Higher Preoperative Pain Predicts Meaningful Function Gains One Year After Hip Replacement

Older adults who go into total hip replacement surgery with more intense preoperative pain are more likely to report a clinically meaningful improvement in physical function one year later, according to a prospective cohort study conducted in a Latin American public healthcare setting. Among 53 patients aged 65 and older who completed a twelve-month follow-up, 66 percent reached the minimal clinically important difference on the physical function subscale of the WOMAC questionnaire, and preoperative pain intensity emerged as the only independent predictor of that achievement.

The study, published in Archives of Orthopaedic and Trauma Surgery, followed consecutive patients with severe hip osteoarthritis awaiting primary total hip arthroplasty between August 2017 and December 2018. Of 85 patients enrolled at baseline, 53, or 62.4 percent, completed the twelve-month assessment. The researchers classified each participant as having achieved the minimal clinically important difference, or MCID, if their WOMAC physical function score improved by at least 26.54 points on a 0-to-100 scale, a threshold validated for the Spanish version of the instrument. Positive change scores indicated improvement relative to each patient’s own preoperative baseline.

At the group level, the contrast in baseline pain was striking. Patients who went on to achieve the MCID reported a median preoperative pain score of 7 on the 0-to-10 Numerical Rating Scale, with an interquartile range of 5 to 9, while those who did not achieve it reported a median of 3, with an interquartile range of 1 to 6. This difference was statistically significant, with a p value of 0.005. Rural residence was also associated with higher MCID achievement in unadjusted comparisons, with 100 percent of rural participants reaching the threshold versus 59.1 percent of urban participants, though the rural subgroup contained only nine patients, making that observation fragile.

To test whether baseline pain retained its association after accounting for other factors, the researchers built a multivariable binary logistic regression model adjusted for age, sex, body mass index, and baseline pain. Preoperative pain intensity was the sole independent predictor, with an odds ratio of 1.42 per one-point increase on the pain scale, a 95 percent confidence interval of 1.11 to 1.82, and a p value of 0.006. In practical terms, each additional point of preoperative pain was associated with a 42 percent increase in the odds of achieving a clinically meaningful functional gain. Age, male sex, and BMI showed no significant associations, with p values of 0.902, 0.971, and 0.810 respectively.

The model showed moderate explanatory power, with a Nagelkerke R-squared of 0.291, meaning it accounted for roughly 29 percent of the variability in MCID achievement, and acceptable discrimination, with an area under the receiver operating characteristic curve of 0.757. Multicollinearity was negligible, with variance inflation factors no higher than 1.084. Because the sample was small, the team also ran an internal validation using non-parametric bootstrapping with 1,000 resamples, estimating bias-corrected and accelerated 95 percent confidence intervals. Baseline pain retained its statistical significance under this procedure, with a bootstrap odds ratio of 1.474 and a confidence interval of 1.058 to 2.204, while secondary covariates showed wide imprecision, particularly male sex, whose bootstrap confidence interval ranged from 0.084 to 9.773.

A secondary parsimonious model including baseline pain, polypharmacy, and assistive device use reinforced the pattern. Baseline pain remained the only significant predictor, with an odds ratio of 1.30, a confidence interval of 1.05 to 1.63, and a p value of 0.019, alongside a bootstrap odds ratio of 1.350 with a confidence interval of 1.002 to 1.805. The authors note this model explained a somewhat larger share of outcome variability, with a Nagelkerke R-squared of 0.324 and an area under the curve of 0.794.

The exploratory ROC analysis identified a baseline pain cutoff of 6.5 on the Numerical Rating Scale as the optimal discriminator, determined by the maximum Youden’s J statistic of 0.46. This threshold yielded 70 percent sensitivity and 77 percent specificity for predicting MCID achievement, with an area under the curve of 0.740 and a 95 percent confidence interval of 0.59 to 0.90. The authors stress, however, that this cutoff was derived and evaluated within the same cohort of 53 patients without internal or external validation, so the sensitivity, specificity, and AUC figures are likely optimistic due to sample-specific overfitting. They frame the threshold strictly as an exploratory benchmark for future risk stratification research rather than an established clinical decision-making tool.

The direction of the association fits what the authors describe as the room-for-improvement phenomenon. The MCID reflects the magnitude of intra-individual change relative to baseline severity, not absolute postoperative performance. Patients starting with severe pain or poor function have greater mathematical headroom to register large improvements, while those already close to an optimal functional state have less perceptible change available even after otherwise successful surgery. The authors emphasize that achieving an MCID and reaching a satisfactory absolute postoperative functional state are distinct outcomes, and they argue that future studies of hip interventions should report both change-based thresholds and absolute functional targets to avoid an incomplete picture of clinical success.

The study faced substantial limitations that the authors address directly. The effective sample at twelve months was small, which constrained statistical power and widened confidence intervals for secondary covariates. The dropout rate was high at 37.6 percent, and although baseline characteristics did not differ significantly between completers and dropouts, the authors caution that low statistical power means differential attrition bias cannot be ruled out. An extreme-case sensitivity analysis across the full enrolled cohort of 85 patients bounded the overall MCID achievement rate between 41.2 percent, assuming every dropout failed to achieve the threshold, and 78.8 percent, assuming every dropout achieved it. Baseline pain remained a strong independent predictor under the best-case assumption, with an odds ratio of 1.418 and a confidence interval of 1.136 to 1.770, and retained a positive trend toward significance even under the worst-case assumption, with an odds ratio of 1.199, a confidence interval of 0.995 to 1.445, and a p value of 0.056.

Further caveats shape how the findings should be read. Baseline WOMAC function was deliberately excluded from the regression models to avoid mathematical coupling, since the baseline score is an intrinsic component of the change-based outcome, so the independent role of pain must be interpreted in light of baseline functional severity. Dichotomizing continuous change scores at a fixed threshold of 26.54 points carries known methodological drawbacks, including reduced statistical power and arbitrary classification near the cutoff, and MCID thresholds are known to vary across populations, baseline function levels, and estimation methods. No Chilean-specific thresholds currently exist, so the Spanish-validated cutoff was used for comparability. The models also left most outcome variability unexplained, pointing to unmeasured factors such as pain catastrophizing, patient expectations, and emotional state, which previous research has linked to postoperative functional outcomes. The authors conclude that independent prospective validation in external cohorts is necessary before the pain threshold of 6.5 can be applied in clinical practice, while noting that the study highlights baseline pain intensity as a prognostically relevant, easily measured variable in an older, publicly insured population that has been underrepresented in arthroplasty outcome research.

Subject of Research: Preoperative pain intensity and achievement of the minimal clinically important difference for physical function one year after total hip arthroplasty

Article Title: Preoperative pain and achievement of the minimal clinically important difference for physical function one year after total hip arthroplasty: a prospective cohort study

Article References: Gajardo-Burgos, R., Arteaga San Martín, R., Fuentes, J., Monrroy Uarac, M., Saavedra-Santiesteban, C., Bascour-Sandoval, C., & de Villarreal, E. S. (2026). Preoperative pain and achievement of the minimal clinically important difference for physical function one year after total hip arthroplasty: a prospective cohort study. Archives of Orthopaedic and Trauma Surgery, 146(1), Article 367. https://doi.org/10.1007/s00402-026-06516-9

Image Credits: AI Generated

DOI: 10.1007/s00402-026-06516-9

Keywords: total hip arthroplasty, preoperative pain, minimal clinically important difference, WOMAC, physical function, prospective cohort study, osteoarthritis, logistic regression, ROC analysis, older adults, patient-reported outcomes, risk stratification