± SD 24.3 ± 3.1 27.1 ± 3.6 0.001 Independent t-test Smoking history 12 (15.0%) 8 (40.0%) 0.013 Fisher's exact test High physical activity level 42 (52.5%) 9 (45.0%) 0.549 Chi-square test Table 2. Comparison of Meniscal Tear Characteristics Between Successful and Failed Repair Groups (N = 100) Parameter Successful Repair (n = 80) Failed Repair (n = 20) p-value Statistical Test Meniscal involvement 0.040 Chi-square test Medial meniscus 44 (55.0%) 16 (80.0%) Lateral meniscus 36 (45.0%) 4 (20.0%) Tear pattern 0.119 Chi-square test Longitudinal 42 (52.5%) 6 (30.0%) Bucket-handle 22 (27.5%) 10 (50.0%) Radial/complex 16 (20.0%) 4 (20.0%) Vascular zone 0.015 Chi-square test Red-red 35 (43.8%) 3 (15.0%) Red-white 36 (45.0%) 12 (60.0%) White-white 9 (11.3%) 5 (25.0%) Tear length (mm), mean ± SD 18.6 ± 5.2 23.4 ± 6.1 0.001 Independent t-test Injury-to-surgery interval (months), mean ± SD 3.2 ± 2.1 5.1 ± 3.0 0.002 Independent t-test Table 3. Comparison of Surgical Characteristics Between Successful and Failed Meniscal Repair Groups (N = 100) Parameter Successful Repair (n = 80) Failed Repair (n = 20) p-value Statistical Test Repair technique 0.730 Fisher––exact test All-inside 50 (62.5%) 15 (75.0%) Inside-out 22 (27.5%) 4 (20.0%) Outside-in 8 (10.0%) 1 (5.0%) Concomitant ACL reconstruction 0.025 Chi-square test Yes 38 (47.5%) 4 (20.0%) No 42 (52.5%) 16 (80.0%) Number of sutures, mean ± SD 3.4 ± 1.1 3.1 ± 1.0 0.270 Independent t-test Injury-to-surgery interval 0.013 Chi-square test ≤3 months 50 (62.5%) 6 (30.0%) >3 months 30 (37.5%) 14 (70.0%) Table 4. Postoperative Outcomes and Characteristics of Meniscal Repair Failure (N = 100) Parameter Value Follow-up duration (months), mean ± SD 18.6 ± 5.2 Repair outcome Successful repair 80 (80.0%) Failed repair 20 (20.0%) Revision procedure among failures (n = 20) Revision meniscal repair 7 (35.0%) Partial meniscectomy 13 (65.0%) Time to failure (months), mean ± SD 11.4 ± 4.6 Timing of failure (n = 20) Within 6 months 3 (15.0%) 6–12 months 10 (50.0%) After 12 months 7 (35.0%) Postoperative infection 2 (2.0%) Postoperative stiffness 5 (5.0%) Table 5. Correlation Between Patient-Related and Injury-Related Variables and Meniscal Repair Failure (N = 100) Variable Correlation Coefficient (r) p-value Statistical Test Age (years) 0.271 0.006 Spearman BMI (kg/m²) 0.318 0.001 Tear length (mm) 0.336 0.001 Injury-to-surgery interval (months) 0.294 0.003 Number of sutures −0.108 0.285 Table 6. Univariate Logistic Regression Analysis of Predictors of Meniscal Repair Failure (N = 100) Predictor Crude OR 95% CI p-value Statistical Test Age >30 years 3.00 1.04–8.66 0.042 Binary logistic regression Male sex 1.14 0.37–3.51 0.822 Smoking history 4.00 1.35–11.87 0.012 Medial meniscal tear 3.27 1.01–10.64 0.049 Injury-to-surgery interval >3 months 3.89 1.35–11.22 0.012 Concomitant ACL reconstruction 0.28 0.09–0.90 0.032

In the present study, the overall success rate following arthroscopic meniscal repair was 80.0%, while failure occurred in 20.0% of patients. The mean follow-up duration was 18.6 ± 5.2 months, and the mean time to failure was 11.4 ± 4.6 months. Similarly, Logan et al. [13] reported a definite failure rate of 24.0% among 45 meniscal repairs during a mean follow-up of 8.5 years. Kurosaka et al. [14] reported recurrent tears in 13 of 90 initially healed repairs, demonstrating that failure may occur despite arthroscopically confirmed healing. These findings highlight the importance of long-term follow-up after meniscal repair.In the present study, patients with failed repair were significantly older (35.6 ± 9.2 versus 29.8 ± 8.4 years; p = 0.008) and had higher BMI (27.1 ± 3.6 versus 24.3 ± 3.1 kg/m²; p = 0.001). Smoking was also more frequent among failed repairs (40.0% versus 15.0%; p = 0.013). Logan et al. [13], however, reported no significant association between age and repair failure (p = 0.626). Kise et al. [15] observed higher preoperative Tegner activity scores among patients requiring reoperation (median 5 versus 4; p = 0.037). These findings suggest that patient-related factors may influence healing, although their associations vary across populations. Medial meniscal involvement was significantly higher among failed repairs (80.0% versus 55.0%; p = 0.040). Tear length was also greater in the failed repair group (23.4 ± 6.1 versus 18.6 ± 5.2 mm; p = 0.001), while less vascular tear locations were associated with failure (p = 0.015). Logan et al. [13] similarly reported higher failure rates following medial compared with lateral meniscal repair (36.4% versus 5.6%; p < 0.05). Kotsovolos et al. [16] evaluated 61 repairs using the FasT-Fix system, primarily involving longitudinal tears within the red-red and red-white vascular zones. These findings emphasize the importance of meniscal location, tear morphology, and vascularity in repair outcomes.In the present study, all-inside repair was performed in 75.0% of failed cases and 62.5% of successful cases, without a statistically significant association between repair technique and failure (p = 0.730). Kise et al. [15], in a randomized trial involving 46 patients, reported significantly higher reoperation rates with Biofix arrows than FasT-Fix sutures (43.0% versus 12.0%; p = 0.018). Kotsovolos et al. [16] reported favourable clinical outcomes using the FasT-Fix all-inside repair system, whereas Kurzweil et al. [17] described unsatisfactory results with meniscal arrow fixation. These findings suggest that fixation-device characteristics may influence repair durability, even when broad surgical technique categories show no significant differences.Concomitant ACL reconstruction was more frequent among successful repairs (47.5%) than failed repairs (20.0%; p = 0.025), with lower crude odds of failure (OR = 0.28; p = 0.032). Kotsovolos et al. [16] reported concomitant ACL reconstruction in 36 of 58 patients (62.0%). In contrast, Logan et al. [13] observed failure rates of 14.3% in ACL-intact knees and 28.9% in ACL-reconstructed knees (p = 0.420). These findings indicate that the association between ACL reconstruction and meniscal healing remains variable across studies. The mean injury-to-surgery interval was significantly longer in failed repairs (5.1 ± 3.0 months) than successful repairs (3.2 ± 2.1 months; p = 0.002). Delayed surgery beyond three months was associated with increased odds of failure (OR = 3.89; p = 0.012). Conversely, Logan et al. [13] reported no significant difference in injury-to-surgery intervals between failed and intact repairs (6.2 versus 8.2 months; p = 0.615). This discrepancy suggests that surgical timing may interact with tear characteristics and associated injuries. Among 20 failed repairs in the present study, 13 patients (65.0%) underwent partial meniscectomy and 7 (35.0%) underwent revision meniscal repair. The mean time to failure was 11.4 ± 4.6 months. Kurosaka et al. [14] demonstrated that recurrent tearing could occur even after confirmed meniscal healing. Logan et al. [13] reported 11 definite failures among 45 repairs (24.0%) during a mean follow-up of 8.5 years, with seven failures associated with subsequent injuries. These findings emphasize the importance of long-term follow-up and prevention of reinjury.

The present study demonstrated an 80% success rate and 20% failure rate following arthroscopic meniscal repair. Older age, higher BMI, smoking, medial meniscal involvement, greater tear length, and delayed surgery were significantly associated with repair failure. Concomitant ACL reconstruction was associated with reduced odds of failure. Early intervention, appropriate patient selection, and individualized surgical planning may improve outcomes. Limitations The retrospective study design and relatively small sample size of 100 patients limit the generalizability of findings. The mean follow-up duration of 18.6 months may not capture long-term repair failures. Reliance on medical records may introduce information bias. Larger prospective multicentre studies with longer follow-up are recommended.

1. Andersson-Molina H, Karlsson H, Rockborn P. Arthroscopic partial and total meniscectomy: a long-term follow-up study with matched controls. Arthroscopy. 2002;18(2):183-9. 2. Paxton ES, Stock MV, Brophy RH. Meniscal repair versus partial meniscectomy: a systematic review comparing reoperation rates and clinical outcomes. Arthroscopy. 2011;27(9):1275-88. 3. Abrams GD, Frank RM, Gupta AK, Harris JD, McCormick FM, Cole BJ. Trends in meniscus repair and meniscectomy in the United States, 2005-2011. Am J Sports Med. 2013;41(10):2333-9. 4. Borden P, Nyland J, Caborn DN, Pienkowski D. Biomechanical comparison of the FasT-Fix meniscal repair suture system with vertical mattress sutures and meniscus arrows. Am J Sports Med. 2003;31(3):374-8. 5. Albrecht-Olsen P, Kristensen G, Burgaard P, Joergensen U, Toerholm C. The arrow versus horizontal suture in arthroscopic meniscus repair: a prospective randomized study with arthroscopic evaluation. Knee Surg Sports Traumatol Arthrosc. 1999;7(5):268-73. 6. Ahn JH, Lee YS, Yoo JC, Chang MJ, Koh KH, Kim MH. Clinical and second-look arthroscopic evaluation of repaired medial meniscus in anterior cruciate ligament-reconstructed knees. Am J Sports Med. 2010;38(3):472-7. 7. Nepple JJ, Dunn WR, Wright RW. Meniscal repair outcomes at greater than five years: a systematic literature review and meta-analysis. J Bone Joint Surg Am. 2012;94(24):2222-7. 8. Petersen W, Karpinski K, Bierke S, Müller Rath R, Häner M. A systematic review about long-term results after meniscus repair. Arch Orthop Trauma Surg. 2022;142(5):835-44. 9. Rönnblad E, Barenius B, Stålman A, Eriksson K. Failed meniscal repair increases the risk for osteoarthritis and poor knee function at an average of 9 years follow-up. Knee Surg Sports Traumatol Arthrosc. 2022;30(1):192-9. 10. Ronnblad E, Barenius B, Engstrom B, Eriksson K. Predictive factors for failure of meniscal repair: a retrospective dual-center analysis of 918 consecutive cases. Orthop J Sports Med. 2020;8(3):2325967120905529. 11. Schweizer C, Hanreich C, Tscholl PM, Blatter S, Windhager R, Waldstein W. Meniscal repair outcome in 3829 patients with a minimum follow-up from 2 years up to 5 years: a meta-analysis on the overall failure rate and factors influencing failure. Am J Sports Med. 2024;52(3):822-31. 12. Cannon WD Jr, Vittori JM. The incidence of healing in arthroscopic meniscal repairs in anterior cruciate ligament-reconstructed knees versus stable knees. Am J Sports Med. 1992;20(2):176-81. 13. Logan M, Watts M, Owen J, Myers P. Meniscal repair in the elite athlete: results of 45 repairs with a minimum 5-year follow-up. Am J Sports Med. 2009;37(6):1131-4. 14. Kurosaka M, Yoshiya S, Kuroda R, Matsui N, Yamamoto T, Tanaka J. Repeat tears of repaired menisci after arthroscopic confirmation of healing. J Bone Joint Surg Br. 2002;84(1):34-7. 15. Kise NJ, Drogset JO, Ekeland A, Sivertsen EA, Heir S. All-inside suture device is superior to meniscal arrows in meniscal repair: a prospective randomized multicenter clinical trial with 2-year follow-up. Knee Surg Sports Traumatol Arthrosc. 2015;23(1):211-8. 16. Kotsovolos ES, Hantes ME, Mastrokalos DS, Lorbach O, Paessler HH. Results of all-inside meniscal repair with the FasT-Fix meniscal repair system. Arthroscopy. 2006;22(1):3-9. 17. Kurzweil PR, Tifford CD, Ignacio EM. Unsatisfactory clinical results of meniscal repair using the meniscus arrow. Arthroscopy. 2005;21(8):905.