Piroxicam

Limits on what this shows

• Null and reverse findings: Burch 1983 [15] (Tier B): carboxylic acid NSAIDs (indomethacin, meclofenamate, naproxen) inhibited calcium uptake into rat and hamster liver mitochondria with ID50 values of 25 to 90 micromolar, whereas piroxicam, a non-carboxylic oxicam, had no effect at concentrations up to 300 micromolar. This is calcium transport, not respiration or uncoupling, so it does not exclude uncoupling.

Regulatory status

No mitochondrial statement in the US label sections retrieved [1].

Population evidence

None in humans. Abdeen 2020 [16] (Tier B, rats, 28 days, piroxicam 7 mg/kg i.p. alone or with coenzyme Q10 10 mg/kg): piroxicam produced severe gastric ulceration and liver and kidney damage with oxidative stress, and the abstract identifies "mitochondrial dysfunction and excess generation of reactive oxygen species" as contributing mechanisms (abstract via summarizing fetch; no mitochondrial endpoint values read). Herminghaus 2020 [27]: an indomethacin study whose discussion states aspirin, naproxen and piroxicam show a similar uncoupling effect on hepatic mitochondria; this is a secondary statement without piroxicam data and is Tier C at most. Moreno-Sanchez 1999 [28] (rat heart and liver mitochondria) and Nadanaciva 2013 [29] (isolated mitochondria, hepatocytes, zebrafish) list piroxicam among the NSAIDs studied in OpenAlex metadata; the Moreno-Sanchez abstract was later opened and its piroxicam result is recorded under ATP production [28], while no Nadanaciva abstract text was obtainable, so no piroxicam result is cited from [29].

Mechanism

Plausible but not shown at therapeutic exposure: protonophore-like uncoupling by acidic NSAIDs is described for the class [27]. Free plasma piroxicam is low because of about 99% protein binding [1].

Counter-evidence held on record

Counter-evidence The rat studies use doses and routes (i.p.) that do not map to human exposure, the cited uncoupling sentence is secondary, and the direct in vitro studies were not opened. Framing caution Mechanistic plausibility only. This entry does not support any mitochondrial or antioxidant supplement (including coenzyme Q10, which appears in the rat study only as an experimental protectant) and is not a deficiency claim.

Limits on what this shows

• Null and reverse findings: Sordelli [18] shows a protective, antioxidant-like activity only when piroxicam is complexed with copper, and copper alone matched it, so this does not show an antioxidant effect of piroxicam itself. No human null study was found.

Regulatory status

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

Population evidence

None in humans. Abdeen 2020 [16] (rats, 28 days, 7 mg/kg i.p.): malondialdehyde raised, glutathione and catalase activity lowered, with CoQ10 attenuating the injury. Yunita 2025 [17] (30 male Wistar rats in 5 groups, piroxicam 20 mg/kg orally, then 14 days of garlic extract in treatment groups; ELISA): the piroxicam-only positive control had lower SOD and higher MDA, which garlic extract reversed (p under 0.05). Sordelli 1993 [18] (mouse and in vitro): piroxicam-copper complexes showed SOD-like activity (inhibition of superoxide-mediated cytochrome c reduction), and copper alone had comparable antioxidant effects.

Mechanism

Proposed in rat tissue injury (reactive oxygen species generation with mitochondrial dysfunction) [16]; not demonstrated in humans.

Counter-evidence held on record

Counter-evidence Both rat studies are toxicology models at high doses, are by research groups testing a protective agent, and measure tissue markers rather than circulating antioxidant nutrients. Framing caution This is a tissue-injury marker finding in animals, not evidence that piroxicam depletes any antioxidant nutrient. It supports no antioxidant supplement and no dose is recommended.

Limits on what this shows

• Null findings: In the volunteer trial the primary covariance analysis showed no significant fecal blood loss difference across groups [10]. Two sources that looked relevant were excluded after opening: a rheumatic disease iron deficiency anemia endoscopy series of 104 patients [13] and a spondyloarthritis anemia study of 107 patients [14] both discussed NSAIDs generally without naming piroxicam (the spondyloarthritis study found only 1 patient with confirmed NSAID gastropathy and anemia tracking inflammatory markers). The Moore 2008 systematic review of fecal blood loss [31] did not name piroxicam in its abstract and is not used.
• At-risk subgroups: Older adults, people with prior ulcer or GI bleeding, people on anticoagulants, antiplatelets, SSRIs or SNRIs (label interaction warnings), longer exposure, and CYP2C9 polymorphism carriers with higher exposure [1][2].

Regulatory status

US label sec 5.2: upper GI ulcers, gross bleeding or perforation in approximately 1% of patients treated for 3 to 6 months and about 2% to 4% treated for one year; sec 5.12: anemia has occurred in NSAID-treated patients, possibly from occult or gross blood loss, fluid retention or an incompletely described effect on erythropoiesis, with hemoglobin or hematocrit monitoring if signs of anemia appear. Iron itself is not mentioned [1]. The UK SmPC states piroxicam is not a first-line NSAID and that observational evidence suggests a high risk of serious GI toxicity relative to other NSAIDs; it contraindicates use with active ulcer or GI bleeding and lists anemia as common [2].

Population evidence

Castellsague 2012 SOS meta-analysis [11]: systematic review and meta-analysis of observational studies (cohort and case-control, MEDLINE 1980 to May 2011), piroxicam pooled RR 7.43 (95% CI 5.19 to 10.63) for upper GI complications, among the highest of the NSAIDs listed (celecoxib 1.45, ibuprofen 1.84, diclofenac 3.34, meloxicam 3.47 in the same list). Values came from a reconstructed abstract through a summarizing fetch; only the piroxicam figure is used. Castellsague 2012 Friuli cohort [12]: 588,827 NSAID users, 3,031 upper GI complications, piroxicam in the highest band (RR 5 or more) with the exact RR not in the abstract. Patoia 1996 [10]: randomized double-blind placebo-controlled trial, 51 healthy men, 4 weeks, piroxicam 20 mg versus meloxicam 7.5 and 15 mg versus placebo; analysis of covariance found no significant fecal blood loss difference between groups, but piroxicam exceeded placebo by Student t-test on weighted means, endoscopy scores exceeded placebo for all active arms (P under 0.01), and 6 piroxicam volunteers withdrew early for poor endoscopic scores while no meloxicam volunteer did. All three measure GI outcomes, not iron.

Mechanism

Present. NSAID inhibition of prostaglandin synthesis damages the gastric and small-bowel mucosa and impairs platelet function (label: increased bleeding risk, anticoagulant and antiplatelet interactions) [1]. The iron loss itself (ferritin, transferrin saturation, serum iron) was not measured in any retrieved piroxicam cohort.

Hard Rule 2 and 6 checks

GI bleeding and anemia are clinical outcomes, not iron deficiency; iron deficiency anemia needs measured iron indices. Confounding by indication is presumed: rheumatoid disease and chronic inflammation lower hemoglobin independently, and the spondyloarthritis study [14] found anemia tracked inflammatory markers. Meloxicam case reports of small-bowel ulceration with iron deficiency anemia (cited in the b013 dossier) are meloxicam-specific and were not carried over.

Counter-evidence held on record

Counter-evidence and confounding The only controlled blood-loss study found no significant difference on its primary analysis, was in young healthy men for 4 weeks, and used piroxicam as an active comparator for meloxicam [10], so it does not test piroxicam against non-use in older arthritis patients. The GI risk estimates are from observational data with channeling bias (higher-risk patients may receive piroxicam). Framing caution This is a toxicity pathway (blood loss), not an absorption or intake deficiency. New iron deficiency or melena on piroxicam calls for evaluating the bleeding source and the prescription, not an iron supplement by default. No iron dose is recommended here.

Limits on what this shows

• Null findings: A 2011 study of 32 rheumatoid arthritis patients [20] found elevated serum copper and reduced zinc and selenium versus 52 controls, and no significant change with NSAIDs as a group (piroxicam not named). Serum copper in these patients is driven by the inflammatory disease (Hard Rule 6), so it cannot be attributed to the drug.

Regulatory status

No copper statement in the label sections retrieved [1][2].

Population evidence

None in humans. Cini 2007 [19] (Tier B, in vitro chemistry): the record text states the oxicam work began from cytotoxic testing at the US National Cancer Institute of [Cu(HPIR)2(DMF)2], where HPIR is piroxicam, with growth inhibition GI50 values as low as 20 micromolar and a mean of 54.4 micromolar across cell lines; the structural work on isoxicam, meloxicam and cinnoxicam copper complexes is separate. Sordelli 1993 [18]: piroxicam alone 0.64 mg/kg and its copper complex equally reduced leukocyte migration and lung myeloperoxidase in mice. Both used preformed complexes.

Mechanism

Present for the laboratory complex only: piroxicam coordinates copper(II) to form [Cu(HPIR)2(DMF)2] [19]. No evidence that piroxicam binds copper in the body at therapeutic exposure.

Counter-evidence held on record

Counter-evidence and confounding Coordination chemistry in a flask or a cell line is not copper handling in a patient, and the one clinical trace-element study did not name piroxicam. Framing caution Mechanistic plausibility only. This is not a copper deficiency or excess claim and supports no copper or zinc supplement.

Limits on what this shows

• Null findings: The human class review is a null for alpha diversity and beta diversity [25]. An NSAID enteropathy 16S review and a Maier 2018 in vitro non-antibiotic drug screen appeared in listings but were not opened, so whether piroxicam was tested there is unknown. No piroxicam C. difficile or dysbiosis report was retrieved.

Regulatory status

No microbiome statement in the label sections retrieved [1][2].

Population evidence

Cohort test result: piroxicam was NOT in the cohort of either source; both are Tier C for this drug and capped at yellow. Hutka 2021 [26] (rats, 4-week oral ketorolac only): loss of Firmicutes families (Paenibacillaceae, Clostridiales Family XIII, Christensenellaceae) and bloom of Enterobacteriaceae in the small intestine, without intestinal damage or delayed peristalsis. Le Bastard 2018 [25] (systematic review of non-antibiotic drugs, Tier C): metformin and NSAIDs were not associated with significant alpha diversity change, and beta diversity changed significantly for all drug classes studied except NSAIDs (reconstructed abstract; individual NSAIDs not named).

Mechanism

Plausible for the class (altered bile acid and mucosal environment) but unstudied for piroxicam; biliary elimination of piroxicam [7] raises the possibility of luminal exposure, which is not evidence of an effect.

Counter-evidence held on record

Counter-evidence The only human source is null, the animal source studied a different NSAID, and neither contained piroxicam. Piroxicam enteropathy could change the microbiome indirectly; no data. Framing caution Class-level plausibility only. No probiotic or prebiotic recommendation follows from this entry.

Limits on what this shows

• Null findings: Burch 1983 [15] found that none of the NSAIDs tested, including piroxicam, inhibited calcium transport in human erythrocyte ghosts, a narrow membrane-function finding, not a measure of oxygen delivery. Piroxicam case reports of hemolytic anemia were searched for and none was retrieved (one search returned only an unrelated DRESS hepatitis report). A methemoglobin-specific query was not run before WebSearch ran out, so absence of a methemoglobin report is tooling-limited.

Regulatory status

US label sec 5.12 (anemia, monitoring) and postmarketing list including anemia, thrombocytopenia, leukopenia, hemolytic anemia, aplastic anemia and eosinophilia [1]. UK SmPC: anemia common; aplastic anemia and hemolytic anemia very rare [2]. Feldene injection label: reduction in hemoglobin and hematocrit not associated with obvious GI bleeding, exceptional cases of hemolytic anemia, rare aplastic anemia [3]. Health Canada Brexidol monograph: anemia may be due to fluid retention, GI blood loss or an effect on erythropoiesis; blood dyscrasias rare [4]. None of the label texts retrieved mentioned methemoglobin.

Population evidence

No human hemoglobin, hematocrit, red-cell mass or oxygenation study naming piroxicam was retrieved. Majeed and Abass 2017 [21] (Tier B, rats): 40 male Sprague Dawley rats, piroxicam 20, 30 or 40 mg/kg for 45 days; significant decreases in hemoglobin and packed cell volume at higher doses (P under 0.05), with liver and kidney damage on histology (summarizing fetch of the journal page; doses far exceed human equivalents and route details were not read). A 2026 rat study of piroxicam against anthracene-induced hematologic change [35] had no piroxicam-specific numeric result in its abstract and is not relied on.

Mechanism

Per label: occult or gross blood loss, fluid retention (hemodilution) or an incompletely described effect on erythropoiesis [1]. Immune hemolysis is a rare idiosyncratic event listed in labels; no mechanism study for piroxicam was found.

Hard Rule 2 check

Anemia is an outcome and is not filed as iron deficiency or as an oxygen-delivery nutrient deficiency. Any effect is mediated by the Iron entry (blood loss) and the Sodium entry (dilution) and is not counted twice.

Counter-evidence held on record

Counter-evidence and confounding Chronic inflammatory disease lowers hemoglobin independently. The label incidence of anemia was not retrievable with a comparator, and the rat doses are high. Framing caution A monitoring prompt (hemoglobin), not a deficiency claim. No supplement is implied.

Limits on what this shows

• Null and reverse findings: No intact-cell or tissue ATP measurement for piroxicam was found, and no human fatigue, lactate or exercise-capacity study. Burch 1983 [15] found no effect of piroxicam on mitochondrial calcium uptake up to 300 micromolar, a different endpoint that does not test ATP synthesis. In the same 1999 study naproxen had no effect on any mitochondrial parameter, so the effect is not class-wide.

Regulatory status

No ATP or energy-metabolism statement in the US label sections retrieved [1].

Population evidence

None in humans. Moreno-Sanchez 1999 [28] (Tier B, in vitro, piroxicam in the tested set; abstract opened through a PubMed listing page in the PubMed text format): the abstract states that diclofenac, piroxicam, indomethacin, naproxen, nabumetone, nimesulide and meloxicam were tested for effects on mitochondrial respiration, ATP synthesis and membrane potential, in preparations of mitochondria, submitochondrial particles, cells and whole heart, at micromolar concentrations. Except for nabumetone and naproxen, the other five drugs, which include piroxicam, "stimulated basal and uncoupled respiration, inhibited ATP synthesis, and collapsed membrane potential in mitochondria"; the authors conclude that "nimesulide, meloxicam, piroxicam, and indomethacin behaved as mitochondrial uncouplers." Which of the four preparation types gave the piroxicam result, the piroxicam concentrations and the effect sizes are not in the abstract and the full text was not opened. Nadanaciva 2013 [29]: OpenAlex metadata lists piroxicam among the drugs studied, but no abstract text was obtainable, so no result is cited from it.

Mechanism

Named and demonstrated in vitro: protonophore-like uncoupling dissipates the proton gradient that drives ATP synthase, so ATP yield per unit oxygen falls in isolated preparations [28]. Whether free piroxicam concentrations in people reach uncoupling levels is not shown; the drug is about 99% protein bound [1].

Hard Rule 2 and 7 note

This is the drug's own in vitro pharmacology, not a nutrient deficiency and not an energy-nutrient claim. It is kept as its own entry, separate from Mitochondrial burden and Oxygen delivery/utilization, and shares its source with the former (see Overlap map).

Counter-evidence held on record

Counter-evidence Isolated-mitochondria uncoupling at micromolar concentrations is a screening-level finding; cells can compensate for modest uncoupling and plasma protein binding lowers free exposure. The abstract does not state that the piroxicam effect occurred in intact cells or whole heart. Framing caution Mechanistic plausibility only, from one in vitro source. It supports no energy, mitochondrial or CoQ10 supplement, implies no fatigue claim, and is not a deficiency claim. No dose is recommended.