Ethambutol

Limits on what this shows

• Null findings: Pommer: no mutation synergism and toxicity only above plasma-relevant concentrations [16]. Hwang: negative LHON genotyping [17]. Yoon: glutamate antagonists did not rescue [12], a conflict with the excitotoxic framing in Heng's title.
• At-risk subgroups: A single patient with an OPA1 mutation in Guillet. Other case reports (a sibling pair, an LHON-carrier case) appeared in listings and were not read.

Direction

Toxicity mechanism candidate (mitochondrial interference). Not a nutrient entry and not a deficiency.

Regulatory status

The labels attribute decreased acuity to optic neuritis and do not name a mitochondrial mechanism [1][2].

Population evidence

Guillet 2010 [14]: fibroblasts from controls and from a man with an OPA1 mutation who developed autosomal dominant optic atrophy after ethambutol; ethambutol produced a mitochondrial coupling defect with a 25 percent reduction in complex IV activity, vacuole formation, lower mitochondrial membrane potential and a more fragmented mitochondrial network. Heng 1999 [15]: rodent retinal cells and adult rats dosed orally for 3 months; cytosolic calcium decreased and mitochondrial calcium and mitochondrial membrane potential increased (no numbers in the retrieved summary). Pommer 2008 [16]: NT2/D1 LHON cybrids versus parental cells; no synergism between the G11778A mutation and ethambutol on viability or mtDNA copy number; short-term toxicity only at micromolar concentrations well above estimated patient plasma peaks. Hwang 2003 [17]: no pathogenic LHON mtDNA mutation in ethambutol optic neuropathy patients tested (0 of 24 for nt 11778, 0 of 15 for nt 14484, 0 of 8 for nt 3460).

Mechanism

Named but contested: complex IV and coupling defect (Guillet); mitochondrial calcium loading (Heng); lysosomal zinc (Chung, see Zinc entry). Mitochondrial membrane potential fell in Guillet and rose in Heng, so the direction is not settled.

Cohort test

Ethambutol-specific Tier B throughout; no human in vivo mitochondrial measurement. The human genetics (Hwang) used ethambutol patients directly. No class extrapolation.

Counter-evidence held on record

Counter-evidence Cell and animal data only; the in vitro concentration problem (Pommer) applies to others as well. Overlaps with ATP production and Copper (see Overlap map). Framing caution Optic neuropathy is a toxicity outcome. This entry does not imply a mitochondrial nutrient deficit or any supplement.

Limits on what this shows

• Null findings: Yoon 2000 [12]: the antioxidant trolox did not attenuate ethambutol vacuolation or neuronal loss in culture.

Direction

Possible pro-oxidant effect in rat retina; not an antioxidant-depletion claim for humans.

Regulatory status

No oxidative-stress statement in the labels [1][2].

Population evidence

Cinici 2016 [18]: Wistar rats, 10 per group, ethambutol 30 mg/kg orally daily for 90 days; retinal malondialdehyde and 8-hydroxyguanine were higher and total glutathione lower than in healthy controls; thiamine pyrophosphate pretreatment brought values closer to normal (no numbers in the abstract). Excluded: Uzar 2014 (rat brain and sciatic nerve oxidative markers with ethambutol) was retracted in 2026 [19] and is not used. Human context: Taha 2010 [9], 40 TB patients on a four-drug regimen including ethambutol 275 mg versus 50 healthy controls (full text opened): mean (SD) malondialdehyde 1.24 (0.17) umol/l in controls, 3.18 (0.58) before and 2.82 (0.60) after 2 months (p below 0.001 versus both); total antioxidant status 1.86 (0.30), 1.06 (0.12) and 1.27 (0.11) mmol/l; CRP fell, with 30 percent of patients under 6 mg/l afterward versus 0 percent before. The authors concluded active TB is associated with oxidative stress that treatment improved.

Mechanism

Redox injury is proposed in the rat study; no named enzymatic mechanism.

Cohort test

Rat row: ethambutol-specific Tier B. Human row: COMBINATION REGIMEN (ethambutol in the cohort, not separable) and the change reflects disease resolution. No class extrapolation.

Counter-evidence held on record

Counter-evidence Single abstract-only rat study; the earlier rat paper is retracted; human data run the opposite way but are confounded by TB itself. Framing caution Not a systemic antioxidant-depletion claim and not a reason to take antioxidants.

Limits on what this shows

• Null findings: Kozak 1998 [6]: 6 M. tuberculosis and 5 M. avium strains; 'Copper did not effect EMB's antimicrobial actions against either species.' This tests only whether copper blunts ethambutol's antimicrobial activity (nutrient to drug direction) and says nothing about whether copper prevents neuropathy. Solecki: overall trace-element balance unaffected [10]. Moraes: no significant change in the copper/zinc ratio over 60 days [8].
• At-risk subgroups: Not defined. No study compared patients who developed optic neuropathy with those who did not for copper status.

Direction

Possible depletion (circulating level), unproven. Separate from the drug's optic neuropathy, which is a toxicity outcome and not a nutrient deficiency (Hard Rule 2).

Regulatory status

Copper is not mentioned in either label [1][2].

Population evidence

Abbasi Nazari 2009 [7]: 60 adults with pulmonary TB, single group, serum copper by atomic absorption at baseline and day 10. Mean (SD) 0.94 (0.24) ug/mL at baseline and 0.64 (0.24) ug/mL on day 10, p below 0.0001; the change did not correlate with age, sex or smoking. Authors attribute it to ethambutol's chelating effect. Regimen, doses, reference range and test not in the abstract. Moraes 2014 [8]: 75 male pulmonary TB patients at baseline, 30 and 60 days versus healthy subjects; at baseline TB patients had higher copper and copper/zinc ratio than healthy subjects; iron, zinc and copper/zinc did not change significantly during treatment; lower copper was associated with early sputum conversion (30 days, not 60 days), and CRP and CRP/albumin were lower in converters. Regimen not in the abstract (substance index lists isoniazid, rifampin, pyrazinamide, ethambutol). Solecki 1984 [10]: ethambutol 400 mg/kg/day for rats (roughly 20 to 27 times the 15 to 20 mg/kg/day human dose cited in [5], not species-adjusted) lowered copper in heart and liver versus pair-fed controls, so not explained by reduced food intake; overall trace-element balance was not affected.

Mechanism

Present in vitro, unproven in vivo. Cole 1981 [11]: ethambutol itself complexes Cu(II) and Zn(II) only to a limited extent in vivo, while its oxidation product EDBA competes effectively. The sources do not state that EDBA is the label's dicarboxylic acid metabolite; that identity is an inference. Kozak 1998 [6] frames chelation of copper impairing cytochrome c oxidase as one proposed mechanism of optic neuropathy; Guillet 2010 [14] found a 25 percent reduction in complex IV activity in fibroblasts but did not measure copper.

Competing explanation

Copper is raised in active TB (Moraes baseline) and tracks early sputum conversion and CRP [8]; CRP falls on four-drug therapy [9]. A day-10 fall from an inflammation-raised baseline is therefore expected without any drug chelation (Hard Rule 6).

Cohort test

COMBINATION REGIMEN. Abbasi names ethambutol in its title but the abstract does not state the regimen; Moraes lists it only in the substance index. Ethambutol is in the cohort but not separable from isoniazid, rifampin, pyrazinamide or the disease. No class extrapolation. Capped at yellow.

Counter-evidence held on record

Counter-evidence No ethambutol-free comparator arm anywhere; the only human copper study is uncontrolled, short (10 days) and abstract-only. A second human series (Elo 1970 [32]) exists by title but its text could not be opened; the OpenAlex record is closed access with no abstract. Repair pass: the Springer page for Abbasi [7] shows the abstract only (regimen, doses, reference range and any comparator sit behind the paywall) and its visible reference list contains no ethambutol-free comparator study; the Springer page for Moraes [8] likewise shows no regimen text (it cites WHO treatment guidelines). Neither regimen could therefore be confirmed, and no human copper study in an ethambutol-free regimen was found. Framing caution This is not a copper deficiency claim and not a reason to take copper. Optic neuropathy on ethambutol is managed by visual monitoring and stopping the drug [1][2]. The Kozak result addresses antimicrobial interference only.

Limits on what this shows

• Null findings: Moraes: no significant iron change over 60 days [8]; the two human cohorts disagree on the iron time course.

Direction

Neutral for ethambutol; disease-linked hypoferremia as context.

Regulatory status

Iron is not mentioned in the labels [1][2].

Population evidence

Taha 2010 [9] (full text on PubMed Central opened): 40 TB patients versus 50 controls on a fixed-dose tablet of isoniazid 75 mg, rifampicin 150 mg, pyrazinamide 400 mg and ethambutol 275 mg for 2 months. Mean (SD) serum iron 99.12 (27.5) ug/dl in controls, 55.73 (16.78) before treatment and 76.95 (28.29) after (p below 0.001 versus both); transferrin saturation 25.58 (8.4), 15.89 (7.0) and 21.37 (8.7) percent; ferritin 183.1 (65.6), 441.76 (319.4) and 185.64 (129.1) ng/ml, back to the control range; TIBC unchanged. The authors state anemia of inflammation was corrected, but the paper reports no hemoglobin or hematocrit. Moraes 2014 [8]: TB patients had lower iron than healthy subjects at baseline and no significant change over 60 days; hemoglobin and albumin rose. Solecki 1984 [10]: iron concentrations rose in kidney, liver, heart and spleen of both ethambutol and pair-fed rats, tied to reduced food intake, with total organ iron comparable to ad-libitum controls.

Mechanism

None named for ethambutol.

Cohort test

COMBINATION REGIMEN; ethambutol not isolated. Confounding by indication is presumed (Hard Rule 6). No class extrapolation.

Counter-evidence held on record

Counter-evidence The signal belongs to the disease, not to ethambutol. Iron overlaps with Oxygen delivery through hemoglobin findings (Overlap map). Framing caution Anemia is not iron deficiency (Hard Rule 2). No iron supplementation is recommended by this entry.

Limits on what this shows

• Null findings: Latent TB itself had no detectable microbiome effect [24]. No ethambutol-alone arm was found in the PubMed query 'ethambutol gut microbiota' (8 records, none testing ethambutol alone).

Direction

Possible dysbiosis of the combination regimen. Not attributable to ethambutol.

Regulatory status

No microbiome statement in the labels [1][2].

Population evidence

Wipperman 2017 [24]: Haiti, 19 patients on HRZE (mean 3.4 months), 19 cured patients (mean 417 days after therapy), 50 IGRA-negative and 25 latent TB controls. Shannon diversity did not differ, observed OTUs were lower during treatment (p=0.0074 versus IGRA-negative), Ruminococcus was down 675-fold, Lactobacillus and Coprococcus more than 200-fold, Bifidobacterium about 20-fold versus latent TB controls; the perturbation persisted at least 1.2 years after cure. Meng 2022 and Zhu 2025 [26]: further HRZE patient cohorts (listing-level only; Veillonella and Bacteroides increased in Meng).

Mechanism

Not named for ethambutol. Namasivayam 2017 [25]: in infected mice, isoniazid, rifampin and pyrazinamide were studied alone and in combination; rifampin was the main driver; ethambutol was not given and the authors say it should be studied in future.

Cohort test

COMBINATION REGIMEN: ethambutol is in the Wipperman and Meng cohorts but not separable. Namasivayam FAILS the cohort test (ethambutol absent) and is not borrowed. No class extrapolation. Capped at yellow.

Counter-evidence held on record

Counter-evidence Cross-sectional design, small groups, no baseline samples; rifampin dominance in the mouse dissection. Framing caution Do not state that ethambutol causes dysbiosis. This is a regimen-level observation.

Limits on what this shows

• Null findings: Pommer: no ethambutol-specific ATP/ADP change in the abstract [16].

Direction

Possible reduced efficiency of ATP generation in vitro; unproven for cellular ATP output.

Regulatory status

No energy-metabolism statement in the labels [1][2].

Population evidence

Guillet 2010 [14]: mitochondrial coupling defect and 25 percent lower complex IV activity; the abstract summary gives no ATP value. Pommer 2008 [16]: ATP/ADP ratios measured by luminometry; cybrid ratios were unexpectedly low but did not correlate with mutation status in drug-treated cells, and no distinct ethambutol-induced change was reported in the abstract; the authors call for neuronally differentiated cybrids that depend fully on oxidative phosphorylation. Heng 1999 [15]: mitochondrial membrane potential increased.

Mechanism

Named only through the Guillet coupling defect. Whether ATP output falls in retinal ganglion cells in vivo is not shown.

Cohort test

Ethambutol-specific Tier B in vitro; no human ATP measurement. Sources are shared with Mitochondrial burden (Overlap map). No class extrapolation.

Counter-evidence held on record

Counter-evidence Pommer's ATP/ADP finding does not isolate ethambutol. No muscle or blood ATP measure in ethambutol patients was found. Framing caution Not a 'low energy' claim. Fatigue in TB patients is explained by the disease and the other regimen drugs and is not attributed to ethambutol here.

Limits on what this shows

• Null findings: The 500-patient cohort and Moraes show hematologic indices improving on regimens that contain ethambutol [38][8]. Anemia in 32 percent of 380 TB patients was attributed to disease severity [28]. Antoniskis 1992 attributes a larger anemia frequency to zidovudine plus the TB drugs with ethambutol not separable [27].
• At-risk subgroups: None defined; one patient. A 2016 case with a fatal outcome makes any new unexplained anemia on therapy a reason to contact the prescriber, but gives no frequency.

Direction

Drug toxicity (immune-mediated red-cell destruction), rare and idiosyncratic. Under Hard Rule 2 this is a clinical outcome and an adverse reaction, not an oxygen-delivery deficiency of the nutrient kind, and no dietary or supplement action follows. Kept in the Oxygen delivery category because the category covers red-cell mass.

Regulatory status

The labels list thrombocytopenia, leukopenia and neutropenia as hematologic reactions; no hemolytic anemia statement was found in the sections retrieved [1].

Population evidence

Nicolini 2016 [37]: case report, International Journal of Tuberculosis and Lung Disease, 7 authors; the abstract (read as an OpenAlex inverted-index record through a summarizing fetch tool) states that this is the first reported case of acute fatal autoimmune hemolytic anemia due to ethambutol, notes that isoniazid and rifampicin (especially intermittent rifampicin) have very rarely been linked to it, and gives no hemoglobin values, antibody results, co-medications or timing. Cohort context: a 500-patient Nigerian cohort on a two-month ethambutol, isoniazid, rifampicin and pyrazinamide phase followed by isoniazid and thiacetazone [38]: packed cell volume and white cell count rose and ESR fell over treatment, 95 percent of blood films were normochromic, and the authors concluded that mild anemia of chronic disease improves without iron or folic acid. Moraes 2014 [8] likewise reports hemoglobin rising over 60 days. Taha 2010 [9] discusses anemia of inflammation but, in the full text opened, reports no hemoglobin or hematocrit values.

Mechanism

Immune-mediated (autoimmune) hemolysis as stated by the case authors; the antibody work-up and drug-dependent antibody evidence were not in the abstract. No oxygen-transport or nutrient mechanism is proposed.

Cohort test

Case report: ethambutol is the attributed drug per its authors, but the abstract does not state whether other antituberculosis drugs were co-administered, so attribution rests on the authors. Tier B. The Nigerian cohort and Moraes are COMBINATION REGIMEN cohorts (ethambutol not separable) and are used only as context. Listing-level only, not opened: Trop Doct 1997 on autoimmune hemolytic anemia improving on an antituberculous regimen with ethambutol, and a 1976 Romanian report linking hemolytic anemia to intermittent rifampicin with ethambutol co-administered. No class extrapolation.

Counter-evidence held on record

Counter-evidence One case, abstract only, no causality assessment retrieved, and the label is silent. Population regimen data run the other way. Overlaps with Iron through hemoglobin findings (Overlap map). Framing caution Not an iron, B12 or other nutrient deficiency and not a "low oxygen carrying capacity" claim. Anemia is an outcome (Hard Rule 2). No supplement is recommended; unexplained fatigue, jaundice or dark urine on therapy should be reported to the prescriber.