Drug Use for Non-Hepatic Associated Conditions in Patients with Liver Cirrhosis

Eur J Clin Pharmacol (2003) 59: 71–76 DOI 10.1007/s00228-003-0586-2

PHARMACOEPIDEMIOLOGY AND PRESCRIPTION

M. Isabel Lucena Æ Rau´l J. Andrade Æ Gianni Tognoni Ramo´n Hidalgo Æ Felipe Sanchez de la Cuesta Æ the Spanish Collaborative Study Group on Therapeutic Management in Liver Diseases Drug use for non-hepatic associated conditions in patients with liver cirrhosis

Received: 1 November 2002 / Accepted: 3 February 2003 / Published online: 9 April 2003 Ó Springer-Verlag 2003

Abstract Aims: To study the prescribing patterns of Keywords Pharmacoepidemiology Æ Drug use Æ Liver practising physicians for the most frequent non-hepatic cirrhosis associated conditions in patients with liver cirrhosis. Methods: A multi-centre prospective observational study carried out in 25 Spanish hospitals. Inpatients admitted to gastrointestinal and liver units with a diagnosis of liver cirrhosis were included in five centrally Introduction assigned index days, between February and June 1999. Information was collected about pharmacological treat- Cirrhosis of the liver is one of the most common ments used on admission and recommended at discharge. chronic diseases that can alter drug disposition and Results: Five hundred and sixty-eight in-patients with a effect [1, 2, 3]. The heterogeneity of the effects of liver diagnosis of liver cirrhosis (44% alcoholic cirrhosis) and disease on different routes of drug metabolism, the an average number of 2.5 co-morbid conditions were variability in the rates of these pathways and the ab- studied: diabetes mellitus (30%), infectious disorders sence of a routinely measurable test of metabolic (24%), cardiovascular disease (20%) and active alco- function of the liver make it difficult to accomplish a holism (15%)—the most common associated conditions. simple rule to individualise drug treatment in cirrhotic Chlormethiazole, amoxicillin–clavulanic acid, paraceta- patients [1, 4]. This is of further concern because mol, gliblenclamide, lorazepam, captopril and tiapride underlying liver dysfunction is one of the most com- were the drugs used most prevalently. The average pre- mon factors that increase the risk of adverse effects in scribed daily dose was <1 defined daily dose per day for medical practice [5]. Unfortunately, scarce or occa- most medication classes hepatically handled except for sionally conflicting data are available on whether the calcium channel blockers. presence of liver disease should influence drug choice Conclusions: The present study expands current knowl- and dosage adjustments for common associated con- edge of prescribing patterns for associated conditions in ditions [6, 7]. Nevertheless, there is general agreement patients with underlying liver cirrhosis. Drug dosing was regarding the hazards to which cirrhotic patients are affected in general by the influence of age and hepatic exposed when taking some drugs, such as non-steroid disease on the disposition of drugs, but knowledge on anti-inflammatory drugs (NSAIDs), or the appropri- drug selection needs further attention. ateness to perform a dosage modification for others (i.e. calcium channel blockers). However, the diverse influence of liver function on drug disposition may M.I. Lucena (&) Æ R.J. Andrade Æ F. Sanchez de la Cuesta lead health-care professionals to inappropriate drug Clinical Pharmacology Service and Liver Unit, selection, to inappropriate dosing of medicines or to Hospital Universitario, School of Medicine, University of Ma´laga, 29071 Ma´laga, Spain some degree of therapeutic nihilism, particularly if E-mail: [email protected] there is only a recommendation that the drug should Tel.: +34-952-131572 be used ‘‘with caution’’. Fax: +34-952-131568 The present study was aimed to assess the G. Tognoni prescribing patterns of practising physicians for the Istituto di Ricerche Pharmacologiche Mario Negri, Milan, Italy most frequently associated conditions (excluding dis- R. Hidalgo orders of the gastrointestinal tract) in hospitalised pa- Centro de Ca´lculo, University of Ma´laga, 29071 Ma´laga, Spain tients with underlying liver cirrhosis. 72

represented 38% (218), and cirrhosis of unknown cause Methods 18% (102). The mean age was 61 years (range 27– 92 years); 30% of the patients was younger than This study was part of a large-scale study carried out in Spain and aimed to describe the prescribing patterns of practising physicians 55 years, while 45% was 65 years or older. The younger for patients with liver cirrhosis and to determine the extent of patients were mainly classified in the diagnostic group of practice variability across participating centres, which represented alcoholic cirrhosis of the liver. Women were older than up to one-quarter of the national health system [8]. men. Of the female sample, 58% was older than A multi-centre prospective observational study was carried out 65 years, while the same age group made up 39% of the in 25 Spanish hospitals with a catchment population of 10.8 million inhabitants. A network was set up under the methodological and male patient population. operational co-ordination of a clinical pharmacology research unit. Ascites was the most prevalent major complication of The methodology followed was described previously [8]. Briefly, all cirrhosis in this population (49%), followed by portal– patients admitted to gastrointestinal and liver units with a diag- systemic encephalopathy (24%) and variceal bleeding nosis of liver cirrhosis on five centrally assigned index days, between February and June 1999, were included. The research (22%). Of the patients, 82% was class B–C according to protocol was approved by the locally appointed ethics committee. Child–Pugh classification. Diabetes mellitus (125 patients), infectious disorders other than spontaneous bacterial peritonitis (110 pa- Data collection tients), active alcoholism (68 patients), chronic respira- tory diseases (49 patients), heart disease (47 patients), In each participating hospital, the physician in charge of the study hypertension (45 patients), pain disorders (37 patients) prospectively collected (with a standardised form) information on and non-liver cancer (16 patients) were the most com- all patients with a diagnosis of liver cirrhosis that were admitted in any of the assigned days. Information consisted of demographic monly associated conditions. One hundred and four variables, present medical condition and pharmacological treat- patients (17%) did not have any associated condition. ments prescribed for non-liver related co-morbid conditions on Fifty-seven patients (10%) died during hospitalisation. admission (refers to drugs that patients were already receiving at Multiple co-morbid conditions were prevalent, as indi- time of hospital admission) and at discharge from hospital. Infor- cated by an average number of 2.5±1.7 diagnoses. The mation was obtained by interviewing patients and from the clinical records. Patients were questioned to identify drugs used for other drugs used for the management of these conditions are medical problems not related to the liver that might not have been shown in Table 1. recorded in the medical record. In addition, family members were Prescribed oral hypoglycaemics accounted for 41% interviewed when patients were not able to collaborate. Medication of the total diabetes prescriptions (47 of 116) on ad- containers, which family members were asked to bring to the hospital, were inspected. A written medication plan was provided mission and 26% at discharge (26 of 99). The most by the patient when available. The name of each drug, daily dose, frequently prescribed sulfonamides at discharge were route and frequency of administration, and diagnosis motivating glibenclamide (69%) followed by gliclazide, glipizide the prescription were recorded. The diagnosis of liver cirrhosis was and chlorpropamide (8% each). usually made before admission to hospital and according to the standardised clinical, ultrasonographic, endoscopic and histologi- A total of 26 benzodiazepine agents were used at cal criteria. The severity of chronic liver disease was assessed using discharge, lorazepam being the most common (50%), the Child–Pugh classification, with score C representing the most followed by oxacepam, bromazepam, clorazepate and advanced disease. Forms were checked for completeness at the co- lormetazepam (8% each). On admission, the rank order ordinating centre before data entry. was lorazepam (34%), clorazepate (10%), diazepam and The prescribed daily dose (PDD) was recorded as a percentage of the defined daily dose (DDD) for the selected main medication oxazepam (7% each) and others (five compounds: lor- classes by dividing the PDD by the DDD [9]. The DDD of a metazepam, bromazepam, bentazepam, alprazolam and substance is established by the Nordic Council on Medicines and flunitrazepam; 3%). Hydroxyzine, a diphenylmethane the World Health Organization (WHO) Drug Utilization Research derivative, was used for the treatment of pruritus related Group and represents the recommended average daily dose of a drug, if used for its main indication in an average adult. Medicines to liver cirrhosis in five patients on admission but only in were classified according to the Anatomical Classification System three patients (3 of 14) at discharge. (ATC) recommended by WHO Europe. Those medications for Overall, 9 patients received drugs for insomnia and 19 which the liver has a great contribution to their metabolism and patients for the management of alcohol deprivation disposition were chosen. The data were analysed using the Statistical Package for Social syndrome at discharge. The corresponding figures on Sciences (SPSS, version 10.0) for Windows software. Variables admission were 18 patients and 13 patients, respectively. were examined using descriptive frequencies. For the management of insomnia and alcoholic deprivation syndrome, chlormethiazole was the drug of ref- erence; prescription rates were 77% on admission and Results 81% at discharge. Benzamides were the most commonly prescribed class of neuroleptics (tiapride, 55%) for The study population consisted of 568 patients (396 the diagnosis of agitation and alcoholic deprivation men, 70%) with a diagnosis of liver cirrhosis, recruited syndrome. from 25 hospitals. Sixteen protocols were found to have Opioids were scarcely used (6 cases), while other an- included ineligible patients or to be unsuitably coded algesics were preferred (31 cases). The most frequently and were therefore excluded. Alcoholic cirrhosis ac- prescribed drugs were paracetamol followed by metam- counted for 44% (248) of the sample, viral cirrhosis izole sodium. Paracetamol was mainly prescribed in 73

Table 1 Drugs prescribed for the treatment of non-hepatic Drug group Admission (n=568) Discharge (n=511) associated conditions in patients with underlying liver Prescriptions Patients Prescriptions Patients cirrhosis. Data are given as the number of drugs, number of Insulins 69 56 (10) 73 61 (12) patients and percentage that Oral blood glucose-lowering drugs 47 43 (8) 26 25 (5) this latter number represents Antithrombotic agents 12 12 (2) 11 11 (2)

Cardiac glycosides 13 13 (2) 14 14 (3) Antihypertensives 62 62 (11) 42 42 (8) Calcium channel blockers 15 15 10 10 ACE and angiotensin-II inhibitors 19 19 12 12 Beta-blockers (cardioselective) 7 7 7 7 Others 21 21 13 13

Corticosteroids for systemic use 14 14 (2) 14 14 (3)

Systemic hormonal preparations 4 4 (1) 7 7 (1)

Antibacterials 17 14 (2) 66 66 (13) Betalactam 5 3 26 26 Fluoroquinolones 5 5 28 28 Others 7 6 12 12

Musculoskeletal system 21 19 (3) 6 4 (1) Anti-inﬂammatory drugs 13 13 1 1 Anti-gout drugs 8 6 5 3

Analgesics 33 31 (5) 44 34 (7) Opioids 4 4 6 6 Other analgesics and antipyretics 29 27 31 28

Anti-epileptics 10 9 (2) 11 9 (2)

Psycholeptics 53 53 (9) 69 69 (14) Anti-psychotics 4 4 11 11 Anxiolytics 27 27 26 26 Hypnotics and sedatives 22 22 32 32

Psychoanaleptics Antidepressants 7 7 (1) 5 5 (1)

Anti-asthmatics 42 25 (4) 51 29 (6) patients with viral cirrhosis (55%, 10/18). NSAIDs were Reserpine, indapamide or xipamide were not prescribed drastically reduced at discharge, with only one prescrip- at discharge. tion given relative to on admission [13 prescriptions, The average PDDs as a percentage of DDD of diclofenac (54%) being the agent most frequently used]. those medication classes prescribed most frequently Among antidepressants, preference was given to se- that undergo hepatic metabolism (except for digoxin) lective serotonin reuptake inhibitors (71% on admission are shown in Table 2. The average PDDs were below 1 vs 60% at discharge), paroxetine being the most fre- DDD per day for all medication groups on admission quently prescribed agent (67%). and at discharge, except for calcium channel blockers For the treatment of infectious disorders, the ﬂuor- (amlodipine and nifedipine), 70% of users receiving oquinolones (excluded norﬂoxacine) (42%) and beta- more than 1 DDD per day at discharge. For parac- lactam antibiotics (39%) were the most frequently used etamol, the average PDD was <1 DDD per day (1 drugs. Of all betalactamics used, the amoxicilllin–cla- DDD is equivalent to 3 g) being used as needed in vulanic acid association was the most prevalent (42%). 30% of users. For the management of hypertension, angiotensin- converting enzyme (ACE) and angiotensin-II inhibitors (29%) and calcium channel blockers (24%) were the Discussion most frequent therapeutic classes recommended at discharge. On admission, the pattern of prescription A large prospective cohort of patients with moderate to followed the same trend, although numerous other severe chronic liver disease whose co-morbid conditions drugs in the same therapeutic class were also used. are managed by specialised physicians in a setting of 74

Table 2 Prescribed daily dose (PDD) as percentage of the deﬁned captopril, 10 mg enalapril, 150 mg irbesartan, 25 mg ramipril, daily dose (DDD) of the main groups of medication in patients 15 mg fosinopril, 3 g acetaminophen, 400 mg tiapride, 576 mg with liver cirrhosis. Data are given as weighted mean ± standard chlormethiazole, 2.5 mg lorazepam, 10 mg diazepam, 1 mg alpra- deviation. (1 DDD is equivalent to, respectively, 10 mg glibencla- zolam, 75 mg bentazepam, 20 mg clorazepate, 10 mg bromazepam, mide, 0.160 mg gliclazide, 10 mg glipizide, 0.375 mg chlorprop- 50 mg oxazepam, 1 mg lormetazepam, 1 mg ﬂunitrazepam). amide, 0.25 mg digoxin, 5 mg amlodipine, 30 mg nifedipine, Combination preparations excluded 240 mg diltiazem, 20 mg nisoldipine, 240 mg verapamil, 50 mg

Medication class Admission Discharge

n Age (years) % PDD/DDD Child-Pugh score n Age (years) % PDD/DDD Child-Pugh score

Oral hypoglycaemics 39 64 79±39 8.1 24 64 91±46 7.3 Cardiac glycosides 13 68 85±24 7.9 14 65 80±24 7.1 Antihypertensives Ca-channel blockers 14 69 93±40 7.3 10 70 129±105 7.6 ACE and angiotensin-II 12 73 80±79 7.6 9 75 74±58 7.6 inhibitors Paracetamol 11 58 33±18 7.7 17 59 57±16 8.0 Tiapride 2 49 63±53 10.5 6 50 75±63 8.0 Benzodiazepines 24 65 49±42 7.6 24 65 38±49 7.9 Chlormethiazole 17 56 67±31 9.4 27 53 58±24 8.3 daily clinical practice offers a favourable framework for the hallmarks of advanced liver failure. However, when a comprehensive assessment of prescribing practices. To antihypertensive drugs are prescribed in patients with our knowledge, these practices have not been previously chronic liver disease, it should be taken into account that investigated in a prospective study. ACE inhibitors and angiotensin-II receptor blockers A lower prescription frequency at discharge was re- counteract the enhanced activity of the renin–angioten- corded for anxiolytics, antidepressants and NSAIDs. sin system; therefore, they do not seem to be a good This might reflect the growing awareness of the likely choice and are not recommended in patients with cir- development of dependence and of the potential to rhosis and ascites [7]. Physicians seemed to disregard worsen encephalopathy by the use of psychoactive drugs these recommendations since all patients receiving these or to cause an impairment of renal function and natri- drugs had Child–Pugh class-B disease. uresis by NSAIDs. Among benzodiazepine use, the most Still under debate is the use of amoxicillin–clavulanic evident trend was a higher preference to drugs with an acid as first-line therapy in the management of infectious intermediate half-life eliminated mainly by glucuroni- disorders in patients with cirrhosis, since this association dation [3]. In general, there was a trend to reduce the is a leading cause of antibiotic-related cholestatic jaun- number of different active compounds within a thera- dice [15]. The risk of developing hepatotoxicity with the peutic class used in the treatment of a particular disorder combination of amoxicillin and clavulanic acid rises to 1 at discharge. This made the therapeutic recommenda- in 1000 in older patients with repeated exposure [16], tions more uniform. circumstances expected to be encountered frequently Interestingly, paracetamol—which is thought to be a among patients with cirrhosis. In fact, amoxicillin–cla- safe analgesic in cirrhotic patients since it undergoes vulanic acid was the antibiotic most frequently impli- direct glucuronidation [10] and the only concern raised cated in the hepatic reactions reported to a registry [17]. about its use is in active alcohol drinkers [11]—was used This combination accounted for 16 of 29 cases of hep- systematically at a reduced dosage and as needed irre- atotoxicity in the antibacterial group (55%), one of spective of the aetiology of the liver cirrhosis. An epi- which was a fatality in a patient with cirrhosis [18]. The demiological study aimed to determine the pattern of prescription of drugs that have been associated with acetaminophen use in alcoholic patients showed that hepatic injury to patients with liver cirrhosis is an area of 37.5% of alcoholics used paracetamol as needed, with continued controversy [5]. one of every ten abusing the drug [12]. Daily prescribed dosages of drugs hepatically han- A noticeable finding in this study was the practice of dled decreased to approximately 68% of the DDD in prescribing tiapride or chlormethiazole in patients with this population, as might be expected on the basis of a alcoholic cirrhosis of the liver. This attitude does not modification of the dynamic response to the drug in the reflect any influence of pharmacotherapeutic innovations presence of liver cirrhosis and to a decline in hepatic in the management of alcohol withdrawal syndrome. clearance [1, 2, 3, 4]. Surprisingly, calcium channel Preferences for particular drug entities may be based blockers that exhibit a high intrinsic clearance limited by more strongly on reputation acquired over the years than liver blood flow—which is impaired in advanced liver on evidence-based medicine and toxicity studies [13, 14]. cirrhosis—were prescribed at a full dose in patients with Hypertension as a complication of liver cirrhosis type B–C Child Pugh classification [1, 2, 3]. For a renally usually occurs in patients with well-preserved liver handled drug such as digoxin, average PDDs were also function, as systemic vasodilation and hypotension are reduced according to an age decline in renal function [3]. 75

Interestingly, in a comparable patients’ collective with Bengoechea, F. Escandi, A. Tejada; Islas Canarias: Hospital Insular normal liver function (age range 65–74 years) [19], the Universitario, Gran Canaria: L. Garcia–Villarreal, E. Jimeńez, A. Monescillo; Hospital Universitario deTenerife: E. Quintero; Can- average PDD value of ACE inhibitors was 76% and the tabria: Hospital Marque´s de Valdecilla, Santander: F. Pons, C. mean PDD value of cardiac glycosides 80%, which Almohalla, B. de las Heras–Arechavala; Catalun˜ a: Hospital Uni- parallel the results obtained in the present study. The versitario Germans Trias i Pujol, Barcelona: R. Planas, R. Duran- rather low benzodiazepine dosage taken by the patients dez, M.C. Pen˜ a; Hospital Vall dHebron, Barcelona: A. Vallano, A. appears to be in accordance with other findings obtained Gonza´lez, J. de la Lama, M. Garcıá; Hospital del Mar, Barcelona: C. Hernańdez–Lo´pez, M. Farre´, R. Sola´; Hospital Clinic i Provin- in nursing home residents (mean age 82 years; average cial, Barcelona: M. Bruguera, M. Sala; Madrid: Hospital Clıńico PDD value of anxiolytics 55%) [20]. In this last study San Carlos, Madrid: O. Mariscal, E. Vargas, A. Moreno, M. Diaz [20], paracetamol was the most frequently prescribed Rubio; Hospital 12 de Octubre, Madrid: T. Mun˜ oz–Yagu¨ e, I. Sal- analgesic (average PDD values 39%). Notwithstanding ces–Franco, J.A. Solı´s–Herruzo; Fundacioń Hospital Alcorcoń, Alcorcoń, Madrid: J. A. Gonza´lez–Martin, C. Fernandez, G. Ca- the problems of comparing heterogeneous populations, cho–Acosta; Comunidad Valenciana: Hospital General de Valen- these findings are consistent with the knowledge of a cia: M. Diago, P. Carbonell, V. Garcıá; Hospital Arnau de reduced oxidative metabolising activity and an altered Vilanova, Valencia: O. Gonza´lez Lo´pez, R. Giner Durań, C. pharmacodynamic response in the older elderly patients Martıńez Lapiedra and in those with liver disease [3]. Several limitations need to be considered. Since the sample was designed to over-represent patients with References moderate or severe liver disease, the results might be less applicable to patients with mild liver disease. Another 1. Reichen J (1997) Prescribing in liver disease. J Hepatol concern could be that the data presented were applicable 26[Suppl 1]:36–40 2. Williams RL (1984) Drugs and the liver: clinical applications. to Spanish practice only. However, taking into account In: Pharmacokinetic basis for drug treatment. Benet LZ, the high number of participating hospitals widely dis- Massoud N, Gambertoglio JG (eds) Raven Press, New York, tributed we assume that the drug profiles are represen- pp 63–75 tative of those commonly recommended for associated 3. Rowland M, Tozer TN (1995) Clinical pharmacokinetics. conditions in patients with liver cirrhosis by specialised Concepts and applications, 3rd edn. Williams and Wilkins, Philadelphia, pp 248–263 physicians in Spain. It is worth mentioning that, al- 4. Rodighiero V (1999) Effects of liver disease on pharmacoki- though some patients with liver cirrhosis may be admit- netics. An update. Clin Pharmacokinet 37:399–431 ted to other units apart from gastrointestinal and liver 5. Lewis JH (2002) The rational use of potentially hepatotoxic units, the gastrointestinal and liver disease specialists are medications in patients with underlying liver disease. Expert Opin Drug Saf 1:159–172 usually contacted for medical advise. Furthermore, as 6. Westphal J-F, Brogard J-M (1997) Drug administration in patients were referred to different hospitals, the data are chronic liver disease. Drug Saf 17:47–73 unlikely to be influenced by an idiosyncratic hospital 7. Sokol SI, Cheng A, Frishman WH, Kaza CS (2000) Cardio- drug profile and referral pattern. vascular drug therapy in patients with hepatic diseases and patients with congestive heart failure. J Clin Pharmacol 40:11– In conclusion, the present study expands current 30 knowledge on prescribing patterns for associated con- 8. Lucena MI, Andrade RJ, Tognoni G, Hidalgo R, Sanchez de la ditions in patients with underlying liver cirrhosis. Drug Cuesta F, the Spanish Collaborative Study Group on Thera- dosing was affected in general by the influence of age peutic Management in Liver Disease (2002) Multicenter hos- and hepatic disease on the disposition of drugs, but drug pital study on prescribing patterns for prophylaxis and treatment of complications of cirrhosis. Eur J Clin Pharmacol selection for extrahepatic conditions in the setting of 58:435–440 liver cirrhosis needs further attention. 9. World Health Organisation Collaborating Centre for drugs statistics methodology (1997) Anatomical Therapeutic Chemi- Acknowledgements This study was in part supported by a research cal (ATC) classification index including defined daily dose grant from Ministerio de Sanidad y Consumo, Fondo de Investi- (DDDs) for plain substances. World Health Organisation, gacioń Sanitaria, FISS No. 00/1109 and from the Agencia Espan˜ ola Oslo del Medicamento. K. Shashok checked the use of English in the 10. Whitcomb DC, Block GD (1994) Association of acetamino- manuscript. Participating centres assembled by regions: Andalucia: phen hepatotoxicity with fasting and ethanol use. JAMA Hospital Clıńico Universitario, Ma´laga: J.M. Fraile, M.R. Cabello, 272:1845–1850 R. Alcańtara, M. Garcıá–Corte´s—Coordinating centre. Hospital 11. Zimmerman HJ, Maddrey EC (1995) Acetaminophen (parac- Torreca´rdenas, Almeria: M.C. Fernandez, G. Pelaez, J.L. Vega, M. etamol) hepatotoxicity with regular intake of alcohol. Analysis Casado; Hospital Ciudad de Jaen, Jaen: E. Baeyens, M. M. Col- of instances of therapeutic misadventure. Hepatology 22:767– menero; Hospital Puerta del Mar, Cadiz: F. Diaz, P. Rendon, M. 773 Macias; Hospital Reina Sofia, Cordoba: M. de la Mata; Hospital 12. Seifert CF, Lucas DS, Vondracek TG, Kastens DJ, McCarty Virgen de las Nieves, Granada: R. Martıń–Vivaldi, F. Nogueras; DL, Bui B (1993) Patterns of acetaminophen use in alcoholic Hospital Puerto Real, Cadiz: J. Perez–Moreno, J.M. Puertas; patients. Pharmacotherapy 13:391–395 Hospital Nuestra Sra de Valme, Sevilla: Dr. M. Romero–Gomez, A. 13. Carvajal A, Martin–Arias LH (1996) Antipsychotic drugs. In: Guil, L.Grande; Hospital General de la Axarquia, Velez–Ma´laga, Meylers side effects of drugs. Aronson JK, van Boxtel CJ (eds) Ma´laga: F. Santalla; Hospital Clıńico Universitario S. Cecilio, Elsevier, Holland, pp 40–60 Granada: M.A. Lopez–Garrido, E. Cerilla, J. Salmeroń; Aragon: 14. Mayo–Smith MF (1997) Pharmacological management of al- Hospital Clıńico Universitario de Zaragoza: M. Simoń, M. Villar, cohol withdrawal. A meta-analysis and evidence-based practice A. Lago; Asturias: Hospital Central de Asturias, Oviedo: L. guideline. JAMA 278:144–151 Rodrigo, A. Alvarez–Alvarez, M. Garcıá–Espiga; Pais Vasco. 15. George DK, Crawford DHG (1996) Antibacterial–induced Hospital Nuestra Sra de Aranzazu, San Sebastian. M. Garcia hepatotoxicity. Drug Saf 15:79–85 76

16. Garcı´a–Rodrı´guez LA, Stricker BH, Zimmerman HJ (1996) Risk unrecognized problem. Lessons from a fatal case related to of acute liver injury associated with the combination of amoxi- amoxicillin/clavulanic acid. Dig Dis Sci 46:1416–1419 cillin and clavulanic acid. Arch Intern Med 156:1327–1332 19. van Kraaij DJ, Jansen RWMM, de Gier JJ, Gribnau FWJ 17. Lucena MI, Camargo R, Andrade RJ, Perez-Sanchez CJ, (1998) Prescription patterns of diuretics in Dutch community- Sanchez de la Cuesta F (2001) Comparison of two clinical scales dwelling elderly patients. Br J Clin Pharmacol 46:403–407 for causality assessment in hepatotoxicity. Hepatology 33:123–130 20. van Dijk KN, de Vries CS, van den Berg PB, Brouwers JRBJ, 18. Andrade RJ, Lucena MI, Fernandez MC, Vega JL, Camargo R de Jong-van den Berg LTW (2000) Drug utilization in Dutch (2001) Hepatotoxicity in patients with cirrhosis, an often nursing homes. Eur J Clin Pharmacol 55:765–771