Salsalate

Limits on what this shows

• Null findings: Human muscle mitochondrial coupling did not change on salsalate 4 g per day for 4 days [10]. No human study of mitochondrial DNA, biogenesis, membrane potential or respiration in salsalate users beyond that crossover was found.
• At-risk subgroups: Not defined. Salicylate poisoning (overdose) is the only setting where uncoupling is described as harmful [22].

Regulatory status

No mitochondrial statement in the label sections read; the overdose section lists salicylism and disruption of electrolyte balance and blood pH [1].

Population evidence

Smith 2016 [11] (Diabetes; mouse study with salsalate, abstract via reconstructed OpenAlex record): salsalate increased oxygen consumption, lowered fasting glucose and cut liver lipid by about 55% in wild-type and AMPK knockout mice; mechanism was increased oligomycin-insensitive respiration and greater mitochondrial proton conductance at clinical concentrations, tightly correlated with suppressed de novo lipogenesis; brown adipose respiration rose independent of UCP1. van Dam 2015 [12] (Diabetes; mice and cultured brown adipocytes): salsalate raised uncoupled respiration and Ucp1 expression via PKA signaling. Hawley 2012 [13] (Science; abstract via reconstructed OpenAlex record): salicylate at concentrations achievable in blood after salsalate treatment activates AMPK by allosteric binding and by blocking Thr172 dephosphorylation; the abstract names aspirin and salsalate as salicylate precursors and does not say salsalate itself was the test compound. Day 2021 [14] (Mol Metab, mice, abstract read in full via OpenAlex): salsalate lowered macrophage proliferation and plaque lipid through AMPK alpha2 beta1 pathways. Human: Meex 2011 [10] (J Clin Endocrinol Metab; abstract via reconstructed OpenAlex record): 9 participants in a crossover, salsalate 4,000 mg per day for 4 days, then clamp testing; energy expenditure rose about 18% basal and 16% insulin-stimulated, fueled by more fat oxidation (fasting) and more oxidative glucose disposal (insulin-stimulated); salsalate "did not affect mitochondrial coupling", and substrate shifts were not attributable to skeletal muscle mitochondrial capacity.

Mechanism

Named in animals: mitochondrial uncoupling (proton leak) and AMPK activation by the salicylate released from salsalate [11][12][13]. The same uncoupling is the described mechanism of salicylate poisoning at toxic exposure [22], which does not name salsalate.

Hard Rule 2 note

AMPK activation and a rise in energy expenditure are the drug's own pharmacology and the proposed basis of its glucose and lipid effects [10][11][13]. They are not a nutrient deficiency and are not evidence of mitochondrial injury.

Class cohort test

Salsalate itself was in the cohorts of [10][11][12][14]. [13] is moiety-level (salicylate); salsalate is named only as a precursor, so it is read as supporting context. [22] is a salicylate toxicity chapter that does not mention salsalate: CLASS EXTRAPOLATION, toxic exposure only.

Counter-evidence held on record

Counter-evidence and confounding Mouse exposure and cell concentrations are not equated with the labeled human dose here. The only human mitochondrial measurement is a 9-person, 4-day study of muscle coupling [10]. Authors of [11] present uncoupling as the beneficial mechanism, so the finding is not a toxicity signal. Framing caution Not a mitochondrial toxicity claim and not a deficiency claim. No supplement is implied. Shares its mechanism with ATP production and the AMPK activity noted above (see Overlap map).

Limits on what this shows

• Null findings: No human study of glutathione, MDA, F2-isoprostanes or antioxidant enzymes in salsalate users was found. A PubMed listing on salsalate with oxidative and antioxidant terms returned trial and review records that mention oxidative stress only as disease background (titles seen; not opened).
• At-risk subgroups: Not defined.

Regulatory status

No oxidative-stress statement in the label sections read [1].

Population evidence

Hüttl 2023 [15] (Front Pharmacol; non-obese hereditary hypertriglyceridemic rats, salsalate 200 mg/kg per day for 6 weeks; abstract via reconstructed OpenAlex record): inflammatory markers, lipoperoxidation products and methylglyoxal were significantly decreased, described as reduced oxidative dicarbonyl stress. Trnovska 2017 [16] (PLoS One; 15-month-old SHR rats expressing human CRP and controls, salsalate 200 mg/kg per day for 4 weeks; abstract via reconstructed OpenAlex record): TNF-alpha and MCP-1 fell and oxidative stress in liver and kidney was reduced in the transgenic strain. Both from the same Czech research network (shared authors Pravenec, Silhavy), so not fully independent.

Mechanism

Proposed by the authors as secondary to lower inflammation [16] and to altered lipid metabolism [15]. Direct radical scavenging by salicylate was not tested in the abstracts read.

Class cohort test

Salsalate was the test drug in both rat studies [15][16]. No class extrapolation used.

Counter-evidence held on record

Counter-evidence and confounding Rodent models, small and short, disease-model specific, with no human confirmation. Salicylate at poisoning-level exposure disturbs oxidative metabolism [22], but no redox marker was measured for salsalate at human exposure. Framing caution Not an antioxidant benefit claim and not a deficiency claim. Salsalate should not be described as depleting antioxidants or as an antioxidant supplement.

Limits on what this shows

• Null findings: No excess fecal blood loss versus placebo in [4]; 2 of 42 abnormal in [5]. Ferritin, transferrin saturation, iron absorption: no salsalate study found.
• At-risk subgroups: Prior ulcer or GI bleeding (label: greater than 10-fold bleed risk with NSAIDs), anticoagulant users (synergistic GI bleeding with warfarin), older adults, long courses [1].

Regulatory status

Boxed warning for NSAID class GI bleeding, ulceration and perforation; "Anemia is sometimes seen in patients receiving NSAIDs"; post-marketing reports include G.I. bleeding; salsalate "does not inhibit platelet aggregation" in the clinical pharmacology section, while the hematologic precautions say NSAIDs prolong bleeding time in some patients [1]. Iron is not listed as a depletion. Opposite-direction items are not described.

Population evidence

Cohen 1979 [4] (J Clin Pharmacol; abstract via reconstructed OpenAlex record): 20 healthy volunteers housed 23 days, placebo-controlled, salsalate 3,000 mg per day versus aspirin 3,900 mg per day; aspirin produced significant fecal blood loss over control and over salsalate (P under 0.01), and "blood loss with salsalate was not different than placebo". Mielants 1981 [5] (Scand J Rheumatol; reconstructed abstract): 42 patients on salsalate 5 g per day, radiolabeled fecal blood loss; abnormal loss in 2 of 42. Cryer 1990 [6] and Scheiman and Elta 1990 [7] (endoscopy in volunteers; text taken from a PubMed listing page summarized by the fetch tool, because OpenAlex held no abstract): minimal mucosal injury with salsalate or placebo and considerable injury with aspirin (P under 0.001) [6]; less gastroduodenal damage with salsalate than enteric-coated aspirin [7]. Case report Davies 1996 [19]: transfusion-dependent anemia for 8 years on indomethacin plus naproxen; after both were stopped and salsalate plus metronidazole started, hemoglobin stayed normal for at least 1 year (confounded). Class data: ASPREE secondary analysis [20], aspirin 100 mg daily versus placebo, 19,114 adults aged 70 or older, anemia HR 1.20 (95% CI 1.12 to 1.29), with ferritin analysis in a subset.

Mechanism

Gastric and small-intestinal mucosal injury from prostaglandin inhibition is plausible for any NSAID [1]. Salsalate is insoluble in gastric acid and hydrolyzes in the small intestine [1]; the retrieved sources do not test why blood loss is lower. Irreversible platelet inhibition by aspirin is absent for salsalate per the label [1].

Class cohort test

CLASS EXTRAPOLATION for [20] and the aspirin ferritin figures reported in [21]: cohort drug is aspirin only; salsalate was not in it. Differs in acetylation, platelet effect and measured blood loss. Capped at yellow; it does not raise the grade. [4][5][6][7][19] are salsalate-specific (Tier A or B) and measure blood loss or endoscopic injury, not iron.

Counter-evidence held on record

Counter-evidence and confounding The blood-loss studies are short (23 days; weeks) and several date from 1978 to 1990; they measure blood loss, not iron status. Hard Rule 2: anemia is not iron deficiency, and blood loss is not iron deficiency. The 1996 case report is one patient with two changes at once [19]. The 2016 trial of salsalate in older adults with anemia stopped early for renal toxicity and its hemoglobin result was not in the abstract record [18]. Framing caution Not an iron deficiency claim for salsalate and no iron supplementation is implied. New iron deficiency in an adult warrants evaluation for GI blood loss rather than assuming a drug cause. Do not import the aspirin anemia signal onto salsalate.

Limits on what this shows

• Why no grade: Dual condition met: no Tier B or C evidence on copper for salsalate was found that could support a yellow claim, and no circulating consumer claim about salsalate and copper was found to rebut (the consumer-claims search could not be run). This is a disclosed gap, not a finding of safety. Copper is kept independent of Iron (Hard Rule 7). Re-check when PubMed access is restored: read the plain "salsalate copper" listing first, then the aspirin and copper complexing lead (Brumas et al., acetylsalicylic acid, Tier C at best) and copper-salicylate complex literature, which would be moiety-level only and capped at yellow if salsalate is absent from the cohort.

What was searched

PubMed search listings were attempted with salsalate AND (zinc OR copper OR ceruloplasmin OR iron OR ferritin) and a second copper and zinc phrasing: both returned an "Invalid form" error page, or were rejected with 429 on repeat. The Renaud 2024 aspirin and warfarin review [21], read via summarizing fetch, reports one study on aspirin and copper (Brumas et al.): aspirin "can bind with transition elements such as copper", which was found to potentiate anti-inflammatory effect. That primary was not opened, concerns acetylsalicylic acid, and no copper status, absorption or excretion outcome is reported in the summary, so it does not support even a theoretical copper effect for salsalate. The label sections read list no copper effect [1].

Channel outcomes

PubMed listing: "Invalid form" x3 for mineral and vitamin phrasings, plus 429s. OpenAlex works-search: 429. Crossref and Semantic Scholar searches: 429. WebSearch: budget exhausted. Europe PMC not reachable under the org policy. Tooling-limited rather than a proven literature zero.

Repair re-check 2026-10-07

After waiting several minutes between attempts, three PubMed search pages that loaded returned zero results: "salsalate ceruloplasmin" and "salsalate cupric" (each page stated the search was processed without automatic term mapping because it retrieved zero results). The plain "salsalate copper" PubMed search was attempted about six times (listing and format=pubmed variants) and returned 429 or a proxy rejection every time, so it was never read; "salsalate copper complex" and "salsalate Cu" were rejected with 429; "salsalate trace elements" returned a reCAPTCHA page. WebSearch was still budget-exhausted, a Crossref title search returned 429, and an OpenAlex works-search returned 429. No Tier B or Tier C source for salsalate and copper was found or opened, so no upgrade was possible. The zero-result pages support a narrow literature gap for the ceruloplasmin and cupric phrasings only; the broad copper phrasing remains tooling-limited.

Counter-evidence held on record

Counter-evidence and confounding No competing explanation can be assessed without a copper measurement in salsalate users. Framing caution Not a reassurance and not a deficiency or excess claim. No copper supplementation or avoidance is implied.

Limits on what this shows

• Null findings: No human salsalate microbiome, 16S or C. difficile study found in the listing queried. Diarrhea in the older-adult anemia trial [18] is a clinical event, not a microbiome measurement.
• At-risk subgroups: Not defined.

Regulatory status

Label adverse reactions include diarrhea among post-marketing reports; no microbiome statement [1].

Population evidence

Zhang 2020 [17] (Front Pharmacol; Zucker diabetic fatty rats on a high-fat diet with or without salsalate; abstract via reconstructed OpenAlex record, dose and group sizes not in the record): all placebo-treated rats developed diabetes versus 10% of salsalate-treated rats with impaired glucose tolerance; salsalate lowered plasma glucagon and insulin and improved insulin sensitivity and beta-cell function; the proposed route was increased microbial diversity, amelioration of gut dysbiosis, restored intestinal epithelial junctions, less endotoxin entry into blood and less inflammation.

Mechanism

Proposed by the authors as a gut-barrier and inflammation effect; not shown to be causal.

Class cohort test

Salsalate was the test drug [17]. No class extrapolation used.

Counter-evidence held on record

Counter-evidence and confounding Single rodent model of diet-induced diabetes, high-fat feeding, abstract only. Rodent flora and exposure do not map to people. Label-listed diarrhea and the diarrhea in [18] leave a human GI-flora question unmeasured. Framing caution Not a probiotic indication and not a deficiency claim. The finding points in the protective direction in rats only.

Limits on what this shows

• Null findings: Muscle mitochondrial coupling unchanged and no capacity effect that explained the fuel shift [10]. Fatigue is not among the label adverse reactions read; tinnitus, nausea, hearing impairment, rash and vertigo are [1].

Regulatory status

No energy-metabolism statement in the label sections read [1].

Population evidence

Same sources as Mitochondrial burden, read here for energy output. Humans [10]: energy expenditure up about 18% (basal) and 16% (insulin-stimulated) after 4 days at 4,000 mg per day, n = 9, crossover. Mice [11]: higher whole-body oxygen consumption and liver lipid down about 55% with uncoupled respiration; [12]: brown adipocyte uncoupled respiration up. None reports ATP content, ATP synthesis rate or P/O ratio in the abstracts read.

Mechanism

Proton leak (uncoupling) is named in mice [11][12]. Whether it lowers cellular ATP at human exposure is not shown; the human coupling measurement was null [10].

Hard Rule 7 note

Kept independent of Mitochondrial burden and of Oxygen delivery/utilization. It shares the same sources and mechanism as the mitochondrial entry (Overlap map).

Class cohort test

Salsalate was in the cohorts [10][11][12]. No class extrapolation is used. The poisoning-level uncoupling reference [22] does not name salsalate and is context only.

Counter-evidence held on record

Counter-evidence and confounding Energy expenditure is an indirect proxy; a higher expenditure can coexist with normal ATP. The mouse results used pharmacologic exposures and the authors present them as benefit [11]. Framing caution Not a "low energy" claim for salsalate. No source supports a fall in ATP production; this entry records a mechanism that was not tested at the ATP level.

Limits on what this shows

• Null findings: G6PD: the salsalate label is silent [1]; the StatPearls G6PD deficiency chapter lists dapsone, methylene blue, pegloticase, phenazopyridine, primaquine, rasburicase, tafenoquine and toluidine blue as high-risk and nitrofurantoin and primaquine (medium dose) as medium-risk, and does not mention aspirin, salicylates or NSAIDs [23]. A PubMed search for salicylate with G6PD and hemolysis returned only an "Invalid form" error page (tooling failure, not a zero result). Methemoglobin: the one salsalate-term listing hit was a 1949 in vitro paper on salicylate and methemoglobin denaturation (title seen only, not opened).

Regulatory status

Label hematologic precaution: "Anemia is sometimes seen in patients receiving NSAIDs" (class wording) [1]. The label does not mention hemolysis, G6PD deficiency or methemoglobinemia in the sections read [1].

Population evidence

Goldfine 2013 [8] (Ann Intern Med; reconstructed OpenAlex abstract; journal page returned 403): 286 randomized, 21 sites, salsalate 3.5 g per day versus placebo for 1 year; anti-inflammatory effects included lower leukocyte and lymphocyte counts and "increased adiponectin and hematocrit"; uric acid and triglycerides fell, weight and LDL rose, kidney function was stable. The abstract record gives no hematocrit magnitude or hemoglobin value. Cohen HJ 2016 [18] (J Am Geriatr Soc; reconstructed abstract): phase II trial of salsalate in older adults with unexplained anemia, stopped after 11 participants because of adverse events including acute renal failure and diarrhea (creatinine 1.3 to 5.7 mg/dL in one participant; recurrence on rechallenge in another); the abstract record reports no hemoglobin result. Davies 1996 [19]: see Iron. Class: ASPREE [20] anemia HR 1.20 for aspirin, see Iron.

Mechanism

None established for any oxygen-delivery effect. The hematocrit rise is unexplained in the record read; the authors place it with anti-inflammatory effects [8]. Salicylate at toxic exposure disturbs oxidative metabolism [22], but nothing ties that to oxygen delivery at therapeutic exposure.

Hard Rule 2 note

Hematocrit and anemia are clinical or laboratory outcomes, not nutrient states. Any anemia effect is expected to run through blood loss or inflammation and is covered under Iron; it is not counted here as independent evidence (Overlap map). Renal failure in [18] is drug toxicity.

Class cohort test

[8] and [18] are salsalate cohorts. [20] is CLASS EXTRAPOLATION (aspirin only; salsalate not in cohort), yellow cap.

Counter-evidence held on record

Counter-evidence and confounding Hematocrit moved in the direction opposite to an anemia concern in [8], but the record lacks magnitude and a mechanism and this is a secondary outcome in a trial designed for glycemia. The anemia trial [18] was in a population with anemia already, so confounding by indication is presumed (Hard Rule 6) and it cannot show the drug caused or cured anemia. Anemia in older adults has many causes. Framing caution Not an oxygen-delivery deficiency and not a hemolysis claim. G6PD status should not be inferred from this dossier; the question is unanswered rather than cleared.