Assessing Renal Outcomes After Switching from Tenofovir Disoproxil Fumarate to Tenofovir Alafenamide in People with HIV: Proteinuria Changes and Limits of Causal Attribution.
D'Anna Tommaso T, Trizzino Marcello M, Belmonte Beatrice B, Gioè Claudia C et al.
Background: The total urinary protein-to-creatinine ratio (PCr) measures proteinuria but does not establish tubular injury or its recovery. Methods: We examined PCr changes after switching from tenofovir disoproxil fumarate (TDF) to tenofovir alafenamide (TAF) in a single-centre retrospective cohort of 364 adults with human immunodeficiency virus (HIV) infection. Paired PCr values were available for 159 patients (43.7%). The primary descriptive endpoint was a PCr reduction below 20%, an operational, nonvalidated threshold. Analyses comprised paired rank-based comparisons, baseline-adjusted regression, threshold and influence sensitivity analyses, and nested cross-validation of ridge logistic regression using three predictor sets. Results: The median paired PCr change was -4.0 mg/g (interquartile range -41.5 to 11.5; signed-rank Benjamini-Hochberg-adjusted q = 0.0098); 99/159 patients (62.3%) had a reduction below 20%. Baseline log(1 + PCr) was positively associated with post-switch log(1 + PCr), while its coefficient in the corresponding change model was exactly one unit lower, illustrating mathematical coupling. The cross-validated area under the curve was 0.709 for baseline PCr alone, 0.475 for the other covariates, and 0.701 for all predictors. Conclusions: PCr changes were heterogeneous, and additional covariates did not improve discrimination over the baseline PCr alone. Missing outcomes, unavailable specimen dates, the absence of tubular-specific biomarkers and the uncontrolled design preclude conclusions about irreversible tubular injury, causal treatment effects or clinical prediction utility.