Over-The-Counter Analgesic Use

CLINICAL COMMENTARY www.jasn.org

Over-the-Counter Analgesic Use

Marc E. De Broe and Monique M. Elseviers

Laboratory of Pathophysiology and Department of Nursing Sciences, University of Antwerp, Antwerp, Belgium

ABSTRACT Chronic analgesic nephropathy, particularly chronic interstitial nephritis and renal or chronic interstitial nephritis. Still, papillary necrosis, results from daily use for many years of mixtures containing at other investigators have used this vali- least two analgesics and caffeine or dependence-inducing drugs. Computed to- dated diagnostic test in an epidemiologic mography scan can accurately diagnose this disease even in the absence of reliable study related to late referral of patients information on previous analgesic use. The occasion to moderate regular use of with severe renal failure toward renal aspirin and nonsteroidal anti-inflammatory drugs is without renal risk when renal units. function is normal. Paracetamol use is less clear although the risk is not great. The A study of patients with ESRD (Na- continued use of non–phenacetin-combined analgesics with or without nonsteroi- tional Analgesic Nephropathy Study dal anti-inflammatory drugs is associated with faster progression toward renal [NANS]) in the United States evaluated impairment. As long as high-risk analgesic mixtures are available over the counter, the value of the non–contrast-enhanced analgesic nephropathy will continue to be a problem. CT as diagnostic test for AN.4 It turns out that in comparison with previous stud- J Am Soc Nephrol 20: 2098–2103, 2009. doi: 10.1681/ASN.2008101097 ies, the sensitivity of the noncontrast CT scan for the detection of analgesic-associated kidney injury was too low to be Chronic analgesic nephropathy (AN) is a test to diagnose or rule out AN as a pos- used as a routine test; that is, a specificity slowly progressive renal disease resulting sible cause of renal disease in patients of Ͼ95%, comparable with the earlier from daily use for many years of mix- with ESRD as well as in patients with reports, was found. The low prevalence tures containing at least two analgesics stages 3 to 4 chronic kidney disease of AN (clearly Ͻ5%) found in the United (e.g., aspirin, paracetamol, pyrozolones, (CKD), even in the absence of reliable States, however, precluded calculating phenacetin) and caffeine, codeine, information on the use of combined an- the clinically relevant sensitivity of the and/or barbiturates, which may lead to algesics (Figure 1).3 test. psychological dependence and overuse. Known causes of renal papillary ne- The disease is characterized by capillary crosis/calcifications other than com- sclerosis, renal cortical atrophy, chronic bined analgesic use are diabetes, sickle WHAT KIND OF ANALGESICS ARE interstitial nephritis, and papillary scle- cell disease, and pyelonephritis, the INVOLVED IN THE GENERATION rosis/necrosis/calcifications. In a num- differential diagnosis of the two first OF AN? ber of cases, not related to the progres- conditions being straightforward. sion or stage of renal failure, the Chronic pyelonephritis is associated Phenacetin was present in all products uroepithelia can develop transitional cell with progressive renal scarring, in the used by patients described in the early carcinoma.1,2 context of intrarenal reflux of infected reports of AN. This finding is the sole The lack of reliable validated criteria, urine. Calcified papillary necrosis ob- argument to support the generally ac- the high prevalence of AN in 1980 served in the majority of the papilla in cepted “phenacetin kidney’ concept. Ex- through 1990 in Belgium (Ͼ15% of the both kidneys of patients with AN is not dialysis population), and the underesti- seen in chronic pyelonephritis. Published online ahead of print. Publication date mation of the disease in Europe led us to Several studies using this validated di- available at www.jasn.org. perform studies defining and validating agnostic test showed either the absence Correspondence: Dr. Marc E. De Broe, University of the criteria for AN. On the basis of the or low prevalence of AN, whereas others Antwerp, Laboratory of Pathophysiology, Univer- siteitsplein 1, B-2610, Wilrijk/Antwerpen, Belgium. results of three multicenter multina- confirmed the underestimation of AN in Phone: ϩ32-3-820-2599; Fax: ϩ32-3-820-2592; tional studies, computed tomography their country and showed a substantial E-mail: [email protected]

(CT) scanning without contrast medium number of patients with AN exist in the Copyright ᮊ 2009 by the American Society of turned out to be a valuable and validated cohort of patients with unknown cause Nephrology

2098 ISSN : 1046-6673/2010-2098 J Am Soc Nephrol 20: 2098–2103, 2009 www.jasn.org CLINICAL COMMENTARY

Figure 1. Diagnostic criteria of analgesic nephropathy. (A) Macroscopic aspect of an AN kidney from a patient with ESRD. (B) Diagnostic criteria used. (C) CT scans without contrast material of an individual with normal kidneys and patients with AN and with stage 3 CKD and ESRD. Adapted from De Broe and Elseviers.2 perimental studies revealed, however, any analgesic mixtures containing two and Morlans et al.16 resulted in comparable that aspirin and phenazone derivatives— analgesic components combined with odds ratios (ORs) between 2 and 3, despite the drugs invariably taken with phenace- caffeine and/or codeine.9 differences in study design. McCredie’s11 tin—all produce experimental nephro- Clinical observations performed study showed a considerably higher OR by toxicity more readily than phenacetin.5 within the framework of the diagnostic using the more specific lesion of renal pap- Whether phenacetin is the sole re- criteria of AN document the nephrotox- illary necrosis as the primary identifier of sponsible ingredient causing AN still is a icity of the combinations of salicylic acid disease. matter of intense debate. Phenacetin be- with paracetamol, pyrazolones, paracet- A solid demonstration of the associa- lievers argue it was the ban of phenacetin amol-pyrazolones, or two pyrazolones tion between analgesic use and renal failure that caused the decline in AN incidence associated with dependence-producing is also provided by two prospective, con- observed in Switzerland6 and in other drugs.10 Additional data from Australia trolled, cohort studies performed of pa- countries.7 In contrast, researchers in the and Belgium support the suggestion that tients with excessive use of analgesic mix- field of AN have joined forces to identify withdrawal of phenacetin is not solely re- tures in Switzerland and Belgium with a the nephrotoxic potential of all analgesic sponsible for the decline in AN.2 Only follow-up of 10 and 6 yr, respectively.21,22 mixtures with or without phenacetin. In after the drastic decrease in the sales of all Although both studies differed substan- 1995, the ad hoc committee of the Na- analgesic mixtures was a substantial de- tially with respect to study populations, an- tional Kidney Foundation in the United cline in the incidence of AN observed. algesics consumed, and length of follow- States examined the available informa- Many epidemiologic studies have in- up, the increased ORs were remarkably tion from hundreds of peer-reviewed ar- vestigated the risk for renal failure related similar. In contrast, the observational studies ticles and stated that habitual consump- to the prolonged excessive consumption of of Kurth et al.23 and Curhan et al.24 could tion of both phenacetin-containing analgesics (Figure 2). In case-control stud- not demonstrate an increased risk after the mixtures and non–phenacetin-contain- ies, the overall risk after any analgesic con- (nonexcessive) consumption of any anal- ing mixtures is associated with AN.8 One sumption ranged from 1.02 (95% confi- gesic in a population of healthy US male year later, European nephrologists prof- dence interval 0.80 to 1.30) to 17.20 (95% physicians or female nurses, respectively. fered a similar viewpoint, asking for the confidence interval 8.50 to 34.70). The It is inherent to the case-control design prohibition of over-the-counter sales of studies of Sandler et al.13, Pommer et al.15, that the observed association between

J Am Soc Nephrol 20: 2098–2103, 2009 Over-the-Counter Analgesic Use 2099 CLINICAL COMMENTARY www.jasn.org

tain, necessitating the use of special inter- view techniques. Interviews by tele- phone13,14,17 or written self-reports of analgesic consumption in the past23,24 were not the methods of choice to detect hidden consumption of analgesics (information bias). Half of the case-control studies as well as all of the cohort studies also suffer from indication (protopathic) bias; that is, not being able to distinguish between analgesic consumption preceding de novo development of renal disease and the analgesic consumption because of symptoms of other diseases (e.g., diabetes) that predispose patients to renal failure. Finally, it is well established that patients with AN used analgesics on a daily basis and at least for a period of 5 yr, resulting in a total analgesic consumption of at least 1500 U.4,8 As shown in Figure 2, several studies suf- fered from dosage bias because they used a total amount that was far below the mini- mum consumption described in patients with AN.13,19,20 The epidemiologic literature concern- ing the role of different substances that cause AN is limited and controversial. Sub- stances combined in analgesic mixtures are invariably taken together. Most case-control studies suffer from ingredient bias because they present risk ratios based on analyses without making a distinction between analgesics used as single ingredient or in combination. Mainly because of ingredient bias, the potential nephrotoxicity of paracetamol remains a matter of debate, with six studies showing an increased risk and three studies not (Figure 3). The safety of aspirin used as a single ingredient is eas- ier to evaluate. From seven case-control studies, only three showed an increased Figure 2. Overview of epidemiologic studies investigating the renal risk of analgesic risk. In addition, two robust, observational consumption. (A) Description of methodologic details used in the included studies. cohort studies reported even slightly de- (B) Presentation of the overall risk (OR with 95% confidence interval) associated with the creased OR for the use of aspirin (Figure 3). consumption of “any analgesic” exceeding the mentioned dosage. (C) Presentation of the In both studies, calculated ORs were based ORs with 95% confidence interval published in the included epidemiologic studies on hundreds of regular users of aspirin, focusing separately on the ingredients: Aspirin, paracetamol, and NSAIDs. used as a single-ingredient analgesic.23,24

CKD and analgesic consumption does not ulation11,20 and significant differences in establish cause and effect. Moreover, seri- gender and race between case patients and DOES ANALGESIC USE ous flaws in study design or analysis of data control subjects17 can seriously influenced EXACERBATE THE PROGRESSION have to be considered and were discussed results. Moreover, most heavy users tend OF CKD? in several reviews (Table 1). Most case- to deny their analgesic consumption. Ret- control studies suffer from selection bias. rospective investigation of a precise history Some epidemiologic studies generate the Lack of randomization of the control pop- of analgesic consumption is difficult to ob- hypothesis that habitual analgesic use in-

2100 Journal of the American Society of Nephrology J Am Soc Nephrol 20: 2098–2103, 2009 www.jasn.org CLINICAL COMMENTARY

Table 1. Sources of bias in the epidemiologic studiesa Selection Information Indication or Protopathic Ingredient Dosage Studies Bias Bias Bias Bias Bias Case-control studies McCredie et al.,11 Australia, 1982 Yes No Yes Yes No Murray et al.,12 United States, 1983 Yes No Yes Yes Yes Sandler et al.,13,14 United States, 1989 and 1991 No Yes Yes Yes No Pommer et al.,15 West Berlin, 1989 No No Yes No No Morlans et al.,16 Barcelona, 1990 Yes No No Yes Yes Perneger et al.,17 United States,1994 Yes Yes Yes Yes No Fored et al.,18 Sweden, 2001 No No No Yes Yes Ibanez et al.,19 Barcelona, 2005 Yes No No Yes Yes Van der Woude et al.,20 Austria, Germany, 2007 Yes No No No Yes Prospective controlled cohort studies Dubach et al.,21 Switzerland, 1983 No Yes Yes Yes No Elseviers and De Broe,22 Belgium, 1995 No No Yes No No Observational cohort studies Kurth et al.,23 United States, 2004 No Yes Yes Yes No Curham et al.,24 United States, 2004 No Yes Yes Yes No aSelection bias, random selection of controls failed or the chosen control population is biased; information bias, methods used to obtain information about analgesic consumption were doubtful; indication or protopathic bias, failure to control for analgesic intake preceding the development of renal failure; ingredient bias, failure to entangle the use of particular ingredients either as single analgesic or as one of the ingredients of analgesic mixtures; dosage bias, definition of analgesic use far below the amount consumed by patients with analgesic nephropathy. fluences the progression of CKD. Fored To evaluate a possible effect of analge- There were no significant differences et al.18 performed a carefully designed sics on the progression of CKD, one needs between the two groups in terms of racial study of patients with moderate to severe at least two time points of renal function background, gender, age, or proteinuria renal failure. They tried to exclude sev- measurement after randomization of pa- at presentation. The proportion of pa- eral biases, including indication bias. In- tients with incipient or moderate degrees tients in each group with a history of deed, the use of analgesics and the risk for of renal failure and documented pro- smoking, hypertension, or vascular dis- CKD was not consistently stronger nounced analgesic use and groupings ease was not significantly different, ei- among patients with underlying diseases into those who stopped analgesic intake ther. Despite these similarities, the renal causing frequent aches and pain.18 They and those who maintained a substantial function of those who continued to take analyzed their results on the basis of consumption of analgesics. The study analgesics declined 3.5 ml/min per yr three different periods of exposure, try- that realized the best possible approach faster than the patients who stopped taking to separate analgesic use as a causal toward this important design issue is that ing all analgesics. In addition, continuing factor from the possibility that renal dis- of Mackinnon et al.25 The authors stud- analgesics conferred a six-fold increase in eases or their prodromal symptoms ini- ied patients with CKD of undetermined the risk for death or progression to tiated analgesic use (protopathic bias). In cause associated with important use of ESRD. Despite some flaws in this study, addition, they avoided the previously de- analgesics in which the stoppage or on- it supports the contention that the con- scribed excess risk for ESRD to be greater going analgesic use was registered regu- tinued use of non–phenacetin-com- in those whose paracetamol use occurred larly along with several measurements of bined or single-agent analgesics is associ- within a 5-yr period closest to the initia- renal function of a median follow-up pe- ated with faster progression of renal tion of treatment for ESRD, explained by riod of 58 mo. Patients were considered impairment and an increased risk for the advice of their physicians to avoid to have AN when they had a history of reaching a combined end point of death taking nonsteroidal anti-inflammatory analgesic ingestion on a daily basis (ex- or ESRD in patients with AN. drugs (NSAIDs) or aspirin. All of these cluding low-dosage aspirin as an anti- analyses result in only minor reduction platelet agent) for at least 3 yr and no in estimates of relative risk. Nevertheless, other explanation for their renal impair- IS THE PROLONGED USE OF caution is still required because, as the ment could be found. NSAIDs, analgesic NSAIDS ASSOCIATED WITH authors pointed out, “it is impossible to mixtures, and paracetamol, all or not in RENAL PAPILLARY NECROSIS OR rule out bias caused by consumption of combination, were used. No patient gave CKD? these analgesics for symptoms of the a history of phenacetin ingestion. During condition that predispose patients to re- follow-up, 27 patients were judged to The most common renal disorder asso- nal failure,” the latency time between the have ceased all analgesic intake, whereas ciated with NSAIDs is acute kidney in- exposure to analgesics and CKD being the remaining 51 continued using one or jury, largely reversible, as a result of the unknown. more of the preparations listed. inhibition of renal vasodilatory prosta-

J Am Soc Nephrol 20: 2098–2103, 2009 Over-the-Counter Analgesic Use 2101 CLINICAL COMMENTARY www.jasn.org glandins in the clinical setting of a stim- scanning without contrast medium, a 3. Elseviers MM, De Schepper A, Corthouts R, ulated renin-angiotensin system. Older simple validated test, even in the absence Bosmans JL, Cosyn L, Lins RL, Lornoy W, Matthys E, Roose R, Van Caesbroeck D, et age, hypertension, concomitant use of of reliable information on previous anal- al.: High diagnostic performance of CT scan diuretics or aspirin, preexisting renal gesic use. for analgesic nephropathy in patients with failure, diabetes, and plasma-volume AN is invariably caused by compound incipient to severe renal failure. Kidney Int contraction are known risk factors for re- analgesic mixtures containing depen- 48: 1316–1323, 1995 nal failure after the ingestion of NSAIDs. dent-inducing substances regardless of 4. Henrich WL, Clark RL, Kelly JP, Buckalew VM, Fenves A, Finn WF, Shapiro JI, Kimmel Rarely, NSAIDs cause acute interstitial the presence or absence of phenacetin as PL, Eggers P, Agodoa LE, Porter GA, Sha- nephritis (AIN) with proteinuria. one of the active ingredients. In healthy piro S, Toto R, Anderson T, Cupples LA, A retrospective study of all cases of individuals with normal renal function, Kaufman DW: Non-contrast-enhanced com- AIN (reviewing 1068 renal biopsies from the occasional to moderate regular use of puterized tomography and analgesic-re- 1968 to 1997) noted analgesics, particu- aspirin and NSAIDs is without renal risk. lated kidney disease: Report of the national analgesic nephropathy study. JAmSoc larly NSAIDs, as risk factors for sustained The case with paracetamol is less clear, Nephrol 17: 1472–1480, 2006 renal insufficiency. AIN was found in although the described risk is at the most 5. Prescott LF: Analgesic nephropathy: A reas- 6.5% of all biopsies; infection-induced in modest. The acute renal effects of sessment of the role of phenacetin and 10%, idiopathic in 4%, and drug-in- NSAIDs and the frequent irreversibility other analgesics. Drugs 23: 75–149, 1982 duced in 85% of the cases (antibiotics in of the NSAID-induced AIN are now well 6. Mihatsch MJ, Khanlari B, Brunner FP: Obit- uary to analgesic nephropathy: An autopsy 13 cases, analgesics in 17, NSAIDs in 16, documented. How much the chronic use study. Nephrol Dial Transplant 21: 3139– diuretics in five, and various other drugs of NSAIDs may induce renal papillary 3154, 2006 in seven). Renal insufficiency was revers- necrosis, hence evolution toward ESRD, 7. Michielsen P, De Schepper P: Trends of an- ible in 69% and permanent in 31% (12% is far from clear. algesic nephropathy in two high endemic partially reversible, 19% irreversible). The continued use of non–phenace- regions with different legislations. JAmSoc Nephrol 12: 550–556, 2001 Drug-related AIN turned out to be main tin-combined analgesics with or without 8. Henrich WL, Agodoa LE, Barrett B, Bennett cause of permanent renal insufficiency in NSAIDs is associated with faster progres- WM, Blantz RC, Buckalew VM Jr, D’Agati 36% with a maximum of 56% in NSAID- sion of renal impairment and increased VD, De Broe ME, Duggin GG, Eknoyan G: induced cases.26 risk for reaching ESRD. Analgesics and the kidney: Summary and In contrast to the well-characterized In patients with stages 3 to 4 CKD and recommendations to the Scientific Advisory Board of the National Kidney Foundation acute effects of NSAIDs on the kidney, ESRD, paracetamol is the analgesic of from an Ad Hoc Committee of the National chronic effects are less well documented. choice in the short-term treatment of Kidney Foundation. Am J Kidney Dis 27: Renal papillary necrosis has been in- mild to moderate pain. NSAIDs may be 162–165, 1996 duced experimentally by NSAIDs in ani- used for short-term management, taking 9. De Broe ME, Elseviers MM, Bengtsson U, mals; the severity of the effects varies into account the list of widely known risk Mihatsch MJ, Molzahn M, Pommer W, Ritz E, Schwarz A: Analgesic nephropathy. from one product to another and in- factors for acute, frequently irreversible, Nephrol Dial Transplant 11(12): 2407– creases with caffeine. Although renal deterioration of renal function. Regular 2408, 1996 papillary necrosis and chronic renal fail- renal function monitoring is mandatory. 10. Elseviers MM, De Broe ME: Combination ure can occur after the prolonged use of As long as analgesic mixtures containing analgesic involvement in the pathogenesis NSAIDs, the actual risk for these serious more than one ingredient associated of analgesic nephropathy: The European perspective. Am J Kidney Dis 28[Suppl 1]: complications in humans is not known. with centrally acting, dependence-in- S48–S55, 1996 ducing drugs are available over the 11. McCredie M, Stewart JH, Mahony JF: Is counter, AN will continue to be a prob- phenacetin responsible for analgesic ne- CONCLUSIONS lem. phropathy in New South Wales? Clin Nephol 17: 134–140, 1982 12. Murray TG, Stolley PD, Anthony JC, Schin- AN is a slowly progressive renal disease nar R, Hepler-Smith E, Jeffreys JL: Epidemi- resulting from the daily use, heavy or DISCLOSURES ologic study of regular analgesic use and not, for many years of preparations con- None. end-stage renal disease. Arch Intern Med taining at least two analgesics (aspirin, 143: 1687–1693, 1983 paracetamol, phenacetin, pyrazolones) 13. Sandler DP, Smith JC, Weinberg CR, Buck- alew VM Jr, Dennis VW, Blythe WB, Bur- and central-acting dependence-inducing REFERENCES gess WP: Analgesic use and chronic renal substances (caffeine, codeine, and/or disease. N Engl J Med 320: 1238–1243, barbiturates). AN is a well-defined clini- 1. Kincaid-Smith P, Nanra RS: Lithium-induced 1989 copathologic entity, characterized by and analgesic-induced renal diseases. In: 14. Sandler DP, Burr FR, Weinberg CR: Nonste- chronic interstitial nephritis and renal Diseases of the Kidney, 5th Ed., edited by roidal anti-inflammatory drugs and the risk ‘for chronic renal disease. Ann Intern Med papillary necrosis and calcifications. This Schrier RW, Gottschalk CW, Boston, Little, Brown and Company, 1993, pp 1099–1130 115: 165–172, 1991 disease can be accurately diagnosed or 2. De Broe ME, Elseviers MM: Analgesic ne- 15. Pommer W, Bronder E, Greiser E, Helmert excluded, at any stage of CKD, by CT phropathy. N Engl J Med 338: 446–452, 1998 U, Jesdinsky HJ, Klimpel A, Borner K, Mol-

2102 Journal of the American Society of Nephrology J Am Soc Nephrol 20: 2098–2103, 2009 www.jasn.org CLINICAL COMMENTARY

zahn M: Regular analgesic intake and the Laporte JR: Case-control study of regular Buring JE, Stampfer MJ, Hennekens CH, Ga- risk of end-stage renal failure. Am J Nephrol analgesic and nonsteroidal anti-inflamma- ziano JM: Analgesic use and change in kid- 9: 403–412, 1989 tory use and end-stage renal disease. Kid- ney function in apparently healthy men. 16. Morlans M, Laporte JR, Vidal X, Cabeza D, ney Int 67: 2393–2398, 2005 Am J Kidney Dis 42: 234–244, 2003 Stolley PD: End-stage renal disease and non- 20. van der Woude FJ, Heinemann LA, Graf H, 24. Curhan GC, Knight EL, Rosner B, Hankinson narcotic analgesics: A case-control study. Br J Lewis M, Moehner S, Assmann A, Kuhl- SE, Stampfer MJ: Lifetime nonnarcotic anal- Clin Pharmacol 30: 717–723, 1990 Habich D: Analgesics use and ESRD in gesic use and decline in renal function in 17. Perneger TV, Whelton PK, Klag MJ: Risk of younger age: A case-control study. BMC women. Arch Intern Med 164: 1519–1524, kidney failure associated with the use of Nephrol 8: 15, 2007 2004 acetaminophen, aspirin, and nonsteroidal 21. Dubach UC, Rosner B, Pfister E: Epidemio- 25. Mackinnon B, Boulton-Jones M, McLaughlin antiinflammatory drugs. N Engl J Med 331: logic study of abuse of analgesics containing K: Analgesic-associated nephropathy in the 1675–1679, 1994 phenacetin: Renal morbidity and mortality West of Scotland: A 12-year observational 18. Fored CM, Ejerblad E, Lindblad P, Fryzek JP, (1968–1979). N Engl J Med 308: 357–362, study. Nephrol Dial Transplant 18: 1800– Dickman PW, Signorello LB, Lipworth L, 1983 1805, 2003 Elinder CG, Blot WJ, McLaughlin JK, Zack 22. Elseviers MM, De Broe ME: A long-term 26. Schwarz A, Krause PH, Kunzendorf U, Keller MM, Nyreń O: Acetaminophen, aspirin, and prospective controlled study of analgesic F, Distler A: The outcome of acute interstitial chronic renal failure. N Engl J Med 345: abuse in Belgium. Kidney Int 48: 1912– nephritis: Risk factors for the transition from 1801–1808, 2001 1919, 1995 acute to chronic interstitial nephritis. Clin 19. Ibań˜ ez L, Morlans M, Vidal X, Martıńez MJ, 23. Kurth T, Glynn RJ, Walker AM, Rexrode KM, Nephrol 54: 179–190, 2000

J Am Soc Nephrol 20: 2098–2103, 2009 Over-the-Counter Analgesic Use 2103