In the clinical conditions of decreased renal perfusion, such as nephritic syndrome, liver cirrhosis, heart failure, volume depletion, or aging kidneys, renal prostaglandin (PG) production mediated primarily by cyclooxygenase-1 (COX-1) and possibly by clooxygenase-2 (COX-2) plays a major role in maintaining renal hemodynamics. Inhibition of the synthesis of renal PGs by conventional non-steroidal anti-inflammatory drugs (NSAIDs), which inhibit both COX-1 and COX-2, may result in acute renal complications such as sodium and water retention, hyperkalemia, acute renal failure, and acute papillary necrosis, or chronic renal effects such as nephrotic syndrome, chronic papillary necrosis, and analgesic nephropathy1-3). NSAIDs may also influence blood pressure in treated hypertensive individuals. The COX-2 specific inhibitors were first thought to have much less renal adverse effects by sparing homeostatic COX-1 activity in the kidney. However, emerging evidence is disappointing at this point and indicates that the cardiorenal risks associated with the use of COX-2 specific inhibitors and conventional NSAIDs are comparable. Regardless of cyclooxygenase specificity, up to approximately 5% of individuals exposed to NSAIDs may develop one or more forms of renal adverse effects, although the incidence of nephrotoxicity associated with the use of NSAIDs is relatively low in the normal population. Regarding the consumption of these compounds which account for approximately 2.5% of all prescription dollars and the populations that use NSAIDs frequently (old age and who have underlying diseases), the number of at-risk individuals should be high (Fig. 1)4).

This review will summarize the possible mechanisms by which fluid and electrolyte abnormalities, acid-base disturbances, and acute renal failure associated with NSAIDs use evolve.

PGs play active roles in the local regulation of vascular tone and salt and water homeostasis by modulating the glomerular hemodynamics and also by adjusting the function of the distal nephron5). In the early 1990s, the existence of two isoforms of cyclooxygenase (COX-1 and COX-2) was confirmed6). COX-1 isoform, which is expressed constitutively in many tissues and catalyzes the PG synthesis, is believed to have physiologic functions including gastric mucosal defense and platelet aggregation. COX-1 is abundantly expressed across species in the collecting ducts, renal vasculature, glomeruli, and papillary interstitial cells7). In contrast, basal COX-2 expression in the kidney is less apparent and exhibits some inter-species differences in its localization. COX-2 is constitutive in some tissues and is markedly induced by bacterial endotoxins, cytokines, and growth factors and catalyzes the synthesis of pro-inflammatory PGs.

In addition to their role in preserving renal perfusion, renal PGs promote the secretion of renin, impair sodium reabsorption in the loop of Henle and cortical collecting tubule, and partially antagonize the effect of antidiuretic hormone to increase water reabsorption in the collecting tubules1, 8). Locally generated PGs may also mediate part of the natriuretic effect of dopamine and of one of the atrial natriuretic peptides9).

These roles of PGs are not very important in normal subjects in whom basal renal PG production is relatively low. However, they may become important when prostaglandin synthesis is stimulated by underlying renal disease or by vasoconstrictors such as angiotensin II or norepinephrine, the secretion of which is increased in states of effective volume depletion; true volume depletion due to gastrointestinal or renal losses (as with diuretic therapy), heart failure, or cirrhosis (Fig. 2).

NSAIDs inhibit cyclooxygenase, impairing transformation of arachidonic acid to prostaglandins, prostacyclin, and thromboxanes, although the extent of enzyme inhibition varies among the different NSAIDs. Administration of cyclooxygenase-inhibiting NSAIDs in the setting of volume depletion interferes with local vasodilator effects of PGs and may result in a catastrophic decline in glomerular filtration rate (GFR), resulting in overt renal failure.

Prostaglandin (PGE₂) either stimulates or inhibits epithelial solute and water transport along the nephron10). When added to vasopressin-prestimulated collect ing ducts, PGE₂ potently inhibits water absorption11). At least three distinct effects of basolateral PGE₂ on transport have been described: stimulation of basal water reabsorption; inhibition of vasopressin-stimulated water reabsorption; and inhibition of sodium reabsorption.

The mechanism by which PGE₂ contributes to natriuresis may involve changes in resistance of the renal medullary microcirculation. Although expression of COX-2 in the macula densa and cortical thick ascending limb (cTAL) is increased by low salt diet, medullary interstitial COX-2 expression is increased by high salt diet. Increased PGE₂ directly dilates the descending vasa recta and increases medullary blood flow, which may contribute to increased interstitial pressure and medullary hypotonicity, leading to enhanced salt excretion (Fig. 3).

PGE₂ directly inhibits salt and water absorption in the in-vitro microperfused thick ascending limb (TAL) and collecting duct cells12), of which mechanisms are still uncertain. PGE₂ directly inhibits chloride absorption in mouse or rabbit medullary TAL from either the luminal or basolateral surfaces. PGE₂ also inhibits hormone-stimulated cyclic adenosine monophosphate (cAMP) generation in TALs. Because cAMP stimulates TAL transport, inhibition of cAMP generation through a Gi-coupled PGE₂ receptor likely contributes to the inhibitory effects of PGE₂ on TAL transport. In the collecting duct, PGE₂ inhibits both vasopressin-stimulated osmotic water absorption and cAMP generation13). Intrarenal infusion of PGE₂ is natriuretic, and although its effects on intrarenal hemodynamics undoubtedly play an important role, direct effects of PGE₂ on epithelial transport are equally important. In contrast to PGE₂-mediated inhibition of vasopressin-timulated water reabsorption, its capacity to inhibit sodium reabsorption is insensitive to pertussis toxin. Instead, PGE₂ inhibits collecting duct sodium reabsorption via a calcium-dependent mechanism13).

The renal medulla is a major site of prostaglandin synthesis and abundant COX-1 and COX-2 expression. The two cyclooxygenase isoforms exhibit different expression within the medulla, with COX-1 predominating in the medullary collecting duct cells and COX-2 predominating in medullary interstitial cells14). After dehydration, renal medullary COX-2 mRNA and protein expression are significantly induced, primarily in medullary interstitial cells15). In contrast, COX-1 expression in the mouse kidney is unaffected by water deprivation. Although hormonal factors may also contribute to COX-2 induction, shifting cultured renal medullary interstitial cells to hypertonic media (using either NaCl or mannitol) is sufficient to induce COX-2 expression. Osmotic induction of COX-2 expression in medullary interstitial cells plays an important role in the survival of these cells in hypertonic milieu after dehydration8).

Cultured medullary interstitial cells produce abundant PGE₂, which has been shown to directly dilate vasa rectae, counteracting the constricting effect of angiotensin II and endothelin and thereby helping to maintain renal medullary blood flow16). Regulation of renal medullary blood flow has significant implications for regulating salt excretion and systemic blood pressure as previously described. Reduced medullary interstitial pressure increases renal sodium reabsorption17) and could attribute to sodium retention in the setting of NSAIDs use. Medullary interstitial prostaglandins may modulate epithelial solute and water reabsorption not only via hemodynamic effects, but also through direct effects on epithelial sodium reabsorption by the thick ascending limb and collecting duct. Loss of the tonic inhibitory effect of COX-2-derived PGE₂ on salt absorption by these segments may contribute to the sodium retention seen with NSAIDs18).

In the kidney, COX-2 is expressed in the macula densa/cTAL and medullary interstitial cells. The macula densa is involved in regulating afferent arteriolar tone and renin release by sensing alterations in luminal chloride concentration via changes in the rate of Na⁺-K⁺-2Cl^- cotransport. Administration of non-specific cyclooxygenase inhibitors bluntly increases renin release mediated by macula densa sensing of decreases in luminal NaCl. High renin states such as salt deficiency, angiotensin converting enzyme (ACE) inhibitors or angiotensin receptor blockers, diuretic administration or experimental renovascular hypertension, are associated with increased macula densa/cTAL COX-2 expression. Furthermore, angiotensin II and/or aldosterone inhibits COX-2 expression19). In angiotensin II type 1 receptor knockout mice, COX-2 expression is increased similar to increases seen with ACE inhibitors or angiotensin II type 1 receptor blockers.

In vivo studies in rats indicated that increased renin release in response to low-salt diet, ACE inhibitor, loop diuretics, or aortic coarctation could be inhibited by administration of COX-2 inhibitors. In mice with genetic deletion of COX-2, ACE inhibitors or low-salt diet failed to increase renal renin expression, although renin was significantly increased in wild type mice. In contrast, in COX-1 null mice there were no significant differences in either the basal or ACE inhibitor-stimulated levels of renal renin activity from plasma or renal tissue compared with wild type mice.

By inhibiting the production of PGs that contribute to maintaining vasodilation of adjacent afferent arterioles, COX-2 inhibition may contribute to the decline in GFR observed in patients taking NSAIDs or selective COX-2 inhibitors. In healthy human subjects, COX-2 inhibitors may have minimal effects on renal hemodynamics. However, COX-2 inhibitors significantly decrease GFR in salt-depleted subjects. Further evidence of an important role for COX-2 in regulation of renal hemodynamics and renin production, acute renal failure, and hyperkalemia/type IV renal tubular acidosis (RTA) has been reported as acute nephrotoxic effects of COX-2 inhibitors, especially in the elderly.

In the settings of enhanced PGs production, inhibition of PGs synthesis with NSAIDs could, in selected patients, lead to various electrolytes, acid-base disturbances such as hyponatremia, sodium retention, hyperkalemia, or renal tubular acidosis, which are usually reversible with discontinuation of the therapy. This is mediated in part by reduction in the release of renin and aldosterone, which also occurs with selective COX-2 inhibitors. In addition, vasoconstriction due to decreased synthesis of vasodilator PGs can exacerbate underlying hypertension and heart failure.

Selective COX-2 inhibitors also adversely affect renal function in such at risk patients who include those with volume depletion, heart failure, cirrhosis, intrinsic renal disease (eg, diabetic nephropathy), and hypercalcemia27). In a review of the literature, acute renal failure and/or severe electrolyte disturbances (particularly hyperkalemia and metabolic acidosis) were clearly precipitated by either celecoxib or rofecoxib. A study of 15 similar patients found that a single dose of high amounts of rofecoxib (250 mg) had significant and similarly detrimental effects on renal function as a single dose of indomethacin (75 mg)28). These observations suggest that selective COX-2 inhibitors should be avoided in patients with chronic renal insufficiency, severe heart disease, volume depletion, and/or hepatic failure.

Compared with celecoxib, the incidence of adverse events related to hypertension and edema were significantly higher with ibuprofen (but not diclofenac), and a decline in renal function was more apparent with diclofenac than with ibuprofen. These data provide a greater understanding of the relative effects of these agents on cardiorenal function and homeostasis29).

Both COX-1 and COX-2 are constitutively expressed in the kidney. Prostaglandin enhances the renin secretion in the macula densa. PG induces natriuresis and diuresis by various mechanisms. Inhibition of renal prostaglandin synthesis by NSAIDs causes various electrolyte and acid-base disturbances including sodium retention (edema, hypertension), hyponatremia, and hyperkalemia/type IV RTA. NSAIDs associated electrolyte and acid-base disturbances are not uncommon in some clinical situations. Adverse renal effects of NSAIDs are generally associated with the PGs dependent states; volume-contracted states, low cardiac output, or other conditions that tend to compromise renal perfusion. Both non-selective and selective COX-2 inhibitors have similar renal adverse effects. In view of the many NSAIDs users' susceptibility to renal adverse effects due to their underlying disease or condition, physicians should be cautious in prescribing NSAIDs to susceptible patients.