Proteinuria Reporting Signals With ICI–VEGFR-TKI Regimens in RCC
Key Takeaways
- In renal cell carcinoma (RCC) adverse-event reports in FAERS, three ICI–VEGFR-TKI regimens had significantly greater VEGF-axis reporting than nivolumab plus ipilimumab.
- Proteinuria accounted for 75%–95.8% of target-regimen VEGF-axis events reported in FAERS, a fraction of reports rather than treated patients.
- JADER showed significantly positive VEGF-axis reporting patterns for all three regimens in the same directions as FAERS.
- Exclusive Tier-A AKI was not enriched for any target regimen, whereas a separate narrow acute-renal-failure query was elevated for pembrolizumab plus lenvatinib.
The analysis retrospectively examined the Food and Drug Administration Adverse Event Reporting System (FAERS), including its earlier Adverse Event Reporting System (AERS) files, from 2004Q1–2026Q2, and the Japanese Adverse Drug Event Report database (JADER) 202607 release, covering reports through 2026Q1. After retaining the latest FAERS report versions, investigators identified 84,369 reports in a broader strict-RCC cohort, with the indication linked to a regimen drug on the same drug record rather than elsewhere in the case; membership in a named doublet was not required.
They compared pembrolizumab plus axitinib, pembrolizumab plus lenvatinib, and nivolumab plus cabozantinib separately against nivolumab plus ipilimumab within smaller, regimen-specific common-support cohorts, excluding reports matching multiple named regimens. Author-defined exact-term exclusive Tier-A families captured immune-interstitial nephritis, vascular endothelial growth factor (VEGF)-axis events comprising proteinuria, nephrotic syndrome, or renal-limited thrombotic microangiopathy, and nonspecific AKI or acute renal failure; reports spanning families were classified as mixed. Year-stratified Mantel–Haenszel reporting odds ratio (MH-ROR) estimates compared each phenotype with otherwise eligible RCC reports, with multiplicity adjustment across the primary comparisons. JADER was analyzed separately for directional agreement. A later temporal evaluation tested recovery of rule-derived labels from non-anchor report features with or without regimen indicators.
VEGF-axis reporting in FAERS exceeded that for nivolumab plus ipilimumab for pembrolizumab plus axitinib (MH-ROR, 5.31; 95% confidence interval, 2.43–11.59), pembrolizumab plus lenvatinib (MH-ROR, 19.48), and nivolumab plus cabozantinib (MH-ROR, 4.59). All three associations remained positive after multiplicity adjustment. Proteinuria dominated the reported phenotype events; removing it left sparse comparisons for the remaining VEGF-axis terms.
JADER showed significant VEGF-axis associations in the same directions, although its estimates were not pooled with FAERS. The pembrolizumab–axitinib AKI and nivolumab–cabozantinib immune-interstitial JADER comparisons did not meet evaluability criteria. In FAERS, exclusive Tier-A AKI was not significantly enriched for any of the three regimens. A separately reconstructed, narrower acute-renal-failure query instead showed elevated reporting for pembrolizumab plus lenvatinib (MH-ROR, 1.56; 95% confidence interval, 1.20–2.02). Excluding the final reporting year preserved positive FAERS VEGF-axis directions; excluding Japan-linked FAERS reports also preserved positive point directions, but intervals crossed one. Adding regimen indicators modestly improved temporal recovery of rule-defined labels, with immune-interstitial labels hardest to recover.
These spontaneous-report contrasts describe relative reporting, not treated-patient incidence, comparative clinical risk, causation, or confirmed renal mechanisms. Without treated-person denominators or biopsy adjudication, differing urine surveillance could favor detection of proteinuria, while selected exact terms could miss nephritis coded only as AKI. The checkpoint inhibitor–VEGFR-TKI versus dual-checkpoint comparator cannot isolate a TKI add-on effect, and available baseline kidney-function and treatment-line information did not support clinical-population matching. Sparse comparator VEGF-axis events and concentration among report senders limit interpretation of signal magnitude; top-sender exclusions were not evaluable. FAERS includes U.S. reports, whereas JADER reflects Japanese reporting; the systems may share reporting pathways without case-level linkage. The narrower renal query did not reproduce the primary cohort definition, and label recovery by the classifier was not diagnostic accuracy.
Li and colleagues identified a proteinuria-dominant VEGF-axis reporting pattern with directional agreement between databases. Their findings also distinguish exclusive AKI labels from the differently defined acute-renal-failure signal.
Clinician Questions
Which FAERS and JADER drug-role criteria qualified ICI combinations for the RCC renal-reporting comparison?
For renal cell carcinoma adverse-event reports, FAERS required both components of pembrolizumab–axitinib, pembrolizumab–lenvatinib, nivolumab–cabozantinib, or comparator nivolumab–ipilimumab to be designated primary or secondary suspects, with RCC linked to a target drug on the same drug-sequence record. The main JADER analysis required both components to be primary suspects and RCC to be indicated on the DRUG row. These eligibility rules do not establish drug causation.
Did generic thrombotic microangiopathy qualify as a VEGF-axis kidney event in the RCC combination reports?
In Li and colleagues’ renal cell carcinoma spontaneous-report comparison of ICI–VEGFR-TKI regimens, the author-defined VEGF-axis exact-term family included proteinuria, nephrotic syndrome, and specifically renal-limited thrombotic microangiopathy. Generic thrombotic microangiopathy alone did not satisfy the renal-limited anchor; the rule neither confirms pathology nor supplies a clinical diagnostic criterion.