Clindamycin

Limits on what this shows

• Null findings: No study showing preserved mitochondrial function with clindamycin was found. In the same Xiao panel ampicillin and sulfamethoxazole did not cause apoptosis.
• At-risk subgroups: None established. Neuronal vulnerability is an in vitro observation only.

Direction

Drug toxicity (burden). Not a nutrient deficiency and not a basis for any supplement.

Regulatory status

No mitochondrial statement in the label text read; StatPearls describes only reversible binding to the bacterial 50S ribosomal subunit [1][2][6].

Population evidence

No human data. Xiao 2019 [21] (Tier B): six antibiotic classes in human primary neurons and neuronal cell lines; metronidazole, tigecycline, azithromycin and clindamycin induced apoptosis at clinically relevant concentrations, ampicillin and sulfamethoxazole did not; clindamycin, tigecycline and azithromycin each impaired mitochondria through effects on different respiratory complexes and disrupted mitochondrial membrane potential (paraphrase of the abstract); antioxidant pretreatment reversed the harm. The concentrations, the complexes affected for clindamycin and the effect sizes were not retrieved (abstract via the PubMed record and an OpenAlex reconstruction; the Wiley full text returned 403).

Mechanism

Proposed, not demonstrated for clindamycin: structural similarity between bacterial and mitochondrial ribosomes can make mitochondrial translation a target of ribosomal antibiotics (Boettger 2001 review, abstract level) [23]. The Xiao abstract reports respiratory complex impairment but does not say that it arises from inhibition of mitochondrial translation. PubMed listings for clindamycin with mitochondrial translation, the mitoribosome, and lincosamides with the mitochondrial ribosome returned no clindamycin experiment [23].

Cohort test

Clindamycin was a tested compound in Xiao 2019 (Tier B in vitro). Studies of other antibiotics on mitochondria cited in sibling dossiers (Kalghatgi 2013, Sailer 2025) were not opened in this review and are not used.

Counter-evidence held on record

Counter-evidence and confounding In vitro concentrations may not reflect tissue exposure; neurons are one cell type. The same source supports the oxidative stress and ATP entries (see Overlap map). Framing caution Mechanistic, in vitro, and not a human finding. No antioxidant or mitochondrial supplement is suggested by this entry.

Limits on what this shows

• Null findings: Lipid peroxidation was unchanged in the Abad 2025 mouse model; no study with unchanged antioxidant markers in clindamycin users was found.
• At-risk subgroups: None established.

Direction

Mixed: increased oxidative damage in several non-human or in vitro models (drug effect, possibly through dysbiosis), and a protective reverse-direction finding in one rat model. Not a nutrient deficiency.

Regulatory status

No oxidative stress statement in the label text read [1][2].

Population evidence

No human marker study. Xiao 2019 [21] (Tier B): clindamycin, tigecycline and azithromycin caused oxidative (ROS-linked) damage in human neurons; antioxidant pretreatment reversed the effects. Animal listings [22] (abstract level, numbers mostly not retrieved): Abad 2025, mice with clindamycin dysbiosis, protein carbonyls up almost 5-fold versus control, lipid peroxidation unchanged, IL-6 and TNF-alpha up 1- and 2-fold; Aabed 2019, hamsters on 30 mg/kg for 3 days, brain glutathione and vitamin C changed; Al-Orf 2018, hamsters given a single 30 mg/kg dose, lipid peroxidation and catalase up, GST activity and glutathione down. Reverse direction: Ibrahim 2020, rats given clindamycin 300 mg/kg intraperitoneally for 5 days before doxorubicin, clindamycin prevented doxorubicin-driven glutathione depletion and catalase reduction.

Mechanism

Proposed: mitochondrial respiratory impairment generating ROS [21], and dysbiosis-linked injury in the hamster and mouse models. Neither is demonstrated in humans.

Cohort test

Clindamycin was the exposure in every study cited (Tier B in vitro and animal). Several animal studies use high mg/kg doses and a dysbiosis or toxin-combination design, so they are weak evidence for typical human dosing.

Counter-evidence held on record

Counter-evidence and confounding The rat result runs opposite to the others and is a co-treatment design, so it does not describe the drug alone. Dysbiosis, not clindamycin chemistry, may drive the hamster and mouse changes; see the Overlap map. Framing caution Not a deficiency claim. No antioxidant supplement is recommended from this entry.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C source was found to support even a yellow grade, and no circulating consumer claim on iron and clindamycin was found to rebut. This is a disclosed gap, not a finding of safety. Note the hematologic outcome signals are filed under Oxygen delivery/utilization, not here.

What was searched

Two DailyMed label copies (summarizer-read): iron appears only as black iron oxide, a capsule colorant, and no iron interaction or iron-related reaction is listed [1][2]. PubMed listing "clindamycin iron absorption OR clindamycin ferritin OR clindamycin copper": ten records, none a nutrient-status study of clindamycin; the one iron-related hit was a Leuconostoc hemophagocytic syndrome case in which clindamycin was treatment (ferritin 1,068; hemoglobin 88 falling to 39 g/L), which is not a drug effect [32]. PubMed listing "clindamycin serum iron OR iron deficiency OR ceruloplasmin": ten records shown from a very large noisy result set, none mentioning clindamycin [32]. Drugs.com food-interaction page: "no known alcohol/food interactions" [30]. Hemolysis and anemia listings are covered in the Oxygen delivery entry [33].

Channel outcomes

PubMed listings opened on the phrasings above; many other PubMed phrasings returned 429 and were retried on new wording after waits of 2 to 5 minutes. The web-search budget was exhausted, so no consumer-claim search for "clindamycin and iron" could be run; Europe PMC, OpenAlex search and PMC article pages were rate-limited or redirected. This is a mixed result: nothing found on the channels that worked, plus tooling limits on others. The label and the PubMed listings make a genuine gap likely for human data, but a later re-check with working Europe PMC and web search is warranted.

Limits on what this shows

• Why no grade: Dual condition met: the in vitro and in silico copper records concern antimicrobial or gene-set effects, not copper nutrient status, so no Tier B or C source supports even a yellow grade; and no circulating consumer claim on copper and clindamycin was found to rebut. Copper is kept separate from Iron per Hard Rule 7. This is a disclosed gap, not a finding of safety.

What was searched

Two label copies: no copper statement [1][2]. PubMed listing "clindamycin iron absorption OR clindamycin ferritin OR clindamycin copper": copper appeared only as an antimicrobial partner or comparator, not as a nutrient. Copper nanoparticles with clindamycin acted synergistically against periodontal pathogens in vitro (fractional inhibitory concentration 0.312 to 0.5); copper complexes outperformed clindamycin in an in vitro antimicrobial comparison; clindamycin was one of ten agents predicted in silico to target cuproptosis-related genes in ulcerative colitis data; and clindamycin and copper resistance genes co-occur on bacterial plasmids [32]. None measures copper status in a person or animal given clindamycin. PubMed listing "clindamycin serum iron OR iron deficiency OR ceruloplasmin": no clindamycin record [32]. Drugs.com: no mineral interaction listed [30].

Channel outcomes

As for Iron: the PubMed listings that opened returned nothing on topic; web search, Europe PMC, OpenAlex search and PMC pages were rate-limited or unavailable. Mixed result, likely a genuine gap for human data, with a re-check recommended.

Limits on what this shows

• Null and contradicting findings: ARMORD found no clindamycin effect on Enterobacteriaceae (p = 0.31) or Enterococcus, whereas the systematic review reports more non-E. coli Enterobacteriaceae [12][13]: the taxa-level picture is not uniform, but both agree on lower diversity and fewer anaerobes. Baldanzi found no richness association at 4 to 8 years [11]. Zaura fecal shifts differed from placebo for up to 4 months, not at 12 [9]. No Tier A study with clindamycin in the exposed group reported no change in diversity or anaerobes. Cohort A has 4 exposed people.
• At-risk subgroups: Adults over 60 for the colitis consequence (label) [1]; people with recent repeated courses (registry exposure variable) [11]; hospitalized patients. Baldanzi reported sex and age interactions for some species (22 species stronger in women, 9 in men; 31 stronger in younger, 22 in older people) [11].

Direction

Disturbance of the gut microbial community (risk entry). It is not a deficiency, and nothing here recommends a probiotic, fiber, vitamin or any supplement.

Regulatory status

Boxed warning: therapy "has been associated with severe colitis which may end fatally", so the drug is reserved for serious infections. Label: C. difficile-associated diarrhea "has been reported with use of nearly all antibacterial agents", may occur more than two months after treatment, and "Treatment with antibacterial agents alters the normal flora of the colon, leading to overgrowth of C. difficile" [1][2].

Population evidence

Cohort A, Jernberg 2007 [7] and Lofmark 2006 [8] (same 8 volunteers): 4 healthy adults given clindamycin for 7 days versus 4 unexposed controls, 9 fecal samples over 2 years. Bacteroides clonal diversity fell sharply, the Bacteroides community "never returned to its original composition" in 2 years (T-RFLP), levels of specific resistance genes rose dramatically and stayed raised, and control variation was minor; 1,306 Bacteroides isolates were typed [8]. Abstract level only: dose, diversity values and fold changes not retrieved (full text paywalled). Cohort B, Zaura 2015 [9] and Rashid 2015 [10] (treated as one randomized placebo-controlled trial arm; same drugs, follow-up and overlapping authors): clindamycin 150 mg four times daily for 10 days versus placebo; 9 clindamycin completers [9][10]. Fecal Shannon diversity was significantly below baseline for up to 4 months; butyrate producers (Faecalibacterium, Subdoligranulum, Roseburia, Coprococcus, uncultured Ruminococcaceae and Lachnospiraceae) decreased; at 1 month 74 of 4,606 predicted functions rose and 438 fell (about 11%); fecal ermB rose 240-fold in the one subject with a full metagenome [9]. Rashid: 1 to 12 months to normalize depending on species, largest shift at month 1, diversity changes persisting to month 12, and by pyrosequencing clindamycin had more impact than ciprofloxacin; no C. difficile colonization or infection occurred [10]. P values by timepoint and the n randomized in the clindamycin arm were not retrieved. Cohort C, Baldanzi 2026 [11]: 14,979 Swedish adults with fecal metagenomes (1,340 species) linked to prescription registers. Within 1 year of sampling each clindamycin course was associated with about 47 fewer detected species (q = 2.1 x 10^-17), versus about 20 per fluoroquinolone and 21 per flucloxacillin course; 296 species were associated within 1 year and 349 at 1 to 4 years (208 lower, 141 higher); clindamycin accounted for 37.9% of all FDR-significant species associations; higher abundance of Enterocloster bolteae, E. citroniae, Flavonifractor plautii, Ruminococcus gnavus and Eggerthella lenta; richness effects were seen at 1 to 4 years but not at 4 to 8 years, although clindamycin, fluoroquinolones and flucloxacillin together still altered 10 to 15% of species at 4 to 8 years. Users per window and regression coefficients were not retrieved. Cohort D, ARMORD, Peto 2025 [12]: Oxfordshire inpatients, outpatients and volunteers; 225 adults in the cross-sectional analysis. Oral clindamycin: Shannon diversity -2.2 (95% CI -3.5 to -0.8; p = 0.002), reduced anaerobes (Actinobacteria lower), more macrolide/clindamycin and aminoglycoside resistance genes (p up to 0.02). Longitudinal data were insufficient for clindamycin; the number of exposed participants was not reported. Stratified systematic review, Zimmermann and Curtis 2019 [13]: 129 studies, 2,076 antibiotic-treated participants, 301 controls; for clindamycin: strong decrease in anaerobes, more non-E. coli Enterobacteriaceae (Citrobacter, Enterobacter, Klebsiella), less E. coli, longest persistence 2 years. Abstract level; it likely draws on Cohort A, which was not verified.

Mechanism

Named and present: antibacterial activity against susceptible anaerobes with overgrowth of resistant organisms, stated in the label [1]. In mice a single 200 microgram intraperitoneal dose removed about 87 to 90% of cecal taxa for 4 weeks (Shannon about 5 to under 1, then 2 or less), mice stayed susceptible to C. difficile for at least 10 days, and 40 to 50% died after challenge [14] (Tier B). Loss of the bile-acid 7-alpha-dehydroxylating bacterium Clostridium scindens, and so of secondary bile acids, is linked to C. difficile susceptibility in mice, with a human transplant-patient dataset not specific to clindamycin [15] (Tier B/C).

Clinical consequence (outcome, not a nutrient)

C. difficile infection: pooled community-associated odds ratio for clindamycin 16.80 (95% CI 7.48 to 37.76; 7 observational studies, no-antibiotic comparator) [16]; hospital-acquired pooled OR 2.86 (2.04 to 4.02; 6 studies, review total 15,938 patients), second only to third-generation cephalosporins [17]. Listing-level abstracts of more recent studies: a pharmacopoeia-wide case-cohort in older adults gave clindamycin OR 16.83 (15.53 to 18.24); a pediatric case-control (65 cases, 130 controls) gave adjusted OR 14.842 (4.496 to 49.000) [18]. Clindamycin-specific attack rates were not found: a Ghana pilot reported antibiotic-associated diarrhea in 11 of 70 patients (15.7%), 73% of whom had received clindamycin [18].

Cohort test

Clindamycin was the studied exposure in A, B, D, the clindamycin stratum of the review, and a named drug in C [7][9][10][11][12][13]. Elvers 2020 [19] was opened and does not include clindamycin among its antibiotics: not used. Ramirez 2020 [20] was opened and has no clindamycin data: not used. Buffie 2015 [15] human data come from transplant patients on mixed antibiotics: Tier C for clindamycin, used only for mechanism context.

Independence

A (Sweden, healthy volunteers, 2000s), B (randomized trial arm, Karolinska), C (population registry cohort) and D (Oxfordshire clinical cohort) are different people. Lofmark and Jernberg are one cohort; Zaura and Rashid are treated as one; the review is not counted as a separate cohort. Observational cohorts C and D are confounded by indication (the infection and the setting), which A and B, with healthy volunteers, are not.

Why red

At least two independent Tier A human cohorts with clindamycin in the exposed group (four are on record) measured the microbiome directly, agree on direction, show a clinically meaningful and in places lasting change, have a named mechanism, and no Tier A source contradicts the direction. The ceiling is bounded by evidence quality, not by the grade logic: full-text effect sizes were not retrieved for A, B and C, and A and B are small.

Counter-evidence held on record

Counter-evidence and confounding Infection, hospitalization, diet and co-medication confound the registry and clinical cohorts; the volunteer cohorts are small (4 and 9 exposed). The microbiome shift is not itself a measured nutrient deficit, and effect on vitamin K, short-chain fatty acid or other nutrient status was not measured in any of these cohorts. Most reads were of abstracts or through summarizing fetch tools. Framing caution This is a disturbance and infection risk, not a deficiency. No probiotic, yogurt, fiber or vitamin supplement is recommended by this entry, and nothing here says every course causes lasting harm. New watery or bloody diarrhea during or up to two months after a course needs clinical assessment for C. difficile.

Limits on what this shows

• Null findings: None located for clindamycin.
• At-risk subgroups: None established.

Direction

Possible reduction in cellular energy output as a drug toxicity effect. Not a nutrient deficiency.

Regulatory status

No statement in the label text read [1][2].

Population evidence

No human data. Xiao 2019 [21]: the abstract (paraphrased) states that tigecycline, azithromycin and clindamycin disrupted mitochondria, lowering membrane potential and leading to an energy deficit; whether ATP was measured directly, at what concentration and by how much were not retrieved (full text 403).

Mechanism

Proposed, same chain as the mitochondrial entry: respiratory complex impairment lowering membrane potential. Not demonstrated in vivo.

Cohort test

Clindamycin was a tested compound (Tier B in vitro). No class-level source was used.

Counter-evidence held on record

Counter-evidence and confounding This entry adds no evidence independent of the mitochondrial entry (see Overlap map). It is kept separate per the category rules. Framing caution Not a human finding and not a "low energy" claim for people taking clindamycin.

Limits on what this shows

• Null findings: A PubMed listing for "clindamycin induced hemolysis" (15 results, first 10 read) found no human clindamycin hemolysis case report or series [33]. Gray 1974, an animal toxicity paper with hemolysis in its indexing, has no abstract in the record and was not usable.
• At-risk subgroups: None established for clindamycin. G6PD-related hemolysis in the combination arm is a primaquine context, not a clindamycin finding.

Direction

Possible adverse hematologic outcome (anemia). It is not iron deficiency, and it is not a nutrient entry (Hard Rule 2).

Regulatory status

Label hematologic reactions: transient neutropenia (leukopenia), eosinophilia, and "Reports of agranulocytosis and thrombocytopenia"; hemolytic anemia is not mentioned in the label text read. Periodic blood counts are advised during prolonged therapy [1][2].

Population evidence

Hoyer 2020 [33] (animal case, listing-level): in 8 captive wallabies treated for oral disease, one developed a suspected adverse reaction to prolonged intravenous clindamycin with immune-mediated anemia. Safrin 1996 [33] (human randomized trial in mild to moderate Pneumocystis pneumonia, listing-level): serious hematologic toxicities (neutropenia, anemia, thrombocytopenia or methemoglobinemia) were more frequent in the clindamycin-primaquine arm (P = 0.01); the arm combines clindamycin with primaquine and the abstract does not apportion the effect. Smilack 1991 [33] attributes anemia to chloramphenicol, not clindamycin. No hemoglobin, hematocrit or red-cell mass study of clindamycin alone was found.

Mechanism

Immune-mediated hemolysis is the proposed mechanism in the single animal report; not demonstrated in humans. For the combination arm the mechanism cannot be assigned to clindamycin.

Cohort test

Hoyer: clindamycin is the suspected agent (Tier B, n = 1 animal). Safrin: clindamycin in cohort only in combination with primaquine, so Tier C for clindamycin alone and capped at yellow.

Counter-evidence held on record

Counter-evidence and confounding One animal report cannot establish a human risk, and the primaquine arm is confounded by design. The label lists no red-cell reaction. Framing caution An adverse-outcome flag, not an iron or oxygen-delivery deficiency. Do not suggest iron or other supplements from this entry.