Portal de Periódicos da CAPES

Acute kidney injury following proton pump inhibitor therapy

2006; Elsevier BV; Volume: 71; Issue: 6 Linguagem: Inglês

10.1038/sj.ki.5002038

1523-1755

Ursula C. Brewster, Mark A. Perazella,

Acute Kidney Injury Research

A 63-year-old male with a past medical history significant for hypertension for 12 years, stage 3 chronic kidney disease (baseline serum creatinine concentration, 1.4 mg/dl; estimated glomerular filtration rate, 54 ml/min) secondary to hypertensive nephrosclerosis presented with generalized malaise and weakness for 2 weeks. He was referred for admission to the hospital for evaluation of renal dysfunction. Three months before this admission, he was treated with naproxen 500 mg per day for bilateral knee pain owing to degenerative joint disease. After 2 weeks of therapy, he developed epigastric pain with anorexia, vomiting, and hematemesis. He underwent upper endoscopy, which revealed esophageal, gastric, and duodenal erosions and ulcerations. Serum creatinine concentration was 2.3 mg/dl. Naproxen was discontinued, esomeprazole 40 mg per day was initiated as anti-acid therapy and a liter of normal saline was administered intravenously. Two days later, the serum creatinine concentration returned to baseline (1.4 mg/dl) and the patient was discharged home on esomeprazole and tramadol/acetaminophen. On presentation to the hospital, blood pressure was 142/88 mm Hg, pulse 100/min, temperature 98.9 degrees F, and respiratory rate 18/min. The conjunctivae were pale with anicteric sclera. Heart and lung examination was unremarkable except for a grade 1/6 systolic ejection murmur over the left sternal border. The abdomen had minimal epigastric tenderness, and there was trace edema in the ankles. There was no skin rash, petechiae, or purpura. Medications included amlodipine 10 mg daily, esomeprazole 40 mg daily, diazepam 2.5 mg daily, tramadol/acetaminophen 100 mg twice daily, and calcitriol 1 mcg every other day. Admission laboratory data are noted in Table 1. The remaining laboratory values, including liver function tests and coagulation profile were within normal range. There was no peripheral eosinophilia. Urinalysis was notable for specific gravity, 1.015; pH, 6.0; trace blood; 1+ protein; and trace leukocyte esterase. Examination of the urine sediment revealed 2–5 dysmorphic red blood cells/high-power field, 2–5 white blood cells/high-power field; few renal tubular epithelial cells; no cellular or granular casts were present. Urine eosinophils (Wright stain) were negative. Fractional excretion of sodium was 1.8% whereas spot urine protein/creatinine ratio was 0.65 mg/g. A sonogram of kidneys demonstrated slightly echogenic but normal-sized kidneys without hydronephrosis. The patient was administered 2 Liters of normal saline and all medications except amlodipine were held. Despite this, the serum creatinine concentration rose the next day to 5.2 mg/dl, whereas urine output was 550 ml for the day. A kidney biopsy was performed for non-oliguric acute kidney injury.Table 1Serum laboratory resultsLaboratory testConcentrationBlood urea nitrogen56 mg/dlCreatinine4.9 mg/dlPotassium5.2 meq/lBicarbonate18 meq/lAlbumin3.7 g/dlCalcium8.8 mg/dlWhite blood cell count7.6 × 103/μlHemoglobin12.5 g/dl Open table in a new tab Light microscopic examination revealed 15 glomeruli, none of which were sclerotic. The glomeruli showed no significant abnormalities (Figure 1). No cellular or fibrous crescents were present. The blood vessels revealed mild intimal and medial thickening. There was a diffuse cellular infiltrate, which consisted of lymphocytes, neutrophils, and eosinophils (Figure 2). Tubular atrophy and tubulitis were also present. Immunofluorescence and electron microscopy of glomeruli did not reveal immune staining or electron dense deposits, respectively. On electron microscopy, the glomerular endothelial cell fenestrations were normal and foot processes were intact without effacement.Figure 2Light microscopy demonstrates diffuse cellular infiltrate within the interstitium. The infiltrate consists of eosinophils (arrowhead) and lymphocytes. Tubulitis is also present (arrow) (hematoxylin and eosin stain).View Large Image Figure ViewerDownload (PPT) Acute interstitial nephritis secondary to esomeprazole. Esomeprazole was discontinued in view of the diagnosis of associated acute interstitial nephritis (AIN). Given the severity of the kidney dysfunction a short course of oral prednisone (40 mg/day) was started and within 6 days the serum creatinine concentration began to fall. The steroids were tapered off after 2 weeks; subsequently, his serum creatinine concentration leveled off at 1.8 mg/dl. Proton pump inhibitors (PPIs) are widely prescribed to treat numerous acid-related gastrointestinal disorders. In 1989, omeprazole was introduced as the first effective PPI; subsequently other drugs in this class have been synthesized and released into clinical practice. There are five PPIs currently available in the US market. In 2005, over 43 million prescriptions were written for anti-ulcer therapy in the United States, and over 8 billion dollars were spent on PPIs. These drugs work by specifically binding to the H+/K+ ATPase at the secretory surface of gastric parietal cells, decreasing both basal and stimulated acid secretion. They are metabolized by the cytochrome P450 system and their inactive metabolites are largely excreted in the urine. As less than 1% of active drug is excreted in the urine, the need for dose adjustment in the setting of a decreased glomerular filtration rate and kidney disease is eliminated.1.Delhotal-Landes B.F.B. Duchier J. Molinie P. et al.Pharmacokinetics of lansoprazole in patients with renal or liver disease of varying severity.Eur J Clin Pharmacol. 1993; 45: 367-371Crossref PubMed Scopus (57) Google Scholar The fact that no active drug is excreted in the urine explains the lack of renal effects of these medications. When studied, omeprazole had no effect on renal electrolyte balance or urine pH in healthy male subjects,2.Howden C.W.R.J. Omeprazole, a gastric 'proton pump inhibitor': lack of effect on renal handling of electrolytes and urinary acidification.Eur J Clin Pharmacol. 1984; 26: 639-640Crossref PubMed Scopus (24) Google Scholar explaining the absence of renal tubular acidosis in patients treated with PPIs. In general, renal complications are rare, although the widespread use of these drugs requires recognition of complications when they do occur. To date, the only direct adverse renal effects described are the rare development of hyponatremia (nine cases) and the increasing number of cases of PPI-AIN. This article will focus on PPI-induced AIN. Acute interstitial nephritis is a relatively infrequent cause of acute kidney injury, and in several large series, represents approximately 2–3% of all renal biopsies.3.Buysen J.G. Houthoff H.J. Krediet R.T. Arisz L. Acute interstitial nephritis: a clinical and morphological study in 27 patients.Nephrol Dial Transplant. 1990; 5: 94-99Crossref PubMed Scopus (113) Google Scholar, 4.Clarkson M.R. Giblin L. O'Connell F.P. et al.Acute interstitial nephritis: clinical features and response to corticosteroid therapy.Nephrol Dial Transplant. 2004; 19: 2778-2783Crossref PubMed Scopus (220) Google Scholar, 5.Davison A.M. Jones C.H. Acute interstitial nephritis in the elderly: a report from the UK MRC Glomerulonephritis Register and a review of the literature.Nephrol Dial Transplant. 1998; 13: 12-16Crossref PubMed Scopus (45) Google Scholar In undiagnosed acute kidney injury with normal-sized kidneys on ultrasonography, AIN is more common as noted in up to 27% of cases on kidney biopsy.6.Farrington K. Levison D.A. Greenwood R.N. et al.Renal biopsy in patients with unexplained renal impairment and normal kidney size.QJ Med. 1989; 70: 221-233PubMed Google Scholar In a review of the medical literature, 88 case reports of PPI-induced AIN are found. Of these, 76 cases are biopsy proven. This suggests a class effect as all five of the PPIs clinically available are implicated, with 75 cases attributed to omeprazole, six to pantoprazole, three to esomeprazole, two to lansoprazole, and two to rabeprazole. The higher total number of AIN cases in omeprazole-treated patients likely represents the longer availability of this agent, although other factors, such as inherent immunogenicity of the molecule cannot be ruled out. The sentinel case report of AIN from omeprazole was published in 1992.6.Farrington K. Levison D.A. Greenwood R.N. et al.Renal biopsy in patients with unexplained renal impairment and normal kidney size.QJ Med. 1989; 70: 221-233PubMed Google Scholar Subsequently, several case reports and series over the next 12 years confirmed the association between omeprazole and AIN (n=29).7.Post A.T.V.G. Zanen A.L. Reversible renal failure after treatment with omeprazole.Neth J Med. 2000; 57: 58-61Crossref PubMed Scopus (22) Google Scholar, 8.Myers R.P. McLaughlin K. Hollomby D.J. Acute interstitial nephritis due to omeprazole.Am J Gastroenterol. 2001; 96: 3428-3431Crossref PubMed Google Scholar, 9.Portera-Cailliau C.S.A. Acute interstitial nephritis due to omeprazole.Ann Long-Term Care. 1999; 7: 206-207Google Scholar, 10.Geetha D. Omeprazole-induced acute interstitial nephritis.Am J Gastroenterol. 1999; 94: 3375-3376Crossref PubMed Google Scholar, 11.Wall C.A. Gaffney E.F. Mellotte G.J. Hypercalcaemia and acute interstitial nephritis associated with omeprazole therapy.Nephrol Dial Transplant. 2000; 15: 1450-1452Crossref PubMed Scopus (16) Google Scholar, 12.Torregrosa E. Rovira R.E. Calvo C. et al.Acute interstitial nephritis associated with omeprazole therapy.Nefrologia. 2004; 24: 61-63PubMed Google Scholar In 2004, other PPIs were implicated in causing AIN. Two large case series describing the association of PPIs and AIN are worth noting. In 2004, Torpey et al.13.Torpey N. Barker T. Ross C. Drug-induced tubulo-interstitial nephritis secondary to proton pump inhibitors: experience from a single UK renal unit.Nephrol Dial Transplant. 2004; 19: 1441-1446Crossref PubMed Scopus (81) Google Scholar published a description of all cases of biopsy-proven AIN diagnosed at their hospital (United Kingdom) over a 4-year period from 1995 to 1999. Of a total of 24 identified cases of AIN, 14 cases were drug-induced AIN (60%), of which eight were due to PPIs (omeprazole=6, lansaprazole=2). The rest of the AIN cases were caused by an non-steroidal anti-inflammatory drugs (n=4), amoxacillin (n=1), and warfarin (n=1). Most patients presented with nonspecific symptoms and all recovered kidney function to some degree, but most were left with some level of chronic kidney disease. Recently, 15 cases of AIN associated with PPIs (14 omeprazole, one pantoprazole) were published by investigators in the Auckland region of New Zealand over a 3-year period (2002–2005).14.Simpson I.J.M.M. Pilmore H. Manley P. et al.Proton pump inhibitor and acute interstitial nephritis: report and anlysis of 15 cases.Nephrology. 2006; 11: 381-385Crossref PubMed Scopus (95) Google Scholar Twelve out of 15 cases were biopsy proven; time from initiation of drug to AIN ranged from 10 days to 18 months. Although all of the patients recovered kidney function, 12/14 patients were left with chronic kidney disease. Additionally, based on data from the Centre for Adverse Reactions Monitoring (CARM) in New Zealand, PPIs are now the most common cause of drug-induced AIN in their region (32%), with an incidence of one in 12 500 patient years. The largest report to date documenting the association of PPIs with AIN was recently published by Geevasinga et al.15.Geevasinga N. Coleman P.L. Webster A.C. Roger S.D. Proton pump inhibitors and acute interstitial nephritis.Clin Gastroenterol Hepatol. 2006; 4: 597-604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar These authors examined the issue of PPI-induced AIN by two methods. First they retrospectively evaluated all cases of biopsy-proven AIN noted in two Australian teaching hospitals over a period of 10 years (1993–2003). The cases were evaluated for demographics, clinical parameters (symptoms/signs, type of PPI, concurrent medications, level of renal function, time to development of AIN) and histopathology (independently reviewed by pathologists). Second, they queried the Therapeutic Goods Administration (TGA) database (1991–2004), which is a unit of the Australian Government Department of Health and Aging that records spontaneously reported adverse drug effects after they have been reviewed by the TGA Advisory Committee. Cases were identified and reviewed for clinical features including symptoms/signs, type of PPI, level of renal function, and time to AIN/renal failure (excluding those with concurrent multiorgan involvement and hospitalized cases). Following data review, cases were classified as 'biopsy proven AIN', 'suspected interstitial nephritis', 'unexplained acute renal failure' and 'renal impairment'. The retrospective review of hospital data found 28 cases of biopsy-proven AIN, of which 18 (64%) were due to a PPI (omeprazole=11, pantoprazole=3, esomeprazole=3, rabeprazole=1). AIN developed at a mean of 11 weeks following PPI administration; nonspecific symptoms were common. The TGA database review noted 34 cases of 'biopsy-proven AIN', 10 cases of 'suspected interstitial nephritis', 20 cases of 'unexplained acute renal failure', and 26 cases of 'renal impairment' associated with PPI therapy. Although the TGA database suggests an association between PPI therapy and AIN/AKI, it does not prove cause and effect and therefore was not used in the calculation of the total number of cases of PPI-induced AIN (n=88). However, the combination of these data when examined in the context of previously reported cases strongly supports a relationship between PPI therapy and development of AIN. Classically, AIN associated with methicillin presents with hypersensitivity as manifested by the triad of symptoms including fever, rash, and eosinophilia. However, as with other non-methicillin drugs, this presentation appears to be less common in patients with PPI-induced AIN. Where reported, only about 10% had the classic triad of hypersensitivity in the cases published to date. Fewer then 50% of patients had fever, less then 10% had rash, and approximately 33% had eosinophilia. Most patients had nonspecific complaints including weakness, fatigue, malaise, and anorexia on presentation. In the Geevasinga paper, fatigue and nausea were present in 39% of patients, whereas 22% complained of weakness. Figure 3 demonstrates the clinical and laboratory findings noted with AIN associated with methicillin, other non-methicillin type drugs (excluding non-steroidal anti-inflammatory drugs), and PPI-associated AIN (all cases combined).16.Rossert J. Drug-induced acute interstitial nephritis.Kidney Int. 2001; 60: 804-817Abstract Full Text Full Text PDF PubMed Scopus (320) Google Scholar It is notable that the clinical and laboratory presentation of PPI-induced AIN is similar to that of the non-methicillin drug-associated AIN. The timing from initiation of drug to presentation with kidney involvement appears to be quite variable, occurring anywhere from 1 week to 9 months. When combining all of the cases published, symptoms, and/or identification of AIN occurred on average 9.9 weeks after starting PPI therapy, not very different than the time frame (11 weeks) described by Geevasinga et al.15.Geevasinga N. Coleman P.L. Webster A.C. Roger S.D. Proton pump inhibitors and acute interstitial nephritis.Clin Gastroenterol Hepatol. 2006; 4: 597-604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar In patients who were re-challenged (most often accidental or unsuspected) with the drug after suspected PPI-induced AIN, symptoms occurred much more rapidly, usually developing kidney injury within days.17.Ruffenach S.J. Siskind M.S. Lien Y.H. Acute interstitial nephritis due to omeprazole.Am J Med. 1992; 93: 472-473Abstract Full Text PDF PubMed Scopus (91) Google Scholar, 18.Christensen P.B. Albertsen K.E. Jensen P. Renal failure after omeprazole.Lancet. 1993; 341: 55Abstract PubMed Google Scholar, 19.Assouad M. Vicks S.L. Pokroy M.V. Willcourt R.J. Recurrent acute interstitial nephritis on rechallenge with omeprazole.Lancet. 1994; 344: 549Abstract PubMed Scopus (30) Google Scholar, 20.Gronich J.H. Snipes E.R. Stein H.D. et al.Omeprazole related acute interstitial nephritis (AIN) with omeprazole (Abstract).JASN. 1994; 5: 394Google Scholar Diagnosis of PPI-induced AIN should be made based on renal biopsy given the infrequency of classic signs and symptoms. Biopsy typically shows an interstitial infiltrate with or without tubulitis. In the largest case series to date where biopsy findings were recorded, 88% of patients, but not all, had eosinophils in the tubulointerstitium.15.Geevasinga N. Coleman P.L. Webster A.C. Roger S.D. Proton pump inhibitors and acute interstitial nephritis.Clin Gastroenterol Hepatol. 2006; 4: 597-604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar Glomeruli in most cases were normal, unless there was another unrelated kidney lesion also present. Management of PPI-induced AIN is similar to that of other causes. Key to treatment is the cessation of the offending agent. In addition to drug withdrawal, most cases of PPI-associated AIN were treated with some form of corticosteroid therapy. The debate about the use of steroids in this disorder remains active and suffers from a lack of prospective randomized controlled trials. In the largest retrospective study to date, data from 2598 patients with AIN was reviewed and there was no statistical difference in final serum creatinine concentration in those patients treated with steroids and those who were not. There is some thought, however, that steroids may play a role in hastening renal recovery, albeit a similar recovery, compared to no therapy.3.Buysen J.G. Houthoff H.J. Krediet R.T. Arisz L. Acute interstitial nephritis: a clinical and morphological study in 27 patients.Nephrol Dial Transplant. 1990; 5: 94-99Crossref PubMed Scopus (113) Google Scholar,21.Pusey C.D. Saltissi D. Bloodworth L. et al.Drug associated acute interstitial nephritis: clinical and pathological features and the response to high dose steroid therapy.QJ Med. 1983; 52: 194-211PubMed Google Scholar Steroids are typically dosed as oral prednisone at 1 mg/kg/day for 1–2 months and tapered rapidly. Mycophenylate mofetil has recently been shown to successfully treat AIN from multiple causes (including drugs) in steroid resistant/dependent cases.22.Preddie D.C.M.G. Radhakrishnan J. Nickolas T.L. et al.Mycophenolate mofetil for the treatment of interstitial nephritis.Clin J Am Soc Nephrol. 2006; 11: 718-722Crossref Scopus (77) Google Scholar Thus, mycophenylate mofetil represents another potential therapeutic option for patients who are steroid dependent or resistant. When recognized early, the prognosis of PPI-induced AIN seems to be quite good. Of note, in all the case reports published, very few patients required renal replacement therapy, and none developed end-stage renal disease. Many patients, however, were left with some renal impairment and chronic kidney disease. In the study by Geevasinga et al.,15.Geevasinga N. Coleman P.L. Webster A.C. Roger S.D. Proton pump inhibitors and acute interstitial nephritis.Clin Gastroenterol Hepatol. 2006; 4: 597-604Abstract Full Text Full Text PDF PubMed Scopus (125) Google Scholar patients recovered kidney function (Figure 4), however, mean calculated creatinine clearance was 15.9 ml/min and 11.5 ml/min lower than baseline at 3 and 6 months, respectively. Simpson and colleagues also noted incomplete recovery of kidney function as the mean serum creatinine concentration increased from a baseline value of 0.94–1.57 mg/dl after PPI-induced AIN. In other reported cases, patients also recovered to lower levels of kidney function (serum creatinine concentration, 1.3 mg/dl) as compared with baseline (serum creatinine concentration, 1.1 mg/dl). In conclusion, PPIs are widely prescribed as therapy for acid-related gastrointestinal diseases. They are clearly associated with the development of AIN, with a clinical presentation typical of non-methicillin type drugs. PPI-induced AIN should be suspected as a potential cause of acute kidney injury in patients without an otherwise obvious cause of kidney dysfunction. Importantly, despite recovery of kidney function in the majority of cases following drug withdrawal (±steroids), most patients are left with some level of chronic kidney disease.