Naproxen

Limits on what this shows

• Null findings: Naproxen did not affect any measured mitochondrial parameter in [19]. Nadanaciva 2013 [38] (isolated mitochondria, hepatocytes, zebrafish) was opened and does not include naproxen among the NSAIDs tested, so it does not bear on this drug.

Regulatory status

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

Population evidence

None in humans. Moreno-Sanchez 1999 [19] (Tier B, in vitro, naproxen tested): the abstract states that, except for nabumetone and naproxen, the other drugs (diclofenac, piroxicam, indomethacin, nimesulide, meloxicam) stimulated respiration, inhibited ATP synthesis and collapsed membrane potential, and that naproxen did not affect any of the mitochondrial parameters measured; source tissue and concentrations were not in the abstract text. Salimi 2019, Drug Research [20] (Tier B, isolated rat heart mitochondria; naproxen, diclofenac and celecoxib): all increased ROS formation, collapsed membrane potential, caused mitochondrial swelling and lipid peroxidation, and depleted glutathione and ATP, with protection by permeability-transition-pore sealers and antioxidants; concentrations were not retrieved. An Int Pharm Acta 2018 abstract [21] with the same three drugs and heart mitochondria reports similar endpoints and may be an earlier report from the same line of work, so it is not counted as independent.

Mechanism

Contested. Uncoupling and respiratory chain effects that are well shown for several other NSAIDs were not seen with naproxen in [19]; ROS-mediated permeability transition was reported in [20][21].

Counter-evidence held on record

Counter-evidence and confounding The two positive studies and the one negative study use different preparations and unspecified concentrations, and the label contains no mitochondrial toxicity signal. Naproxen hepatotoxicity in people is a clinical outcome, not a mitochondrial measurement. Framing caution Mechanistic plausibility only, with a direct null. This entry does not support any mitochondrial or energy supplement and is not a deficiency claim.

Limits on what this shows

• Null and conflicting findings: SOD and catalase fell in Ilyas [15] but rose in Ahmad [23]; the human study showed an antioxidant-direction trend [24]. The rodent studies used doses and concentrations not matched to human free plasma levels (not retrieved).

Regulatory status

No oxidative-stress or antioxidant statement in the naproxen label sections retrieved [1].

Population evidence

Human: Orhan 1999 [24] (Tier A for these markers, in vivo): 16 healthy volunteers and 35 patients received naproxen, tiaprofenic acid, acemetacin or paracetamol for 15 days with a 1-week washout; erythrocyte catalase tended to increase and plasma malondialdehyde decreased somewhat; per-drug breakdown for naproxen and the naproxen group size were not in the abstract text retrieved. Rodent and in vitro: Yokoyama 1998 [22] (rat hepatocytes) oxidized glutathione (GSSG) rose during naproxen metabolism before cell injury, and glutathione depletion increased injury; Ahmad 2018 [23] (male Wistar rats; dose not retrieved) decreased GSH with increased SOD, catalase and lipid peroxidation, raised liver enzymes and comet-assay DNA damage; Ilyas 2020 [15] decreased red-cell SOD, catalase and glutathione peroxidase in vitro; Salimi 2019 [20] reported ROS, lipid peroxidation and glutathione depletion in isolated rat heart mitochondria. Reverse direction: Costa 2006 [25] (in vitro chemistry) found naproxen had the highest hydroxyl-radical scavenging among the arylpropionic NSAIDs tested and was also active against peroxynitrite and nitric oxide.

Mechanism

Named: oxidized glutathione (GSSG) formation during naproxen metabolism in hepatocytes [22]; mitochondrial ROS in isolated mitochondria [20]; direct radical scavenging [25]. None demonstrated at therapeutic human exposure.

Counter-evidence held on record

Counter-evidence and confounding Direction is inconsistent across models, the human study is small per drug and uncontrolled for naproxen alone, and marker changes in rodents often reflect high-dose toxicity. Framing caution Mixed or theoretical. Not a reason to take antioxidant supplements, and antioxidant supplementation has no support in these data. No dose is implied.

Limits on what this shows

• Null findings: Racz 2013 [6] (Tier A, naproxen 1000 mg/day, uncontrolled open-label pilot of patients with naproxen-induced small-bowel enteropathy, 10 completers): mean hemoglobin 131.1 +/- 5.8 g/L at screening and 127.1 +/- 7.8 g/L after 4 weeks, with no significant anemia developing. Average controlled fecal blood loss with naproxen is a few mL/day [5]. Both are short, small studies and do not test years of use.
• At-risk subgroups: Older adults, prior peptic ulcer or GI bleeding, anticoagulant (label: synergistic bleeding effect with warfarin) or aspirin co-use, and long or high-dose exposure [1]. In an NSAID-class capsule endoscopy cohort of 113 long-term users, age 65 or older and acid-suppressant use were associated with severe small-bowel damage and lower hemoglobin (naproxen not named) [39].

Regulatory status

Label sec 5.2: NSAIDs cause an increased risk of serious GI events including bleeding, ulceration and perforation; upper GI ulcers occurred in approximately 1% of patients treated for 3 to 6 months and about 2% to 4% at one year; elderly patients and those with prior ulcer or GI bleeding are at greater risk. Sec 5.12: anemia has occurred in NSAID-treated patients; monitor hemoglobin or hematocrit if signs or symptoms of anemia appear. Sec 5.14: consider periodic CBC and chemistry profile in long-term use, and periodic hemoglobin when initial hemoglobin is 10 g or less. Iron is not mentioned [1]. Two summarizer passes gave inconsistent anemia frequency wording (1% to 10% versus under 1%), so no anemia incidence is quoted.

Population evidence

Moore systematic review of 51-chromium fecal blood loss trials [5] (Tier A, naproxen in the cohort, fewer than 100 naproxen participants): naproxen mean 2.0 mL/day against placebo 0.76 mL/day (172 participants); the review noted a small proportion of individuals lose more than 5 or 10 mL/day on aspirin or NSAIDs (summarized text). McAlindon 1995 [7]: 13 healthy volunteers, naproxen 750 mg on day 1 and day 7; gastric microbleeding rose from 0.60 (placebo) to 2.15 and 1.75 microL per 10 min (p under 0.05), no adaptation by day 7. Goldstein 2005 [8]: 356 randomized subjects, capsule endoscopy; small-bowel mucosal breaks in 55% (mean 2.99 per subject) on naproxen plus omeprazole versus 16% on celecoxib and 7% on placebo (P under.001). Castellsague SOS meta-analysis [9]: naproxen pooled RR 4.10 (95% CI 3.22-5.23) for upper GI complications versus non-use, 28 observational studies. Bhala 2013 [10]: in 280 placebo-controlled NSAID trials (124,513 participants) naproxen upper GI complication ratio 4.22 (2.71-6.56). These are bleeding and injury outcomes, not iron measures.

Mechanism

Proposed, partly contested. COX-1 and COX-2 inhibition (label sec 12.1 [1]) is the class route to mucosal injury, and ulcers can bleed; the label text retrieved does not spell out the mucosal mechanism, and a 2026 mouse study argues prostaglandin suppression alone does not cause NSAID enteropathy [30]. Iron loss itself (ferritin, serum iron, transferrin saturation) was not measured in any naproxen study retrieved.

Hard Rule 2 and 6 checks

GI bleeding and anemia are clinical outcomes, not iron deficiency; iron deficiency needs measured iron indices. Confounding by indication is presumed (rheumatologic disease, aspirin or anticoagulant co-use, age). Direction is loss through bleeding, a drug-toxicity route, not an absorption defect. Goldstein 2011 hemoglobin analysis [11] was opened and does not qualify: naproxen was not in the analyzed cohort (celecoxib versus diclofenac).

Counter-evidence held on record

Counter-evidence and confounding Average controlled blood loss is low, the only naproxen hemoglobin series showed no significant drop, the label itself lists several possible anemia causes in the NSAID class, and no study compares ferritin before and after starting naproxen. GI-REASONS [12] included naproxen (21% of the nonselective comparator users) but the retrieved text reported no hemoglobin outcome. Framing caution This is a toxicity pathway (blood loss), not a dietary or absorption deficiency. New iron deficiency or dark stools on naproxen 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: No copper, ceruloplasmin or zinc-to-copper ratio data with naproxen found; rheumatoid arthritis serum zinc and copper search results reflected the disease, not the drug, and none was opened.

Regulatory status

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

Population evidence

None in humans. Hasan and Begum 2015 [33] (Tier B, in vitro chemistry): naproxen complexes with copper, cobalt, iron, silver and zinc were synthesized and tested for antimicrobial, cytotoxic and anthelmintic activity; the work is about complex properties, not copper physiology.

Mechanism

Chelation chemistry is a present mechanism class (A) in vitro. Label: greater than 99% albumin-bound at therapeutic levels [1], so the free drug available to chelate copper in vivo is not established.

Counter-evidence held on record

Counter-evidence and confounding Chemistry-only evidence at the very lowest end of yellow; no depletion measurement exists. Rheumatoid disease alters copper and zinc independent of treatment. Framing caution No copper supplement or restriction is implied. Copper is graded as its own entry and not combined with Iron.

Limits on what this shows

• Null findings: No human fecal 16S or metagenomic study dosing naproxen was retrieved (see Sources access note). The Cox-knockout result implies prostaglandin suppression alone does not cause the injury, which complicates a simple COX-mediated reading [30].

Regulatory status

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

Population evidence

None in humans for naproxen specifically. Blackler 2015, Am J Physiol Gastrointest Liver Physiol [27] (Tier B, rat): naproxen caused significant intestinal damage and inflammation, and the enteric microbiota of naproxen-treated rats was distinct from vehicle-treated rats; proton pump inhibitor administration caused dysbiosis. Blackler 2015, Br J Pharmacol [28] (Tier B, abstract-level): a hydrogen sulfide donor prevented or reversed naproxen-induced changes in the intestinal microbiota and reduced naproxen-induced increases in bile cytotoxicity. Wallace 2011, Gastroenterology [29] (Tier B, rat, naproxen and celecoxib studied): PPIs produced marked jejunal dysbiosis (about 80% reduction in Actinobacteria and Bifidobacteria) and exacerbated NSAID small-intestinal injury; the dysbiosis reported is the PPI effect. Barekat 2026, J Clin Invest [30] (Tier B mouse plus human cohort; abstract only): Cox-1 and Cox-2 double-knockout mice developed dysbiosis, cohousing with controls rescued dysbiosis and delayed naproxen-induced GI bleeding, and in the UK Biobank and All of Us cohorts coadministration of antibiotics with NSAIDs was associated with more GI bleeding (naproxen-specific human estimate not retrieved).

CLASS EXTRAPOLATION

Meta-analysis of 8 observational studies (over 39,000 subjects) of NSAIDs and C. difficile-associated diarrhea, pooled OR 1.41 (95% CI 1.06-1.87) [32]: Tier C, individual NSAIDs not analyzed, so whether naproxen was in the constituent cohorts is not determinable from the abstract; capped at yellow. Maseda 2019, mBio [31]: indomethacin in mice altered the microbiota and worsened C. difficile colitis; naproxen was not in the cohort, so it is Tier C for naproxen and was not relied on for any statement above.

Mechanism

Named in animals: naproxen-induced injury with altered bile cytotoxicity and a microbiome that amplifies injury, not explained by prostaglandin suppression alone [28][30]. Causal direction in people on therapeutic naproxen is not established.

Counter-evidence held on record

Counter-evidence and confounding Rodent doses and exposure durations were not retrieved, microbiota change coexists with mucosal injury (cause versus consequence not separated), and the human antibiotic-plus-NSAID signal is class-wide. Framing caution This is a mechanistic and context entry. It does not support a probiotic, fiber or antibiotic decision; GI symptoms, black stools or persistent diarrhea on naproxen belong with the prescriber.

Limits on what this shows

• Null findings: Racz 2013 [6]: hemoglobin did not fall significantly over 4 weeks of naproxen 1000 mg/day in enteropathy patients. No arterial oxygenation or red-cell oxygen-affinity study with naproxen was found.

Regulatory status

Label sec 5.12 anemia statement and hemoglobin or hematocrit monitoring [1]; sec 6 postmarketing lists hemolytic anemia, aplastic anemia, eosinophilia, leukopenia, granulocytopenia and thrombocytopenia [1] (summarized text).

Population evidence

No human measurement of hemoglobin oxygen affinity, red-cell mass or tissue oxygenation with naproxen was found. Mele 2023 [13] (Tier B, case report): 68-year-old with Evans syndrome after two weeks of increased naproxen intake, hemoglobin 6.1 g/dL, direct antiglobulin test positive for IgG and C3, improved after corticosteroids, transfusion and stopping naproxen. Lee 2014 [14] (Tier B, case report, abstract-level only): 42-year-old woman with dyspnea and dizziness, methemoglobinemia attributed to naproxen sodium taken with methocarbamol, alcohol, lorazepam, trazodone and paroxetine; the methemoglobin level and outcome were not retrieved and the exposure is heavily confounded. Ilyas 2020 [15] (Tier B, in vitro, species not stated in the retrieved text): naproxen sodium 1 to 25 microM for 48 h decreased erythrocyte superoxide dismutase, catalase and glutathione peroxidase activity at 10 to 25 microM (described by the authors as therapeutic) and triggered membrane blebbing, eryptosis and hemolysis. Iron-pathway anemia is covered in the Iron entry.

Mechanism

Three separate routes, none proven to lower oxygen delivery at therapeutic dosing: blood loss anemia (Iron entry), rare immune hemolysis with drug-dependent antibodies (case level), and in vitro oxidative injury to red cells.

Hard Rule 2 check

Anemia is an outcome, not a nutrient deficiency, and a drug-induced hemolysis case is an adverse reaction, not oxygen-delivery deficiency. Naproxen has no intended hemodynamic oxygen-demand effect to clarify. This entry is a context and monitoring flag only.

Counter-evidence held on record

Counter-evidence and confounding Hemolysis is rare (case reports and label postmarketing lists), the in vitro concentrations are not matched to free plasma levels, the methemoglobin case involved five co-ingested agents, and the label states no mechanism for anemia causation beyond blood loss, fluid retention and incompletely described effects. Framing caution Do not read this as an oxygen-carrying deficiency that supplements can fix. Unexplained anemia, jaundice or dark urine on naproxen is a prescriber question, and no iron, folate or antioxidant supplement is implied.

Limits on what this shows

• Null findings: ATP synthesis unchanged with naproxen in [19]. No tissue ATP, phosphocreatine, lactate or exercise-capacity study with naproxen was found.

Regulatory status

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

Population evidence

None in humans. ATP synthesis was an explicit endpoint in Moreno-Sanchez 1999 [19], where naproxen did not inhibit ATP synthesis or collapse membrane potential; Salimi 2019 [20] reported ATP depletion in isolated rat heart mitochondria exposed to naproxen, with the 2018 abstract [21] listing ATP concentration decrease. All three are the same evidence base as the Mitochondrial burden entry.

Mechanism

Contested: no uncoupling in [19]; ROS-mediated permeability transition with ATP loss in [20].

Hard Rule 7 and 9 note

Kept as its own entry, separate from Mitochondrial burden and Oxygen delivery/utilization; it is not independent evidence (see Overlap map).

Counter-evidence held on record

Counter-evidence and confounding Conflicting preparations and unspecified concentrations; no human measurement. Framing caution Mechanistic plausibility only. Fatigue or low energy on naproxen is not explained by these data, and no supplement is implied.