Doxycycline

Limits on what this shows

• Null findings: The conflicting survival and lifespan findings [19] are the nearest counter-evidence. Experimental concentrations and doses were not available in the abstracts read, so comparison with human plasma levels (for example Cmax 0.5 to 0.6 mcg/mL on 40 mg and 3.0 mcg/mL after a single 300 mg dose [1][2]) is not possible here.
• At-risk subgroups: Not defined. A 2024 smooth-muscle cell study (Yap, Vascul Pharmacol) was seen in a PubMed listing but not opened [23].

Regulatory status

The labels read do not mention mitochondria [1][2][3].

Population evidence

No human study with a mitochondrial endpoint was located. Moullan 2015 [17] (Cell Reports; abstract only, via OpenAlex; full text blocked): tetracyclines such as doxycycline disrupt mitochondrial translation, causing mitonuclear protein imbalance; even low doses trigger mitochondrial proteotoxic stress and alter nuclear gene expression and mitochondrial dynamics and function across cell types and organisms including worms, flies, mice and plants. Wüst 2021 [18] (abstract only): in H9C2 rat cardiomyoblasts doxycycline caused mitonuclear protein imbalance, lower maximal respiration, fragmented mitochondria and a shift from oxidative phosphorylation toward glycolysis; contractility was impaired in rat cardiomyocytes and fly hearts; mice showed reduced complex I and OXPHOS capacity and diastolic dysfunction. Kozicz 2022 [19]: an editorial noting conflicting reports, including lifespan extension in worms and better survival in a Leigh syndrome mouse model versus impaired mitochondrial function in mice, possibly dependent on dose and metabolic rewiring.

Mechanism

Present. Tetracyclines inhibit bacterial protein synthesis and, given the evolutionary link between mitochondria and bacteria, disrupt mitochondrial translation [17][18].

Counter-evidence held on record

Counter-evidence and confounding Research-grade concentrations in culture can exceed clinical tissue levels, and the Moullan paper itself frames the issue as a caution for laboratory use of doxycycline-inducible gene systems [17]. Whether human cells were among the cell types in [17] could not be verified from the abstract. Framing caution Do not state that doxycycline damages mitochondria in patients. This is a preclinical signal with a reversed-benefit counter-signal, and no nutrient was measured.

Limits on what this shows

• Null findings: None for people. A fish cell-line report of doxycycline-induced oxidative stress (Shan 2022) was seen in a listing and not opened [23].

Regulatory status

No oxidative stress statement in the labels read [1][2][3].

Population evidence

Clemens 2018 [20] (abstract read via OpenAlex; PubMed listing note adds HEK293 cells): in a cell-free system doxycycline scavenged reactive oxygen species generated during malondialdehyde-acetaldehyde (MAA) adduct formation, directly scavenged superoxide but not hydrogen peroxide by electron paramagnetic resonance, reduced MAA-protein adducts and inhibited MAA-induced Nrf2 activation. The abstract has no numbers. A PubMed title search combining doxycycline with antioxidant, redox and free radical terms (126 results, first 10 reviewed) returned no human study of oxidative stress markers in doxycycline-treated people [23].

Mechanism

Present, reverse direction: direct radical scavenging [20]. A competing, unmeasured possibility is increased mitochondrial reactive oxygen species from translation inhibition (see mitochondrial burden); no ROS measurement in doxycycline-exposed mammalian cells was retrieved.

Counter-evidence held on record

Counter-evidence and confounding Cell-free scavenging does not predict in vivo antioxidant capacity at 0.5 to 3 mcg/mL plasma levels [1][2]. The conflicting mitochondrial signals mean the net redox direction in people is unknown. Framing caution Not an antioxidant depletion claim and not a claim that doxycycline is a dietary antioxidant. Reverse-direction chemistry only.

Limits on what this shows

• Null findings: None found. No iron status measurement after doxycycline was located in PubMed title searches combining doxycycline with iron, anemia, hemolysis and hemoglobin.
• At-risk subgroups: Not defined by any retrieved data.

Regulatory status

Labels mention iron only as an agent that lowers doxycycline absorption [1][2][3]. They do not report an effect of doxycycline on iron.

Population evidence

None for doxycycline. Records titled "Tetracycline inhibits iron absorption in man" (Heinrich 1973, 1974) exist in PubMed but had no abstract available, name tetracycline and not doxycycline, and were not opened; they cannot support a grade [10]. Hematologic listings in the doxycycline labels name hemolytic anemia, not iron deficiency [1][2].

Mechanism

Present as chemistry. Korać Jaćić 2023 [11] (abstract read via a publisher page summary; full text paywalled): Fe3+-doxycycline complex forms best at pH 4, ratio 6, 1:1 by mass spectrometry, two binding sites by Raman spectroscopy. The authors suggest this could impede duodenal absorption of doxycycline and raise intestinal drug concentrations. Fe2+ was not studied and no humans were involved.

Class cohort test

No class extrapolation is made. The tetracycline iron-absorption records were not opened and cannot be assigned a cohort.

Counter-evidence held on record

Counter-evidence and confounding Binding in a test tube does not show that the amount of iron absorbed from food changes, and the chelation shown in humans runs from iron to drug [8]. Chronic doxycycline users are not shown to have different iron status. Framing caution This is a plausibility note, not an iron deficiency claim. Do not recommend iron supplementation from this entry; if iron and doxycycline are both prescribed, see the timing entry above.

Limits on what this shows

• Null findings: A PubMed title search combining doxycycline with copper, zinc and magnesium (78 results, first 10 listed) returned only complex chemistry, COVID prophylaxis trials that co-administered zinc, and biomaterial studies [23].

Regulatory status

Copper is not mentioned in the labels read [1].

Population evidence

None in humans or animals. Abosede 2016 [12] (abstract read via OpenAlex): mixed-ligand copper(II) complexes with doxycycline and polypyridyl co-ligands were synthesized, characterized by X-ray crystallography (doxycycline binding in two coordination modes), cleaved plasmid DNA, inhibited MMP-2 and were cytotoxic to cancer cell lines. This is coordination chemistry with added co-ligands, not a physiological test.

Mechanism

Present as chemistry only: tetracyclines chelate divalent metals; relevance to dietary copper absorption is not shown.

Counter-evidence held on record

Counter-evidence and confounding Synthetic complexes with phenanthroline-type ligands are unlike the lumen of the gut. Copper and iron are separate entries, but both rest on the same chelation chemistry (see the overlap map). Framing caution Plausibility note only. Not a copper deficiency claim; no copper supplementation is implied.

Limits on what this shows

• Null findings: Both studies found minimal compositional change [13][14]. The results of [13] also report metatranscriptome alpha diversity higher at month 6 in standard of care than in the doxycycline arm (adjusted P 0.045), and no significant change in Chlamydia trachomatis abundance (P 0.06).
• At-risk subgroups: People with repeated intermittent exposure (the dose-response was in doses taken). Older adults and people with prior antibiotic exposure are general C. difficile risk groups per the label class warning, not a doxycycline-specific finding.

Regulatory status

Labels carry Clostridioides difficile associated diarrhea warnings ("alters the normal flora of the colon leading to overgrowth of C. difficile") and a warning that doxycycline "may result in overgrowth of non-susceptible organisms, including fungi" [1]. No microbiome statement beyond this.

Population evidence

Chu 2025 [13]: substudy of the DoxyPEP randomized trial (one 200 mg dose after condomless sex versus standard of care). 150 participants were selected (100 doxycycline, 50 standard of care); 127 analyzable DNA sequencing samples from 89 participants and 86 RNA sequencing samples from 70. Self-collected rectal swabs at enrollment and month 6. Median doses by month 6: 42 (IQR 27 to 64) versus 0 (IQR 0 to 7). Alpha diversity did not differ between day 0 and month 6 in the doxycycline arm; no genus was consistently differentially abundant. The tetracycline resistance gene proportion rose from 46% to 51% in the metagenome (adjusted P 0.030) and from 4% to 15% in the metatranscriptome (adjusted P 3.6 x 10^-6); in paired samples DNA 45% to 51% (adjusted P 0.020) and RNA 6% to 26% (adjusted P 0.016). Dose correlation: Spearman rho 0.23 (DNA) and 0.55 (RNA). No other resistance gene class rose comparably. Knodel 2025 [14]: open-label pilot randomized trial, 41 HIV-negative gay and bisexual men on HIV PrEP, immediate versus deferred doxycycline PrEP, 48 weeks, 16S rRNA sequencing of rectal samples. Composition did not change significantly in either arm; phylum-level alpha diversity fell by under 10% in the immediate arm only.

Mechanism

Present. Doxycycline is excreted unchanged in feces and bile [1][2], so gut bacteria are exposed to drug. The antibacterial action is the intended pharmacology; the resistance gene rise is selection. An in vitro study proposes that iron-complexed doxycycline could raise luminal drug exposure, framed by its authors as a possibility only [11].

Other human data

A rosacea trial of 60 adults gave all participants a 2-week doxycycline course and compared probiotic, placebo and control arms; it reports decreased gut diversity (group not specified in the abstract) and fewer tetracycline resistance genes in the probiotic group, but has no untreated comparator, so it does not isolate a doxycycline effect [15]. An endometrial microbiota study (52 women, 100 mg twice daily for 14 days) compared treated chronic endometritis with non-endometritis and is confounded by the disease and is not gut [16].

Counter-evidence held on record

Counter-evidence and confounding Both cohorts used intermittent prophylactic dosing, rectal swabs rather than stool, and men who have sex with men, so they do not describe a continuous 100 to 200 mg daily treatment course in the general population. Knodel did not report resistome data and Chu reported composition and resistome from selected subsets (89 of 150 participants analyzable for DNA), so the two studies do not measure the same endpoint. The resistance gene finding is not a nutrient effect and is not a measure of harm. A source discrepancy exists in [13]: the methods state 150 of 510 participants selected while the results cite 501 enrolled. Why not red Two independent human cohorts exist, but both show minimal compositional change, only one measured the resistome, and neither measured a nutrient. The magnitude is not shown to be clinically meaningful. Framing caution This is a bounded finding about resistance gene content under intermittent prophylaxis. It is not a claim of dysbiosis, not a nutrient deficiency, and not a reason to add probiotics or any supplement.

Limits on what this shows

• Null findings: The editorial [19] cites conflicting animal outcomes. No measurement showing unchanged or reduced ATP in doxycycline-exposed human cells or people was found.

Regulatory status

No mention in the labels read [1][2][3].

Population evidence

None in humans. Wüst 2021 [18]: lower maximal respiration (particularly with complex I substrates) and an OXPHOS-to-glycolysis flux shift in H9C2 cells; reduced complex I and OXPHOS capacity in mice. Surrogate endpoints (respiration, OXPHOS capacity) are used here as the ATP-relevant measures.

Mechanism

Present as for mitochondrial burden: reduced mitochondrial translation lowers respiratory chain components [17][18]. A cell can compensate with glycolysis, so lower respiration does not equal lower ATP.

Counter-evidence held on record

Counter-evidence and confounding The same cell and animal sources underlie this entry and the mitochondrial burden entry (see the overlap map). It is one line of evidence, not two. Framing caution Not a claim of low energy or low ATP in patients. A "low energy" statement about doxycycline users has no support in the sources read.

Limits on what this shows

• Null findings: No oxygen saturation or hemoglobin effect reported in any source read.

Regulatory status

Lupin label sec 6 and 40 mg label sec 6: "Hemolytic anemia, thrombocytopenia, neutropenia, and eosinophilia have been reported"; no frequency in the sections retrieved. Section 5.11 recommends periodic hematopoietic, renal and hepatic laboratory evaluation in long-term therapy [1][2].

Population evidence

No hemoglobin, hematocrit or red-cell mass study of doxycycline users was located. PubMed title searches combining doxycycline with anemia, hemolytic, iron and hemoglobin returned a case of doxycycline-associated acquired hemophilia A (listing only, not opened), veterinary and chemistry papers, and no hemoglobin outcome study [23].

Mechanism

Absent for oxygen delivery as a nutrient pathway. Hemolysis is an idiosyncratic or immune adverse reaction.

Counter-evidence held on record

Counter-evidence and confounding Anemia is not iron deficiency and hemolysis is not an oxygen-delivery nutrient finding (Hard Rules 2 and 7). Any hemoglobin effect would run through the iron entry, which has no human data. Framing caution Adverse-reaction note for clinicians. It does not support any nutrient or iron supplementation claim.