Skip to content
The Injury Atlas
TMA

Thrombotic Microangiopathy

Endothelial injury with microvascular thrombi, hemolysis and thrombocytopenia — gemcitabine, mitomycin C, anti-VEGF.

13signature agents

Where it strikes

Vasculature / Endothelium

Glomerular & peritubular capillaries

Glomerulus

Filtration barrier (podocytes + endothelium)

See it on the nephron

Agents’ overall severity

Each agent’s whole-drug severity grade, not the severity of this lesion specifically — an agent whose signature injury is elsewhere can still be graded severe here.

Severe· 5Moderate· 6Mild· 2

Agents’ overall reversibility

Often irreversible· 1Partially reversible· 2Variable· 8Reversible· 2
Does the kidney recover? Cross-drug outcomes

Agents’ onset window

How soon each agent’s kidney toxicity typically appears — a whole-drug tempo, not specific to this lesion.

Acute (days)· 1Subacute (weeks)· 1Delayed (weeks–months)· 5Variable· 6

Real-world reporting for this lesion

FAERS across all lesions →

How disproportionately thrombotic microangiopathy is reported for each agent in FAERS, against every other drug (reporting odds ratio, 95% CI; a signal needs the lower bound above 1, on at least 50 total reports for the agent) — a spontaneous-reporting signal, not incidence or proven causation. FAERS carries reporting and indication biases and has no denominator.

26Corroborated

Documented in the atlas profile and carrying a FAERS signal — the strongest claim the atlas makes.

16Documented, FAERS-silent

Documented in a profile with no reporting signal. 13 of 16 sit on a reportable term with enough reports, so FAERS could have flagged the lesion and did not; the rest are expected silences.

3Not attributable

A real reporting signal that is not evidence for this drug-lesion pair: no MedDRA term names it, the naming subset asked alone came back flat, or the 2026-08 review attributed the reporting to the population, co-therapy, or class-level literature.

51Reported by name

The terms that name this lesion are disproportionate, but no profile documents it for that agent. Most are echoed by a sibling agent in the same class. Leads for review, never lesions the atlas claims.

Order
  1. Moxetumomab pasudotoxImmunotoxin (anti-CD22 PE38)341.06 (190.22–611.55) · n=14 Reporting signal.
  2. PentostatinPurine analog (ADA inhibitor) · FAERS only61.13 (40.26–92.84) · n=23 Reporting signal.
  3. DinutuximabAnti-GD2 antibody40.50 (27.78–59.04) · n=28 Reporting signal.
  4. SirolimusmTOR inhibitor · FAERS only39.43 (35.74–43.50) · n=423 Reporting signal.
  5. ThiotepaAlkylator · FAERS only36.92 (32.79–41.58) · n=286 Reporting signal.
  6. BusulfanAlkylator32.27 (29.20–35.66) · n=406 Reporting signal.
  7. GemcitabineNucleoside analog32.13 (30.32–34.04) · n=1,290 Reporting signal.
  8. AmsacrineTopoisomerase II inhibitor (acridine) · FAERS only26.75 (14.31–50.02) · n=10 Reporting signal.
  9. CarfilzomibProteasome inhibitor26.14 (24.04–28.42) · n=582 Reporting signal.
  10. FludarabinePurine analog · FAERS only25.91 (24.11–27.84) · n=798 Reporting signal.
  11. MelphalanAlkylator · FAERS only25.67 (23.48–28.06) · n=512 Reporting signal.
  12. Denileukin diftitoxImmunotoxin (IL-2–diphtheria) · FAERS only23.92 (3.31–172.68) · n=1 Reporting signal, fragile (few reports).
  13. IsatuximabAnti-CD38 antibody · FAERS only19.27 (15.68–23.70) · n=92 Reporting signal.
  14. Mitomycin CAntitumor antibiotic19.05 (14.66–24.77) · n=57 Reporting signal.
  15. LifileucelTumor-infiltrating lymphocyte (TIL) therapy · FAERS only12.76 (4.09–39.82) · n=3 Reporting signal, fragile (few reports).
  16. Inotuzumab ozogamicinAntibody-drug conjugate (CD22/calicheamicin) · FAERS only12.58 (8.83–17.93) · n=31 Reporting signal.
  17. RamucirumabAnti-VEGFR2 antibody8.05 (5.94–10.90) · n=42 Reporting signal.
  18. EverolimusmTOR inhibitor7.78 (6.96–8.71) · n=314 Reporting signal.
  19. BevacizumabAnti-VEGF antibody6.10 (5.63–6.61) · n=638 Reporting signal.
  20. Carmustine (BCNU)Nitrosourea alkylator5.63 (3.70–8.56) · n=22 Reporting signal.
  21. OxaliplatinPlatinum agent4.75 (4.24–5.34) · n=295 Reporting signal.
  22. BortezomibProteasome inhibitor4.22 (3.77–4.72) · n=310 Reporting signal.
  23. DecitabineHypomethylating agent4.07 (2.56–6.46) · n=18 Reporting signal.
  24. 5-FluorouracilPyrimidine analog3.57 (3.13–4.07) · n=231 Reporting signal.
  25. RuxolitinibJAK1/2 inhibitor3.28 (2.85–3.78) · n=198 Reporting signal.
  26. DoxorubicinAnthracycline2.98 (2.62–3.38) · n=238 Reporting signal.
  27. PonatinibBCR-ABL TKI2.98 (2.00–4.45) · n=24 Reporting signal.
  28. SunitinibVEGFR TKI2.87 (2.34–3.52) · n=93 Reporting signal.
  29. DocetaxelTaxane2.68 (2.29–3.15) · n=151 Reporting signal.
  30. BendamustineAlkylator2.45 (1.85–3.25) · n=49 Reporting signal.
  31. CapecitabinePyrimidine analog (oral 5-FU)2.21 (1.90–2.57) · n=168 Reporting signal.
  32. FruquintinibVEGFR TKI2.14 (1.02–4.48) · n=7 Reporting signal, fragile (few reports).
  33. PembrolizumabPD-1 checkpoint inhibitor2.10 (1.80–2.46) · n=161 Reporting signal.
  34. SorafenibVEGFR TKI1.86 (1.31–2.63) · n=32 Reporting signal.
  35. ErlotinibEGFR TKI1.85 (1.29–2.65) · n=30 Reporting signal.
  36. IxazomibProteasome inhibitor1.46 (1.05–2.04) · n=35 Reporting signal.
  37. PazopanibVEGFR TKI1.38 (0.97–1.97) · n=31 No reporting signal.
  38. NivolumabPD-1 checkpoint inhibitor1.17 (0.96–1.43) · n=96 No reporting signal.
  39. IpilimumabCTLA-4 checkpoint inhibitor1.04 (0.76–1.43) · n=38 No reporting signal.
  40. LenalidomideImmunomodulatory drug (IMiD)1.02 (0.91–1.13) · n=360 No reporting signal.
  41. Trastuzumab emtansine (T-DM1)Antibody-drug conjugate (HER2/DM1)0.71 (0.29–1.71) · n=5 No reporting signal.
  42. NintedanibVEGFR/FGFR/PDGFR TKI0.66 (0.41–1.04) · n=18 No reporting signal.
  43. Ziv-afliberceptVEGF trap0.59 (0.36–0.96) · n=16 No reporting signal.
  44. OlaparibPARP inhibitor0.39 (0.19–0.82) · n=7 No reporting signal.
  45. TislelizumabPD-1 immune checkpoint inhibitor0.32 (0.08–1.27) · n=2 No reporting signal.
  46. Lutetium-177 DotatateRadioligand therapy (PRRT)0.21 (0.03–1.46) · n=1 No reporting signal.
  47. AxitinibVEGFR TKI0.06 (0.01–0.43) · n=1 No reporting signal.

reporting odds ratio (log) — no signal left of the dashed line at 1.0

No ratio — no TMA reports, or too few (count in brackets): Doxifluridine, Tegafur-uracil (UFT)

No ratio — under 50 FAERS reports in total: Carmofur (HCFU), Daraxonrasib, Ivonescimab

Not matched in FAERS (the name cannot be told apart from other products): Interferon-α

All 43 agents documented with TMA in the atlas are shown whether or not FAERS flags them, plus the 54 agents FAERS flags that the atlas does not document (marked “FAERS only”; the strongest 10 until you show all). Class composites (mTOR inhibitors (everolimus · temsirolimus); VEGFR TKIs (sunitinib · sorafenib · pazopanib · axitinib)) are not asked of FAERS; their member agents are ranked individually. 80 carry a reporting signal (filled dot: 95% CI lower bound above 1, on at least 50 reports for the agent); a signal on fewer than 10 reports of the disorder is hollow, on a dashed whisker. Grey marks are ratios that do not clear it. Signals are listed first, each group by ROR. n = reports of the disorder for the agent. openFDA FAERS, as of 2026-11-15. A reporting signal, not incidence or causation, and no signal is not evidence of safety. Layout after the single-event rankings of Klomjit, Gupta et al.

§Mechanism, step by step

How anti-cancer drugs cause thrombotic microangiopathy

One lesion, two mechanistic routes — direct endothelial toxicity and VEGF withdrawal — converging on the glomerular capillary. Scroll to follow the injury from the endothelial surface to the sheared red cells on the smear.

How anti-cancer drugs cause thrombotic microangiopathyA schematic glomerular tuft with a magnified cross-section of one capillary loop. Direct endothelial toxins and VEGF withdrawal injure the fenestrated endothelium, which swells and denudes; platelet–fibrin microthrombi narrow the lumen; red cells shear into schistocytes; and the glomerulus becomes ischemic.Direct toxicityVEGF withdrawalPodocyte foot processeslumen occludedGlomerular tuftEndotheliumLumenGlomerular ischemia → proteinuria, ↓GFR
Schematic · not to scale

TMA is a final common pathway — the workup turns on the cause. Narrow the differential, then estimate the pretest probability of TTP before reaching for plasma exchange. After hematopoietic cell transplant, see the dedicated TA-TMA page.

What's driving this TMA?

Thrombotic microangiopathy is a final common pathway — the treatment turns entirely on the cause. Select a driver.

ADAMTS13
Normal / near-normal
How to tell
A temporal link to a culprit nephrotoxin. VEGF inhibitors give a renal-limited, proteinuric picture; gemcitabine, mitomycin C, carfilzomib, and calcineurin inhibitors give more systemic disease.
The trigger
The offending drug — dose-dependent/cumulative (VEGF, mitomycin, gemcitabine) or immune (quinine). Gemcitabine's mechanism is unsettled; its best-documented risk is cumulative-dose, so it is grouped with the dose-dependent agents here.
Directed treatment
Stop the offending agent — the single most important step. Blood-pressure control; complement blockade only for severe, systemic, or refractory disease. Not plasma exchange as reflex.
VEGF-inhibitor TMA — the deep dive

Pretest probability of severe ADAMTS13 deficiency (TTP) in an adult with thrombotic microangiopathy. Answer each feature — a Yes scores one point. The score appears once all seven are answered.

PLASMIC score criteria
Platelet count < 30 × 10⁹/LThe profound thrombocytopenia typical of TTP.
Hemolysis presentReticulocytes > 2.5%, undetectable haptoglobin, or indirect bilirubin > 2.0 mg/dL.
No active cancerNo cancer treated within the past year (active malignancy points away from TTP).
No solid-organ or stem-cell transplantNo history of transplant (transplant points toward TA-TMA instead).
MCV < 90 fLMean corpuscular volume below 90 femtoliters.
INR < 1.5Below 1.5 counts, even if above the normal range: a near-normal INR argues against DIC as the cause.
Creatinine < 2.0 mg/dLBelow 2.0 mg/dL counts, even if above normal: marked renal failure points toward aHUS or another TMA.

Answer all seven criteria to see the score and risk band (0/7 answered).

Educational aid only — not medical advice. The PLASMIC score supplements, and does not replace, ADAMTS13 activity testing and clinical judgment.

Management approach

Full framework →

Stop the culprit drug — the single most important step — with supportive care; complement-pathway-directed therapy for severe, refractory, or complement-mediated cases.

Drug-level levers

  • Discontinue the offending agent promptly; re-exposure risk is high.
  • Distinguish dose-dependent/toxic TMA (e.g. mitomycin C, gemcitabine, VEGF inhibitors, carfilzomib) from complement-mediated TMA — the latter is the one most likely to respond to complement blockade.
  • Switch out of class for ongoing therapy — drug-induced TMA tends to recur on rechallenge.
  • Control hypertension aggressively.

Pharmacologic toolkit

  • C5 inhibition (eculizumab, ravulizumab) — First-line complement blockade for severe or drug-discontinuation-refractory TMA; eculizumab achieved renal recovery in ~80% of reported drug-induced cases (gemcitabine, carfilzomib, bevacizumab). Ravulizumab is the longer-acting C5 alternative. Cover/vaccinate against encapsulated organisms before dosing.
  • Narsoplimab (anti-MASP-2, lectin pathway) — Targets the lectin pathway driving transplant-associated (HSCT) TMA; its pivotal single-arm trial reported a 61% response. FDA-approved in late 2025 as Yartemlea (narsoplimab-wuug) — the first therapy indicated for HSCT-associated TA-TMA, in adults and children ≥2 years — though access/cost considerations apply and responses vary.
  • Emerging complement inhibitors — Iptacopan (factor B), pegcetacoplan (C3), crovalimab and danicopan are under investigation; second-line selection after a suboptimal C5-inhibitor response is increasingly guided by complement biomarkers (e.g. sC5b-9).
  • Supportive care — Transfusion, blood-pressure control, and dialysis as needed.
  • Plasma exchange — Generally ineffective for drug-induced TMA (unlike TTP) and not routinely recommended.

When to biopsy

Consider when the diagnosis is unclear; but the peripheral smear (schistocytes), LDH, haptoglobin, and platelet count usually establish TMA noninvasively.

Educational use only. Educational synthesis of the published literature — not a treatment protocol, dosing guide, or medical advice. Regimens and agents shown are illustrative of what the literature describes; verify against current guidelines (ASON / KDIGO / ASCO / NCCN) and individualize to the patient. Using this site creates no clinician–patient relationship.

What the guidelines say

All guidelines →

Society and consensus recommendations that speak to thrombotic microangiopathy.

Cited incidence across agents

Where the literature gives a representative thrombotic microangiopathy figure, the agents ranked highest first. Hover a dot for its cited note.

0%5%10%Mitomycin C: 10% — Mitomycin-associated (cancer-associated) HUS/TMA occurs in roughly 4-15% of treated patients; risk is cumulative-dose related, most cases after total dose >60 mg. (PMID 2497229)Mitomycin C10%Moxetumomab pasudotox: 7.5% — Hemolytic uremic syndrome 7.5% (treatment-related serious AE) in the pivotal relapsed/refractory HCL trial; boxed warning (PMID 30030507)Moxetumomab pasudotox7.5%Carfilzomib: 5% — TMA in 6 of 114 (5%) patients in a real-world relapsed/refractory myeloma cohort; renal complications not related to myeloma progression in 17%. (PMID 33149167)Carfilzomib5%Gemcitabine: 0.31% — Cumulative incidence 0.31% (8 of 2,586 patients) in the largest single-institution series, higher than the earlier 0.015% estimate; new or worsening hypertension preceded the diagnosis in 7 of 9 cases. (PMID 15197810)Gemcitabine0.31%
Representative per-agent thrombotic microangiopathy incidence where a published figure is citable — open an agent's name for its profile and cited source, or hover its dot for the source inline. Agents without a citable figure are omitted; a tier without a number is not the same as a low number.

Signature offenders

13

Agents for which thrombotic microangiopathy is the defining renal lesion.

GemcitabineDelayed — months of cumulative exposure.Estimates span three orders of magnitude by ascertainment. The only figure with a real denominator is the manufacturer safety-database review: 12 cases against 78,800 patient exposures, a crude 0.015% (range 0.008–0.078%). The product characteristics give 0.01%. Against that, a single center reported 2.7% — four of its five cases on nab-paclitaxel/gemcitabine, where a pharmacokinetic interaction is suspected. Historically high mortality: of 23 cases with reported outcome in the literature tally, 11 died within a few weeks, death directly attributed to HUS in two.SevereMitomycin CDelayed — after cumulative dosing, sometimes after therapy ends.Dose-dependent, generally 4–15%; nearly every case in the landmark 85-patient registry had received a cumulative total dose above 60 mg. >50% historical mortality.SevereCarfilzomibVariable — frequently early (first cycles), but TMA can also appear later, sometimes after a treatment break and re-escalation.Renal complications are common and a recognized class concern: in a 114-patient real-world cohort, ~17% had renal complications not related to myeloma progression (TMA in 6 patients, ~5%; albuminuria >1 g/day ~6%; otherwise-unexplained grade >=3 AKI in another 6 patients), occurring mostly early and unpredictably. On the prospective CARDAMON trial 8 patients experienced TMA (6 of 8 hypertensive at presentation, 7 of 8 with AKI); after a protocol amendment adding aggressive hypertension management, carfilzomib step-up dosing at the start of maintenance and dexamethasone premedication, the rate fell from 4.2 to 1.6 events per 1,000 patient-cycles with no further maintenance events. Pharmacovigilance (FAERS) shows carfilzomib carries by far the strongest TMA signal among proteasome inhibitors.SevereMoxetumomab pasudotoxCycle-related: CLS within the first days of a cycle; HUS often during/after cycles 2-3.Boxed warning for capillary-leak syndrome (CLS) and hemolytic-uremic syndrome (HUS)/TMA. In the pivotal phase 3 trial HUS occurred in ~7.5% and CLS in ~5%; reviews cite roughly 9% each. Fatal CLS has been reported (in a pediatric ALL case).SevereTegafur-uracil (UFT)Variable; TMA/HUS typically emerges after weeks to months of cumulative exposure.Direct renal injury from UFT is rare and largely class-level. Fluoropyrimidine-associated thrombotic microangiopathy / hemolytic-uremic syndrome is a rare, mostly case-report-level event, frequently in combination regimens (e.g., with mitomycin C). No reliable drug-specific incidence rate is established.SevereBusulfanWeeks after conditioning/transplant.Transplant-associated thrombotic microangiopathy (TA-TMA) with renal involvement complicates a subset of conditioning regimens; high-dose busulfan (e.g., 16 mg/kg) has been identified as an independent risk factor for post-transplant TMA, and high busulfan exposure also drives sinusoidal obstruction syndrome/VOD.ModerateTrastuzumab emtansine (T-DM1)Variable; reported after initiation, over weeks to months of therapy.Renal toxicity is rare and case-level. FDA adverse-event analyses flagged renal events with trastuzumab-based therapy, and biopsy-proven cases (collapsing focal segmental glomerulosclerosis; TMA-like microvascular injury) have been reported. Incidence is not quantified.ModerateBortezomibRare adverse renal events can occur at any point during therapy, from weeks to several months after initiation; the beneficial light-chain reduction is often rapid.Bortezomib is generally renal-friendly and often improves renal function in myeloma by rapidly reducing light-chain-driven cast nephropathy; it requires no renal dose adjustment. Thrombotic microangiopathy and glomerular microangiopathy are rare, case-level events, and pharmacovigilance shows a far weaker TMA signal than carfilzomib.ModerateIxazomibVariable; reported after weeks to months of therapy, including with cumulative exposure. Class median time to TMA onset ~8 days from a triggering cycle.Drug-induced thrombotic microangiopathy (TMA) is a rare, case-level event for ixazomib and is not reliably quantified. In a pharmacovigilance analysis of the FDA Adverse Event Reporting System (FAERS, 2004-2023), proteasome inhibitors as a class were significantly associated with TMA (225 cases in 213 exposed individuals; ROR 1.71), but carfilzomib dominated the signal (58.7% of exposed individuals, 125 of 213; ROR 17.97), with bortezomib and ixazomib contributing far fewer reports. The ixazomib signal is therefore extrapolated largely from the class.ModerateDoxifluridineVariable; class-level TMA typically after prolonged cumulative exposure.No drug-specific nephrotoxicity incidence is established. Renal risk is inferred at the fluoropyrimidine-class level (rare TMA/HUS); direct doxifluridine renal injury reports are sparse.ModerateCarmofur (HCFU)Variable; class-level TMA after cumulative exposure.No established drug-specific renal incidence. Renal risk is class-level (rare fluoropyrimidine TMA/HUS). Carmofur's characteristic serious toxicity is leukoencephalopathy, not nephrotoxicity.ModerateOxaliplatinAcute if immune-mediated (often on re-challenge).Lowest nephrotoxic potential of the platinums; AKI is rare and case-level, in some cases immune-mediated.Mild5-FluorouracilTMA often delayed (weeks to months); prerenal effects acute with GI toxicity.Intrinsic nephrotoxicity is low; the recognized vascular renal complication is thrombotic microangiopathy/hemolytic-uremic syndrome, classically with mitomycin C, with 5-FU as a frequent co-agent. In TTP/HUS series the most common antecedent chemotherapy is mitomycin C plus 5-FU.Mild