The clinical efficiency of nonsteroidal antiinflammatory drugs (NSAIDs) has made their use very popular for multiple medical indications. Their adverse effects have been evaluated and are well known. These include fluid and electrolyte abnormalities (1), renal failure (2), gastropathy and gastrointestinal bleeding (3), uncontrolled hypertension (4), and aggravation of congestive heart failure (5). Other side effects involving the small bowel and colon, liver, central nervous system, and hematologic system do exist but are less frequent.
The principal mechanism of action (and of adverse reactions) of NSAIDs is inhibition of cyclooxygenase (COX or prostaglandin synthase), which impairs the conversion of arachidonic acid to prostaglandins, prostacyclins, and thromboxanes. However, many other mechanisms of action have also been proposed to explain the diversity of their effects (6). Two isoforms of the COX enzymes have been described. COX-1 produces prostaglandins ubiquitously and regulates cellular housekeeping functions, such as gastric cytoprotection, vascular homeostasis, and kidney function. COX-2, the inducible isoform, was previously thought to be expressed only in inflamed tissues. As a result, the selective inhibition of COX-2 has been proposed to be safer than nonselective NSAIDs, maximizing effectiveness and minimizing toxicity. However, because COX-2 has also been localized in noninflamed tissues, including the kidney, it has been suggested that this isoform has a renal homeostatic function (7). Therefore, inhibition of COX-2 might compromise kidney function, just as do nonselective NSAIDs.
A review of clinical trials conducted to evaluate the renal effects of COX-2 inhibitors, particularly rofecoxib and celecoxib, evaluated data on their qualitative effects on renal prostaglandins, glomerular filtration rate, sodium excretion, and their clinical effects on equilibration of blood pressure and occurrence of leg edema. The authors concluded that COX-2 inhibitors had the same relative potential as nonselective NSAIDs for adverse renal effects (8). Because of the known fragility of kidney function in the elderly and the high prevalence of osteoarthritis in this population, many clinical trials have evaluated the renal effects of COX-2 inhibitors in the elderly; these trials also concluded they had the same acceptable safety profile as the nonselective NSAIDs (9–12). Renal hemodynamic effects of the COX-2 inhibitors rofecoxib and celecoxib, leading to acute renal failure, have been analyzed and reviewed (13). In contrast, the occurrence of COX-2 inhibitor-induced interstitial nephritis has been described very rarely. Herein we describe a case of acute renal failure due to severe interstitial nephritis in a young woman treated with rofecoxib, a selective COX-2 inhibitor. This report illustrates the potential for similar renal adverse effects in both selective and nonselective COX inhibitors, and emphasizes the possibility of severe adverse renal effects of COX-2 inhibitors by several different mechanisms, even in young healthy individuals.
A 23-year-old otherwise healthy woman, with a history of lumbar backaches due to a herniated disc at the level of L4-L5, was treated with 25 mg/day of rofecoxib for a period of ∼1 year, with partial relief. During that period, 2 epidural steroid injections were performed without any improvement. Because of increasing pain and symptomatic compression of the right L5 root she underwent a laminectomy and discectomy of L4-L5. During her year of treatment with rofecoxib, she complained of headache and visual disturbances and was referred to the emergency room. Physical examination revealed an afebrile, overweight young woman with a body mass index of 34 kg/m2, heart rate of 86 beats/minute, and blood pressure of 140/80 mm Hg. On retinal examination, bilateral papilledema was noted. The rest of the physical examination was otherwise normal. Blood test results (Table 1) showed renal failure, elevated inflammatory markers, and no peripheral eosinophilia in the leukocyte differential. Urinalysis revealed proteinuria of 1.6 gm/24 hours and unremarkable microscopy. The creatinine clearance was 13 ml/minute. Lumbar puncture (following normal brain computed tomography) showed cerebral spinal fluid pressure of 330 mm H2O, and cultures and cytology were negative.
Serologies for infections: CMV, Parvovirus B19, EBV, HSV, Q fever, Brucella, HAV, HBV, HCV, and HIV
Immunologic and various infectious serologic blood tests (Table 1) were negative, as were blood and urine cultures. A transesophageal echocardiogram ruled out endocarditis, and brain magnetic resonance venography ruled out venous thrombosis as a cause of pseudotumor cerebri. An abdominal skin biopsy did not stain with Congo red, ruling out amyloidosis. Renal ultrasound examination showed the right kidney to be 11.6 cm and the left, 12.6 cm, with no evidence of hydronephrosis.
To exclude a potentially reversible prerenal condition, the patient was treated with intravenous hydration, but there was no significant improvement in her renal function. A kidney biopsy showed normal glomeruli and blood vessels. The most striking findings were edema and dense polymorphonuclear, lymphohistiocytic, and plasma cell infiltrates involving the tubular interstitium. Several eosinophils were also seen. In many foci, inflammatory cells infiltrated the tubular epithelium, causing tubular destruction. Silver and trichrome staining revealed no excess extracellular matrix or collagen deposition in the biopsy material. Immunofluorescence microscopy revealed peritubular fibrin staining and trace peritubular and periarteriolar C3 deposition. The above findings were consistent with severe interstitial nephritis (Figure 1).
High-dose intravenous corticosteroid therapy was initiated with improvement in kidney function in the first few days of treatment. Sixteen weeks later, her creatinine was back to normal and retinal examination showed no papilledema. The patient is not currently receiving any medications.
Renal side effects of NSAIDs might be classified in a simple manner according to the duration of treatment with the drug (14): hemodynamically mediated acute renal failure and electrolyte disorders occur within hours after the initial introduction of the NSAIDs; in contrast, acute tubulointerstitial nephritis typically occurs within 2–18 months after the introduction of NSAIDs. Analgesic nephropathy, manifested by papillary necrosis and chronic interstitial nephritis, is usually due to many years of NSAID use. It has been suggested that the new selective COX-2 inhibitors should be considered to have equivalent nephrotoxicity as nonselective COX inhibitors (9–13, 15). Thus, the description of COX-2 inhibitor-induced interstitial nephritis is not surprising. However, to our knowledge, only 1 case involving the COX-2 inhibitor rofecoxib (16) and a few involving celecoxib (17–19) have been well documented with renal biopsies. As a result, the exact prevalence of COX-2 inhibitor-induced interstitial nephritis is not known and cannot be compared with that of nonselective NSAID-induced interstitial nephritis, which is thought to occur in 1 of 5,000–10,000 treated patients (14).
Interstitial nephritis is an inflammatory process of the interstitium, which can involve the tubules; it is then a tubulointerstitial nephritis, and it may be acute or chronic (20). Depending on the etiology, mechanisms involved can be either direct through cytotoxicity, or indirect through inflammatory or immunologic reactions (21). On the basis of several clinical and immunohistologic findings, mechanisms involved in drug-induced interstitial nephritis are considered cell mediated. The interstitial infiltrate in drug-induced acute interstitial nephritis is distinguishable from the pathologic patterns seen in other causes (20, 21): One can see edematous interstitium containing eosinophils, mononuclear cells, and plasma cells. Granulomata, similar to those seen in sarcoidosis interstitial nephritis, without fibrosis, may be seen. Another relatively specific pattern of the drug-induced acute interstitial nephritis is the tubulitis, characterized by lymphocytes invading the tubules with variable ranges of tubular epithelial cell damage. Such a picture can be rarely seen in other clinical settings, such as idiopathic acute interstitial nephritis, lupus nephritis, acute interstitial nephritis associated with anti–tubular basement membrane antibodies, and TINU syndrome (tubulo interstitial nephritis-uveitis). Thus, our patient presented typical pathologic features of drug-induced acute interstitial nephritis. Moreover, the pathologic findings in our patient ruled out the possibility of an analgesic-induced interstitial nephritis because the specific pathologic patterns of this latter entity, including papillary necrosis, tubular atrophy, and interstitial fibrosis, were all absent.
Because NSAID-induced interstitial nephritis with or without minimal-change glomerulopathy can be severe enough to lead to dialysis and to irreversible end-stage renal disease, some authors propose to treat such patients with steroids (20). This is what we did in our patient after she had no spontaneous improvement in her renal function test results.
Drug-induced interstitial nephritis often leads to signs of hypersensitivity, with skin rash, fever, eosinophilia, and eosinophiluria (21), but those signs are inconstant and present together in only 5% of cases (20). However, nonselective NSAID-induced interstitial nephritis syndromes were distinguished from other drug-induced interstitial nephritides by their particular association with higher-range proteinuria, onset occurrence after longer exposure to the drug, and infrequent association with hypersensitivity signs (2, 22, 23), as was demonstrated in the present case. This milder expression of hypersensitivity was suggested to be related to the direct antiinflammatory effect of NSAIDs (24), but could be related to other factors. Thus, COX-2 inhibitor-induced interstitial nephritis might be a particular entity without hypersensitivity signs, as has been proposed for nonselective NSAID-induced interstitial nephritis.
Rofecoxib use has been shown to reduce the glomerular filtration rate by ∼12% in elderly patients with mild renal impairment (12). However, our patient had no age- or illness-related risk factors for the development of NSAID- or selective COX inhibitor-induced nephropathy, apart perhaps for the long-term use of rofecoxib. Indeed, it might be of interest to study the influence of the duration of such treatment on the incidence of such renal side effects as interstitial nephritis because there are no available data on this specific point.
Another interesting aspect of this report is the pseudotumor cerebri syndrome that resolved together with the interstitial nephritis. We cannot exclude an association with the use of rofecoxib but other causes, such as the obesity observed in this patient (25), are possible.
In summary, acute renal failure and interstitial nephritis were associated with prolonged use of rofecoxib in a healthy young woman. Due to widespread use of COX-2 inhibitors, and until specific studies are published, an awareness of the possibility of renal adverse effects is required in all classes of patients. In cases of COX-2 inhibitor-related interstitial nephritis, as in other drug-induced interstitial nephritides, corticosteroid therapy may be indicated.