Total knee replacement (TKR) is one of the most successful and cost-effective operations in orthopaedics, reliably relieving pain and restoring function in advanced arthritis. The number of procedures performed worldwide has risen steadily, and in the United States alone more than three million primary knee replacements per year were projected by 2030 [1]. In India, osteoarthritis of the knee is highly prevalent, with community studies reporting figures of about 28% in adults [2,3], and the number of knee replacements is increasing rapidly, as shown by the Indian Society of Hip and Knee Surgeons (ISHKS) joint registry [4]. Despite generally good outcomes, a proportion of patients are dissatisfied, and restricted motion is a major reason. Stiffness following TKR has been reported in about 1.3% to more than 10% of patients depending on the definition used and the duration of follow-up [5–7]. No single definition is universally accepted. Most authors accept a flexion of less than 90° (or a loss of more than 10° of extension) at six to twelve weeks as indicating a clinically relevant early stiffness, whereas others define stiffness by a total arc of motion of less than 75° to 80° [8]. Stiffness impairs stair climbing, rising from a chair and gait, and is among the commonest reasons for dissatisfaction and for early re-intervention. The functional consequences of reduced flexion are greater in India and other non-Western societies. Activities of daily life such as squatting, cross-legged sitting on the floor, kneeling for prayer and use of Indian-style toilets require knee flexion of roughly 110° to more than 130° [9,10], considerably beyond the 90° to 110° needed for chair-based activities in Western settings. A knee that would be considered "acceptable" in a Western patient may therefore be functionally inadequate in an Indian patient. Indian surgeons also operate on a population with distinct features: a high proportion of neglected, severe varus and stiff or ankylosed knees [11,12], anatomical differences in femoral and tibial morphology in Asian patients [13,14], a younger age at arthroplasty, and variable access to supervised physiotherapy. The pathogenesis of postoperative stiffness is multifactorial. Preoperative range of motion has been consistently shown to be the strongest predictor of postoperative motion [15–18]. Other factors reported include previous knee surgery, diabetes, obesity, poor pain control, surgical technique and implant positioning, infection, and inadequate rehabilitation [19,20]. However, most studies were retrospective, were conducted in Western cohorts, used heterogeneous definitions, and rarely included postoperative variables such as pain scores and exercise adherence, which are potentially modifiable. Prospective data from Indian patients, in whom both the demands on the knee and the risk profile may differ, are limited. The present study was therefore undertaken to identify clinical, radiological, surgical and rehabilitation-related predictors of knee stiffness in a prospective cohort of patients undergoing primary TKR. 1.1 Aim and objectives The primary objective was to determine the incidence of knee stiffness (flexion < 90° at 12 weeks) after primary TKR. Secondary objectives were (i) to identify independent preoperative, intraoperative and postoperative predictors of stiffness, (ii) to describe the treatment of stiffness and its results, and (iii) to compare functional outcome and patient satisfaction at 12 months between stiff and non-stiff knees.

2.1 Study design, setting and ethics This was a single-centre, prospective, observational cohort study conducted in the Department of Orthopaedics of a tertiary-care teaching hospital in India. Consecutive patients who underwent primary TKR between January 2020 and December 2020 were enrolled and followed for 12 months, the last follow-up being completed in December 2020. The study was approved by the Institutional Ethics Committee, was conducted in accordance with the Declaration of Helsinki [21], and written informed consent was obtained from all participants. 2.2 Participants Inclusion criteria: adults aged 40 years or more undergoing primary unilateral cemented TKR for symptomatic Kellgren–Lawrence grade III or IV osteoarthritis [22] or rheumatoid arthritis, who were willing to attend all follow-up visits. Exclusion criteria: revision TKR; simultaneous bilateral TKR; post-traumatic or post-infective arthritis; neuromuscular disorders or spasticity; preoperative flexion of less than 60° or ankylosed knees requiring quadricepsplasty or tibial tubercle osteotomy; need for constrained or hinged implants; and refusal of consent. 2.3 Sample size Assuming an anticipated incidence of early stiffness of 10%, an absolute precision of 4.2% and a 95% confidence level, the minimum sample required was 196 patients. To allow for an acceptable number of events for multivariable modelling [23] and for loss to follow-up, 212 patients were enrolled, and 200 completed follow-up and were analysed. 2.4 Surgical technique and perioperative protocol All operations were performed by one of three fellowship-trained arthroplasty surgeons using a standard midline skin incision, medial parapatellar arthrotomy and a thigh tourniquet. Measured-resection technique with intramedullary femoral and extramedullary tibial alignment guides was used, aiming for neutral mechanical alignment, balanced flexion and extension gaps and a posterior tibial slope of 3° to 7°. Cruciate-retaining (CR) or posterior-stabilised (PS) cemented implants were used according to surgeon preference and the status of the posterior cruciate ligament; patellar resurfacing was selective. Spinal (with or without epidural) or general anaesthesia was used. A uniform multimodal analgesic protocol (paracetamol, an NSAID if not contraindicated, tramadol, and a peri-articular local anaesthetic infiltration) and low-molecular-weight heparin thromboprophylaxis were followed. All patients followed the same rehabilitation protocol. Physiotherapy was started on the first postoperative day with ankle pumps, static quadriceps exercises, active-assisted knee flexion and assisted weight-bearing walking with a walker. Continuous passive motion was used in the hospital stay only as an adjunct. At discharge (day 4 to 5), patients received a written and illustrated home exercise programme aimed at a flexion of 90° by two weeks and 110° or more by six weeks, with outpatient supervised physiotherapy twice a week for six weeks [24,25]. 2.5 Data collection and study variables Data were recorded prospectively on a structured proforma at the preoperative visit, during hospitalisation, and at 2 weeks, 6 weeks, 3 months, 6 months and 12 months. Patients were examined by an independent observer (a senior resident not involved in the surgery) using a long-arm goniometer, with the patient supine and the measurement taken as the maximum active flexion; the average of two readings was used. The following variables were collected. Demographic and clinical: age, sex, body mass index (BMI; Asian-specific cut-offs were used for categories [26]), rural or urban residence, diabetes mellitus, hypertension, previous knee surgery, side, diagnosis, coronal deformity, preoperative flexion, preoperative fixed flexion deformity (FFD), preoperative Knee Society knee score [27], and the patient's expectation of floor-sitting after surgery. Radiological and intraoperative: Kellgren–Lawrence grade, postoperative tibial slope on a standardised lateral radiograph, implant design (CR or PS), patellar resurfacing, type of anaesthesia and duration of surgery. Postoperative: pain on a 0–10 visual analogue scale (VAS) during active knee exercises on postoperative day (POD) 3, flexion on POD 2, superficial wound complications, time of starting physiotherapy, and adherence to the home exercise programme. Poor adherence was defined as performing less than 50% of the prescribed sessions over the first six weeks, as recorded in a patient exercise diary and verified at each visit. 2.6 Outcome definition The primary outcome was knee stiffness, defined as active knee flexion of less than 90° at 12 weeks after surgery. Patients were thereby divided into a "stiff" group and a "non-stiff" group. Secondary outcomes at 12 months were flexion, residual FFD, KSS (knee and function scores), Oxford Knee Score [28], ability to sit cross-legged, and patient satisfaction (five-point Likert scale, with the categories "satisfied" and "very satisfied" combined). 2.7 Management of stiffness Knees with flexion of less than 90° at 6 weeks were reviewed more closely. Intensified supervised physiotherapy was continued, and manipulation under anaesthesia (MUA) was offered for persistent stiffness, ideally within 6 to 12 weeks, as recommended in the studies [29,30]. Arthroscopic arthrolysis was reserved for knees that failed MUA, and revision surgery for component malposition. 2.8 Statistical analysis Data were analysed using IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA). Continuous variables are presented as mean ± standard deviation and were compared with the Student's t-test (or the Mann–Whitney U test if non-normal); categorical variables are presented as numbers (percentages) and were compared with the chi-square or Fisher's exact test. The incidence is given with an exact 95% confidence interval. Variables with p < 0.10 on univariate analysis were entered into a multivariable binary logistic regression model, with backward likelihood-ratio elimination; collinear variables (correlation coefficient > 0.6) were not entered together. Results are presented as odds ratios (OR) with 95% confidence intervals. Model calibration was assessed with the Hosmer–Lemeshow test [31] and discrimination with the area under the receiver operating characteristic curve (AUC). A two-sided p-value < 0.05 was taken as statistically significant.

3.1 Study population During the study period, 232 patients were assessed for eligibility. Twenty were excluded (11 did not satisfy the inclusion criteria, 5 declined consent and 4 underwent simultaneous bilateral TKR), and 212 were enrolled. Twelve patients (10 lost to follow-up and 2 deaths unrelated to the surgery) did not complete the 12-month follow-up, leaving 200 patients for the final analysis (follow-up rate 94.3%). The baseline characteristics of the cohort are shown in Table 1. The mean age was 64.3 ± 7.8 years, and women accounted for 69.0% of the patients. Most knees (93.0%) had primary osteoarthritis with a varus deformity (82.0%), and the mean preoperative flexion was 104.7 ± 14.2°. Table 1. Baseline characteristics of the study population (N = 200) Characteristic Value Age (years), mean ± SD (range) 64.3 ± 7.8 (45–82) Sex: female / male, n (%) 138 (69.0) / 62 (31.0) Body mass index (kg/m²), mean ± SD 28.9 ± 4.6 BMI ≥ 27.5 kg/m² (Asian high-risk cut-off), n (%) 124 (62.0) Residence: urban / rural, n (%) 118 (59.0) / 82 (41.0) Diabetes mellitus, n (%) 58 (29.0) Hypertension, n (%) 96 (48.0) Previous knee surgery, n (%) 20 (10.0) Side: right / left, n (%) 108 (54.0) / 92 (46.0) Diagnosis: primary osteoarthritis / rheumatoid arthritis, n (%) 186 (93.0) / 14 (7.0) Kellgren–Lawrence grade III / IV, n (%) 38 (19.0) / 162 (81.0) Coronal deformity: varus / valgus / neutral, n (%) 164 (82.0) / 18 (9.0) / 18 (9.0) Preoperative flexion (°), mean ± SD 104.7 ± 14.2 Preoperative fixed flexion deformity (°), mean ± SD 8.4 ± 6.0 Preoperative KSS knee score, mean ± SD 41.2 ± 9.7 Implant: posterior-stabilised / cruciate-retaining, n (%) 113 (56.5) / 87 (43.5) Patellar resurfacing, n (%) 100 (50.0) Anaesthesia: spinal ± epidural / general, n (%) 176 (88.0) / 24 (12.0) SD, standard deviation; KSS, Knee Society Score. 3.2 Incidence of stiffness Knee stiffness (flexion < 90° at 12 weeks) was present in 24 of 200 knees, an incidence of 12.0% (95% CI 7.8–17.3%). The mean flexion in the stiff group was 78.4 ± 8.1° at 12 weeks, compared with 112.6 ± 10.2° in the non-stiff group. A residual FFD of more than 10° at 12 weeks was seen in 9 of the 24 stiff knees (37.5%). 3.3 Univariate analysis of predictors Preoperative and demographic factors are compared in Table 2. Patients who developed stiffness had a significantly higher BMI (31.6 ± 4.9 vs 28.5 ± 4.4 kg/m², p = 0.006), were more often diabetic (50.0% vs 26.1%, p = 0.03), and more often had undergone previous knee surgery (25.0% vs 8.0%, p = 0.02). They also had a significantly lower preoperative flexion (92.3 ± 13.5° vs 106.4 ± 13.4°, p < 0.001) and a greater FFD (12.9 ± 6.4° vs 7.8 ± 5.7°, p = 0.001). Age, sex, type of deformity, preoperative KSS knee score and expectation of floor-sitting did not differ significantly between the groups. Table 2. Preoperative and demographic factors: stiff versus non-stiff knees Variable Stiff (n = 24) Non-stiff (n = 176) p-value Age (years), mean ± SD 62.9 ± 8.1 64.5 ± 7.7 0.37 Female sex, n (%) 19 (79.2) 119 (67.6) 0.35 BMI (kg/m²), mean ± SD 31.6 ± 4.9 28.5 ± 4.4 0.006* Diabetes mellitus, n (%) 12 (50.0) 46 (26.1) 0.03* Previous knee surgery, n (%) 6 (25.0) 14 (8.0) 0.02* Varus deformity, n (%) 21 (87.5) 143 (81.3) 0.58 Preoperative flexion (°), mean ± SD 92.3 ± 13.5 106.4 ± 13.4 <0.001* Preoperative FFD (°), mean ± SD 12.9 ± 6.4 7.8 ± 5.7 0.001* Preoperative KSS knee score, mean ± SD 38.2 ± 9.1 41.6 ± 9.8 0.10 Expects floor-sitting after TKR, n (%) 17 (70.8) 109 (61.9) 0.50 *Statistically significant (p < 0.05). Student's t-test for continuous variables; Fisher's exact test for categorical variables. FFD, fixed flexion deformity. Intraoperative and early postoperative factors are shown in Table 3. The implant design (PS or CR), patellar resurfacing and type of anaesthesia were not associated with stiffness. In contrast, a longer operating time, a lower postoperative tibial slope, superficial wound complications, lower flexion on POD 2, a higher POD-3 pain score during exercise, delayed start of physiotherapy and poor adherence to the home programme were significantly more frequent in the stiff group. Poor adherence was recorded in 58.3% of stiff knees compared with 19.9% of non-stiff knees (p < 0.001). Table 3. Intraoperative and postoperative factors: stiff versus non-stiff knees Variable Stiff (n = 24) Non-stiff (n = 176) p-value Intraoperative / radiological Duration of surgery (min), mean ± SD 96.4 ± 14.0 88.2 ± 12.6 0.01* Spinal anaesthesia, n (%) 19 (79.2) 157 (89.2) 0.18 Posterior-stabilised design, n (%) 15 (62.5) 98 (55.7) 0.66 Patellar resurfacing, n (%) 10 (41.7) 90 (51.1) 0.51 Postoperative tibial slope (°), mean ± SD 3.1 ± 1.8 5.2 ± 1.9 <0.001* Postoperative Superficial wound complication, n (%) 4 (16.7) 6 (3.4) 0.02* Flexion on POD 2 (°), mean ± SD 58.4 ± 11.2 72.6 ± 10.5 <0.001* Pain VAS on POD 3 (during exercise), mean ± SD 6.8 ± 1.3 5.1 ± 1.4 <0.001* Physiotherapy started within 24 h, n (%) 15 (62.5) 148 (84.1) 0.02* Poor adherence to home exercises, n (%) 14 (58.3) 35 (19.9) <0.001* *Statistically significant (p < 0.05). POD, postoperative day; VAS, visual analogue scale (0–10). 3.4 Recovery of knee flexion over time The recovery of flexion in the two groups is shown in Table 4 and Figure 1. The stiff group lagged behind from the first postoperative fortnight, and the difference between the groups remained significant at every time point (p < 0.001). Importantly, the separation was already evident by 2 weeks (62.1° vs 78.3°) and 6 weeks (71.2° vs 101.8°), suggesting that the early recovery trajectory may be used to identify patients at risk. Table 4. Knee flexion (degrees, mean ± SD) at follow-up Time point Overall (N = 200) Stiff (n = 24) Non-stiff (n = 176) p-value 2 weeks 76.4 ± 12.3 62.1 ± 10.4 78.3 ± 11.2 <0.001 6 weeks 98.1 ± 15.3 71.2 ± 9.3 101.8 ± 11.9 <0.001 3 months 108.5 ± 14.9 78.4 ± 8.1 112.6 ± 10.2 <0.001 6 months 112.4 ± 13.8 86.4 ± 10.9 116.0 ± 9.8 <0.001 12 months 114.2 ± 11.5 98.1 ± 12.0 116.4 ± 9.6 <0.001 Student's t-test between the stiff and non-stiff groups. Figure 1. Recovery of knee flexion (mean ± SD) in stiff and non-stiff knees over 12 months. The dashed line marks the 90° threshold. 3.5 Multivariable analysis Thirteen variables with p < 0.10 on univariate analysis were entered into the logistic regression model (previous surgery, BMI, diabetes, preoperative flexion, preoperative FFD, preoperative KSS knee score, duration of surgery, tibial slope, wound complication, POD-3 pain, delayed physiotherapy, poor adherence and POD-2 flexion). POD-2 flexion was excluded because of collinearity with preoperative flexion and POD-3 pain. After backward elimination, six variables remained independently associated with stiffness (Table 5, Figure 2). Poor adherence to the home exercise programme had the strongest association (adjusted OR 4.35; 95% CI 1.45–13.05), followed by diabetes mellitus (OR 3.12; 1.08–9.01). Each 10° reduction in preoperative flexion increased the odds of stiffness by 62%, and each 5° increase in preoperative FFD by 48%. Previous knee surgery, tibial slope, wound complication, operating time and delayed physiotherapy were not retained in the final model. The model was well calibrated (Hosmer–Lemeshow p = 0.62), explained 46% of the variance (Nagelkerke R²), and showed good discrimination (AUC 0.87; 95% CI 0.79–0.94; optimism-corrected AUC after 1000 bootstrap samples 0.84). Table 5. Multivariable logistic regression for predictors of stiffness at 12 weeks Predictor Crude OR Adjusted OR 95% CI p-value Poor adherence to home exercises (yes vs no) 5.64 4.35 1.45–13.05 0.009 Diabetes mellitus (yes vs no) 2.83 3.12 1.08–9.01 0.036 Preoperative flexion (per 10° decrease) 2.18 1.62 1.15–2.28 0.006 Preoperative FFD (per 5° increase) 2.12 1.48 1.05–2.09 0.026 POD-3 pain VAS (per 1-point increase) 2.38 1.55 1.08–2.22 0.017 BMI (per 1 kg/m² increase) 1.17 1.13 1.02–1.26 0.023 OR, odds ratio; CI, confidence interval; FFD, fixed flexion deformity; POD, postoperative day; VAS, visual analogue scale. Hosmer–Lemeshow p = 0.62; Nagelkerke R² = 0.46; AUC = 0.87. Figure 2. Forest plot of adjusted odds ratios (95% CI) for independent predictors of knee stiffness at 12 weeks. 3.6 Management of stiffness The treatment of the 24 stiff knees is summarised in Table 6. Three knees (12.5%) recovered beyond 90° with intensified physiotherapy alone. Seventeen knees (70.8%) underwent MUA at a mean of 8.4 ± 2.1 weeks (range 5–12 weeks); flexion improved from 74.5 ± 8.2° before manipulation to 112.4 ± 8.9° immediately afterwards, but 12-month flexion averaged 99.4 ± 10.1° because some of the gain was lost. Three knees required arthroscopic arthrolysis after failed MUA, and one knee with an internally rotated tibial component required revision at 30 weeks. There were no periprosthetic fractures or extensor mechanism ruptures after MUA. Table 6. Treatment of knee stiffness and flexion achieved (n = 24) Treatment n (%) Timing (weeks) Flexion before (°) Flexion at 12 months (°) Intensified physiotherapy alone 3 (12.5) – 82.7 ± 4.2 101.3 ± 3.1 Manipulation under anaesthesia 17 (70.8) 8.4 ± 2.1 74.5 ± 8.2 99.4 ± 10.1 Arthroscopic arthrolysis (after failed MUA) 3 (12.5) 18.7 ± 5.5 69.3 ± 6.1 88.7 ± 5.5 Revision of malpositioned component 1 (4.2) 30 62 95 Values are mean ± SD where applicable. MUA, manipulation under anaesthesia. 3.7 Functional outcome and satisfaction at 12 months At 12 months, stiff knees had significantly poorer flexion, a greater residual FFD, lower KSS and Oxford Knee Scores, a much lower ability to sit cross-legged and lower patient satisfaction than non-stiff knees (Table 7). Only 12.5% of patients in the stiff group could sit cross-legged on the floor compared with 55.7% in the non-stiff group. Table 7. Outcomes at 12 months in stiff versus non-stiff knees Outcome Stiff (n = 24) Non-stiff (n = 176) p-value Flexion (°), mean ± SD 98.1 ± 12.0 116.4 ± 9.6 <0.001 Residual FFD (°), mean ± SD 4.6 ± 3.2 1.2 ± 1.8 <0.001 KSS knee score, mean ± SD 71.4 ± 9.6 88.2 ± 6.9 <0.001 KSS function score, mean ± SD 62.5 ± 11.2 82.6 ± 9.8 <0.001 Oxford Knee Score, mean ± SD 31.2 ± 6.4 41.6 ± 5.2 <0.001 Able to sit cross-legged, n (%) 3 (12.5) 98 (55.7) <0.001 Satisfied or very satisfied, n (%) 13 (54.2) 161 (91.5) <0.001 KSS, Knee Society Score; FFD, fixed flexion deformity.

In this prospective cohort of 200 patients undergoing primary TKR, 12.0% had flexion of less than 90° at 12 weeks. Six independent predictors were identified, of which four (poor adherence to exercises, diabetes, higher BMI and higher early pain) are potentially modifiable, while the other two (preoperative flexion and FFD) reflect the severity of the baseline joint disease. Stiffness had a substantial effect on function and satisfaction at one year, particularly on the ability to sit cross-legged, which is a key requirement for Indian patients. The incidence we found is higher than that reported in large series, in which stiffness necessitating intervention occurred in about 1.3% to 6% of patients [32], but lies within the range reported in the studies when broader definitions of stiffness are applied and all patients, not only those who were treated, are counted. Our early (12-week) time point captures knees that may later improve, and the definition we have selected is stricter than the thresholds used in many Western studies, so direct comparison should be done with caution. The variation in definitions underlines the need for a uniform definition of stiffness in arthroplasty research. Reduced preoperative flexion and a larger FFD were strong predictors, which supports the well-established principle that the preoperative arc of motion is the best predictor of the postoperative arc [33]. Gandhi et al. identified lower preoperative motion among the main risk factors for stiff knees, while Ritter et al. found preoperative range of motion to be the leading determinant of postoperative motion, along with other clinical and radiographic factors. The same trend was found in neglected, stiff Indian knees: Rajgopal and colleagues reported good pain relief and functional improvement after TKR in stiff and ankylosed knees, although the achievable range of motion remains limited in such knees. Patients with severe preoperative contracture should be counselled that gains in flexion after TKR are limited, and surgical release and gap balancing should be meticulous in this group. Diabetes was an independent predictor (OR 3.12) in our series. Diabetic patients are believed to have an altered healing response with a tendency towards fibrosis, and have been shown to have more postoperative complications after TKR [34]. Arthrofibrosis is characterised by an exaggerated, myofibroblast-driven fibrotic response of the joint and the periarticular soft tissues [35–37]; this may explain why systemic metabolic factors, a prolonged inflammatory state caused by pain and swelling, and a delayed mobilisation increase the risk. A higher BMI was also associated with stiffness in our patients. Obesity has been linked to poorer flexion, due to mechanical soft-tissue impingement, a greater operative difficulty, and a lower activity level [38]. In our population, BMI is of particular importance because Asian patients have a higher body-fat percentage at a given BMI, and a lower BMI threshold is recommended for defining risk. Optimisation of glycaemic control (for instance, HbA1c below 7.5% to 8%) and weight reduction before elective TKR are reasonable measures to reduce this risk. Adherence to the exercise programme and early postoperative pain were the most important modifiable predictors. Poor control of pain inhibits quadriceps activation and active knee movement, causes reflex splinting and effusion, and triggers a cycle of immobility and fibrosis [6,8]. High-intensity early rehabilitation has been shown to improve functional outcome after TKR, whereas the benefit of continuous passive motion appears to be small and short-lived. Poor adherence was observed in a quarter of our patients, which may reflect patient factors that are particularly relevant in India, such as long travel distance to the hospital, limited availability of physiotherapists in rural areas, fear of pain, and the lack of family understanding of the importance of exercises. A structured pre-discharge education, a written and pictorial exercise sheet in the local language, regular telephonic reminders and involvement of a family caregiver are simple and low-cost strategies that could be implemented in Indian hospitals. In our series, implant design (PS versus CR) and patellar resurfacing did not influence stiffness, in agreement with several studies that failed to find an influence of the prosthesis design on postoperative range of motion [39,40]. A lower postoperative tibial slope and a longer operating time were significant on univariate analysis only, probably because of their association with pain and tissue trauma in the early postoperative phase. Tibial slope, posterior condylar offset and the balance of the flexion gap are known to influence deep flexion [41,42]. Component malposition, particularly an internally rotated tibial or femoral component, was the cause of stiffness in one of our patients, a finding consistent with revision series that identified malrotation and oversizing as an important cause of the stiff, painful knee [43,44]. Differences in the anatomy of Asian knees, including a greater inherent varus and tibial plateau inclination, should be considered while planning bone cuts and the need for deep flexion in patients who require floor sitting should be considered at the planning stage [45]. An Indian comparison of fixed- and mobile-bearing designs likewise found similar clinical outcomes at a minimum follow-up of 4.5 years [46]. In our series, MUA improved flexion substantially, but some of the gain was lost over time, and the results were the best when manipulation was performed within 12 weeks, in line with previous studies [47,48]. Arthroscopic arthrolysis was helpful in a small number of knees, and revision for a malpositioned component was reserved for selected cases, as final outcomes after revision for stiffness are less predictable. These results reinforce that prevention and early detection, rather than late treatment, should be the main strategy. Since a large proportion of Indian patients need 110° or more of flexion for squatting and cross-legged sitting, a knee with a "satisfactory" flexion of 90° may still impair daily life and religious or social activities, and patient dissatisfaction is closely related to restricted motion and unmet expectations [49]. With the number of TKRs growing in India, and with revision for stiffness and other causes forming a significant part of the revision burden [50], a risk-based approach is practical: (i) identify high-risk patients preoperatively (BMI, diabetes, flexion < 95° or FFD > 10°); (ii) counsel them regarding realistic goals; (iii) optimise glycaemic control and weight; (iv) use enhanced multimodal analgesia; (v) provide closer supervised physiotherapy with documented adherence; and (vi) review at 2 and 6 weeks, with an early decision for MUA if flexion is below 90° at 6 weeks. Strengths and limitations The strengths of this study are its prospective design, the uniform surgical and rehabilitation protocols, a standardised goniometric assessment by an independent observer, a high rate of follow-up (94.3%), the inclusion of modifiable postoperative variables and a patient-relevant outcome (cross-legged sitting). The limitations are as follows. First, it was a single-centre study with a modest sample size and only 24 events, so the number of events per variable in the multivariable model was low and confidence intervals are wide, particularly for poor adherence and diabetes; the estimates should therefore be regarded as hypothesis-generating and validated in larger multicentre Indian cohorts. Second, we used a short-term definition of stiffness at 12 weeks, and some knees might have improved with time. Third, adherence was measured by diary and self-report and may be overestimated. Fourth, we did not measure glycaemic control (HbA1c), vitamin D status, or the inflammatory markers, which may modify the risk. Finally, patients with ankylosed knees and those needing constrained implants were excluded, which limits generalisability to the most severe cases.

Approximately one in eight patients developed knee stiffness at 12 weeks after primary TKR in this prospective series. Poor adherence to the home exercise programme, diabetes mellitus, reduced preoperative flexion, preoperative flexion contracture, higher early postoperative pain and higher BMI independently predicted stiffness. Because most of these factors are modifiable or can be identified before surgery, a risk-stratified pathway of preoperative optimisation, realistic counselling, effective analgesia, supervised physiotherapy and early manipulation when needed may improve outcomes, particularly in patients who require high flexion for activities of daily living. Larger multicentre studies are needed to validate these findings and to develop a simple risk score for use in Indian practice. Acknowledgements The authors thank the physiotherapy and nursing staff of the Department of Orthopaedics for their help in patient follow-up. Conflict of Interest and Funding The authors declare no conflict of interest. No external funding was received for this study.

1. Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am. 2007;89(4):780-5. 2. Pal CP, Singh P, Chaturvedi S, Pruthi KK, Vij A. Epidemiology of knee osteoarthritis in India and related factors. Indian J Orthop. 2016;50(5):518-22. 3. Mahajan A, Verma S, Tandon V. Osteoarthritis. J Assoc Physicians India. 2005;53:634-41. 4. Pachore JA, Vaidya SV, Thakkar CJ, Bhalodia HK, Wakankar HM. ISHKS joint registry: a preliminary report. Indian J Orthop. 2013;47(5):505-9. 5. Kim J, Nelson CL, Lotke PA. Stiffness after total knee arthroplasty. Prevalence of the complication and outcomes of revision. J Bone Joint Surg Am. 2004;86-A(7):1479-84. 6. Bong MR, Di Cesare PE. Stiffness after total knee arthroplasty. J Am Acad Orthop Surg. 2004;12(3):164-71. 7. Yercan HS, Sugun TS, Bussiere C, Ait Si Selmi T, Davies A, Neyret P. Stiffness after total knee arthroplasty: prevalence, management and outcomes. Knee. 2006;13(2):111-7. 8. Scranton PE Jr. Management of knee pain and stiffness after total knee arthroplasty. J Arthroplasty. 2001;16(4):428-35. 9. Mulholland SJ, Wyss UP. Activities of daily living in non-Western cultures: range of motion requirements for hip and knee joint implants. Int J Rehabil Res. 2001;24(3):191-8. 10. Hemmerich A, Brown H, Smith S, Marthandam SS, Wyss UP. Hip, knee, and ankle kinematics of high range of motion activities of daily living. J Orthop Res. 2006;24(4):770-81. 11. Rajgopal A, Ahuja N, Dolai B. Total knee arthroplasty in stiff and ankylosed knees. J Arthroplasty. 2005;20(5):585-90. 12. Mullaji AB, Padmanabhan V, Jindal G. Total knee arthroplasty for profound varus deformity: technique and radiological results in 173 knees with varus of more than 20 degrees. J Arthroplasty. 2005;20(5):550-61. 13. Mullaji AB, Marawar SV, Mittal V. A comparison of coronal plane axial femoral relationships in Asian patients with varus osteoarthritic knees and healthy knees. J Arthroplasty. 2009;24(6):861-7. 14. Shetty GM, Mullaji A, Bhayde S, Nha KW, Oh HK. Factors contributing to inherent varus alignment of lower limb in normal Asian adults: role of tibial plateau inclination. Knee. 2014;21(2):544-8. 15. Ritter MA, Stringer EA. Predictive range of motion after total knee replacement. Clin Orthop Relat Res. 1979;(143):115-9. 16. Harvey IA, Barry K, Kirby SP, Johnson R, Elloy MA. Factors affecting the range of movement of total knee arthroplasty. J Bone Joint Surg Br. 1993;75(6):950-5. 17. Lizaur A, Marco L, Cebrian R. Preoperative factors influencing the range of movement after total knee arthroplasty for severe osteoarthritis. J Bone Joint Surg Br. 1997;79(4):626-9. 18. Schurman DJ, Matityahu A, Goodman SB, Maloney W, Woolson S, Shi H, et al. Prediction of postoperative knee flexion in Insall-Burstein II total knee arthroplasty. Clin Orthop Relat Res. 1998;(353):175-84. 19. Gandhi R, de Beer J, Leone J, Petruccelli D, Winemaker M, Adili A. Predictive risk factors for stiff knees in total knee arthroplasty. J Arthroplasty. 2006;21(1):46-52. 20. Ritter MA, Harty LD, Davis KE, Meding JB, Berend M. Predicting range of motion after total knee arthroplasty. Clinical factors and radiographic measurements. J Bone Joint Surg Am. 2003;85-A(7):1278-85. 21. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310(20):2191-4. 22. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis. 1957;16(4):494-502. 23. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996;49(12):1373-9. 24. Bade MJ, Stevens-Lapsley JE. Early high-intensity rehabilitation following total knee arthroplasty improves outcomes. J Orthop Sports Phys Ther. 2011;41(12):932-41. 25. Harvey LA, Brosseau L, Herbert RD. Continuous passive motion following total knee arthroplasty in people with arthritis. Cochrane Database Syst Rev. 2014;(2):CD004260. 26. WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157-63. 27. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res. 1989;(248):13-4. 28. Dawson J, Fitzpatrick R, Murray D, Carr A. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br. 1998;80(1):63-9. 29. Daluga D, Lombardi AV Jr, Mallory TH, Vaughn BK. Knee manipulation following total knee arthroplasty. Analysis of prognostic variables. J Arthroplasty. 1991;6(2):119-28. 30. Parvizi J, Tarity TD, Steinbeck MJ, Sharkey PF, Hozack WJ, Rothman RH. Management of stiffness following total knee arthroplasty. J Bone Joint Surg Am. 2006;88 Suppl 4:175-81. 31. Hosmer DW, Lemeshow S. Applied Logistic Regression. 2nd ed. New York: John Wiley & Sons; 2000. 32. Nelson CL, Kim J, Lotke PA. Stiffness after total knee arthroplasty. J Bone Joint Surg Am. 2005;87 Suppl 1(Pt 2):264-70. 33. Kawamura H, Bourne RB. Factors affecting range of flexion after total knee arthroplasty. J Orthop Sci. 2001;6(3):248-52. 34. Meding JB, Reddleman K, Keating ME, Klay A, Ritter MA, Faris PM, et al. Total knee replacement in patients with diabetes mellitus. Clin Orthop Relat Res. 2003;(416):208-16. 35. Magit D, Wolff A, Sutton K, Medvecky MJ. Arthrofibrosis of the knee. J Am Acad Orthop Surg. 2007;15(11):682-94. 36. Unterhauser FN, Bosch U, Zeichen J, Weiler A. Alpha-smooth muscle actin containing contractile fibroblastic cells in human knee arthrofibrosis tissue. Arch Orthop Trauma Surg. 2004;124(9):585-91. 37. Abdul N, Dixon D, Walker A, Horabin J, Smith N, Weir DJ, et al. Fibrosis is a common outcome following total knee arthroplasty. Sci Rep. 2015;5:16469. 38. Amin AK, Patton JT, Cook RE, Brenkel IJ. Does obesity influence the clinical outcome at five years following total knee replacement for osteoarthritis? J Bone Joint Surg Br. 2006;88(3):335-40. 39. Gatha NM, Clarke HD, Fuchs R, Scuderi GR, Insall JN. Factors affecting postoperative range of motion after total knee arthroplasty. J Knee Surg. 2004;17(4):196-202. 40. Dennis DA, Komistek RD, Scuderi GR, Argenson JN, Insall JN, Mahfouz MR, et al. Factors affecting flexion after total knee arthroplasty. Clin Orthop Relat Res. 2007;464:53-60. 41. Bellemans J, Banks S, Victor J, Vandenneucker H, Moemans A. Fluoroscopic analysis of the kinematics of deep flexion in total knee arthroplasty. J Bone Joint Surg Br. 2002;84(1):50-3. 42. Massin P, Gournay A. Optimization of the posterior condylar offset, tibial slope, and condylar roll-back in total knee arthroplasty. J Arthroplasty. 2006;21(6):889-96. 43. Fisher DA, Dierckman B, Watts MR, Davis K. Looks good but feels bad: factors that contribute to poor results after total knee arthroplasty. J Arthroplasty. 2007;22(6 Suppl 2):39-42. 44. Haidukewych GJ, Jacofsky DJ, Pagnano MW, Hanssen AD. Functional results after revision of well-fixed components for stiffness after primary total knee arthroplasty. J Arthroplasty. 2005;20(2):133-8. 45. Kurosaka M, Yoshiya S, Mizuno K, Yamamoto T. Maximizing flexion after total knee arthroplasty: the need and the pitfalls. J Arthroplasty. 2002;17(4 Suppl 1):59-62. 46. Bhan S, Malhotra R, Kiran EK, Shukla S, Bijjawara M. A comparison of fixed-bearing and mobile-bearing total knee arthroplasty at a minimum follow-up of 4.5 years. J Bone Joint Surg Am. 2005;87(10):2290-6. 47. Mauerhan DR, Mokris JG, Ly A, Kiebzak GM. Relationship between length of stay and manipulation results after total knee arthroplasty. J Arthroplasty. 1998;13(8):896-900. 48. Issa K, Banerjee S, Kester MA, Khanuja HS, Delanois RE, Mont MA. The effect of timing of manipulation under anesthesia to improve range of motion and functional outcomes following total knee arthroplasty. J Bone Joint Surg Am. 2014;96(16):1349-57. 49. Bourne RB, Chesworth BM, Davis AM, Mahomed NN, Charron KD. Patient satisfaction after total knee arthroplasty: who is satisfied and who is not? Clin Orthop Relat Res. 2010;468(1):57-63. 50. Dalury DF, Pomeroy DL, Gorab RS, Adams MJ. Why are total knee arthroplasties being revised? J Arthroplasty. 2013;28(8 Suppl):120-1.