Prediction models for postoperative renal function after living donor nephrectomy: a systematic review

Area of Science:

• Nephrology
• Transplant Surgery
• Biostatistics

Background:

• Living-donor nephrectomy (LDN) is a vital source of organs for kidney transplantation.
• Current preoperative evaluations lack formal models to predict post-LDN renal function decline.
• Validated prediction models are needed to aid donor-patient decision-making.

Purpose of the Study:

• To systematically review and summarize existing models predicting mid- to long-term renal function after LDN.
• To identify models utilizing only preoperative variables for predicting renal function post-LDN.
• To inform clinical practice and patient counseling regarding LDN outcomes.

Main Methods:

• Systematic literature review adhering to EAU guidelines and PRISMA 2020 recommendations.
• Inclusion of studies with models using only preoperative variables for qualitative analysis.
• Protocol registered with PROSPERO (ID: CRD42022380198).

Main Results:

• Six models from six studies met inclusion criteria, all based on retrospective cohorts with high risk of bias (PROBAST).
• Models incorporated 2-4 preoperative variables, with donor age being most frequent (83%).
• Significant heterogeneity existed in model development, outcomes, and performance; only three models underwent external validation.

Conclusions:

• Few externally validated models exist to predict renal function after LDN using preoperative data.
• Current evidence is insufficient for routine clinical adoption of these models.
• Future research should focus on developing robust, user-friendly models using large, multicenter datasets.