Gastrointestinal (GI) toxicity is the most common clinically significant adverse effect of NSAID (non-steroidal anti-inflammatory drug) therapy and the primary driver of NSAID-related morbidity and healthcare costs. The mechanisms are now well understood, risk stratification is evidence-based, and preventive co-therapy with proton pump inhibitors (PPIs) or selective COX-2 (cyclooxygenase-2) inhibitors substantially reduces but does not eliminate the risk. Understanding who is at high risk and why guides rational gastroprotective prescribing.
Mechanism of NSAID Gastropathy. The gastric mucosa is protected from luminal acid by an interconnected set of prostaglandin (PG)-dependent mechanisms collectively termed the mucosal defense barrier. Prostaglandin E2 (PGE2) and prostacyclin (PGI2), synthesized constitutively by COX-1 (cyclooxygenase-1) in gastric epithelial cells, mucus-secreting cells, and submucosal blood vessels, stimulate mucus and bicarbonate secretion from surface epithelial cells, maintain mucosal blood flow through vasodilation of submucosal arterioles, inhibit acid secretion by parietal cells via EP3 (prostaglandin E receptor subtype 3) receptors, and promote epithelial restitution after injury. When NSAIDs suppress COX-1-derived prostaglandins, all four components of mucosal defense are simultaneously impaired: mucus thins, bicarbonate output falls, mucosal blood flow decreases, and epithelial repair slows. The result is a mucosa rendered vulnerable to injury by luminal acid, bile salts, and Helicobacter pylori. This mechanism is systemic, not topical: even parenterally administered NSAIDs or enteric-coated formulations cause equivalent rates of mucosal injury as oral uncoated preparations, because the gastropathy is driven by prostaglandin suppression in the mucosa rather than by direct topical acid injury from the drug particle.1
Spectrum of GI Injury. NSAID-associated GI injury spans a spectrum from mucosal erosions and subepithelial hemorrhages (visible endoscopically in up to 40% of regular NSAID users but mostly asymptomatic) to clinically significant peptic ulcers, ulcer complications (bleeding, perforation, obstruction), and lower GI complications (small bowel ulceration and stricture, colonic ulceration). Upper GI bleeding is the most clinically important complication, carrying a case fatality rate of approximately 5 to 10% in older series, though outcomes have improved with widespread PPI (proton pump inhibitor) use and endoscopic hemostasis. The clinical deceptiveness of NSAID gastropathy lies in the frequent absence of dyspeptic symptoms preceding serious complications: studies consistently show that a substantial proportion of patients who experience NSAID-associated GI bleeding had no premonitory upper abdominal pain, partly because NSAIDs themselves have analgesic properties that mask ulcer pain. This makes symptom-based risk assessment unreliable and underscores the importance of prophylactic co-therapy in high-risk patients regardless of symptom status.2
GI Risk Stratification. Multiple evidence-based risk factors for NSAID-associated GI complications have been identified and validated in large epidemiological studies. Prior peptic ulcer disease or upper GI bleeding is the single strongest risk factor, increasing the relative risk of a serious GI event by approximately 5-fold; prior ulcer complicated by bleeding confers even higher risk. Age above 65 years is an independent risk factor, with risk increasing progressively with advancing age, partly because of reduced mucosal regenerative capacity and higher prevalence of comorbidities. Concomitant use of anticoagulants (warfarin, direct oral anticoagulants) multiplies the bleeding risk multiplicatively rather than additively when combined with NSAID-induced mucosal injury. Concurrent corticosteroid use roughly doubles the risk of GI complications compared to NSAIDs alone; the combination of corticosteroids plus NSAIDs is particularly hazardous and should trigger routine PPI co-prescription. High doses of NSAIDs and use of two NSAIDs simultaneously (including the combination of a standard NSAID with low-dose aspirin) confer higher risk than standard monotherapy doses. H. pylori infection is an independent risk factor for peptic ulcer disease that synergizes with NSAID use; testing for and eradicating H. pylori before initiating long-term NSAID therapy is recommended in high-risk patients and reduces but does not eliminate NSAID-associated ulcer risk.23
Preventive Strategies. Three pharmacological strategies reduce NSAID-associated GI toxicity: selective COX-2 inhibition (celecoxib), PPI co-therapy, and misoprostol. Celecoxib at therapeutic doses causes significantly less endoscopic gastric and duodenal ulceration than non-selective NSAIDs, as demonstrated in the CLASS (Celecoxib Long-Term Arthritis Safety Study) trial; however, its GI protective advantage is attenuated or lost in patients taking concomitant low-dose aspirin, because aspirin suppresses COX-1 in the gastric mucosa and negates the gastroprotective benefit of COX-2 selectivity. PPIs are the most effective and best-tolerated co-therapy for reducing symptomatic NSAID ulcers and upper GI bleeding risk; they reduce the relative risk of endoscopic ulcers by approximately 75% and are the preferred gastroprotective agent in most clinical guidelines. Misoprostol, a synthetic PGE1 (prostaglandin E1) analogue that directly replaces the mucosal prostaglandin depleted by NSAIDs, was the first proven gastroprotective agent and is the only strategy that directly addresses the underlying mechanism; however, its clinical use is limited by a high rate of dose-dependent GI side effects including diarrhea and abdominal cramping that reduce tolerability and adherence. For high-risk patients, the combination of a COX-2 selective inhibitor plus a PPI provides the greatest reduction in GI risk.3
High GI risk (any one): prior peptic ulcer/GI bleed; age above 65; concurrent anticoagulant; concurrent corticosteroid; high-dose NSAID; two NSAIDs simultaneously (including low-dose aspirin). Moderate GI risk (2 or more moderate factors): age 60–65, dyspepsia, H. pylori positive. Low GI risk: none of the above. High-risk strategy: celecoxib OR non-selective NSAID + PPI. Highest-risk strategy: celecoxib + PPI. Test and treat H. pylori in patients requiring long-term NSAIDs. PPIs: preferred over misoprostol due to better tolerability; H2 antagonists less effective than PPIs for NSAID ulcer prevention.
The cardiovascular risk of each NSAID (non-steroidal anti-inflammatory drug) is a class-wide phenomenon driven by the disruption of prostanoid-mediated vascular homeostasis, but it is not uniform across agents. The cardiovascular signal is greatest with selective COX-2 (cyclooxygenase-2) inhibitors, intermediate with most non-selective agents, and lowest with naproxen. Understanding the mechanistic basis of this risk and the key trial evidence is essential for safe NSAID prescribing in patients with or at risk for cardiovascular disease.
Mechanism of Cardiovascular Risk. The cardiovascular risk of NSAIDs arises from disruption of the balance between two opposing prostanoids: prostacyclin (PGI2), produced by vascular endothelial cells predominantly via COX-2, and thromboxane A2 (TXA2), produced by platelets exclusively via COX-1 (cyclooxygenase-1). Under physiological conditions, endothelial PGI2 inhibits platelet aggregation, promotes vasodilation, and exerts antiproliferative effects on vascular smooth muscle, while platelet TXA2 promotes aggregation and vasoconstriction. These two prostanoids maintain a functional equilibrium that restrains both thrombosis and excessive vasodilation. Selective COX-2 inhibition suppresses endothelial PGI2 production while leaving platelet TXA2 synthesis (a COX-1 product) entirely intact. The resulting PGI2 deficiency in the setting of unopposed TXA2 activity creates a prothrombotic and vasoconstrictive vascular environment that elevates the risk of myocardial infarction (MI), ischemic stroke, and sudden cardiac death. This mechanism was proposed by FitzGerald and colleagues on pharmacological grounds before the clinical evidence emerged, making the subsequent trial findings a prospectively predicted pharmacological consequence rather than an unexpected toxicity.4
Key Trial Evidence. The cardiovascular risk of selective COX-2 inhibitors was first identified in the VIGOR (Vioxx Gastrointestinal Outcomes Research) trial, a randomized controlled trial of rofecoxib versus naproxen in patients with rheumatoid arthritis (RA). VIGOR demonstrated a 5-fold increase in the rate of myocardial infarction (MI) with rofecoxib compared to naproxen; at the time of trial publication, the excess risk was attributed to a cardioprotective effect of naproxen rather than a harmful effect of rofecoxib. The APPROVe (Adenomatous Polyp Prevention on Vioxx) trial, designed to evaluate rofecoxib for colorectal polyp prevention, provided unequivocal evidence of cardiovascular harm: rofecoxib doubled the risk of serious cardiovascular events compared to placebo after 18 months of use. This demonstrated that the increased cardiovascular risk was an intrinsic property of the drug rather than an artifact of comparison to a cardioprotective reference agent; rofecoxib was voluntarily withdrawn from the market in September 2004.4
Valdecoxib was subsequently withdrawn in 2005 after demonstrating increased rates of cardiovascular events in coronary artery bypass graft (CABG) surgery patients. Celecoxib, the remaining approved selective COX-2 inhibitor, carries a black box warning for cardiovascular risk and was evaluated in the PRECISION (Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen or Naproxen) trial, which found celecoxib non-inferior to ibuprofen and naproxen for cardiovascular safety in a high-cardiovascular-risk arthritis population, suggesting that at moderate doses celecoxib does not carry substantially greater cardiovascular risk than the non-selective comparators studied.45
Non-Selective NSAIDs and Cardiovascular Risk. The cardiovascular risk of non-selective NSAIDs, while generally lower than that of selective COX-2 inhibitors, is real and clinically significant at high doses and in patients with underlying cardiovascular disease. The CNT (Coxib and traditional NSAID Trialists) meta-analysis, pooling data from over 280 randomized trials with 124,513 participants, found that high-dose diclofenac (150 mg/day) and high-dose ibuprofen (2,400 mg/day) each increased the rate of major vascular events by approximately one-third compared to placebo, a magnitude comparable to the selective COX-2 inhibitors. High-dose diclofenac and ibuprofen doubled the rate of vascular death. Naproxen (1,000 mg/day) did not significantly increase major vascular events compared to placebo in the CNT meta-analysis, and observational studies consistently show naproxen has the most favorable cardiovascular profile among commonly used NSAIDs. The proposed mechanism for naproxen's relative cardiovascular safety involves its long half-life (12 to 17 hours) providing sustained platelet COX-1 inhibition with incomplete recovery between doses, producing a partial aspirin-like antiplatelet effect that partially counterbalances the endothelial PGI2 suppression.5
Clinical Management of Cardiovascular Risk. The US Food and Drug Administration (FDA) requires a class-wide black box warning for all prescription NSAIDs stating that they increase the risk of serious cardiovascular events including MI and stroke. Several clinical principles follow from the evidence. NSAIDs should be avoided in patients with established cardiovascular disease (prior MI, stroke, coronary artery disease) whenever alternative analgesics suffice; acetaminophen, topical NSAIDs, and short courses of low-dose NSAIDs are preferable in this population. When an oral NSAID is required in a cardiovascular-risk patient, naproxen is preferred based on its favorable cardiovascular profile. Celecoxib should not be used in patients with established cardiovascular disease or multiple cardiovascular risk factors. All NSAIDs can attenuate the antihypertensive effects of angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), beta-blockers, and diuretics through sodium retention and vasoconstriction mechanisms, and can precipitate worsening heart failure in patients with reduced cardiac reserve. The concurrent use of NSAIDs with aspirin for cardiovascular prevention does not eliminate the cardiovascular risk of the NSAID and adds to the gastrointestinal (GI) toxicity burden.5
VIGOR (rofecoxib vs naproxen): 5-fold increase in MI with rofecoxib; initially misattributed to naproxen cardioprotection. APPROVe (rofecoxib vs placebo): 2-fold increase in serious CV events vs placebo after 18 months; rofecoxib withdrawn 2004. CNT meta-analysis: high-dose diclofenac and ibuprofen each increase major vascular events by ~1/3 vs placebo; naproxen does not significantly increase vascular events. PRECISION (celecoxib vs ibuprofen vs naproxen): celecoxib non-inferior for CV safety at moderate doses. Class-wide principle: CV risk is proportional to COX-2 selectivity and dose; naproxen is the preferred NSAID in high CV-risk patients.
Renal toxicity is a class-wide property of each NSAID (non-steroidal anti-inflammatory drug) regardless of COX (cyclooxygenase) selectivity, because both COX-1 (cyclooxygenase-1) and COX-2 (cyclooxygenase-2) contribute to prostaglandin synthesis in the kidney. The clinical manifestations range from reversible acute kidney injury (AKI) in susceptible patients to sodium and water retention, hyperkalemia, and exacerbation of hypertension. Hepatotoxicity is predominantly a concern with diclofenac and sulindac rather than the class as a whole.
Renal Prostaglandin Physiology. Under basal conditions in euvolemic healthy individuals, renal prostaglandins play a relatively minor role in maintaining glomerular filtration rate (GFR) and renal blood flow; NSAIDs in such patients generally cause only trivial and clinically insignificant reductions in GFR. However, in physiologically stressed states including actual or effective volume depletion (hemorrhage, dehydration, vomiting, diarrhea, excessive diuretic use), heart failure, hepatic cirrhosis with ascites, nephrotic syndrome, and chronic kidney disease (CKD), the kidney becomes heavily dependent on prostaglandin-mediated vasodilation of the afferent arteriole to maintain adequate renal perfusion pressure. In these states, elevated angiotensin II and catecholamine levels cause intense systemic and renal vasoconstriction; the compensatory vasodilatory response in the afferent arteriole is mediated by locally produced prostaglandin E2 (PGE2) and prostacyclin (PGI2), synthesized by both COX-1 and COX-2 in the glomerulus, afferent arteriole, and medullary interstitium. NSAID-mediated suppression of these renal prostaglandins removes the vasodilatory buffer, allowing unopposed angiotensin II to constrict the afferent arteriole, sharply reducing GFR and renal blood flow and precipitating AKI.7
Clinical Presentations of Renal Toxicity. NSAID-associated renal toxicity presents through several distinct mechanisms. Hemodynamic AKI is the most common presentation, characterized by a rapid rise in serum creatinine within days of NSAID initiation in a susceptible patient; it is generally reversible with drug discontinuation and volume repletion. Sodium and water retention occur because renal prostaglandins normally oppose sodium reabsorption in the collecting duct and loop of Henle; NSAID-mediated prostaglandin suppression enhances sodium and water reabsorption, causing edema, weight gain, and worsening of hypertension, effects that can unmask latent hypertension and counteract antihypertensive medications. Hyperkalemia results from two mechanisms: reduced aldosterone release (because aldosterone secretion is partly prostaglandin-dependent), and in patients taking potassium-sparing diuretics or renin-angiotensin-aldosterone system (RAAS) inhibitors, additive blockade of the same pathway. Analgesic nephropathy from mixed analgesic abuse (involving phenacetin, aspirin, and caffeine) is a historically recognized cause of chronic interstitial nephritis and papillary necrosis but is rarely seen with contemporary single-agent NSAID use at standard doses. Membranous nephropathy and minimal change disease are rare idiosyncratic immune-mediated renal complications of NSAIDs, particularly with fenoprofen, that present with nephrotic syndrome and respond to drug discontinuation.7
Risk Factors and Clinical Management. The patients at highest renal risk from NSAIDs are those with effective circulating volume depletion (regardless of cause), pre-existing CKD (estimated GFR (eGFR) below 60 mL/min/1.73m²), heart failure (where renal prostaglandins are essential for perfusion), hepatic cirrhosis with ascites, age above 65 years (reduced renal reserve and higher prevalence of subclinical CKD), and concurrent use of RAAS inhibitors or diuretics. The combination of an NSAID, a RAAS inhibitor (ACE inhibitor or ARB), and a diuretic constitutes the well-recognized triple whammy, associated with a markedly elevated risk of AKI that exceeds the risk of any two agents used without the third. NSAIDs should be avoided entirely in patients with an eGFR below 30 mL/min/1.73m² and used with caution and close monitoring in patients with eGFR 30 to 60 mL/min/1.73m². Baseline renal function, serum creatinine, electrolytes, and blood pressure should be checked before and within 1 to 2 weeks of initiating NSAID therapy in patients with any risk factors. Renal function generally recovers within days to weeks of NSAID discontinuation in hemodynamic AKI, provided the drug is stopped promptly.78
Hepatic Toxicity. Among the NSAIDs, diclofenac carries the clearest and best-characterized hepatotoxicity signal, though the overall incidence of clinically significant hepatic injury remains uncommon relative to the frequency of NSAID use. Diclofenac causes asymptomatic transaminase elevations in up to 15% of patients at standard therapeutic doses (75 to 150 mg/day); elevations exceeding three times the upper limit of normal (ULN) occur in approximately 1 to 3% of patients on prolonged therapy, and rare cases of severe drug-induced liver injury (DILI) including acute hepatic failure have been reported. The mechanism involves CYP2C9 (cytochrome P450 2C9) and CYP3A4 (cytochrome P450 3A4)-mediated formation of a reactive acyl glucuronide metabolite (diclofenac-1-O-acyl glucuronide) that is protein-reactive and can trigger immune-mediated hepatocellular injury. Liver function tests should be monitored during prolonged diclofenac therapy, and the drug should be discontinued if transaminase levels exceed three times the ULN. Sulindac carries a secondary hepatotoxicity signal through an unclear mechanism and has been associated with cholestatic hepatitis and granulomatous hepatitis in case reports. Other NSAIDs cause hepatotoxicity rarely; hepatotoxicity is not considered a class effect at the population level, though all NSAIDs should be used cautiously in patients with pre-existing hepatic disease given reduced metabolic capacity and the hepatic synthesis of coagulation factors.9
Avoid NSAIDs in: eGFR below 30 mL/min/1.73m²; decompensated heart failure; hepatic cirrhosis with ascites; volume depletion. Triple whammy (NSAID + ACE inhibitor/ARB + diuretic): avoid; if unavoidable, monitor creatinine and potassium within 1–2 weeks of initiation. For diclofenac: monitor transaminases at baseline and periodically during prolonged use; discontinue if >3× ULN. Topical diclofenac (1% gel, 1.5% solution): systemic absorption minimal; renal and hepatic risk substantially lower than oral formulations; preferred in patients with renal or hepatic compromise who require local NSAID delivery.
Each NSAID (non-steroidal anti-inflammatory drug) affects platelet function through reversible inhibition of COX-1 (cyclooxygenase-1)-dependent thromboxane A2 (TXA2) synthesis, in contrast to aspirin's irreversible acetylation. Aspirin-exacerbated respiratory disease (AERD, Samter triad) is a distinct hypersensitivity syndrome affecting approximately 10 to 20% of adults with asthma and up to 30% with chronic rhinosinusitis and nasal polyposis, in which COX-1 inhibition triggers a potentially life-threatening respiratory reaction through leukotriene shunting. Prostaglandin E2 (PGE2) normally suppresses mast cell and eosinophil activation in the airway; COX-1 inhibition removes this restraint while redirecting arachidonic acid (AA) to the lipoxygenase (LOX) pathway.
Platelet Effects of Non-Aspirin NSAIDs. Non-aspirin NSAIDs inhibit platelet COX-1 reversibly, meaning platelet TXA2 synthesis recovers as drug levels fall below the inhibitory threshold. The clinical duration of platelet inhibition therefore corresponds to the drug's plasma half-life plus the time required for TXA2 synthesis to recover to hemostatic levels. For short-acting agents such as ibuprofen (half-life 1.8 to 2 hours), platelet function recovers within 4 to 6 hours of the last dose; for longer-acting agents such as naproxen (half-life 12 to 17 hours) or piroxicam (half-life 30 to 86 hours), clinically significant platelet inhibition may persist for 24 to 48 hours after cessation. This is clinically important in perioperative settings: non-aspirin NSAIDs should be held before elective surgery for a period corresponding to approximately 5 half-lives to ensure platelet function recovery. For the majority of NSAIDs with half-lives below 6 hours (ibuprofen, diclofenac, indomethacin, ketorolac), holding for 24 hours before surgery is sufficient. Naproxen should be held for 3 to 5 days before high-bleeding-risk procedures. Aspirin, because its inhibition is irreversible, should be held for 7 to 10 days before procedures where platelet function is critical (neurosurgery, ophthalmic surgery), though the decision to hold aspirin must weigh bleeding risk against the risk of a cardiovascular event in patients on antiplatelet therapy for secondary prevention.10
Aspirin-Exacerbated Respiratory Disease. AERD (Samter triad) is defined by the clinical triad of asthma, chronic rhinosinusitis with nasal polyposis, and acute respiratory reactions triggered by aspirin or any COX-1 inhibiting NSAID. It affects approximately 10 to 20% of adults with asthma, rising to 30% in patients with both asthma and nasal polyps. The pathophysiology involves constitutive overproduction of cysteinyl leukotrienes (LTC4, LTD4, LTE4) in the respiratory mucosa due to upregulation of 5-lipoxygenase (5-LOX) and LTC4 synthase, combined with deficient PGE2-mediated suppression of mast cells and eosinophils. COX (cyclooxygenase)-1 inhibition by any NSAID acutely removes even the baseline PGE2 restraint on these cells and simultaneously redirects AA flux from the COX pathway to the LOX pathway (leukotriene shunting), causing a surge in cysteinyl leukotrienes that triggers bronchoconstriction, rhinorrhea, and urticaria within 30 to 180 minutes of ingestion. Severe reactions can include bronchospasm requiring emergency treatment. The diagnosis is clinical, supported by the characteristic history; the oral aspirin challenge (performed in a monitored medical setting) remains the definitive diagnostic procedure when the history is equivocal.11
COX-2 Inhibitors in AERD. Selective COX-2 (cyclooxygenase-2) inhibitors are pharmacologically distinct from COX-1 inhibitors in AERD because they do not inhibit platelet or mucosal COX-1 at therapeutic doses, and thus do not trigger the leukotriene shunting mechanism that underlies AERD reactions. Celecoxib at standard doses does not precipitate bronchoconstriction in AERD patients and is therefore the preferred oral NSAID when analgesic or anti-inflammatory therapy is required in a patient with confirmed or suspected AERD. However, COX-2 inhibitors must be introduced with caution and under medical supervision in AERD patients because some patients with severe AERD have exhibited cross-reactions to celecoxib in challenge testing, though this appears to be uncommon at standard doses. Acetaminophen (paracetamol) at doses up to 1,000 mg is generally tolerated in AERD patients at doses below 1,000 mg, but can trigger mild reactions in some patients at higher doses through weak COX-1 inhibition.11
Aspirin Desensitization. For AERD patients who require aspirin for cardiovascular indications or who have refractory nasal polyposis and rhinosinusitis, aspirin desensitization is a validated therapeutic option. The procedure involves administration of incrementally increasing aspirin doses under medical supervision until a reaction occurs, then continuing aspirin until the reaction abates, and then administering further incremental doses; after successful desensitization, patients tolerate aspirin and all other NSAIDs by a mechanism that involves downregulation of the cysteinyl leukotriene pathway. Long-term aspirin therapy after desensitization reduces nasal polyp burden, improves sense of smell, and reduces the frequency of sinus infections. Desensitization must be performed in a specialized setting with resuscitation capability and maintained through continuous daily aspirin dosing; the tolerance does not persist if aspirin is held for more than 72 hours.11
Diagnosis: clinical triad (asthma + nasal polyps + NSAID-triggered respiratory reaction); oral aspirin challenge in monitored setting if diagnosis equivocal. Avoid: all COX-1 inhibiting NSAIDs including aspirin, ibuprofen, naproxen, indomethacin, ketorolac, diclofenac. Use instead: celecoxib (COX-2 selective; does not trigger LOX shunting) at standard doses under supervision; acetaminophen at doses up to 650 mg (avoid 1,000 mg doses in severe AERD). Treat underlying disease: inhaled corticosteroids, leukotriene receptor antagonists (montelukast), biologics (dupilumab for nasal polyps and asthma). Desensitization: appropriate for patients with cardiovascular aspirin indication or refractory rhinosinusitis; specialist procedure only.
NSAIDs (non-steroidal anti-inflammatory drugs) participate in clinically important drug interactions through both pharmacokinetic mechanisms (albumin displacement, CYP2C9 competition) and pharmacodynamic mechanisms (additive gastrointestinal (GI) bleeding, additive renal vasoconstriction, reduced antihypertensive efficacy). Several of these interactions carry serious safety consequences and require proactive management strategies rather than simple dose adjustment. Key co-prescribing hazards involve anticoagulants, renin-angiotensin-aldosterone system (RAAS) inhibitors, lithium, high-dose methotrexate (MTX), and selective serotonin reuptake inhibitors (SSRIs).
NSAIDs and Anticoagulants. The combination of NSAIDs with anticoagulants constitutes one of the highest-risk drug interactions in ambulatory medicine. The interaction operates through two synergistic mechanisms: first, NSAIDs impair platelet function by reversibly inhibiting COX-1 (cyclooxygenase-1)-dependent thromboxane A2 (TXA2) synthesis, impairing the platelet contribution to primary hemostasis; second, NSAIDs cause GI (gastrointestinal) mucosal erosions and ulcerations that provide a bleeding site. Together, these mechanisms sharply amplify the bleeding risk associated with anticoagulant therapy. Warfarin has an additional pharmacokinetic interaction: NSAIDs that are strong CYP2C9 inhibitors (particularly fluconazole in co-administration scenarios, but phenylbutazone and others) can reduce the metabolism of the more pharmacologically potent S-enantiomer of warfarin, raising the international normalized ratio (INR). Among the direct oral anticoagulants (DOACs), the pharmacodynamic GI bleeding interaction applies equally to all agents; there are no significant pharmacokinetic interactions between standard NSAIDs and DOACs. The absolute risk of GI bleeding with the combination of an NSAID (non-steroidal anti-inflammatory drug) and an anticoagulant (warfarin or DOAC) is approximately 2 to 4 times higher than with the anticoagulant alone. This combination should be avoided whenever possible; when it cannot be avoided, proton pump inhibitor (PPI) co-therapy and the shortest necessary NSAID duration are mandatory.3
NSAIDs and Antihypertensives. NSAIDs attenuate the blood pressure-lowering efficacy of most antihypertensive drug classes through a common mechanism: suppression of renal prostaglandin synthesis enhances renal sodium and water reabsorption, increasing plasma volume and peripheral vascular resistance, and directly counteracting the vasodilatory and natriuretic mechanisms of antihypertensive agents. The antihypertensive classes most significantly affected are ACE (angiotensin-converting enzyme) inhibitors, ARBs (angiotensin receptor blockers), thiazide and loop diuretics, and to a lesser extent beta-blockers; calcium channel blockers are relatively resistant to NSAID interaction because their mechanism does not depend on prostaglandin pathways. The magnitude of blood pressure increase with NSAID co-administration averages 3 to 5 mmHg systolic in meta-analyses but can be substantially larger in individual patients, particularly those with salt-sensitive hypertension or heart failure. The triple whammy interaction between an NSAID, a RAAS (renin-angiotensin-aldosterone system) inhibitor (ACE inhibitor or ARB), and a diuretic deserves special emphasis: this combination causes AKI (acute kidney injury) through additive reduction of GFR (glomerular filtration rate), with the NSAID reducing renal prostaglandin-dependent afferent arteriolar dilation, the RAAS inhibitor blocking angiotensin II-dependent efferent arteriolar constriction (reducing the pressure gradient that sustains GFR), and the diuretic causing volume depletion that further reduces renal perfusion pressure. Studies have demonstrated approximately 31-fold increased risk of AKI hospitalization with this combination compared to patients on none of these drugs.78
NSAIDs and Lithium. Lithium (Li+) is cleared almost exclusively by the kidney through a process that is partially sodium-dependent: when renal tubular sodium reabsorption increases (as occurs with NSAID use), lithium reabsorption also increases, reducing lithium clearance and raising plasma lithium concentrations. NSAIDs can increase lithium levels by 10 to 60% depending on the agent and dose, with indomethacin showing the largest effect and sulindac showing relatively less effect due to its renal-sparing properties. Because lithium has a narrow therapeutic index (therapeutic range 0.6 to 1.2 mEq/L; toxic above 1.5 mEq/L), even modest increases in plasma lithium concentration can produce toxicity (coarse tremor, confusion, ataxia, and in severe cases cardiac arrhythmias and seizures). Lithium levels should be measured within 5 to 7 days of initiating or changing an NSAID in a patient on lithium, and again at NSAID cessation. Sulindac is sometimes cited as the NSAID least likely to raise lithium levels due to its renal prostaglandin-sparing properties, but this relative advantage does not eliminate the interaction, and lithium monitoring is still required.12
NSAIDs and Methotrexate. The interaction between NSAIDs and methotrexate (MTX) is dose-dependent and clinically significant primarily at the higher MTX doses used in oncology (greater than 50 to 100 mg/m² per cycle) rather than at the low weekly doses used for rheumatoid arthritis (RA) and psoriasis (7.5 to 25 mg/week). At high oncology doses, NSAIDs reduce renal MTX excretion through two mechanisms: competitive inhibition of organic anion transporter (OAT) proteins that secrete MTX into the proximal tubular lumen, and reduced renal prostaglandin synthesis decreasing renal blood flow and GFR, further impairing MTX clearance. The result is prolonged and elevated MTX exposure, potentially causing severe myelosuppression, mucositis, and nephrotoxicity. At low weekly rheumatology doses, the interaction is clinically less severe, but caution and monitoring are still appropriate; some rheumatology guidelines recommend withholding NSAIDs for 24 to 48 hours around the weekly MTX dose in patients receiving doses above 15 mg/week. In all cases, NSAIDs should not be used during or within 24 hours of high-dose MTX infusions in oncology patients.12
NSAIDs and Selective Serotonin Reuptake Inhibitors. The combination of NSAIDs with SSRI (selective serotonin reuptake inhibitor) agents increases the risk of upper GI bleeding beyond that attributable to either agent alone, through a pharmacodynamic interaction. SSRIs deplete platelet serotonin stores by blocking serotonin reuptake transporter (SERT) in the platelet membrane; since platelets cannot synthesize serotonin de novo, SSRI use substantially reduces platelet serotonin content and impairs platelet activation by serotonin-dependent pathways. NSAIDs simultaneously impair COX-1-dependent TXA2 synthesis. The concurrent suppression of two independent platelet activation pathways produces additive platelet dysfunction and a disproportionate increase in bleeding risk at GI mucosal injury sites. Epidemiological studies show a 3 to 15-fold increase in upper GI bleeding risk with the SSRI-NSAID (SSRI plus NSAID) combination compared to either agent alone, and this risk is substantially reduced by PPI co-therapy. Patients requiring both an SSRI and an NSAID chronically should receive PPI gastroprotection.3
NSAIDs + anticoagulants (warfarin/DOACs): 2–4× GI bleed risk; avoid if possible, use PPI if unavoidable; monitor INR with warfarin. NSAIDs + ACE inhibitor/ARB + diuretic (triple whammy): markedly elevated AKI risk; avoid combination; monitor creatinine and potassium if combination unavoidable. NSAIDs + antihypertensives: 3–5 mmHg average BP rise; monitor BP and adjust antihypertensive therapy. NSAIDs + lithium: raise lithium levels 10–60%; monitor lithium within 5–7 days of initiating NSAID. NSAIDs + high-dose MTX: avoid within 24 h of oncology MTX doses; caution at rheumatology doses above 15 mg/week. NSAIDs + SSRIs: additive GI bleeding; use PPI co-therapy.
Several patient populations face disproportionate NSAID (non-steroidal anti-inflammatory drug) toxicity risk that requires either avoidance, dose modification, or substitution with safer alternatives. Pregnancy, advanced age, chronic kidney disease (CKD), and hepatic cirrhosis each present distinct pharmacological vulnerabilities that translate into specific clinical management decisions.
Pregnancy. NSAID (non-steroidal anti-inflammatory drug) use during pregnancy carries gestational-age-dependent risks. During the first trimester, NSAID use has been associated in some epidemiological studies with increased rates of spontaneous miscarriage and rare reports of cardiac septal defects, though the data are not fully consistent and the absolute risk from short-term exposure is low. During the second trimester (14 to 27 weeks), the US Food and Drug Administration (FDA) issued a Drug Safety Communication in 2020 warning that NSAID use beginning at 20 weeks or later may cause fetal renal dysfunction leading to oligohydramnios (reduced amniotic fluid from decreased fetal urine output) and, with prolonged use, neonatal renal impairment; ultrasound monitoring of amniotic fluid volume is recommended if NSAID therapy is continued beyond 20 weeks.13
During the third trimester, NSAID use is strongly contraindicated due to the risk of premature closure of the ductus arteriosus (DA): prostaglandin E2 (PGE2) maintains DA patency through EP4 (prostaglandin E receptor subtype 4) receptor-mediated vasodilation of the ductal wall, and COX (cyclooxygenase) inhibition from week 28 onward can cause fetal ductal constriction that, if severe, results in right ventricular pressure overload and fetal hydrops. Acetaminophen remains the preferred analgesic during pregnancy at all trimesters. Low-dose aspirin (81 mg/day) is an exception: it is recommended throughout pregnancy under obstetric supervision for preeclampsia prevention in high-risk patients, based on strong evidence of benefit that outweighs the theoretical risk at this antiplatelet dose.13
Elderly Patients. Patients aged 65 years and older face substantially elevated risks from NSAID use relative to younger adults, for several reasons that converge in this population. Renal reserve declines with age even in the absence of diagnosed CKD; the Cockcroft-Gault equation consistently reveals GFR (glomerular filtration rate) values in elderly patients that underestimate renal vulnerability when serum creatinine alone is used. Gastric mucosal regenerative capacity decreases with age, mucosal blood flow diminishes, and H. pylori prevalence is higher, collectively amplifying gastrointestinal (GI) toxicity risk. Concurrent polypharmacy is common in elderly patients, increasing the probability of the triple whammy combination and NSAID-anticoagulant co-prescription. Cognitive effects of indomethacin and other highly central nervous system (CNS)-penetrant NSAIDs (confusion, dizziness, somnolence) are more pronounced in elderly patients and can precipitate falls. The American Geriatrics Society Beers Criteria designates all oral non-COX-selective NSAIDs as potentially inappropriate medications in adults aged 65 and older, recommending avoidance unless other alternatives are not effective and the patient can take a gastroprotective agent concomitantly.6 When an NSAID cannot be avoided in an elderly patient, topical diclofenac gel for localized musculoskeletal pain provides effective analgesia with minimal systemic exposure and substantially lower GI, cardiovascular, and renal risk.13
Chronic Kidney Disease. NSAIDs are potentially harmful at all stages of CKD and should be avoided in patients with eGFR below 30 mL/min/1.73m². In patients with eGFR 30 to 60 mL/min/1.73m² (CKD stage G3), short-term NSAID use at the lowest effective dose is sometimes necessary but requires close monitoring of renal function, blood pressure, and fluid status. In patients with eGFR 60 to 90 mL/min/1.73m² (CKD stage G2), NSAIDs can generally be used with standard caution. It is important to recognize that serum creatinine alone is an insensitive indicator of underlying CKD, particularly in elderly, cachectic, or malnourished patients whose reduced muscle mass generates little creatinine despite markedly reduced GFR; eGFR calculation using the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equation is mandatory before prescribing NSAIDs in patients where renal function is uncertain. NSAIDs can accelerate the progression of CKD through hemodynamic and direct tubular mechanisms when used chronically; patients with diabetic nephropathy, hypertensive nephrosclerosis, and other proteinuric nephropathies are particularly vulnerable.7
Hepatic Cirrhosis. Patients with hepatic cirrhosis and portal hypertension present particular challenges for NSAID prescribing. Cirrhotic patients have marked activation of the renin-angiotensin-aldosterone system and the sympathetic nervous system to maintain systemic blood pressure in the face of splanchnic vasodilation; renal perfusion in this state is heavily dependent on prostaglandin-mediated afferent arteriolar vasodilation, making these patients exquisitely vulnerable to NSAID-induced acute kidney injury (AKI) and hepatorenal syndrome precipitation. NSAIDs also impair platelet function in patients whose platelet count is already reduced by hypersplenism, and can worsen the coagulopathy of liver failure by impairing prostaglandin-mediated platelet activation. GI toxicity risk is amplified in cirrhotic patients who often have gastric varices, portal hypertensive gastropathy, and reduced mucus secretion. NSAIDs are generally contraindicated in patients with Child-Pugh B or C cirrhosis, and should be used only with extreme caution and close monitoring in Child-Pugh A patients when no alternative is available. Acetaminophen at reduced doses (maximum 2 g/day in patients with active alcohol use or advanced cirrhosis) is generally a safer analgesic choice in this population than NSAIDs.7
Absolute avoidance: pregnancy ≥20 weeks (FDA warning); eGFR <30 mL/min/1.73m²; decompensated heart failure; Child-Pugh B/C cirrhosis; active peptic ulcer. High caution (use lowest dose, shortest duration, with monitoring): age >65; eGFR 30–60 mL/min/1.73m²; concurrent anticoagulant; concurrent RAAS inhibitor + diuretic; prior GI ulcer/bleed. AERD: avoid all COX-1 inhibiting NSAIDs; use celecoxib if NSAID required. Perioperative: hold NSAIDs for 5 half-lives before high-bleeding-risk procedures; aspirin 7–10 days for irreversible COX-1 inhibition. Topical diclofenac: preferred in elderly and in patients with renal, cardiovascular, or GI risk for localized OA pain.
Lanas A, Chan FKL. Peptic ulcer disease. Lancet. 2017;390(10094):613-624.
doi:10.1016/S0140-6736(16)32404-7Sostres C, Gargallo CJ, Lanas A. Nonsteroidal anti-inflammatory drugs and upper and lower gastrointestinal mucosal damage. Arthritis Res Ther. 2013;15(Suppl 3):S3.
doi:10.1186/ar4175Lanza FL, Chan FK, Quigley EM; Practice Parameters Committee of the American College of Gastroenterology. Guidelines for prevention of NSAID-related ulcer complications. Am J Gastroenterol. 2009;104(3):728-738.
doi:10.1038/ajg.2009.115FitzGerald GA. Coxibs and cardiovascular disease. N Engl J Med. 2004;351(17):1709-1711.
doi:10.1056/NEJMp048288Bhala N, Emberson J, Merhi A, et al; Coxib and traditional NSAID Trialists' (CNT) Collaboration. Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs: meta-analyses of individual participant data from randomised trials. Lancet. 2013;382(9894):769-779.
doi:10.1016/S0140-6736(13)60900-9American Geriatrics Society 2023 updated AGS Beers Criteria for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023;71(7):2052-2081.
doi:10.1111/jgs.18372Whelton A. Nephrotoxicity of nonsteroidal anti-inflammatory drugs: physiologic foundations and clinical implications. Am J Med. 1999;106(5B):13S-24S.
doi:10.1016/S0002-9343(99)00113-8Lapi F, Azoulay L, Yin H, Nessim SJ, Suissa S. Concurrent use of diuretics, angiotensin converting enzyme inhibitors, and angiotensin receptor blockers with non-steroidal anti-inflammatory drugs and risk of acute kidney injury: nested case-control study. BMJ. 2013;346:e8525.
doi:10.1136/bmj.e8525Boelsterli UA. Diclofenac-induced liver injury: a paradigm of idiosyncratic drug toxicity. Toxicol Appl Pharmacol. 2003;192(3):307-322.
doi:10.1016/S0041-008X(03)00368-5Patrono C, Baigent C. Role of aspirin in primary prevention of cardiovascular disease. Nat Rev Cardiol. 2019;16(11):675-686.
doi:10.1038/s41569-019-0225-yStevenson DD, Sanchez-Borges M, Szczeklik A. Classification of allergic and pseudoallergic reactions to drugs that inhibit cyclooxygenase enzymes. Ann Allergy Asthma Immunol. 2001;87(3):177-180.
doi:10.1016/S1081-1206(10)62221-1Rainsford KD. Anti-inflammatory drugs in the 21st century. Subcell Biochem. 2007;42:3-27.
doi:10.1007/1-4020-5688-5_1Bloor M, Paech M. Nonsteroidal anti-inflammatory drugs during pregnancy and the initiation of lactation. Anesth Analg. 2013;116(5):1063-1075.
doi:10.1213/ANE.0b013e31828a4b54