Gastrointestinal damage and bleeding from non-steroidal anti-inflammatory drugs. I. Clinical and epidemiological aspects

Massive Gastric Hemorrhage after Indomethacin Therapy: A Rare Presentation and Critical Management in an Extremely Preterm Infant

Indomethacin has been widely used in preterm infants with hemodynamically significant patent ductus arteriosus (PDA). Gastrointestinal complications of indomethacin have been reported in 5% of treated neonates. However, massive gastric mucosa hemorrhage is a rarely reported complication. To the best of our knowledge, the infant in this report is the smallest reported in the literature to have undergone successful surgery for such a complication. A male preterm infant weighing 566 g was born at 25^2/7 weeks of gestational age without a complicated maternal history. Soon after birth, he received nasal noninvasive respiratory support and minimal feeding. PDA was observed since the first day of life (DOL), treatments were initiated on the second DOL for the hemodynamical significance, and PDA was closed after two courses of indomethacin therapy (0.2 mg/kg). At midnight on the seventh DOL, generalized pallor, bloody gastric drainage, and a distended stomach were observed. Massive gastric bleeding was suspected. He suffered from intermittent hypotension, which was corrected with blood products and fluid resuscitation under monitoring with a radial arterial line. Gastric lavage with cooling saline was performed twice but in vain. Prior to surgical consultation, intravascular volume transfusion was given twice. An exploratory laparotomy was arranged after obtaining the parents' consent. Blood oozing from the gastric mucosa was observed through gastrostomy and was successfully stopped via epinephrine-soaked gauze compression. After the operation, his clinical course remained uneventful, and he was discharged without neurological anomaly at two-year follow-up. Physicians need to be cautious of indomethacin's effect on platelet dysfunction in preterm infants with multiple predisposing factors. The tendency for mucosal bleeding should be continuously monitored after indomethacin therapy.

Gastrointestinal mucosal injury following repeated daily oral administration of conventional formulations of indometacin and other non-steroidal anti-inflammatory drugs to pigs: a model for human gastrointestinal disease

Non-steroidal anti-inflammatory drugs (NSAIDs) vary in their propensity to cause damage in different regions of the gastrointestinal (GI) tract in laboratory animals and humans. This may depend on the type of drug formulation as well as the intrinsic pharmacological properties of the drugs. The purpose of this study was to determine the effects of NSAIDs, with cyclooxygenase 1 and 2 inhibitory activity but with different potency as inhibitors of prostaglandin production, when given orally as tablet/capsule formulations of NSAIDs for 10 days to pigs, a species that has close resemblance in structure and function of the tract to that in humans. Three capsule or tablet formulations of NSAIDs were given orally to pigs for 10 days. GI bleeding was measured by determination of radioactive iron in the faeces from (59)Fe-pre-labelled red blood cells. The blood loss was compared with the pathological changes in the GI mucosa observed at autopsy, mucosal myeloperoxidase (MPO) activity as an index of leucocyte infiltration, and plasma and mucosal concentrations of the drugs at termination assayed by high-performance liquid chromatography. Mucosal damage and bleeding varied according to the type of NSAID. Gastroduodenal ulcers and lesions occurred with the cyclooxygenase inhibitors indometacin (indomethacin) (Indocid capsules 10 or 5 mg kg(-1) day(-1) b.i.d.), aspirin (USP tablets 150 mg kg(-1) day(-1) b.i.d) and naproxen (Apotex tablets 50 or 75 mg kg(-1) day(-1) b.i.d.), and there was an increase in the cumulative (i.e. 10-day) blood loss at higher doses of indometacin and naproxen, and with aspirin. There was no statistically significant increase in gastric or intestinal mucosal MPO activity in the non-damaged mucosa with these drugs and this was confirmed by histological observations in non-lesioned areas of the mucosa. Indometacin produced focal ulcers in the caecum but this was not observed with the other drugs. All the NSAIDs produced significant blood loss coincident with gastric ulceration but no increase in gastric or intestinal MPO activity. Plasma concentrations of the non-aspirin NSAIDs were within the range encountered therapeutically in humans. The mucosal concentrations of indometacin in the gastric and intestinal mucosa correlated with mucosal injury. These findings show that: (i) NSAIDs vary in their propensity to produce mucosal injury in different regions of the GI tract according to their pharmacological properties and formulation; (ii) mucosal injury from some NSAIDs may not directly relate to blood loss at low doses of NSAIDs and this may depend on inhibition of platelet aggregation; and (iii) the occurrence of caecal ulcers uniquely observed with indometacin treatment may be relevant to the development of intestinal pathology (e.g. diaphragm-like structures) seen occasionally in humans. These results suggest that the pig model employed in the present studies may be useful for investigations of GI damage from NSAID tablets/capsules, especially in regions that are generally inaccessible to routine endoscopic investigations in humans (e.g. the proximal regions of the large intestine).

The ever-emerging anti-inflammatories. Have there been any real advances?

Gastrointestinal (GI) Adverse Drug Reactions (ADRs) from the NSAIDs are a major cause of morbidity and mortality in arthritic patients taking these drugs. The recent much heralded development of COX-2 selective drugs (celecoxib, rofecoxib), the objective of which has been to spare inhibition of the production of COX-1 derived mucosal protective prostaglandins, may have represented an advance in reducing the risk of serious ADRs--ulcers and bleeding--but does not appear to have reduced the incidence of symptomatic side-effects (nausea, vomiting, epigastric pain/heartburn, abdominal discomfort) which are a major reason for withdrawal from NSAID therapy, especially in the long term. The rationale of COX-2 selectivity from these newer drugs is controversial since there may be pharmacokinetic differences from established carboxylate-NSAIDs that accounts for their apparent lower ulcerogenicity. Moreover, concerns have been recently expressed that as COX-2 is important in ulcer healing, control of prostacyclin production and renal function that they may have adverse reactions from these effects. Indeed, recent reports of enhanced risk of congestive heart failure with rofecoxib are of importance and may relate to impaired prostacyclin production. Moreover, there are other therapeutic strategies that have yielded equally low ulcerogenic NSAIDs (e.g. the prodrug, nabumetone; the established COX-2 inhibitory drug, nimesulide) and even the well-established NSAIDs ibuprofen and diclofenac have relatively low upper GI ulcerogenicity and have been used as benchmark standards in comparative trials of the newer "Oxib" drugs (celecoxib, rofecoxib). Much research interest has centred on the nitric oxide-donating NSAIDs (NO-NSAIDs). The rationale for donating NSAIDs being to counteract the vasoconstriction effects of NSAIDs but this has yet to be fully evaluated. It is not certain that this "antidote" approach will be acceptable as there may also be systemic effects of the nitrobutoxyl--or other NO-donors that may have toxicological consequences. Another strategy is the development of mixed COX-5 lipoxygenase (LOX) inhibitors--the progenitors of which were benoxaprofen and BW-755C. The rationale of reducing the potential for lipoxygenase mediated actions in the stomach (e.g. vasoconstriction, leucocyte accumulation). Clearly, the need to develop newer NSAIDs with lower risks of ulcers and bleeding as well as symptomatic ADRs is still representing a major challenge.