Immunotherapy & the loss of renal tolerance

Checkpoint-inhibitor acute interstitial nephritis

Release the brakes on the immune system to fight the cancer and, weeks to months later, the same unleashed T cells can turn on the kidney's interstitium — a delayed, often steroid-responsive nephritis that hides behind a bland urine and a slowly rising creatinine.

16.5%: develop AKIShare of 309 immune-checkpoint-inhibitor-treated patients who developed acute kidney injury (KDIGO stage 1/2/3 in 53/22/25%) in a single-center cohortPMID 32601079
14 weeks: median onsetMedian time from checkpoint-inhibitor initiation to AKI (IQR 6–37) — later than most other immune-related adverse eventsPMID 31896554
93%: tubulointerstitial nephritisAcute tubulointerstitial nephritis was the dominant lesion in 93% of the 60 biopsied patients in the landmark multicenter seriesPMID 31896554
64%: renal recoveryRenal recovery in a 429-patient, 30-site consortium; steroids within 14 days raised the odds of recovery (adjusted OR 2.64)PMID 34625513
1.84×: risk with a PPIPooled odds ratio for checkpoint-inhibitor AKI with concurrent proton-pump-inhibitor use across 14 studies (12,694 patients) — a modifiable risk factorPMID 38471492

Acute Interstitial Nephritis Acute Tubular Necrosis Nivolumab profile Pembrolizumab profile Atezolizumab profile Ipilimumab profile Durvalumab profile Cemiplimab profile irAE atlas · 11 organs

Teaching case · illustrative composite, not a real patient

A 63-year-old man with metastatic melanoma, about three and a half months into combination ipilimumab plus nivolumab, is found on routine labs to have a creatinine risen from 0.9 to 2.0 mg/dL. He is taking a proton-pump inhibitor for reflux and recently had checkpoint-inhibitor colitis. The urine shows subnephrotic proteinuria and sterile pyuria without significant hematuria or dysmorphic red cells.

Because the picture is a subacute AKI with a bland-to-tubular sediment in a patient on immunotherapy, checkpoint-inhibitor tubulointerstitial nephritis is suspected. Prerenal and obstructive causes are excluded; the proton-pump inhibitor and other potential interstitial-nephritis triggers are stopped. A kidney biopsy shows acute tubulointerstitial nephritis. The checkpoint inhibitor is held and prompt corticosteroids (prednisone ~1 mg/kg/day) are started, with a taper as creatinine improves. Kidney function recovers substantially, though not fully, over the following weeks. Rechallenge is weighed against the risk of recurrence — about one in five in the founding cohorts, roughly one in seven in the most recent synthesis.

Teaching point — Checkpoint-inhibitor nephritis is a delayed (weeks-to-months) tubulointerstitial nephritis, not an acute tubular necrosis: recognize it as a slowly rising creatinine with subnephrotic proteinuria and sterile pyuria in a patient on immunotherapy, often alongside other immune-related adverse events and a concurrent nephritis-causing drug such as a PPI. Management is holding the drug and prompt corticosteroids — earlier steroids improve recovery — plus stopping the modifiable co-triggers. Most patients recover renal function, and rechallenge is possible but carries a real risk of recurrence.

How it happens

The pathophysiology as a cascade — select a step to follow the mechanism.

PD-1/PD-L1 and CTLA-4 are brakes that keep self-reactive T cells in check; blocking them to unleash anti-tumor immunity also lowers the threshold for T cells to attack self tissue — the kidney interstitium among them.
PMID 31896554 (opens PubMed in a new tab)
The resulting lesion on biopsy is an acute tubulointerstitial nephritis — a lymphocytic infiltrate in the interstitium with tubulitis — the dominant finding in the large biopsy series, rather than tubular necrosis or a glomerular lesion.
PMID 34625513 (opens PubMed in a new tab)
Many patients are concurrently taking a drug that itself can prime interstitial nephritis (PPIs, NSAIDs, antibiotics); the leading hypothesis is that checkpoint blockade breaks tolerance to these previously-tolerated drug antigens — consistent with PPI use being an independent, modifiable risk factor.
PMID 38471492 (opens PubMed in a new tab)
Unlike a direct toxin, the injury is delayed (a median of ~14–16 weeks) and often accompanies extrarenal immune-related adverse events, reflecting an evolving autoimmune process rather than a dose-dependent poisoning.
PMID 31896554 (opens PubMed in a new tab)

How we learned it

2020
Meraz-Muñoz, Kitchlu, and colleagues report that 16.5% of 309 checkpoint-inhibitor-treated patients developed AKI, associated with other immune-related adverse events and hypertension, and not with worse survival.
Established that AKI is a common complication of checkpoint inhibitors and tied kidney injury to the broader immune-related-adverse-event syndrome.
PMID 32601079 (opens PubMed in a new tab)
2020
Cortazar et al. assemble a multicenter series of 138 checkpoint-inhibitor-associated AKI cases: onset at a median of 14 weeks, acute tubulointerstitial nephritis in 93% of those biopsied, and complete/partial/no recovery in 40/45/15%.
The landmark characterization — risk factors (low baseline eGFR, PPIs, combination therapy), the delayed onset, the interstitial-nephritis lesion, and steroid-responsiveness.
PMID 31896554 (opens PubMed in a new tab)
2021
The ICPi-AKI Consortium (Gupta et al.) reports 429 patients from 30 sites in 10 countries: acute tubulointerstitial nephritis in 82.7% of biopsies, renal recovery in 64%, and better recovery when corticosteroids were started early (within 3 days, adjusted OR 2.09).
Confirmed the picture at scale and showed that the timing of steroids matters — earlier treatment, better kidney recovery.
PMID 34625513 (opens PubMed in a new tab)
2024
Mohan et al. meta-analyze 14 studies (12,694 patients) and find proton-pump-inhibitor use nearly doubles the odds of checkpoint-inhibitor AKI (pooled OR 1.84).
Elevated PPIs from an observation to a quantified, modifiable risk factor worth de-prescribing where possible.
PMID 38471492 (opens PubMed in a new tab)
2025
The American Society of Onco-nephrology publishes a dedicated position statement on checkpoint-inhibitor nephrotoxicity, favoring biopsy for stage ≥2 AKI and prompt empiric corticosteroids where biopsy is not feasible.
Codified nephrology-specific guidance for diagnosis, steroids, and rechallenge — the field's consensus reference.
PMID 39455026 (opens PubMed in a new tab)

The landmark studies

Multicenter case-control, 138 cases + 276 controls

Clinical features and outcomes of immune checkpoint inhibitor-associated AKI: a multicenter study

Cortazar FB, et al. · J Am Soc Nephrol 2020 · PMID 31896554

Low baseline eGFR, PPI use, and combination therapy were independent risk factors; acute tubulointerstitial nephritis dominated the biopsies; most were steroid-treated, and 85% achieved complete or partial kidney recovery — 40% complete, 45% partial, 15% none.

Onset median 14 wk; TIN in 93% of 60 biopsied; steroids 86%; recovery complete/partial/none 40/45/15%; rechallenge 22%, recurrence 23%

International cohort, 429 cases + 429 controls, 30 sites/10 countries

Acute kidney injury in patients treated with immune checkpoint inhibitors (ICPi-AKI Consortium)

Gupta S, et al. · J Immunother Cancer 2021 · PMID 34625513

Acute tubulointerstitial nephritis was the most common biopsy lesion; corticosteroids — especially started early — were associated with higher odds of renal recovery.

Onset median 16 wk; ATIN 125/151 (82.7%); recovery 64.3%; steroids ≤14d OR 2.64, ≤3d OR 2.09; rechallenge recurrence 16.5%

Single-center cohort, 309 patients

Acute kidney injury associated with immune checkpoint inhibitor therapy: incidence, risk factors and outcomes

Meraz-Muñoz A, et al. · J Immunother Cancer 2020 · PMID 32601079

AKI occurred in 16.5% and was independently associated with other immune-related adverse events and hypertension, but not with worse survival.

AKI 16.5% (KDIGO 1/2/3 = 53/22/25%); other-irAE OR 2.82; not associated with mortality

Systematic review + meta-analysis, 14 studies, 12,694 patients

Association of proton pump inhibitor use and immune checkpoint inhibitor-mediated AKI: a meta-analysis

Mohan A, et al. · Am J Nephrol 2024 · PMID 38471492

PPI use was associated with an ~84% higher likelihood of checkpoint-inhibitor AKI; of those with ICI-AKI, about two-thirds recovered while roughly a third progressed to CKD.

PPI → ICI-AKI pooled OR 1.84 (1.16–2.90); recovery 67%, CKD 32%, death ~36%

Systematic review of all biopsy-proven published cases — 27 articles, 45 patients

A systematic review of immune checkpoint inhibitor-associated glomerular disease

Kitchlu A, et al. · Kidney Int Rep 2020 · PMID 33426386

A glomerular lesion is the minority phenotype but a distinct one, and it behaves worse than AIN: pauci-immune GN/renal vasculitis was commonest, ANCA serology was usually NEGATIVE, and concomitant AIN in 41% means the two compartments are often injured together. The authors argue for baseline urinalysis and quantitative proteinuria before starting an ICI, and for biopsy when features are atypical for AIN.

Pauci-immune GN/vasculitis 27%, podocytopathies 24% (MCD 20%), C3GN 11%; PD-1/PD-L1 implicated in 81%; median onset 93 days (IQR 44–212); baseline creatinine 0.95 mg/dL (0.78–1.2) rising to a peak of 3.4 (2.3–3.5) with proteinuria 8.3 g/day (4.3–13.0); RRT 25%; ANCA positive in only 2 of 12 pauci-immune cases; recovery complete 31% / partial 42%, ESKD 19%, approximately one third died

Pooled review of 53 published patients with ICI-associated glomerular disease

Glomerular diseases after immune checkpoint inhibitors use: what do we know so far?

He X, et al. · Ren Fail 2022 · PMID 36420664

An independent series reproducing the same lesion ranking two years later, and adding that rituximab is particularly useful for the vasculitic pattern.

Pauci-immune GN 28.3%, podocytopathies 26.4%, immune-complex GN 18.9%; anti-PD-1 in 71.4%; ICI discontinued in 89.4% and steroids given in 87.2%

All landmark evidence in this field

What the data says now

How disproportionately each agent's FAERS reports name these phenotypes vs. all other drugs (reporting odds ratio; significant signals only, 95% CI lower bound > 1; as of 2026-09-01). A reporting signal, not incidence or proven causation. A dash means tested without reaching significance, not a phenotype that never occurs — ATN and AIN undercount badly, most true cases filing as generic “acute kidney injury”. Computed by this atlas on the current snapshot — a published disproportionality analysis will not match cell for cell (different window, different term set).

Per-agent FAERS reporting odds ratio for each injury phenotype in this syndrome.
Agent	AIN	ATN
Nivolumab	Nivolumab, Acute Interstitial Nephritis: ROR 6.11, 513 reports	Nivolumab, Acute Tubular Necrosis: ROR 1.66, 75 reports
Pembrolizumab	Pembrolizumab, Acute Interstitial Nephritis: ROR 11.64, 1,024 reports	Pembrolizumab, Acute Tubular Necrosis: ROR 6.92, 329 reports
Atezolizumab	Atezolizumab, Acute Interstitial Nephritis: ROR 4.92, 168 reports	Atezolizumab, Acute Tubular Necrosis: no significant signal
Ipilimumab	Ipilimumab, Acute Interstitial Nephritis: ROR 5.74, 217 reports	Ipilimumab, Acute Tubular Necrosis: ROR 2.09, 42 reports
Durvalumab	Durvalumab, Acute Interstitial Nephritis: ROR 3.73, 67 reports	Durvalumab, Acute Tubular Necrosis: no significant signal
Cemiplimab	Cemiplimab, Acute Interstitial Nephritis: ROR 3.91, 7 reports	Cemiplimab, Acute Tubular Necrosis: no significant signal

16.5%

AKI in checkpoint-inhibitor patients

309-patient single-center cohort (KDIGO 1/2/3 = 53/22/25%)

PMID 32601079

82.7–93%

Tubulointerstitial nephritis on biopsy

Biopsied patients in the multicenter/consortium series

PMID 34625513

~14–16 weeks

Median onset from initiation

Multicenter checkpoint-inhibitor-AKI cohorts

PMID 31896554

OR 1.84

PPI-associated risk

Meta-analysis, 14 studies, 12,694 patients

PMID 38471492

Pauci-immune GN/vasculitis 27–28%, podocytopathies 24–26%

Glomerular (rather than tubulointerstitial) lesion

Two pooled reviews of biopsy-proven ICI glomerular disease — 45 patients (Kitchlu) and 53 (He). These are shares OF the glomerular cases, not of all ICI-AKI, which stays AIN-dominant

PMID 33426386

93 days; peak creatinine 3.4 mg/dL, proteinuria 8.3 g/day

Glomerular disease onset and severity

45 biopsy-proven cases, from a median baseline creatinine of 0.95 mg/dL — normal renal function beforehand did not predict a mild course

PMID 33426386

41%

Concomitant AIN in glomerular cases

45 biopsy-proven glomerular cases — the two compartments are frequently injured together

PMID 33426386

19% (vs 9% in the largest AIN series)

ESKD after glomerular disease

45 biopsy-proven glomerular cases; approximately one third of patients died

PMID 33426386

64.3%

Renal recovery (whole cohort)

429-patient international ICPi-AKI consortium; the paper reports no separate steroid-treated recovery rate, only that steroids within 14 days carried higher odds of recovery (adjusted OR 2.64, 95% CI 1.58-4.41)

PMID 34625513

How it's managed

1
Confirm the lesion — consider kidney biopsy
Because a rising creatinine on immunotherapy has many causes, biopsy (favored for stage ≥2 AKI) confirms tubulointerstitial nephritis and excludes prerenal, obstructive, and other lesions before committing to immunosuppression; where biopsy is not feasible, prompt empiric steroids for clinically suspected disease are recommended.
Society position statement (ASON) + cohort evidence that biopsy characterizes the lesion · PMID 39455026 (opens PubMed in a new tab)
2
Hold the drug and start corticosteroids early
Withhold the checkpoint inhibitor and begin corticosteroids (about 1 mg/kg/day of prednisone-equivalent) promptly; earlier initiation is associated with better renal recovery, tapering as kidney function improves.
Consortium cohort — steroids ≤14 days OR 2.64, ≤3 days OR 2.09 for recovery · PMID 34625513 (opens PubMed in a new tab)
3
Do not assume the lesion is interstitial
A minority of ICI kidney injury is glomerular, and it presents differently: oedema, new hypertension, foamy urine or heavy proteinuria rather than a bland sediment, often from a normal baseline creatinine and later than AIN. It carries worse kidney outcomes than AIN, so the pooled reviews argue for a baseline urinalysis and quantitative proteinuria before starting an ICI, and for biopsy whenever the picture is atypical for AIN — heavy proteinuria, an active sediment, or no response to standard steroid dosing. ANCA is usually negative even in the vasculitic pattern, so a negative serology does not exclude it.
Pauci-immune GN/vasculitis 27–28% and podocytopathies 24–26% of biopsy-proven glomerular cases; ESKD 19% against 9% in the largest AIN series; ANCA positive in only 2 of 12 pauci-immune cases · PMID 33426386 (opens PubMed in a new tab)
4
Stop the modifiable co-triggers
De-prescribe concurrent nephritis-provoking drugs — proton-pump inhibitors above all, plus NSAIDs — since PPI use independently raises the risk and many patients are on a potential tubulointerstitial-nephritis-causing medication at diagnosis.
PPI meta-analysis (OR 1.84); majority on a concurrent TIN-causing drug in the landmark series · PMID 38471492 (opens PubMed in a new tab)
5
Individualize rechallenge
Rechallenge after recovery is possible and does not clearly worsen survival, but recurrent nephritis is a real risk — weigh the oncologic need against it and monitor renal function closely if resumed. The 2026 meta-analysis pools recurrence below the founding cohorts, and reads the fall as modern practice rather than biology; its estimate applies to patients who recovered and stayed treatment-eligible.
Rechallenge recurrence 16.5–23% in the multicenter cohorts; pooled 14.07% (95% CI 10.26–17.89) across the nine studies reporting rechallenge. An age split (10.6% under 65 v 19.1% over) did not reach significance, P-interaction 0.12 · PMID 41816339 (opens PubMed in a new tab)

What the guidelines say

Society and consensus recommendations addressing this syndrome.

Each recommendation below is this atlas's faithful summary of the source, not a quotation from it — follow the PubMed link for the wording the society published. Summaries may be superseded; consult the current full text and individualize to the patient.

ASONDiagnosis and management of immune checkpoint inhibitor-associated nephrotoxicity: a position statement from the American Society of Onco-nephrologyKidney Int 2025 · PMID 39455026ICI-AKI most commonly presents as acute interstitial nephritis; nephrology consultation and kidney biopsy should be considered for stage 2 or higher AKI or suspected glomerular disease, but where biopsy is not feasible, prompt empiric corticosteroids (prednisone ~1 mg/kg/day) should be started for clinically suspected ICI-AKI since early steroids improve renal recovery, with cautious individualized consideration of ICI rechallenge after recovery.ASCOManagement of Immune-Related Adverse Events in Patients Treated With Immune Checkpoint Inhibitor Therapy: ASCO Guideline UpdateJ Clin Oncol 2021 · PMID 34724392For grade 2 or higher ICI-related nephritis/AKI, hold the ICI, exclude alternative causes, and initiate corticosteroids (prednisone 0.5-1 mg/kg/day for grade 2, 1-2 mg/kg/day for grade 3-4) tapered over 4-6 weeks once creatinine improves; permanently discontinue for grade 4 toxicity.ESMOManagement of toxicities from immunotherapy: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-upAnn Oncol 2022 · PMID 36270461For ICI-related AKI, exclude alternative etiologies and stop concomitant nephrotoxins (PPIs, NSAIDs); ESMO permits continuing the ICI for stage 1 AKI with monitoring, but recommends withholding the ICI and starting corticosteroids for stage 2 or higher nephritis, escalating immunosuppression for steroid-refractory disease.SITCSociety for Immunotherapy of Cancer (SITC) clinical practice guideline on immune checkpoint inhibitor-related adverse eventsJ Immunother Cancer 2021 · PMID 34172516Grade ICI-related AKI by CTCAE; for persistent grade 2 or higher renal toxicity, discontinue the ICI, exclude other causes, and treat with corticosteroids with a taper begun once creatinine improves toward grade 1, considering kidney biopsy and additional immunosuppression for refractory cases.

Every citation on this page is a real, PubMed-verified reference. The teaching case is an illustrative composite, not a real patient. Medical-education content — not medical advice.