Minimally invasive pyeloplasty at a tertiary care centre: a retrospective cohort study of robotic, three-dimensional and two-dimensional laparoscopic approaches
TL;DR
Surgical approach was not independently associated with success in an exploratory multivariable model, and in this non-randomised cohort no significant difference in success was demonstrated between the three approaches.
Abstract
Dismembered pyeloplasty is the reference standard for ureteropelvic junction obstruction (UPJO). Two-dimensional (2D) laparoscopy is limited by the loss of stereoscopic depth perception; robotic pyeloplasty and three-dimensional (3D) laparoscopy were developed to address this. Three-arm comparisons including 3D laparoscopy as a distinct intermediate technology, particularly from resource-constrained settings, remain scarce. We retrospectively analysed 109 consecutive patients undergoing primary Anderson–Hynes dismembered pyeloplasty between January 2018 and June 2022, performed by a single surgical team: 2D laparoscopic (n = 27), 3D laparoscopic (n = 27) and robotic (n = 55). Patients were not randomised, and the groups reflect the modality actually used, with robotic procedures performed on the da Vinci Xi and 3D procedures on an Olympus system. Continuous variables were compared by one-way ANOVA with Tukey post-hoc testing, and categorical variables by the chi-square or Fisher–Freeman–Halton exact test. Complications were graded by the Clavien–Dindo classification. Treatment success was a six-month composite of symptom resolution, unobstructed renographic drainage and preserved differential renal function. Multivariable logistic and linear regression were used to adjust for pre-specified confounders. Effect estimates are reported with 95% confidence intervals (CIs). The groups were comparable at baseline (all P > .05); median follow-up was 26 months (interquartile range 15–37). Nursing set-up time (mean difference 27.7 min versus 2D and 31.0 min versus 3D; both P < .001) and surgeon operative time (20.5 min, P = .019; and 27.4 min, P = .001) were longer in the robotic group, with no difference between the 2D and 3D groups (P = .71). Total operating-theatre time was 55.5 min longer than 2D (95% CI 36.3–74.7) and 61.2 min longer than 3D (95% CI 42.0–80.4; both P < .001). Estimated blood loss (P = .11) and complications (20 patients, 18.3%, predominantly Clavien–Dindo grade I–II; P = .46) did not differ. The surgeon-rated ease-of-procedure score, which is subjective and unvalidated, was lowest for 2D laparoscopy (P < .001). Treatment success was 85.2% for 2D laparoscopy, 88.9% for 3D laparoscopy and 96.4% for robotic pyeloplasty (P = .20). There were only nine failures in all, and a difference smaller than about 22% points could not have been detected, so equivalence cannot be inferred. Surgical approach was not independently associated with success in an exploratory multivariable model. In this non-randomised cohort no significant difference in success was demonstrated between the three approaches. With only nine failures, however, the data remain compatible with clinically important differences and cannot be taken as evidence of equivalence. Robotic surgery was associated with longer set-up and operative times. Because group assignment was not randomised, these findings represent associations and should not be interpreted as demonstrating the superiority of any technique.