Lisinopril

Limits on what this shows

• No human measurement at any tier.
• Direction flips with strain and with age in the fly work.
• Exposures do not map onto human therapy: 1 mM in fly medium, and 40 mg/L drinking water in rats (roughly 24 mg/m²/day).

Animal evidence

Ortiz de Choudens 2022: in salt-sensitive rats, radiation reduced maximal respiration and spare respiratory capacity, and lisinopril mitigated both; no adverse mitochondrial effect of lisinopril alone in unirradiated animals is reported. Ederer 2018: state 3 respiration fell about 41 percent in one Drosophila strain at one week, and the mitochondrial-to-nuclear DNA ratio rose about 17 percent in another strain at three weeks.

Mechanism

Plausible in both directions and unresolved. Angiotensin II signalling reduces skeletal muscle mitochondrial content, so blocking it gives a route to improved capacity. No route by which lisinopril would impair oxidative phosphorylation has been described: the drug is not metabolised, induces and inhibits no CYP450 enzyme, and is excreted unchanged.

Null finding

Werrmann and Cohen 1994: isolated working rat heart, high-energy phosphate metabolism essentially unchanged by any treatment regimen, though functional recovery improved.

What this is not

Not a statement that lisinopril damages mitochondria, and not a statement that it protects them. Fatigue reported on lisinopril has label-documented explanations including blood pressure reduction itself. No coenzyme Q10, carnitine or ribose recommendation follows, and none is made.

Limits on what this shows

• One unreplicated study: 32 hypertensive patients, 16 on lisinopril 10 mg monotherapy, measured at 3 and 6 months against their own baseline with no untreated or healthy control arm reported.
• No oxidative damage marker was measured, so no functional consequence is demonstrated.
• Selenium status and reduced glutathione were not measured, so the enzyme change cannot be attributed to substrate or cofactor supply.
• Nothing is known about higher doses in people.

Population evidence

Kaminsky and Kosenko 2010: in the lisinopril monotherapy arm, catalase activity rose 79.3 to 106.5 percent (p < 0.0001) at both timepoints, glutathione peroxidase fell 20.7 to 30.6 percent (p < 0.001) at both timepoints, superoxide dismutase rose only at 6 months, and glutathione reductase did not change.

Mechanism

Plausible in the protective direction: angiotensin II drives NADPH oxidase activity, so blocking its formation lowers superoxide generation and could reduce demand on glutathione peroxidase rather than deplete capacity. Laboratory chemistry (Mira 1992, Fernandes 1996) also runs protective, but is cell-free or ex vivo, not a clinical effect.

Counter-evidence held on record

Akintunde 2016 runs the other way in rats at 28 mg/kg/day for 8 days (depleted superoxide dismutase and catalase, increased lipid peroxidation, airway epithelial damage) — far above the human range of roughly 1.3 mg/kg at the maximum labelled 80 mg/day. Ederer 2018 found lisinopril reduced hydrogen peroxide by around 50 percent in one Drosophila strain, again protective.

Second human study, not counted

Althanoon and Mahmood 2021 studied lisinopril 10 mg daily for three months in 100 hypertensive patients, but the full-text tables could not be retrieved and the direction of change could not be verified, so it is recorded and not counted as a second source.

What this is not

Not a statement that lisinopril causes oxidative stress, and not a statement that it is an antioxidant therapy. Above all, not a selenium or glutathione deficiency and not a reason to supplement either.

Limits on what this shows

• No status measurement at any tier, in any species. A PubMed search on 7 Sep 2026 returned 92 records, none measuring copper in lisinopril-treated subjects.
• All lisinopril copper data are cell-free or gas-phase and carry no dose or duration that maps to a patient.
• Copper is not named anywhere in the lisinopril prescribing information.

Laboratory evidence

Silva 1999: gas-phase metal-ion binding for lisinopril located preferentially at the amine nitrogen, and for captopril at both sulphur and nitrogen. Fernandes 1996: lisinopril moderately inhibited copper-driven albumin oxidation and protected red cells from copper-induced hemolysis.

Mechanism

Weaker than the zinc case for a structural reason. ACE is a zinc metalloenzyme, which gives zinc a target-level rationale; ACE is not a copper enzyme, so no equivalent rationale exists. What remains is generic metal-ion coordination chemistry with no demonstration that it affects copper handling in a living organism.

Null finding

Bartosz 1997 tested captopril and enalapril only and concluded that only sulfhydryl-containing ACE inhibitors show this chemistry. Lisinopril is a non-thiol carboxylate like enalapril, so the finding weighs against a copper-binding effect in the body.

Limits on what this shows

• No iron parameter has been measured at any dose or duration in a lisinopril-treated population.
• The closest study scheduled serum iron but reports change only in hematocrit and erythropoietin — a weak negative observation, not a measured null.
• Iron is not named anywhere in the lisinopril prescribing information.

Mechanism

Absent for iron loss, and present in the opposite direction for iron sparing. Lisinopril causes no gastrointestinal blood loss, no gastric pH change, no bile acid sequestration and no renal iron wasting. Because it suppresses erythropoiesis rather than accelerating red cell turnover, the expected effect on iron demand is a small reduction.

Why the hemoglobin change is not iron

Cole 2000 shows the ACE-blocked phenotype is normocytic anemia with elevated erythropoietin and reduced red cell mass, reversed by angiotensin II. Iron deficiency anemia is microcytic and is not reversed by angiotensin II. D'Amelio 1997 shows erythropoietin falling in lisinopril-treated patients. The mechanism sits upstream of iron entirely.

The most important misreading to avoid

A reader who sees a lower hemoglobin will reach for iron. Iron levels have not been shown to fall on this drug, and taking iron will not correct a hemoglobin change caused by reduced erythropoietic drive. A new or worsening anemia needs to be worked up rather than supplemented on assumption.

Limits on what this shows

• A PubMed search on 7 Sep 2026 for lisinopril and microbiome or microbiota returned three records, none measuring microbiota in a lisinopril-exposed subject.
• Roughly 75 percent of an oral dose is not absorbed, so unabsorbed drug does reach the lower gut — that is exposure, not effect.

Why no grade

Yellow requires mechanistic plausibility at Tier B or class extrapolation where a class member has real supporting data; neither condition is met. White requires a circulating claim to rebut, and none exists for this drug class. The category sits outside all four tiers.

The one lisinopril-in-cohort record, and its direction

Yang 2022 tests the reverse relationship: Coprococcus comes harbours esterase activity that catabolises ester ACE inhibitors and reduced the blood pressure lowering effect of quinapril and ramipril in rats, but “selectively reduced the antihypertensive effects of ester ramipril but not nonester lisinopril”. Lisinopril is the negative control there, so this is filed as an interaction running the other way, and a null one for this drug.

What this is not

Absence of evidence is printed here as absence. It is not a statement that lisinopril is safe for gut bacteria, and not a statement that it harms them.

Limits on what this shows

• Magnitude is small, and under 0.1% of people in clinical trials stopped treatment because of anemia.
• Matters mainly where another cause of anemia already exists, in reduced eGFR or dialysis, and in transplant recipients.
• Not dose-stratified in the label; D'Amelio used 5 mg daily over 30 days. No exposure window is established.
• A drug-specific null (Winnicki 2012, 14 haemodialysis patients) found no link between lisinopril exposure and erythropoiesis-stimulating agent requirement — the main reason this stays at orange rather than red.

Regulatory status

FDA label §6.1, verbatim: "Small decreases in hemoglobin and hematocrit (mean decreases of approximately 0.4 g% and 1.3 vol%, respectively) occurred frequently in patients treated with Lisinopril but were rarely of clinical importance in patients without some other cause of anemia."

Population evidence

The label finding above, plus D'Amelio 1997: 15 renal transplant recipients with symptomatic erythrocytosis on lisinopril 5 mg daily for 30 days showed a significant fall in hematocrit and in serum erythropoietin. Serum iron, folate and B12 were scheduled in that protocol; only hematocrit and erythropoietin are reported as changed.

Mechanism

Angiotensin II supports erythropoiesis through AT1 receptor signalling on erythroid progenitors; removing it lowers red cell production. Cole 2000 establishes this in ACE-knockout mice, which are anemic despite elevated erythropoietin, with hematocrit restored by angiotensin II.

Counter-evidence held on record

Cleland 1993: peak oxygen consumption rose from 19.8 to 21.4 ml/kg/min on lisinopril 10 mg daily versus placebo (p < 0.003). Winnicki 2012 is the second, more direct null. In post-transplant erythrocytosis, lowering hematocrit is the therapeutic goal.

What this is not

Not iron deficiency, and not a reason to start iron. Erythropoietin is falling or the marrow is less responsive to it; iron stores were not measured as low in any lisinopril study located. New or worsening anemia is a conversation with the prescriber, not a supplement decision.

Limits on what this shows

• No human tissue energetics at any dose.
• The two animal results separate cleanly by dose: 1 mg/kg in rats produced no change; 28 mg/kg/day for 8 days produced ATP depletion alongside visible tissue damage.
• A 60 kg adult at the maximum labelled 80 mg/day receives roughly 1.3 mg/kg — nothing in the human dose range has been tested.

Animal evidence

Werrmann and Cohen 1994 (31P spectroscopy, lisinopril 1 mg/kg in vivo): high-energy phosphate metabolism essentially unchanged before, during and after global ischaemia — the better-designed study for this question. Akintunde 2016 (28 mg/kg/day for 8 days): depleted lung ATP alongside depleted superoxide dismutase and catalase and increased lipid peroxidation, reported with frank epithelial damage.

Mechanism

No route by which lisinopril would inhibit ATP synthesis has been described. It is not metabolised, generates no reactive metabolite, does not uncouple oxidative phosphorylation in any assay located, and depletes no named ATP-synthesis cofactor.

Human functional evidence, counted as counter-evidence

Cleland 1993: peak oxygen consumption rose from 19.8 to 21.4 ml/kg/min on lisinopril 10 mg versus placebo (p < 0.003). Cooke 2002: in 12 heart failure patients, aerobic capacity was higher on 5 mg than 20 mg daily (1696 versus 1578 ml/min, p = 0.016). Cooke has no non-use arm, so it establishes a within-drug dose gradient only, and is not a reason to change a prescribed dose.

What this is not

Not a statement that lisinopril lowers cellular energy production. It has never been measured in a person taking the drug, and the human exercise data point the other way at standard doses.