Chapter 15. NSAIDs & the Kidney: Acute Renal Failure

• Nonsteroidal anti-inflammatory drug (NSAID)-associated acute renal failure (ARF) develops predominantly in patients with underlying risk factors.
• Clinical presentations can be asymptomatic or associated with uremic symptoms, edema (pulmonary and peripheral), hypertension, or electrolyte and acid–base disturbances.
• Elevated blood urea nitrogen (BUN) and serum creatinine (Cr) concentrations are present.
• An elevated BUN/Cr ratio (> 20) is often noted.
• Hyponatremia (serum [Na+] < 135 mEq/L), hyperkalemia (serum [K+] > 5.5 mEq/L), and non-anion gap metabolic acidosis (serum [HCO3−] < 20 mEq/L) are common.
• Urine [Na+] < 10–20 mEq/L and FeNa+ < 1.0% characterize NSAID-associated ARF.
• ARF is generally reversible with discontinuation of NSAIDs and treatment of concurrent disease processes.
• Severe ARF from NSAIDs may require dialysis.

Epidemiology of Nonsteroidal Anti-inflammatory Drug Use

NSAIDs are employed widely to treat pain, fever, and inflammation. Other potential uses for these drugs include treatment and prevention of colonic polyposis and Alzheimer-type dementia. The first NSAID discovered was sodium salicylate in 1763. In 1950, phenylbutazone was introduced into clinical practice. It was efficacious but its use faded due to bone marrow toxicity. Subsequently, indomethacin entered the market in the 1960s. More than 20 NSAIDs from seven major classes, including the selective cyclooxygenase (COX)-2 inhibitors (Table 15–1), are available in the United States. In addition to prescription NSAIDs, a large percentage of the general population consumes over-the-counter NSAIDs. Annually, more than 50 million patients ingest these drugs on an intermittent basis, while some 15–25 million people in the United States use an NSAID on a regular basis. Importantly, the elderly who are at risk for multiple complications of NSAID therapy, constitutes a growing population that has a prevalence of NSAID use as high as 15%.

Epidemiology of Nonsteroidal Anti-Inflammatory Drug-Associated Acute Renal Failure

Unfortunately, the price paid for these therapeutic benefits include a number of gastrointestinal (GI) complications and, to a lesser degree, adverse renal effects. It has been estimated that from 5% to 7% of admissions to the hospital occur due to toxicity of NSAIDs, with the major organs involved being the GI tract, kidneys, and nervous system. The nephrotoxicity of NSAIDs, in particular hemodynamic ARF, is a relatively uncommon but important problem. It has been estimated that anywhere from 1% to 5% of patients who ingest NSAIDs will develop nephrotoxicity. Some calculations approximate that 500,000 persons are likely to develop some form of NSAID-associated adverse renal impairment (Table 15–2). Exposure to NSAIDs has been noted to double the risk of hospitalization for ARF in patients with chronic kidney disease (CKD). Patients with a history of heart failure and hypertension, as well as those treated with diuretics are at greatest risk of NSAID-induced ARF. The effect also appeared to be dose related with ARF occurring with higher NSAID dose. Hospitalized patients are at even higher risk. Of the cases of drug-induced ARF that develop in the hospital, it is estimated that nearly 16% are due to NSAID therapy.

As will be discussed in more detail, the healthy general population is at less risk for ARF as compared with patients who possess multiple risk factors associated with NSAID nephrotoxicity. The majority of adverse renal effects are attributable to inhibition of renal prostaglandins by NSAIDs. The selective COX-2 inhibitors (celecoxib, valdecoxib, rofecoxib) appear to have a renal profile similar to other NSAIDs. Therefore, the term NSAIDs will refer to both nonselective NSAIDs and the selective COX-2 inhibitors. Electrolyte and acid–base disorders that occur with these drugs (hyperkalemia, hyponatremia, metabolic acidosis), hypertension that develops or is exacerbated by NSAIDs, and disturbances in sodium balance (edema formation, exacerbation of heart failure) will also be reviewed briefly in the context of ARF.

Prostaglandin Synthesis

It is essential to review the pathway of prostaglandin (PG) production to facilitate an understanding of NSAID efficacy and toxicity. PGs, the major products of COX enzyme metabolism, are produced throughout the body and act at the local organ level in an autocrine and paracrine fashion. The initial step in PG synthesis is the liberation of arachidonic acid from cell membrane phospholipids. This reaction is mediated by phospholipase A2, which is triggered by a number of hormones and mechanical factors. Arachidonic acid is the substrate for COX. Following synthesis, PGs promptly exit the cell to bind PG receptors found on parent or neighboring cells, thereby modulating cellular functions (Figure 15–1).

Two isomers of COX, COX-1 and COX-2, catalyze the synthesis of PGs. These isomers share a similar amino acid sequence and catalytic function. Differences in gene regulation between the COX isomers provide a molecular basis for their purported roles as “constitutive” (COX-1) and “inducible” (COX-2) enzymes. These labels accurately describe the synthesis of COX in most tissues, where COX-1, but not COX-2 is expressed in appreciable levels at baseline. In contrast, abundant expression of COX-2 is demonstrated in macrophages and other cell types in response to inflammatory mediators. However, it is now known that COX-2 is also constitutively expressed and upregulated in the kidney and plays an important role in renal physiologic processes. It is probable that the nephrotoxicity of selective and nonselective NSAIDs results from the inhibition of COX-2 rather than COX-1.

Role of Prostaglandins in Renal Physiology

PG synthesis is of minimal importance in the kidney of healthy individuals with normal volume status. As such, it is not a primary regulator of renal function. Rather, eicosanoids locally modulate the effects of both systemic and locally produced vasoconstrictor hormones. A variety of PGs are synthesized within distinct anatomic locations in the kidney, including PGI2 , PGE2 , thromboxane A2 (TXA2), and PGF2α (Figure 15–2). PGI2 and PGE2 are the predominant mediators of physiologic activity in the kidney. Functionally, PGI2 and PGE2 induce vasodilation in interlobular arteries, afferent and efferent arterioles, and glomeruli.

In the loop of Henle and distal nephron, PGE2 decreases cellular transport of sodium chloride in thick ascending limb cells and collecting duct cells, respectively. An increase in renal sodium excretion and a decrease in medullary tonicity are the direct result of PGE2 action in these nephron segments. PGE2 and PGI2 also stimulate renin secretion in the juxtaglomerular apparatus, ultimately leading to increased angiotensin II and aldosterone synthesis, enhancing sodium retention and potassium excretion in the distal nephron. Finally, PGE2 and PGI2 also inhibit cyclic-AMP synthesis and oppose the action of antidiuretic hormone (ADH), facilitating water excretion.

Risk Factors for Nonsteroidal Anti-Inflammatory Drug-Associated Acute Renal Failure

Since basal PG production is low in healthy persons, the risk of NSAID-associated ARF is negligible. There are, however, risk factors (Table 15–3) that render the kidney PG dependent and, therefore, place patients at risk for the development of ARF when they use NSAIDs. PGs have their major role in the preservation of renal function when pathologic states supervene and compromise physiologic kidney processes. The development of “true” intravascular volume depletion, as seen with vomiting, diarrhea, and diuretic therapy, stimulates PG synthesis to optimize renal blood flow. “Effective” decreases in renal blood flow as seen with congestive heart failure (CHF), cirrhosis, and nephrotic syndrome also stimulate compensatory PG production. PGI2 and PGE2 antagonize the local effects of circulating angiotensin II, endothelin, vasopressin, and catecholamines that would normally maintain systemic blood pressure at the expense of the renal circulation. Specifically, these eicosanoids preserve glomerular filtration rate (GFR) by antagonizing arteriolar vasoconstriction and blunting mesangial and podocyte contraction induced by these endogenous vasopressors. A significant reduction in GFR can occur following administration of an NSAID to a patient with any of these underlying disease states (Figure 15–3).

PG production is also increased in CKD. Upregulation of PG synthesis in CKD is induced by intrarenal mechanisms activated to increase perfusion of remnant nephrons. As a result, impairment of PG production with an NSAID is associated with acute reductions in renal blood flow and GFR. Certain medications, such as drugs that antagonize the renin–angiotensin–aldosterone system (RAAS) such as angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs), increase the risk for ARF when an NSAID is administered. This is typically a problem in patients with true or effective intravascular volume depletion or underlying CKD. In the absence of underlying kidney disease, ARF rarely develops when these patients are euvolemic. Finally, the elderly are at risk for ARF because of unrecognized CKD, intravascular volume depletion, and hypoalbuminemia (which increases levels of free NSAID in the circulation).

Symptoms and Signs

In the absence of severe ARF, most patients with NSAID-associated prerenal azotemia are asymptomatic. The underlying risk factor that predisposes the patient to ARF often determines the clinical presentation. For example, patients with “true” intravascular volume depletion will present with uremic manifestations such as anorexia, nausea and vomiting, weakness, fatigue, inability to concentrate, and possibly GI dyspepsia (from NSAID gastropathy). They will not be hypertensive or edematous. In contrast, patients with “effective” intravascular volume depletion such as CHF, cirrhosis, or nephrotic syndrome will manifest volume overload. In the CHF patient, this can manifest as lung crackles, elevated jugular venous pulsations, and an S3 cardiac gallop from pulmonary edema, as well as peripheral pitting edema. Increased abdominal girth from ascites and worsened peripheral edema can develop in the patient with cirrhosis. Increased peripheral edema and anasarca may develop in the patient with nephrotic syndrome. Patients with CHF and nephrosis will often be hypertensive, whereas the patient with cirrhosis does not develop an increase in blood pressure. Patients with underlying hypertension, especially when under therapy with antihypertensive medications, will often manifest a destabilization of blood pressure and present with worsened hypertension. The patient with CKD will develop acute uremia, severe hypertension, increased peripheral edema, and CHF. Muscle weakness and cardiac arrhythmias (from hyperkalemia) can occur in CKD patients who receive NSAIDs. In addition, patients treated with medications that impair potassium homeostasis (ACE inhibitors, ARBs, spironolactone, eplerenone, calcineurin inhibitors, heparin) may also develop these adverse effects when an NSAID is added to their regimen. The elderly patient can present with any number of these clinical symptoms and signs.

Laboratory Findings

ARF associated with NSAIDs is hemodynamic and, therefore, the laboratory parameters typically reflect a “prerenal” state. Both serum and urine tests support this. BUN and serum Cr concentrations are both elevated, however, the BUN concentration typically increases more than the serum Cr concentration. In general, the BUN/Cr ratio is greater than 20, although this does not occur in all patients. Not uncommonly, several electrolyte disturbances may also be present. They include hyponatremia (serum [Na+] <135 mEq/L) from impaired renal water excretion and hyperkalemia (serum [K+] >5.5 mEq/L) with or without a non-anion gap metabolic acidosis. NSAID-induced hyporeninemic hypoaldosteronism underlies the hyperkalemia and non-anion gap metabolic acidosis. Importantly, these electrolyte and acid–base disturbances can develop even with minimal ARF as they are caused via direct effects of the NSAID on various tubular segments in the nephron.

Urine testing also points to a prerenal state of ARF. The urine specific gravity (SG) is typically greater than 1.015. Urine [Na+] is often less than 10–20 mEq/L. Calculation of the fractional excretion of sodium (FeNa+ = [urine Na+ × serum Cr] ÷ [serum Na+ × urine Cr]) reveals a value less than 1.0% (Table 15–4). This contrasts with a value greater than 3.0%, which would characterize acute tubular necrosis (ATN). Rarely, the specific gravity and FeNa+may be suggestive of ATN (SG ≤1.015, FeNa+ >3.0%) in patients with NSAID-associated ARF. This represents a situation in which some level of ischemic ATN developed, probably due to hypotension and severe renal hypoperfusion. Hyperkalemia is evaluated by calculating the transtubular potassium gradient (TTKG). TTKG is calculated as follows: TTKG = [urine K+÷ (urine osmolality/serum osmolality)] ÷ serum K+. A TTKG <6 in the setting of hyperkalemia reflects impaired renal potassium excretion due to hypoaldosteronism and is consistent with a diagnosis of NSAID nephrotoxicity.

Microscopic examination of the urine in patients with NSAID-associated ARF generally reveals a bland sediment with no cellular elements and sometimes a few hyaline casts. This is consistent with the prerenal azotemia that develops in this setting. Rarely, a few renal tubular epithelial (RTE) cells, RTE cell casts, and granular casts may be present if ischemic tubular injury has also developed. It is uncommon to view red blood cells, white blood cells, or casts containing these cellular elements unless another renal disease coexists. Their presence suggests superimposition of NSAID nephrotoxicity on another renal process, such as acute interstitial nephritis (AIN).

Imaging tests are normal in the absence of another concurrent disease process in the kidneys. Renal ultrasonography reveals normal sized kidneys with normal echotexture and no evidence of hydronephrosis. Computerized tomography (CT) scan and magnetic resonance imaging (MRI) are also normal in pure NSAID-associated ARF. An ACE inhibitor renal perfusion scan would show bilateral decreased uptake of tracer by the kidneys, consistent with prerenal azotemia.

ARF that develops in the setting of NSAID therapy needs to be differentiated from other prerenal states, as well as various intrinsic renal processes such as ATN and AIN. As with other causes of ARF, obstructive uropathy also needs to be considered in the differential diagnosis.

Since both “true” and “effective” intravascular volume depletion cause prerenal ARF and also are risk factors for NSAID-associated ARF, it is very difficult to ascribe renal injury to either alone. This is because the addition of an NSAID can potentially worsen underlying CHF and cirrhosis, making them appear as poorly compensated disease states. Regardless, discontinuation of the drug and therapy directed at the underlying clinical disease are required. Ischemic ATN can develop from the underlying clinical problem (ie, hypotension, sepsis) as well as from the effect of NSAID on renal hemodynamics, which further worsens renal perfusion and GFR in these settings. The urine data distinguish NSAID-associated ARF from ATN as an SG ≤1.015, urine [Na+] >20 mEq/L, FeNa+ >3%, and urine sediment with RTE cells, RTE cell casts, and granular casts characterize ATN.

ARF from AIN may also develop from an NSAID. Patients who do not recover within several days of stopping the NSAID should be evaluated for AIN. Although not always the case, AIN usually is associated with eosinophilia, tubular proteinuria (<1 g/day), pyuria (± eosinophiluria), and hematuria (Table 15–4). AIN from NSAIDs, however, is often devoid of the classic findings of AIN and can present with a bland urine sediment. Finally, ARF from partial urinary tract obstruction can be confused with NSAID-associated ARF because the urine sediment can also be unremarkable. However, the urine SG and electrolytes more often reflect tubular injury and are similar to those seen with ATN.

As with any drug that induces ARF, a number of uremic, metabolic, and intravascular volume-related complications can develop. Severe ARF may be associated with several uremia-related problems. These include central nervous system dysfunction (confusion, agitation, seizure), bleeding from platelet dysfunction, increased infection risk, malnutrition from a catabolic state, and inflammation of serosal surfaces (pericarditis, pleuritis). Hyponatremia, hyperkalemia, and metabolic acidosis can develop from ARF or from direct effects of the NSAID. Hyperkalemia can be more severe than expected for the level of renal dysfunction due to the state of hyporeninemic hypoaldosteronism promoted by the NSAID. Volume overload with pulmonary edema and peripheral edema can complicate NSAID-associated nephrotoxicity in cardiac patients. Precipitation of hepatorenal syndrome is a potentially devastating complication of NSAID therapy in cirrhotic patients. Also, diuretic resistance can develop in patients with CHF and nephrotic syndrome. Recently, increased risk for adverse cardiovascular events has been noted with both selective and nonselective NSAIDs.

The mainstay of therapy is discontinuation of the NSAID. Additional measures include correction of the underlying process that predisposed to nephrotoxicity. In some cases, intravenous normal saline rapidly repairs volume status and facilitates renal recovery in patients with intravascular volume depletion. High-dose intravenous diuretics, sometimes in combination, may be required in cardiac patients with ARF complicated by CHF and in nephrotic patients with severe edema and anasarca. Life-threatening hyperkalemia mandates intravenous calcium gluconate (stabilize cardiac tissue), intravenous insulin plus glucose, high-dose nebulized β2-agonists to shift K+ into cells, and diuretic therapy to enhance renal excretion in patients who maintain reasonable kidney function. It is unusual for metabolic acidosis to be severe enough to warrant sodium bicarbonate administration. Severe ARF complicated by advanced uremia or other life-threatening complications (pulmonary edema, metabolic disturbances) that does not recover within a few days of stopping NSAID therapy requires renal replacement therapy.

Fortunately, discontinuation of the NSAID is often associated with recovery of renal function. In general, renal function returns to baseline within 2–5 days, however, recovery may be delayed in patients with decompensated heart disease, cirrhosis, and underlying CKD. It is extremely unusual that a patient does not recover kidney function. The failure to do so should elicit an evaluation for other causes of ARF such as ATN, AIN, and obstructive uropathy.