Renal Impairment and Late Toxicities Comparing Contemporary Chemotherapy Regimens for Testicular Cancer in a Real-World Setting

41% of testicular cancer survivors had reduced kidney function a decade post-chemo; it cascades into hypertension, hyperlipidemia, and CVD

Journal: Journal of the National Comprehensive Cancer Network | Published: 2026-02-13 | Type: Comparative cohort study | PMID: 41698327 Authors: Kerns SL (Medical College of Wisconsin), Dinh PC (Indiana University), Fung C (University of Rochester), Feldman DR (Memorial Sloan Kettering), Hamilton RJ (Princess Margaret Cancer Centre), and others Funding/COI: Not disclosed

Summary

Cisplatin-based chemotherapy cures most testicular cancer—but this multicenter study of 798 survivors followed for a median of 11 years puts hard numbers on what's left behind. Forty-one percent had reduced kidney function, and the severity tracked with cumulative cisplatin dose. That kidney damage doesn't stay local: it cascades into hypertension, hyperlipidemia, and cardiovascular disease at rates 2 to 20 times higher than survivors with normal kidney function. The two most common NCCN-endorsed regimens (EPx4 and BEPx3) produced similar overall morbidity burden, but EPx4 generated significantly worse renal, auditory, and neurologic toxicity.

Claims

• 41% of 798 testicular cancer survivors had eGFR <90 mL/min/1.73 m² at median 11-year follow-up
• eGFR reduction correlated with cumulative cisplatin dose (r = −0.149, P<.0001)
• eGFR 60–89 mL/min/1.73 m²: 2× higher odds of hypertension (OR 2.01, P=.001)
• eGFR 45–59: 2.84× higher odds of hypertension (P value not individually reported)
• eGFR 30–44: 20× higher odds of hypertension (OR 20.0, P=.001)
• eGFR 30–44 also associated with 6× higher odds of hyperlipidemia (OR 6.10, P=.032) and 7× higher odds of CVD (OR 7.09, P=.023)
• EPx4 vs BEPx3: significantly worse renal impairment (aOR 1.55, P=.035), ototoxicity (aOR 1.48, P=.04), and neuropathy (aOR 1.77, P=.002)
• Overall cumulative burden of morbidity (CBM) scores were comparable for EPx4 vs BEPx3 (aOR 1.04, P=.83), but significantly worse for BEPx4 (aOR 1.77, P=.016) and VIPx4 (aOR 2.24, P=.038)
• BEPx4 associated with higher odds of Raynaud phenomenon vs EPx4 (OR 2.18, P=.003)

Study Quality

This is a real-world multicenter cohort study, not an RCT—patients weren't randomized to regimens, so selection bias is a genuine concern. Patients with more advanced disease likely received more intensive treatment, which could inflate apparent toxicity differences between regimens. Adjusted ordinal logistic regression attempts to account for this, but residual confounding in observational data is never fully eliminated. Notably, baseline eGFR prior to chemotherapy was not available for all patients, making it harder to attribute the entire observed decline to treatment rather than pre-existing factors.

The cisplatin-dose-to-eGFR correlation of r = −0.149, while statistically robust, explains only about 2% of the variance in kidney function—other factors dominate. The downstream cardiovascular associations are more striking clinically, but the severe eGFR category (30–44 mL/min/1.73 m²) is small, making the 20× hypertension odds ratio statistically fragile with wide implied confidence intervals.

Red Flags

• No randomization: treatment assignment reflects clinician decisions, not random allocation—sicker patients may have received different regimens
• Funding and COI not disclosed: a significant gap for a multi-institutional study comparing specific chemotherapy regimens
• Survivor bias: men who died from toxicities or early recurrence are excluded, likely underestimating the true long-term burden
• Weak dose-eGFR correlation: r = −0.149 is statistically significant but explains ~2% of variance
• Small n in severe eGFR category (30–44 mL/min/1.73 m²): the 20× hypertension OR is the most alarming number in the paper but rests on a thin subgroup
• No pre-chemotherapy eGFR baseline for all patients

Strengths

• Large real-world cohort (n=798) with long follow-up (median 11 years; 65% survived ≥10 years)
• Multicenter design spanning multiple major academic cancer centers across the US and Canada
• Quantified cumulative burden of morbidity (CBM) across multiple organ systems simultaneously, not just individual toxicities
• Covers the three dominant NCCN-endorsed regimens, producing clinically actionable comparisons
• Patient-reported global physical health correlated significantly with objective CBM scores (P<.001), validating the composite measure

Verdict

This paper does something genuinely useful: it traces a causal chain—cisplatin dose → kidney damage → cardiovascular disease—in a large survivor cohort at the timeframe when these consequences actually show up. The finding that EPx4 and BEPx3 produce equivalent overall morbidity burden but distinct toxicity profiles (EPx4 worse on kidney, hearing, and nerves) is clinically meaningful for treatment decisions in equivalent-efficacy scenarios. The undisclosed funding and COI are a real problem given the direct comparison between regimens—readers can't assess whether institutional or commercial interests shaped the framing. The study design limits causal inference, but for survivorship research, this kind of real-world long-term data is often the best available. The modest correlation coefficient shouldn't obscure the main signal: once kidney function drops significantly, cardiovascular risk rises sharply and persistently.