Development and internal validation of a new Varicose Vein Recurrence Score for open superficial venous surgery

Recurrence after open superficial venous surgery occurs in 20% to 60% of patients at 5 years, yet no validated composite preoperative scoring system exists to quantify individual risk. We aimed to develop and internally validate a predictive model and a derived clinical risk score-the Varicose Vein Recurrence (VVR) Score-for 5-year recurrence after open surgery of the superficial venous system.

This retrospective single-center cohort study included consecutive patients undergoing saphenofemoral or saphenopopliteal junction ligation with or without trunk stripping and concomitant phlebectomies between January 2006 and December 2025. The primary outcome was duplex-confirmed recurrent reflux with clinically visible recurrent varices at 5 years; reintervention (open, endovenous, or foam) within 5 years served as a prespecified harder secondary end point. The primary analysis was conducted on extremities with complete 5-year outcome ascertainment. We used multivariable logistic regression with cluster-robust standard errors (patient-level clustering) for model development and cluster bootstrap resampling (500 replications) for internal validation. A simplified integer-based score was derived from shrinkage-adjusted coefficients. Prespecified sensitivity analyses included procedure-stratified subgroup models (three mutually exclusive primary subgroups-high ligation + stripping, high ligation alone, and phlebectomy alone-plus an overlapping saphenopopliteal junction-procedure subgroup), continuous calendar-year analyses with time × predictor interactions, treatment choice as a model covariate, inverse probability of censoring weighting, and Cox regression. Reporting followed the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis guidelines (type 1b).

Among 12,480 eligible extremities (9210 patients), 10,142 (81.3%) had complete 5-year ascertainment and constituted the primary cohort. The 5-year recurrence rate was 25.1% (2546/10,142); 74.1% of recurrent extremities underwent reintervention (5-year reintervention rate, 18.6%). Six preoperative predictors were retained: anterior saphenous vein (formerly anterior accessory saphenous vein) reflux (odds ratio [OR], 2.33; 95% confidence interval [CI], 2.03-2.67), incompetent perforators (OR, 2.04; 95% CI, 1.77-2.35), body mass index of ≥30 kg/m² (OR, 1.82; 95% CI, 1.60-2.07), deep venous reflux or obstruction (OR, 1.72; 95% CI, 1.44-2.06), Clinical, Etiologic, Anatomic, and Pathophysiologic classification C4-C6 (OR, 1.62; 95% CI, 1.42-1.86), and great saphenous vein diameter ≥7 mm (OR, 1.41; 95% CI, 1.23-1.62). The optimism-corrected C-statistic was 0.744 (95% CI, 0.726-0.762), with a calibration slope of 0.97 (0.924-1.016) and an intercept of -0.02 (-0.06 to 0.02). Model performance was preserved for the harder reintervention end point (C-statistic, 0.73) and across operative subgroups (apparent C-statistics, 0.70-0.74). The derived VVR Score (0-13 points) stratified patients into low (12.8%), moderate (25.2%), high (41.7%), and very high (66.8%) 5-year recurrence-risk categories. Continuous calendar-year analysis showed no significant time × predictor interactions (smallest P = .21). Sensitivity analyses including inverse probability of censoring weighting on all 12,480 extremities, Cox regression, generalized estimating equations, and treatment-choice-adjusted models produced consistent results.

The VVR Score provides an internally validated tool for estimating 5-year recurrence risk after open superficial venous surgery using six preoperative clinical and duplex parameters. Performance was preserved across operative subgroups and for the harder reintervention end point. Because the score is derived in an open-surgical population, external validation in geographically diverse open-surgical cohorts and rederivation/recalibration before use in endothermal, nonthermal, or foam-sclerotherapy populations are essential before broader clinical adoption.