Elevated kneecap position linked to poorer results after ligament reconstruction in children

Children with recurrent kneecap dislocations who also have an abnormally high-riding patella fare noticeably worse after ligament reconstruction surgery than children with normal kneecap height, according to a retrospective single-center study from China. The research, published in Frontiers in Pediatrics, followed 103 pediatric patients who underwent medial patellofemoral ligament (MPFL) reconstruction at the Affiliated Hospital of Xuzhou Medical University between January 2022 and April 2025. The researchers found that nearly half of the children whose kneecap sat higher than normal before surgery experienced recurrent dislocation, subluxation, or another poor outcome afterward, compared with fewer than one in five children whose kneecap height fell within the typical range.

Lateral patellar dislocation is among the most common acute knee injuries in children and adolescents, with a cumulative incidence that peaks at 147.7 per 100,000 person-years in the 14 to 18 age group. After a first dislocation, recurrence rates reported in the literature range from 17 to 66 percent, and repeated episodes are associated with cartilage damage, anterior knee pain, and long-term functional impairment. MPFL reconstruction, in which a graft is used to rebuild the ligament that normally holds the kneecap against the inner side of the femoral groove, has become the preferred surgical treatment for recurrent instability in both adults and children. In pediatric series, postoperative Kujala scores typically exceed 85 points and redislocation rates run around 5 to 10 percent at short- to mid-term follow-up, but failures persist, and the risk factors behind them remain under investigation.

Patella alta, meaning an abnormally proximal position of the kneecap relative to the femoral trochlea, is a well-established anatomical contributor to instability. A high-riding patella engages the trochlear groove later during early knee flexion, reducing the bony constraint on the joint precisely in the range where the MPFL acts as the primary soft-tissue stabilizer. To quantify kneecap height, the study used the Insall-Salvati index, or ISI, calculated as the ratio of patellar tendon length to the greatest diagonal length of the patella on lateral radiographs taken at 30 degrees of knee flexion. The team prespecified the ISI as the primary height measure because its reference landmarks, the tendon and the patellar outline, remain reliably measurable throughout skeletal growth, whereas the alternative Caton-Deschamps index depends on articular-surface landmarks that change as the knee ossifies.

The cohort comprised 65 girls and 38 boys with a mean age of 13.7 years, ranging from 8 to 18, and a median follow-up of 29.9 months. Using a cutoff ISI of 1.2, drawn from the upper limit of normal ranges reported in adult and pediatric cohorts, the researchers classified 49 patients (47.6 percent) as having patella alta and 54 (52.4 percent) as having normal patellar height. The two groups were broadly comparable at baseline in age, sex, joint hypermobility scores, tibial tubercle to trochlear groove distance, trochlear dysplasia distribution, and preoperative Kujala scores, although body mass index was slightly higher and the preoperative Lysholm score slightly lower in the patella alta group. These differences were later accounted for in multivariable statistical models.

All operations were performed by the same surgical team using a uniform technique with hamstring tendon autografts. In skeletally immature patients, the surgeons used an entirely epiphyseal, physeal-sparing approach, placing the femoral fixation within the distal femoral epiphysis so that neither the tunnel nor the fixation device crossed the growth plate, with fluoroscopic confirmation of the position. Graft tensioning was performed with the knee held at 30 degrees of flexion. Postoperatively, patients wore a hinged brace locked in full extension for four weeks, began range-of-motion exercises at two weeks, and returned to unrestricted sport no earlier than six months after surgery.

Both groups showed significant postoperative reductions in the ISI, but the decrease was markedly larger in the patella alta group, where the mean index fell from 1.301 to 1.223, compared with a drop from 1.060 to 1.033 in the normal group. Despite this radiographic improvement, the clinical picture diverged sharply. At final follow-up, the patella alta group recorded a median Kujala score of 73.0 versus 84.5 in the normal group, and a median Lysholm score of 77.0 versus 85.5. Redislocation occurred in 24.5 percent of the alta group against 7.4 percent of the normal group, subluxation in 34.7 percent against 14.8 percent, and the composite poor outcome, defined as at least one redislocation or subluxation episode, in 44.9 percent against 18.5 percent. Postoperative range of motion was also lower in the alta group, while revision surgery, though numerically more common at 10.2 percent versus 3.7 percent, did not reach statistical significance.

Correlation analysis reinforced the pattern. Preoperative ISI correlated negatively with the final Kujala score, with a Spearman coefficient of −0.457, and more weakly with the final Lysholm score at −0.294, but showed no relationship with postoperative pain scores. In contrast, the Caton-Deschamps index showed no significant correlation with either functional score. After multivariable adjustment for body mass index, preoperative Lysholm score, high-grade trochlear dysplasia, and TT-TG distance, patella alta remained independently associated with poor outcome, with an adjusted odds ratio of 3.83, and each 0.1-unit increase in preoperative ISI carried an adjusted odds ratio of 1.36 for poor outcome and predicted a lower final Kujala score by 4.63 points. Notably, postoperative ISI was not independently associated with poor outcomes at all.

Perhaps the most striking finding concerned normalization. Among the 49 patella alta patients, 24, or 49.0 percent, achieved a postoperative ISI within the normal range, yet their rate of poor outcomes, 45.8 percent, was virtually identical to the 44.0 percent seen in patients whose index remained elevated. The authors interpret this to mean that the radiographic drop in patellar height after reconstruction is likely a secondary biomechanical consequence of restoring the medial check-rein, not the mechanism through which clinical benefit is delivered, and that patella alta may instead act as a marker for a broader constellation of anatomical abnormalities. Supporting this, agreement between the two height indices was essentially absent, with a Cohen’s kappa of −0.017, and only 11 of the 49 ISI-classified alta patients also met the CDI criterion, echoing a near-zero kappa previously reported in healthy adult knees.

Subgroup analyses suggested the adverse association was concentrated in younger and skeletally immature children. Among patients aged 15 or younger, the poor-outcome rate was 51.4 percent in the alta group versus 14.6 percent in the normal group, and among those with an open distal femoral physis it was 54.8 percent versus 16.7 percent, whereas in older and skeletally mature subgroups the differences were no longer significant. The authors caution that these subgroup findings were exploratory and underpowered, with the age-by-group interaction only approaching significance. Other limitations weigh heavily: the retrospective single-center design cannot establish causality, the ISI threshold of 1.2 derives from adult normative data, the minimum follow-up of six months may miss late recurrences, and the count of coexisting anatomical risk factors relied on pragmatic thresholds. The poor-outcome rate did rise with the number of coexisting risk factors, from 15.8 percent with none to 50.0 percent with four, but the trend was not statistically significant.

The study enters a contested field. Prior work in predominantly adult or mixed-age cohorts has variously identified patella alta as a strong predictor of failure, with one series of 211 reconstructions reporting an odds ratio of 4.9 for a Caton-Deschamps index of 1.3 or greater, and a meta-analysis of 807 patients identifying trochlear dysplasia as an independent predictor, while other reports found favorable five-year outcomes after isolated reconstruction regardless of patellar height. The authors suggest that the choice of measurement index may be an underrecognized source of this heterogeneity, and they position their pediatric data as hypothesis-generating rather than definitive. Within this cohort, the Insall-Salvati index tracked clinical outcomes where the Caton-Deschamps index did not, and the team argues that preoperative ISI assessment may prove useful for risk stratification and counseling when planning isolated MPFL reconstruction in children, pending confirmation in larger prospective multicenter studies with longer follow-up.

Subject of Research: The association between preoperative patellar height measured by the Insall-Salvati index and clinical outcomes after medial patellofemoral ligament reconstruction in children with recurrent patellar instability.

Article Title: Impact of Insall-Salvati index on clinical outcomes after medial patellofemoral ligament reconstruction in children: a retrospective single-center study

Article References: Zhang, C., Sun, S., Zhang, X., Deng, B., & Yao, S. (2026). Impact of Insall-Salvati index on clinical outcomes after medial patellofemoral ligament reconstruction in children: a retrospective single-center study. Frontiers in Pediatrics, 14, 1887245. https://doi.org/10.3389/fped.2026.1887245

Image Credits: AI Generated

DOI: 10.3389/fped.2026.1887245

Keywords: patella alta, Insall-Salvati index, MPFL reconstruction, patellar instability, pediatric orthopedics, Kujala score, Caton-Deschamps index, kneecap dislocation, retrospective study, surgical outcomes, trochlear dysplasia, risk stratification