High continuous fragility index values for randomised controlled trials investigating medial patellofemoral ligament reconstruction for patellar instability: A systematic review

TL;DR

This study reviews RCTs on medial patellofemoral ligament reconstruction and finds high statistical fragility, especially in continuous outcomes.

Contribution

The study introduces the use of continuous fragility index in evaluating RCTs for patellar instability.

Findings

The median continuous fragility index was 11.7, indicating high statistical fragility in RCTs.

Loss to follow-up exceeded fragility indices in many outcomes, affecting study reliability.

The reverse fragility index showed moderate fragility in non-significant outcomes.

Abstract

To evaluate the statistical fragility of randomised controlled trials (RCTs) investigating medial patellofemoral ligament reconstruction (MPFLR) of patients with patellar instability.

A search of PubMed, MEDLINE, and EMBASE databases for RCTs investigating surgical management of patients with patellar instability from inception to 25 April 2024. Studies that reported over one significant dichotomous outcome, significant continuous outcome, and/or nonsignificant dichotomous outcome were included for analysis. The fragility index (FI), continuous fragility index (CFI) and reverse fragility index (RFI) were calculated for significant dichotomous outcomes, significant continuous outcomes, and non‐significant dichotomous outcomes, respectively.

Thirteen RCTs were included which reported on a total of 929 patients (64.1% female, mean age of 24.4 years [range: 10–47 years] before any patients were lost to follow‐up. The median FI was 1.0 (interquartile range [IQR], 1; 95% confidence interval [CI], 0.725–2.275; range, 0–4). The number of patients lost to follow‐up at the final time point was greater than the study‐specific FI in 7 out of 8 outcomes from four studies (87.5%). The overall median CFI for included RCTs amongst 31 outcomes from nine studies was 11.7 (IQR, 12.9–95% CI, 8.9–13.9; range 0–30.3). The number of patients lost to follow‐up at the final time point was greater than the study‐specific CFI in thirteen outcomes from six studies (41.9%). The median RFI was 7.0 (IQR, 1.0; 95% CI, 6.5–7.5). The number of participants lost to follow‐up was greater than the study‐specific RFI in a single outcome from one study (16.7%).

This systematic review shows that while FI was low at a median of 1.0, consistent with RCTs in sports medicine, the other indicators of statistical fragility were relatively high, especially CFI (11.7). The level of fragility should be taken into account when assessing practical applicability of RCTs on patellar instability, in combination with other indicators of study rigour.

Level I.

Figures