Hydrochlorothiazide

Limits on what this shows

• Key negative/null findings: No effect was assessed or claimed in cells not exposed to UVA; the entire finding is photoactivation-dependent. No animal or human in vivo mitochondrial measurement, by respirometry or biopsy, has been conducted on hydrochlorothiazide at any dose.
• At-risk subgroups: None defined systemically. By extension of the label's own photosensitivity warning, sun-exposed skin in chronically treated patients is the only tissue context in which this has been demonstrated.

Regulatory status

Mitochondrial function is not named anywhere in the hydrochlorothiazide label. The label's photosensitivity warning is the nearest labeled analog and does not reference mitochondria.

Population evidence

None in humans systemically. Tier B, HCTZ in the cohort: Karkoszka et al., Int J Mol Sci 2024;25(3):1432 (doi 10.3390/ijms25031432): human dermal fibroblasts and epidermal melanocytes exposed to hydrochlorothiazide 0.01–1.0 mM for 24 hours, then UVA irradiation (5 J/cm²) or a dark control, assessed 24 hours later. At the highest hydrochlorothiazide+UVA condition, mitochondrial transmembrane potential (JC-1 dye) fell in roughly 80% of fibroblasts and roughly 60% of melanocytes, versus roughly 35% and 5% respectively for furosemide+UVA at the same concentration.

Mechanism

Named and dose-responsive. Hydrochlorothiazide is photolabile; UVA converts it to unstable photoproducts that generate reactive oxygen species through a Type I, oxygen-dependent phototoxic pathway, collapsing the mitochondrial membrane potential that drives ATP synthase. No dark-condition (non-irradiated) mitochondrial effect is reported in this study.

Counter-evidence held on record

This is in vitro, supraphysiologic-concentration (up to 1.0 mM) cell-culture data in isolated skin cells under an experimental UVA dose; it has not been reproduced in vivo, and no study has looked for a corresponding effect in any other tissue. It is included specifically because it is real, quantified, dose-responsive, hydrochlorothiazide-specific evidence, not because it establishes a systemic mitochondrial burden. Do not state that hydrochlorothiazide damages mitochondria in general. The only demonstrated effect is in irradiated skin cells in a dish, which is mechanistically consistent with, and does not add beyond, the drug's already-labeled photosensitivity and skin-cancer risk. No mitochondrial-support supplement (coenzyme Q10, carnitine, ribose) claim follows from this entry, and none is made.

Limits on what this shows

• Key negative/null findings: The EMINENT study is a clean, drug-specific Tier A null on the one systemic oxidative marker measured in hydrochlorothiazide-treated humans, and is the main reason this entry does not advance past yellow despite the strong in vitro signal.
• At-risk subgroups: None defined systemically. The in vitro finding is bounded to UV-exposed skin, consistent with the drug's separately labeled photosensitivity warning.

Regulatory status

Oxidative stress is not named anywhere in the hydrochlorothiazide label.

Population evidence

Two sources, opposite bound. Tier A, human, systemic: Nakane et al. for the EMINENT Study Investigators, Intern Med 2015;54(6):573–577 (doi 10.2169/internalmedicine.54.2631): open-label, randomized, multicenter trial, 55 hypertensive patients with a history of chronic stroke already on an ACE inhibitor or ARB, randomized to add hydrochlorothiazide or a non-diuretic agent, followed six months; malondialdehyde-modified low-density lipoprotein (MDA-LDL), the sole oxidative marker measured, did not change significantly in either arm. Tier B, in vitro, photoactivated, hydrochlorothiazide in the cohort: Karkoszka et al. 2024 (doi 10.3390/ijms25031432, described in the mitochondrial burden entry above). Reactive oxygen species (H2DCFDA assay) rose roughly 150% in fibroblasts and roughly 100% in melanocytes at the highest hydrochlorothiazide+UVA condition, versus roughly 30% for furosemide in fibroblasts and only a minimal, high-concentration-only rise in melanocytes.

Mechanism

Present but tissue-bounded. UVA-driven photodegradation of hydrochlorothiazide to reactive intermediates, as described above. No route by which hydrochlorothiazide would generate reactive oxygen species systemically, in the absence of UV exposure, has been demonstrated.

Counter-evidence held on record

EMINENT's null result is the only human systemic oxidative marker measured on hydrochlorothiazide, and it found nothing. The in vitro reactive-oxygen-species finding is real but confined to a UVA-exposed dish assay at concentrations well above plasma levels achieved at therapeutic oral doses. Do not state that hydrochlorothiazide causes systemic oxidative stress; the human data available says otherwise. Do not convert the in vitro photoactivated reactive-oxygen-species finding into a general antioxidant-supplement recommendation; no selenium, vitamin C, vitamin E, or glutathione claim is supported by this entry.

Limits on what this shows

• Key negative/null findings: No monotherapy hydrochlorothiazide copper data exists to isolate the drug's own contribution from amlodipine or valsartan. The direction found, an increase, runs opposite to a depletion claim, which further weakens any case for treating this as a burden entry at all.
• At-risk subgroups: None defined. Only geriatric hypertensive patients on this specific triple combination have been studied.

Regulatory status

Copper is not named anywhere in the hydrochlorothiazide label.

Population evidence

Matusz et al., "Beneficial effects of triple antihypertensive therapy with amlodipine, valsartan and hydrochlorothiazide on copper status in a trial of geriatric patients in Southwestern Romania" (journal title not fully verifiable from the accessible excerpt; ISSN 1584-9244 print / 2558-815X online): 115 hypertensive patients aged 65–80 (40 men, 75 women), three-year retrospective/comparative survey (April 2013–May 2016) of fixed-dose triple combination amlodipine/valsartan/hydrochlorothiazide 10/320/25 mg. Serum copper, measured by atomic absorption spectrophotometry, was significantly higher at study end than at baseline (P<0.01); exact absolute values were not available in the accessible excerpt. Hydrochlorothiazide was in the cohort, but only ever as part of the triple combination; no monotherapy arm existed.

Mechanism

Not established for hydrochlorothiazide. Copper handling is predominantly hepatobiliary (ceruloplasmin-mediated), not renal, so hydrochlorothiazide's distal-tubule sodium-chloride mechanism has no obvious route to affect copper status in either direction. Unlike the zinc entry above, which rests on a real renal-wasting pathway, there is no comparable structural or physiological rationale for a copper effect.

Counter-evidence held on record

The single available Tier A source shows the opposite of a depletion effect, and it cannot be isolated to hydrochlorothiazide given the triple-drug regimen. No independent replication, monotherapy arm, or non-Romanian cohort was located. Do not state or imply that hydrochlorothiazide depletes copper. The only human data point runs the other way, and even that is too confounded by combination therapy to support any directional claim about hydrochlorothiazide specifically. No copper testing and no copper supplement is recommended by this entry.

Limits on what this shows

• Key negative/null findings: Not applicable; this is an absence of study, not a measured null.
• At-risk subgroups: None defined for iron specifically. The label's dose-related hematologic reaction subgroup (25 mg and above) applies to the anemia finding itself, graded above, not to any iron parameter.

Regulatory status

Iron is not named anywhere in the hydrochlorothiazide label. The label's hematologic adverse-reaction listing names aplastic anemia, agranulocytosis, leukopenia, hemolytic anemia, and thrombocytopenia, with no iron parameter mentioned.

Population evidence

None located. No study measuring serum iron, ferritin, transferrin saturation, or hepcidin in hydrochlorothiazide-treated subjects was found.

Mechanism

Absent. Hydrochlorothiazide has no chelation mechanism, no renal iron-wasting mechanism (iron is not handled by the distal nephron segments where hydrochlorothiazide acts), and no described gastrointestinal iron-absorption interference, consistent with the pharmacokinetic prescreen's finding of no gut chelation or mineral-transporter-competition mechanism for this drug. Where hydrochlorothiazide does affect red cells, it is through the idiosyncratic immune or marrow-toxicity mechanism described in the oxygen delivery and utilization entry above, a mechanism that does not operate through iron status at all.

Counter-evidence held on record

There is nothing to hold beyond the mechanism argument above; the entry exists specifically to pre-empt a misreading of the oxygen delivery and utilization finding, not to evaluate a genuine depletion candidate. A reader who sees "hydrochlorothiazide can cause anemia" should not reach for an iron supplement. This has never been shown to be iron-mediated on this drug, and a new or unexplained anemia needs clinical workup, not supplementation on assumption. No iron dose is recommended here, and none should be.

Limits on what this shows

• Key negative/null findings: The measured downstream outcome, impaired glucose tolerance and insulin resistance, is a metabolic/clinical outcome, not itself a nutrient-deficiency finding, and must not be filed as one. No study, in this paper or elsewhere, measured a nutrient-synthesis output of the shifted microbiota, such as short-chain fatty acid production or B-vitamin or vitamin K synthesis, so the nutrient-status implication of the dysbiosis remains unmeasured, not established.
• At-risk subgroups: In the mouse model, the metabolic phenotype was reproduced on both normal and high-fat diet and appeared dose-related (5 versus 10 mg/kg). No human subgroup has been identified, because no human study exists.

Regulatory status

Gut microbiota is not named anywhere in the hydrochlorothiazide label.

Population evidence

None in humans. Tier B, hydrochlorothiazide in the cohort, strong causal design: Luo et al., iScience 2023;26(7):107130 (doi 10.1016/j.isci.2023.107130, PMID 37456847): C57BL/6J mice, hydrochlorothiazide 5 or 10 mg/kg in drinking water for 8–12 weeks, on normal or high-fat diet (n=4–8 per group). 16S sequencing showed a tendency toward decreased alpha diversity (Chao, Shannon, Ace, and Simpson indices) and a significant, reproducible enrichment of the Enterobacteriaceae family, with clear beta-diversity separation from controls on principal coordinate analysis. Hydrochlorothiazide significantly raised fasting blood glucose, fasting insulin, and HOMA-IR, and impaired oral glucose tolerance, across both diet groups.

Mechanism

Named and causally tested, not merely correlational. Hydrochlorothiazide raised lipopolysaccharide (LPS) in cecal contents and serum, activating TLR4/MyD88/NF-κB signaling and inducing hepatic macrophage polarization (both M1 marker CD80 and M2 marker CD206 induced in F4/80+ macrophages). Broad-spectrum antibiotic depletion (ampicillin, neomycin, metronidazole, vancomycin) and Gram-negative-selective neomycin alone both abolished the metabolic phenotype. Fecal microbiota transplantation from hydrochlorothiazide-treated mice reproduced elevated insulin, HOMA-IR, glucose intolerance, serum LPS, and inflammatory cytokines (TNF-α, IL-1β, IL-6) in untreated recipient mice. Both a TLR4 pharmacological inhibitor (TAK-242) and Tlr4 -knockout mice blocked the phenotype. This is unusually well-controlled causal evidence for a Tier B animal study.

Counter-evidence held on record

This is mouse, not human, data, and the dose (5–10 mg/kg in drinking water) does not map cleanly onto a human 12.5–50 mg oral tablet dose. No human 16S or shotgun metagenomic study of hydrochlorothiazide-exposed patients exists to confirm or refute the mouse finding. This is real, causally-demonstrated animal evidence of drug-induced dysbiosis, not evidence that hydrochlorothiazide depletes any specific nutrient through the gut, and it says nothing about vitamin or mineral status. The measured downstream harm, insulin resistance and glucose intolerance, is a metabolic outcome consistent with this drug's already-documented, separately excluded-from-grading hyperglycemia effect, not a reason to recommend a probiotic or gut-health supplement.

Limits on what this shows

• Key negative/null findings: The label's own placebo-comparable adverse-event profile at 12.5 mg, the most commonly prescribed dose, bounds this finding to the higher end of the dose range. No registry or cohort study establishing a numeric incidence was found, which is itself why this entry is held at orange rather than red: the mechanism and the label listing are real, but the frequency cannot be quantified from the evidence located.
• At-risk subgroups: Patients on 25 mg or higher; patients with a known sulfonamide hypersensitivity, per the label's general cross-sensitivity caution; case-report literature describes onset both shortly after initiation and after prolonged stable use.

Regulatory status

FDA label, Adverse Reactions, Hematologic: "Aplastic anemia, agranulocytosis, leukopenia, hemolytic anemia, thrombocytopenia," listed among reactions associated with doses of 25 mg and greater. At the 12.5 mg dose used in the lowest-strength formulation, the label states adverse events reported were comparable to placebo.

Population evidence

The regulatory label is the Tier A source with hydrochlorothiazide in the cohort; no population-based cohort study quantifying an incidence rate or hazard ratio for HCTZ-associated anemia was located. Two case reports of hydrochlorothiazide-induced hemolytic anemia were located by title and abstract snippet (a 2021 case report and a 2016 case report); both were behind bot-verification pages on repeated access attempts, so they are named in the sources list as unverified and no specific statistic, lab value, or timeline is drawn from either here.

Mechanism

Idiosyncratic, not the drug's intended mechanism of action. Hemolytic and aplastic reactions to hydrochlorothiazide are consistent with the drug's sulfonamide-derived chemical structure and its labeled sulfonamide cross-sensitivity caution; this is a hypersensitivity/immune or marrow-toxicity mechanism, structurally distinct from the NCC-blockade mechanism responsible for the potassium, magnesium, zinc, calcium, and sodium entries above. It must not be filed as a nutrient-deficiency mechanism.

Counter-evidence held on record

The label's placebo-comparable safety profile at 12.5 mg is itself the main bound on this entry, since that is the dose most commonly prescribed. No case series, registry, or pharmacovigilance cohort quantifying absolute incidence was located during this review, so the finding is recorded as real and label-documented but not as common. Never present this as iron-deficiency anemia or as a reason to take an iron supplement. The mechanism is bone-marrow or immune-mediated red cell destruction or production failure, not iron depletion, and iron levels have not been shown to change on this drug. New or unexplained anemia, bruising, or infection on hydrochlorothiazide needs clinical evaluation, not supplementation. See the Iron entry below for the direct rebuttal of the "reach for iron" misreading this finding invites.

Limits on what this shows

• Why no grade: Yellow requires either mechanistic plausibility demonstrated at Tier B or class extrapolation at Tier C where at least one class member has real supporting data. Neither condition is cleanly met. The mitochondrial membrane potential collapse found in UVA-irradiated skin cells (mitochondrial burden entry, above) is suggestive, since transmembrane potential is the driving force for ATP synthase, but it is not itself a measurement of ATP content or synthesis rate, and under this dossier's overlap rules a mitochondrial-capacity finding is counted once, under mitochondrial burden, and is not double-counted here. The one identified renal enzyme finding, reduced Na-K-ATPase activity along several nephron segments in hypertensive rats given hydrochlorothiazide (PMID 3035268), is the drug's own expected pharmacodynamic consequence of NCC blockade: less distal sodium delivery means less active-transport work for the pump to do. It is not evidence of impaired systemic ATP-generating capacity, and this dossier's governance bars filing a drug's own labeled or expected mechanism of action as a nutrient-status or energy-burden finding. No Tier C class extrapolation was found either: no ATP or cellular-energy-metabolism data was located for chlorthalidone or indapamide, the other major thiazide-type diuretics. White does not apply because no circulating consumer claim naming hydrochlorothiazide together with ATP, cellular energy, or fatigue-as-energy-deficiency was found among the consumer sources already reviewed for this dossier.

What was searched

DailyMed, PubMed, and general web search for hydrochlorothiazide combined with ATP, adenosine triphosphate, cellular energy, oxidative phosphorylation, and mitochondrial respiration. One animal study, "Effect of diuretics on ADP incorporation in kidney mitochondria" (PMID 512910), appeared in search results but could not be opened past a bot-verification page on repeated attempts; it is named in the sources list as unverified and is not used to support any statement in this entry.

Counter-evidence held on record

There is none to hold, which is the finding. The photoactivated mitochondrial-collapse data described above is circumstantial and tissue-bounded, not a direct ATP measurement, and is not treated as counter-evidence here since it does not bear on a systemic claim in either direction. Absence of evidence is not evidence of safety. Do not publish "hydrochlorothiazide does not affect your cellular energy production," and do not attribute patient-reported fatigue on hydrochlorothiazide to an ATP deficit; fatigue on a diuretic has simpler, already-documented explanations, including blood pressure reduction itself, hypokalemia, and hyponatremia, both already graded above. No coenzyme Q10, carnitine, ribose, or similar energy-support supplement recommendation follows from this entry, and none is made.