Drug utilization and morbidity statistics for the evaluation of drug safety in Sweden

For a continuous monitoring and evaluation of drug safety problems in Sweden, the Department of Drugs of the National Board of Health and Welfare has access to a number of computerised patient-, drug-, and disease-oriented registers. The usefulness and limitations of these registers are presented by examples. A recent increase in asthma deaths is presently being analysed by comparing information from death certificates and case records with drug sales and prescription data. A recent analysis of the cancer register showed no increased risk of malignant thyroid tumors after diagnostic or therapeutic doses of I 131. Similarly no increased risk of malformations after occupational exposure to hexachlorophene could be detected by analysing the malformation and medical birth-record registers in relation to hospital hexachlorophene use. The register of patient discharge diagnoses has been repeatedly used to analyse the incidence and pattern of drug induced blood dyscrasias and thromboembolism associated with oral contraceptives (OC). These analyses have resulted i.a. in the withdrawal of dipyrone and tenalidine and a decrease of the estrogen-content of OCs. At the same time about 1/3 of these serious adverse drug reactions (ADR) was found to have been reported to the ADR-register. By combining sales and prescription data with ADR-reports the risk of inducing lactic acidosis was found to be significantly higher for phenformin than for metformin. Also the incidence of tardive dyskinesia from longterm use of metoclopramide was found to be much higher than hitherto recognized. By use of these registers it is possible to obtain valuable information about the safety of drugs. The raw data must, however, be interpreted with care and often be supplemented with in depth studies of the various problems.

Drug-related blood dyscrasias in a Swedish reporting system, 1985-1994

Drug-related blood dyscrasias as reported in Sweden during a 10-yr period have been analysed in relation to sales and prescription data. The number of cases reported were as follows: agranulocytosis 390, thrombocytopenia 391, pancytopenia 50 and aplastic anaemia 36. The annual incidence rates per 10(6) inhabitants were: agranulocytosis 4.8, thrombocytopenia 5.6, pancytopenia 1.1 and aplastic anaemia 0.5. Incidences in the elderly were higher for all dyscrasias except aplastic anaemia. The most commonly reported drugs for all dyscrasias were sulphonamides and diuretics, but when related to sales data the risk of agranulocytosis was high for clozapine, dapsone, mianserin and sulphasalazine, while the risk did not seem to be increased for furosemide. For thrombocytopenia, furosemide, co-trimoxazole and the measles, mumps and rubella vaccine were most commonly reported. The risk for pancytopenia and aplastic anaemia was increased for acetazolamide and co-trimoxazole. As spontaneous reporting systems are primarily set up for signalling purposes, such data must always be interpreted with utmost care.

Drugs in the aetiology of agranulocytosis and aplastic anaemia

Agranulocytosis and aplastic anaemia are rare but serious conditions known to be caused by numerous drugs. Most of what is known or suspected about the aetiology is based on case reports, with only a few formal epidemiological studies that provide quantitative estimates of risk. Updated results have been obtained from a combined analysis of data from 3 case-control studies that used similar methods: the International Agranulocytosis and Aplastic Anemia Study (IAAAS), conducted in Israel and Europe; a study conducted in the northeast US; and a study conducted in Thailand. Totals of 362 cases of agranulocytosis, 454 cases of aplastic anaemia and 6458 controls were included in the analyses. The IAAAS and Thai study were population-based, providing estimates of the incidence of the 2 dyscrasias. The overall annual incidence of agranulocytosis in the ambulatory population was 3.4/10(6) in the IAAAS and 0.8/10(6) in Thailand; by contrast the incidence of aplastic anaemia was 2.0/10(6) in the IAAAS and 4.1/10(6) in Thailand. A total of 21 compounds were significantly associated with an increased risk of agranulocytosis in the IAAAS and US studies. Excess risks ranged from 0.06 to 13 cases/10(6) users/wk; the most strongly associated drugs were procainamide, anti-thyroid drugs and sulphasalazine. An association with drugs that had previously been suspected was also seen in Thailand. The overall aetiologic fractions of agranulocytosis due to drug use were 62% in the IAAAS, 72% in the US and 70% in Thailand. Eleven drugs were significantly associated with an increased risk of aplastic anaemia, with excess risks ranging from 1.4 to 60 cases/10(6) users in a 5-month period. The most strongly associated drugs were penicillamine, gold and carbamazepine. Aetiologic fractions were 27% in the IAAAS, 17% in the US and 2% in Thailand, which paralleled the prevalence of use of associated drugs in the 3 populations. The present results confirm that agranulocytosis is largely a drug-induced disease, with similar proportions accounted for in 3 disparate geographic regions. By contrast, although many of the expected associations were observed for aplastic anaemia, most of the aetiology is not explained by drugs. For all associated drugs, the excess risks are sufficiently low that blood dyscrasias should not figure prominently in the balancing of risks and benefits.

Nifedipine--a survey of adverse effects. Four years' reporting in Sweden

During the period February 1983 to December 1984 the Swedish Adverse Drug Reactions Advisory Committee received 80 reports describing 116 adverse reactions with a possible or probable connection to nifedipine. The most frequently reported reactions are oedema, tachycardia, headache and rash. Confusion and sleep disorders constitute 8 cases. Impaired angina is a potentially serious reaction and one patient developed myocardial infarction.

An evaluation of computer-aided disproportionality analysis for post-marketing signal detection

To understand the value of computer-aided disproportionality analysis (DA) in relation to current pharmacovigilance signal detection methods, four products were retrospectively evaluated by applying an empirical Bayes method to Merck's post-marketing safety database. Findings were compared with the prior detection of labeled post-marketing adverse events. Disproportionality ratios (empirical Bayes geometric mean lower 95% bounds for the posterior distribution (EBGM05)) were generated for product-event pairs. Overall (1993-2004 data, EBGM05> or =2, individual terms) results of signal detection using DA compared to standard methods were sensitivity, 31.1%; specificity, 95.3%; and positive predictive value, 19.9%. Using groupings of synonymous labeled terms, sensitivity improved (40.9%). More of the adverse events detected by both methods were detected earlier using DA and grouped (versus individual) terms. With 1939-2004 data, diagnostic properties were similar to those from 1993 to 2004. DA methods using Merck's safety database demonstrate sufficient sensitivity and specificity to be considered for use as an adjunct to conventional signal detection methods.

Acid-suppressing drugs and gastroesophageal reflux disease as risk factors for acute pancreatitis--results from a Swedish Case-Control Study

PURPOSE

To study risk factors for acute pancreatitis, here with emphasis on gastro-intestinal diseases and their treatments.

METHODS

Population based case-control study covering four areas in Sweden encompassing 2.2 million inhabitants. Included were 462 incident cases of acute pancreatitis aged 20-85 years, hospitalized from 1 January 1995-31 May 1998, and 1,781 unmatched controls randomly selected from the study base using a population register. Information was captured from medical records and structured telephone interviews.

RESULTS

Current use of H(2) antagonists starting within 6 months of index-date was associated with acute pancreatitis with an adjusted OR of 4.9 (95% confidence interval (CI) 1.6-15), and current use of proton pump inhibitors (PPIs) with an adjusted OR of 3.2 (95%CI 1.4-7.4). For both drug classes, the ORs tended to be higher at higher doses. Gastritis/gastro-esophageal reflux disease (GERD) within the last 12 months not treated with PPIs or H(2)-antagonists and inflammatory bowel disease (IBD) not treated with anti-inflammatory or immunosuppressive drugs were associated with development of acute pancreatitis with adjusted odds ratios (OR) of 1.9 (95%CI 1.2-3.0) and 5.1 (95%CI 2.0-13) respectively.

CONCLUSIONS

Current IBD without treatment and gastritis/GERD without treatment were found to be associated with increased risks to develop acute pancreatitis but the nature of the latter association needs to be further evaluated. On balance, we judge that the observed associations between current use of H(2)-antagonists and PPIs and increased risk of acute pancreatitis are unlikely to be explained by bias.

Interviewer variability - quality aspects in a case-control study

Quality assurance and quality control are important for the reliability of case-control studies. Here we describe the procedures used in a previously published study, with emphasis on interviewer variability. To evaluate risk factors for acute pancreatitis, information including previous diagnoses and medication was collected from medical records and by telephone interviews from 462 cases and 1781 controls. Quality assurance procedures included education and training of interviewers and data validity checks. Quality control included a classification test, annual test interviews, expert case validation, and database validation. We found pronounced variations between interviewers. The maximal number of interviews per day varied from 3 to 9. The adjusted average (95% CI) number of diagnoses captured per interview of cases was 4.1 (3.8-4.3) and of controls 3.5 (3.4-3.7) (excluding one deviating interviewer). For drugs, the average (95% CI) number per interview was 3.9 (3.7-4.1) for cases and 3.3 (3.2-3.4) for controls (excluding one deviating interviewer). One of the fourteen interviewers deviated significantly from the others, and more so for controls than for cases. This interviewer's data ;were excluded. Nonetheless, data concerning controls more frequently needed correction and supplementation than for cases. Erroneous coding of diagnoses and medication was also more frequent among controls. Thus, a system for quality control of coding practices is crucial. Variability in interviewers' ability to ascertain information is a possible source of bias in interview-based case-control studies when "blinding" cannot be achieved.

Monitoring outcomes of pregnancy following drug exposure: a company-based pregnancy registry program

Women who discover they are pregnant after exposure to a drug and pregnant women who have a condition that requires continued treatment during pregnancy are told to balance the benefits and risks of the exposure to justify continuation of treatment, discontinuation of treatment or, possibly, pregnancy termination. However, there are limited data available to inform decision-making. The Merck Pregnancy Registry Program is a company-run pregnancy registry whose objective is to acquire and analyse information on drug exposures and pregnancy outcomes to better describe the safety profile of Merck products used during pregnancy. Information is collected from women and healthcare providers who call to report drug exposure during pregnancy. Prospective pregnancies are followed up to outcome and data are collected via questionnaires, telephone calls and a review of medical records. Reports are classified as prospective (information received prior to knowledge of pregnancy outcome) or retrospective (received after the outcome is known). Congenital anomaly reports are assessed for timing of exposure, maternal age and medical history, biological plausibility and concomitant medication exposures. Rates of pregnancy outcomes and birth defects in the prospective cohort are computed and confidence intervals are calculated to reflect the strength of the finding based on the sample size. Rates of pregnancy outcomes in the Pregnancy Registry are compared with the rates of pregnancy outcomes in the general US population and, if available, in subpopulations with the relevant disease states. The limitations of post-marketing surveillance are well known as voluntary reporting of individuals and healthcare professionals is known to be subject to various types of bias. Small sample size is another major limitation. However, the strength of the registry lies in its ability to gather pregnancy outcome reports early in the life of a product and to recognise and analyse unusual birth defects. Our data suggest that pregnancy registries can be used to review human exposure data in a systematic fashion so that useful information can be shared with women and their healthcare providers. The use of the pregnancy registry design has allowed for the collection and analysis of data on the effects of drug exposures on human pregnancies that have otherwise been difficult to obtain.

Dose-response relationships between individual nonaspirin nonsteroidal anti-inflammatory drugs (NANSAIDs) and serious upper gastrointestinal bleeding: a meta-analysis based on individual patient data

AIMS

To define by amalgamation of data obtained in contemporaneous case-control studies, the risks associated with individual nonaspirin nonsteroidal anti-inflammatory drugs (NANSAIDs) according to doses used.

METHODS

Meta-analysis of individual patient data from three retrospective case-control studies using similar data collection protocols was carried out in hospitals in Catalonia, England, Scotland and Sweden. 2472 cases of upper gastrointestinal bleeding and 5877 controls were studied. Main outcome measures were risks associated with individual NANSAIDs according to dose used and the period of time for which they were given.

RESULTS

Ibuprofen showed the lowest odds ratio (OR = 1.7; 95% confidence interval 1.1, 2.5), followed by diclofenac (4.9; 3.3, 7.1), indomethacin (6.0; 3.6, 10.0), naproxen (9.1; 6.0-13.7), piroxicam (13.1; 7.9-21.8) and ketoprofen (34.9; 12.7, 96.5). Striking dose-response relationships were seen with four to eight-fold increases in risk within conventionally used dose ranges for all except ketoprofen, where numbers were too few to allow dose analysis. Across the class, risk was highest during the first week of use (11.7; 6.5, 21.0), decreased thereafter with continuing use (5.6; 4.6, 7.0), and fell to 3.2 (2.1, 5.1) 1 week after discontinuing use. Concurrent use of more than one NANSAID substantially increased risk.

CONCLUSIONS

The risk of upper gastrointestinal bleeding with NANSAIDs varies twenty-fold depending on the drug, and by three to seven-fold depending on the dose chosen. Risk is maximal during the first week and decreases thereafter. Paracetamol (acetaminophen) is not associated with upper gastrointestinal bleeding at any dose and should be the first-line analgesic wherever possible.

A Swedish case-control network for studies of drug-induced morbidity--acute pancreatitis

OBJECTIVE

To evaluate risk factors - notably drugs - for developing acute pancreatitis.

METHODS

A population-based, case-control study, encompassing 1.4 million inhabitants aged 20-85 years from four regions in Sweden between 1 January 1995 and 31 May 1998. A total of 462 cases were hospitalised in surgical departments with their first episode of acute pancreatitis without previously known biliary stone disease. From a population register, 1781 controls were randomly selected. Information was obtained from medical records and through telephone interviews.

RESULTS

Fifty-seven percent of the cases were males. An expert group found evidence for biliary stones in 50% of the cases, alcohol intake in 23%, but in 29% neither of these factors were present. In all, "other" factors, e.g. drugs, could have contributed to the development of acute pancreatitis in 52% of the cases. In a multivariate analysis, the adjusted odds ratios (ORs) for H(2) antagonists were 2.4 (95% CI 1.2-4.8) for proton pump inhibitors (PPIs), 2.1 (1.2-3.4) for non-steroidal anti-inflammatory drugs (NSAIDs), 2.3 (1.3-4.0) for those derived from acetic acid and 1.9 (1.1-3.2) for antibacterials for systemic use. Significant ORs were found for a history of gastrointestinal tract disorders [1.5 (1.1-1.9)] and inflammatory bowel disease (IBD) [3.4 (1.5-7.9)]. Smoking was significantly associated with acute pancreatitis [1.7 (1.2-2.1)] and, for those smoking more than 20 cigarettes per day, the OR was 4.0 (2.2-7.5). Alcohol in moderate amounts did not increase the risk, but for those drinking more than 420 g alcohol per week the OR was 4.1 (2.2-7.5).

CONCLUSION

In addition to cholelithiasis, smoking and heavy alcohol use, drugs may be an important risk factor for acute pancreatitis.

Adverse drug reactions as a cause for admissions to a department of internal medicine

PURPOSE

To assess the occurrence and pattern of adverse drug reactions as a cause for acute hospital admission.

METHODS

In 681 randomly selected patients, acutely admitted to a clinic of internal medicine at a Swedish university hospital, information was collected from their medical records about current symptoms and use of drugs, previous diseases and the results of medical investigations and tests. In addition, a standardized interview according to a questionnaire was carried out. A group of experts in clinical pharmacology assessed the data obtained from the patients' case records and the results of the interviews, and then, according to WHO criteria, judged the probability that an adverse drug reaction could have caused or contributed to the actual admission to hospital.

RESULTS

Out of the 681 cases included, 94 (13.8%) had symptoms and signs that were judged as drug-related and that had caused or contributed to the admission. Eighty-two patients (12.0%) had altogether 99 symptoms that were classified as adverse drug reactions. Of these, 91% were type A reactions. The relationship between the medication and the reaction was judged certain in eight, probable in 17, and possible in 74 cases. The most common adverse drug reactions were cardiovascular (36.3%). Twelve patients (1.8%) had symptoms indicating intoxications.

CONCLUSIONS

The prevalence of drug-related problems causing or contributing to admission to a clinic of internal medicine is high and is dominated by type A reactions, i.e. reactions in principle predictable and preventable. This implies a possibility to increase drug safety by preventive measures.

Identification of sulfonamide-like adverse drug reactions to celecoxib in the World Health Organization database

BACKGROUND

The selective COX-2 inhibitor celecoxib has a sulfonamide structure and is contraindicated for patients with known sulfa allergy. However, there is currently no standard available for the identification of sulfonamide-related adverse drug reactions (ADRs) and the occurrence of such ADRs with celecoxib has not been established. THE AIMS OF THIS STUDY WERE: (1) to identify the typical pattern of sulfonamide ADRs from literature and verify this pattern in the World Health Organization (WHO) ADR database; and (2) to examine whether these sulfonamide ADRs occur more frequently with celecoxib than with the non-sulfonamide, COX-2 inhibitor rofecoxib.

METHODS

A sulfonamide ADR pattern was derived from the most extensive textbook source of ADRs and applied to the WHO database for the three groups of sulfonamide drugs: short- and intermediate-acting sulfonamides, and sulfasalazine. ADRs reported three or more times for each of these groups were included in a 'sulfonamide template' comprising 19 ADRs relating to the skin, the blood, the liver, and anaphylaxis. This template was then applied to celecoxib and rofecoxib.

RESULTS

Overall, the relative reporting rate of a sulfonamide-type ADR with celecoxib was 80% higher than with rofecoxib, whether this was based on total number of reports (RR 1.8, 95% Cl 1.6-1.9) or restricted to reports that listed coxibs as the sole suspected drugs (RR 1.8, 95% Cl 1.6-1.9). There were numerically more ADRs for celecoxib than for rofecoxib in 15 of the 19 terms. Within the ADRs in the sulfonamide template, relative reporting rate of fatal reactions was 80% higher with celecoxib (RR 1.8, 95% Cl 0.9-4.0). Even though serious sulfonamide reactions are rare, their clinical impact on patient safety warrants close monitoring as more data becomes available. Physicians should be aware of possible sulfonamide allergy when prescribing celecoxib.

Polymorphisms of NAT2 in relation to sulphasalazine-induced agranulocytosis

Agranulocytosis is a rare, but serious adverse reaction to sulphasalazine. The polymorphic enzyme N-acetyltransferase 2 (NAT2) plays an important role in the metabolism of sulphasalazine. This study was conducted to analyse whether the risk of sulphasalazine-induced agranulocytosis is increased in slow acetylators. Patients were treated for inflammatory disease, mostly joint disease, with a mean dose of 2 g sulphasalazine daily. Thirty-nine patients reacted with agranulocytosis, while 75 patients had been treated for a minimum of 3 months without haematological side-effects. A population-based control panel of 448 individuals was used for comparison. All subjects were genotyped for NAT2 by polymerase chain reaction followed by restriction enzyme digestion. The six most common allelic variants were analysed: NAT2*4, NAT2*5A, NAT2*5B, NAT2*5C, NAT2*6 and NAT2*7. The proportion of slow acetylators was significantly higher in patients with sulphasalazine-induced agranulocytosis (69%) and population-based controls (64%) compared to patients who tolerated sulphasalazine (45%); odds ratio 2.71 [95% confidence interval (CI) 1.20; 6.15], P = 0.015, and odds ratio 2.17 (95% CI 1.32; 3.56), P = 0.002, respectively. Patients who developed agranulocytosis did not differ from population-based control subjects in the frequency of slow acetylators; odds ratio 1.25 (95% CI 0.62; 2.53), P = 0.535. The risk of agranulocytosis did not appear to be increased in slow acetylators, provided that the difference compared with sulphasalazine-treated control subjects was not due to a predominance of fast acetylators among patients with inflammatory joint disease. Instead, selection bias was suspected since more slow acetylators may have discontinued sulphasalazine therapy because of drug-intolerance.

The risk of acute major upper gastrointestinal bleeding among users of aspirin and ibuprofen at various levels of alcohol consumption

OBJECTIVE

Major upper gastrointestinal bleeding (UGIB) is the most important adverse effect of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs). Alcoholic beverages also precipitate UGIB. This analysis was conducted to evaluate whether the deleterious effects of NSAIDs are further increased among drinkers.

METHODS

An interview-based, case-control study was conducted in the U.S. and Sweden; 1224 patients hospitalized with acute major UGIB due to newly occurring peptic ulcer or gastritis were compared to 2945 neighbor controls.

RESULTS

Compared with those who drank less than one drink/wk, the relative risk of acute UGIB increased with increasing alcohol consumption, rising to 2.8 among those who drank > or = 21 drinks/wk. Among current drinkers, the relative risk of acute UGIB due to the use of aspirin was raised at all levels of alcohol consumption; the estimate for aspirin taken at least every other day (regular use) at doses of > 325 mg among all current drinkers combined was 7.0; for regular use at lower doses, the corresponding estimate was 2.8, and for any occasional use, it was 2.4. All estimates were statistically significant. Data for ibuprofen were more limited, but the relative risk estimates did not appear to vary consistently with level of alcohol consumption. For regular use (all doses combined), the estimate among all drinkers combined was significantly elevated, at 2.7; occasional ibuprofen use was not associated with UGIB (1.2). There were insufficient data to evaluate other NSAIDs according to alcohol consumption.

CONCLUSIONS

The findings suggest that acute UGIB is similarly associated with the use of the two most common nonprescription NSAIDs, aspirin and ibuprofen, at all levels of alcohol consumption. As heavy alcohol intake independently increases the risk, the incidence of UGIB is highest among persons who are both heavy drinkers and users of aspirin or ibuprofen.

Drug-attributed anaphylaxis

Allergic type I reactions to medicines range in their clinical presentation from rhinitis and urticaria to severe bronchoconstriction and anaphylactic shock. We examined all cases of suspected drug induced reactions classified as anaphylactic reactions or shock reported in Sweden between 1972 and 1995 with regard to patient characteristics and drug(s) suspected. Some comparisons with drug sales and prescription data were also made. During the study period of a total of 1338 reports concerned anaphylactic/oid shock or reactions with at least a possible causal relation to medicine giving an overall reporting rate of seven cases per million inhabitants per year of drug-induced anaphylaxis. Of these 1338 patients 51 (3.8%) died from their reactions. Among the non-fatal cases, 460 (34.4%) were diagnosed as shock and 827 (61.8%) as anaphylactic reactions. In total 46.3% of all reports concerned men but men were over-represented among the older patients and among the fatal cases (65%). There were 201 different drugs reported as 'suspected' them most common of which were dextrans (418 reports), X-ray contrast media (161 reports) and antibiotics (153 reports). For dextrans the rate of anaphylactic reactions, shock and fatal cases reported were 128,101 and 21 per million bottles respectively. This decreased to 10.3, 9.8, and 0.4 per million bottles after the introduction of preventive treatment with dextran 1 in 1983.The reporting rate for ionic contrast media were 0.14, 0.13 and 0.02 per 1000 l for reactions, shock and fatal cases respectively whilst for non-ionic contrast media they were 0.7/1000 l for reactions, 0.02/1000 l for shock, but there was no report of a fatal case. For phenoxymethylpenicillin the reported rate of anaphylaxis was 0.14 cases per million defined daily doses and for benzylpenicillin it was 3.7 cases per million defined daily doses. During the study period several drugs have been identified as important causes of anaphylaxis and measures have been taken to decrease the risk of anaphylaxis e.g. the introduction of preventive treatment with dextran 1, the shift from ionic to non-ionic contrast media and the abolition of polyethoxylated castor oil as a solvent. Spontaneous reporting of drug-induced anaphylaxis remains an important surveillance model but needs to be complemented by better quantitative methods.

Clinical aspects of upper gastrointestinal bleeding associated with the use of nonsteroidal antiinflammatory drugs

OBJECTIVE

The aim of this study was to compare the clinical features of major upper gastrointestinal bleeding among patients exposed to nonsteroidal antiinflammatory drugs (NSAID) and those not taking these drugs.

METHODS

Using data from a multicenter international case-control study designed to evaluate the role of drugs in the etiology of major upper gastrointestinal bleeding (UGIB), patients with a confirmed first episode of major UGIB were divided into two groups: those exposed to NSAIDs during the week before the onset of bleeding, and those not exposed. The groups were compared according to age and sex, clinical appearance and site of the bleeding, preceding symptoms, and requirement for transfusion and acute surgery.

RESULTS

The median age was significantly higher and the proportion of women was slightly higher among the NSAID users. There was no significant difference between users and nonusers according to the clinical presentation, the site of the bleeding, or the frequency of preceding symptoms. Forty percent in each group had no symptoms before the onset of bleeding. Slightly more NSAID users received blood transfusions, although the same median amount of blood per transfusion was given in both groups. There was no difference in the frequency of surgical intervention.

CONCLUSIONS

There are no important differences in the clinical presentation of major UGIB according to whether or not an individual is an NSAID user. An important finding is the frequent absence of preceding symptoms in patients with major UGIB, regardless of NSAID use.

European network for the case-population surveillance of rare diseases (Euronet). A prospective feasibility study

OBJECTIVE

Euronet, a case-population surveillance scheme, aims to estimate the risk of certain rare conditions which are commonly iatrogenic, by comparing drug use amongst non-selective cases with overall drug use in the general population.

METHODS

The method is based on three provisos: (1) all incident cases (irrespective of suspected aetiology) should be ascertained and studied; (2) a full drug history should be obtained from cases by direct interview; and (3) drug-use data for the products of interest should be available for this population from which cases are chosen. The feasibility of this problem-oriented approach for the identification of new signals of adverse drug reactions and for risk estimation will be tested in relation to agranulocytosis, Stevens-Johnson syndrome and toxic epidermal necrolysis in four defined areas in Europe, totalling 19 x 10(6) inhabitants, with these latest two outcomes being studied in only three regions. The design, methods and main limitations of this case-population surveillance approach are described.

Seizures and myoclonus associated with antidepressant treatment: assessment of potential risk factors, including CYP2D6 and CYP2C19 polymorphisms, and treatment with CYP2D6 inhibitors

All adverse drug reaction reports labelled seizures or myoclonus during treatment with antidepressants and stored in the Swedish national database for spontaneous reporting of adverse drug reactions were reviewed in order to evaluate possible risk factors. The reporting physicians were contacted and asked for complementary information, and blood samples for determination of the CYP2D6 and CYP2C19 genotypes were obtained from patients available. In total, 25 cases of seizures and 7 cases of myoclonus were studied. The drugs included were maprotiline (n=8), mianserin (n=7), fluvoxamine (n=6), amitriptyline (n=3), clomipramine (n=3), citalopram (n=2), paroxetine (n=2) and lofepramine (n=1). Previously suggested predisposing factors were identified in all but four cases (87%). None of the 11 patients genotyped were found to be poor metabolizers with respect to the enzymes CYP2D6 or CYP2C19. Thus, neither the CYP2D6 nor the CYP2C19 genotype were found to be associated with the occurrence of seizures/myoclonus during treatment with antidepressants. However, 15 patients (47%) were concomitantly treated with drugs with potential inhibitory effects on CYP2D6, such as neuroleptics and dextropropoxyphene, and the patients might thus have been converted from the extensive metabolizer to the poor metabolizer phenotype during this treatment. Concomitant treatment with drugs decreasing the seizure threshold and/or inhibiting the metabolism of antidepressants appeared to be an important risk factor for the occurrence of seizures/myoclonus.

Concepts in risk-benefit assessment. A simple merit analysis of a medicine?

The term 'benefit-risk ratio' is often used as a general term linked to the use of a medicine. To balance risk and benefit is, however, a very complex exercise. For most medicines the benefits are limited to a few indications and for an individual patient there is usually only a single benefit sought but the potential risks are multiple. Perceptions of risks versus benefits are influenced to a great extent by the context in which they occur. Thus, perception of risk may be different to actual risk. In the end in any given situation, the acceptable risk-to-benefit balance is an individual judgement on the part of the patient or the prescriber. For newer medicines, where there is likely to be limited experience, conservative estimates of the overall merit seem preferable so that the prescriber will use the drug critically. Subsequently, re-evaluation of the risk-to-benefit balance is necessary as greater knowledge of efficacy and adverse effects is acquired. It is possible to provide a general 'principle of threes' structure for a merit assessment based upon the concepts of seriousness, duration and incidence as related to disease indication, disease amelioration by a medicine, and the adverse effects ascribed to the medicine. This allows a rapid first comparison of medicines for a given indication. In using this general conceptual model in a transparent fashion for a given hypothesis and context, it is possible to identify the essential data used and assumptions involved that make up a merit statement. The quality and value, particularly of risk data, is problematic. Risk perception is an issue that needs to be clearly identified alongside a merit analysis. A simple merit assessment should pave the way for more focused studies.

Alcohol consumption and the risk of major upper gastrointestinal bleeding

OBJECTIVE

To examine the association between current alcohol consumption and major upper gastrointestinal bleeding.

METHODS

In a case-control study in the United States, Sweden, and Hungary, 1004 incident cases with upper gastrointestinal bleeding without predisposing factors were compared with 2446 controls. Relative risks for categories of alcohol consumption (based on the number of drinks currently consumed/wk) were estimated using logistic regression; the potential confounding effects of cigarettes, nonsteroidal anti-inflammatory drugs, and other factors were controlled simultaneously.

RESULTS

Compared with drinkers of < one drink/wk, the relative risks among other current drinkers ranged from 0.8 for 1-6 drinks/wk to 6.3 for > or = 35 drinks; the trend was statistically significant (p < 0.001). A significantly increased relative risk was seen for the heaviest consumption category within various subgroups: gastric and duodenal hemorrhage; males and females; age < 60 yr and > or = 60 yr; and those who consumed beer, wine, liquor, or a combination of beverages.

CONCLUSIONS

These findings provide evidence that consumption of alcohol increases the risk of major gastric and duodenal bleeding in nonpredisposed individuals.

Acetaminophen, some other drugs, some diseases and the risk of transitional cell carcinoma. A population-based case-control study

The purpose of this study was to investigate the risk of transitional cell carcinoma among subjects with an intake of acetaminophen, aspirin, some other drugs and with some intercurrent diseases. The source person-time ('study base') included subjects living in Stockholm in 1985-1987. The study included 325 subjects with a transitional cell carcinoma of the urinary tract and 393 controls randomly selected from the source person-time. Data were obtained by a postal questionnaire supplemented by a telephone interview. A relative risk (with a 95% confidence interval) of 1.6 (1.1-2.3) was obtained after an intake of acetaminophen, adjusted for age, aspirin, gender and smoking. Conversely, a 30% decrease in risk was obtained after an intake of aspirin. No details in the exposure substantiated the finding for acetaminophen. The inherent validity problems of observational studies, and the weak evidence in this and previous studies of the association between acetaminophen and transitional cell carcinoma, makes available epidemiological evidence insufficient to regulate the use of this commonly ingested analgesic. Increased risks were, in addition, found for tetracyclines, nitrofurantoin and a history of allergic asthma and a decreased risk found for rheumatic symptoms. The findings stress the nonepidemiological data concerning the potential carcinogenicity of acetaminophen and may be a foundation for future research of some other drugs and diseases.

Adverse experiences during treatment with zimeldine on special licence in Sweden

Adverse experiences during licensed treatment with the antidepressant serotonin (5-HT) reuptake inhibitor zimeldine in Sweden are presented. Data were obtained from a written inquiry of 694 patients and 67 reports to the Medical Products Agency. The spectrum of adverse symptoms was in agreement with those reported in previous studies on zimeldine. The most frequent adverse experiences were headache, nausea, myalgia, signs of liver function disturbance, arthralgia, neurological symptoms, fever and insomnia. No new case of the Guillain-Barré syndrome was found. The estimated frequency of the zimeldine-induced hypersensitivity syndrome (HSS), comprising fever, myalgia and/or arthralgia and signs of liver function disturbance, ranged from 1.4% to 13% in the inquiry and from 0.63% to 3.4% in the report part of the study. Adverse experiences usually had a considerably higher incidence during the first 6 weeks of zimeldine treatment than thereafter. This is in agreement with the clinical experience that most of the adverse reactions occur early during zimeldine treatment. However, a number of adverse experiences did occur with a later onset. This may justify a prolongation of the compulsory 4 weeks' testing of liver function that is required during licensed treatment. There were significantly fewer patients who developed fever among the patients who had experienced previous zimeldine treatment than among those who had not. Otherwise there was no statistically significant difference in frequency of adverse symptoms between these two groups. Consequently zimeldine treatment per se does not seem to predispose to development of an HSS or other types of adverse reactions during subsequent therapy.

Dyspnoea, asthma, and bronchospasm in relation to treatment with angiotensin converting enzyme inhibitors

OBJECTIVE

To evaluate the occurrence of asthma and dyspnoea precipitated or worsened by angiotensin converting enzyme inhibitors.

DESIGN

Summary of reports of adverse respiratory reaction in relation to treatment with angiotensin converting enzyme inhibitors that were submitted to Swedish Adverse Drug Reactions Advisory Committee and to World Health Organisation's international drug information system until 1992. Sales of angiotensin converting enzyme inhibitors in Sweden were also summarised.

SUBJECTS

Patients receiving angiotensin converting enzyme inhibitors who reported adverse respiratory reactions.

MAIN OUTCOME MEASURES

Clinical characteristics of adverse reactions of asthma, bronchospasm, and dyspnoea.

RESULTS

In Sweden 424 adverse respiratory reactions were reported, of which most (374) were coughing. However, 36 patients had adverse drug reactions diagnosed as asthma, bronchospasm, or dyspnoea. In 33 of these cases the indication for treatment with angiotensin converting enzyme inhibitors was hypertension, in only three heart failure. The respiratory symptoms occurred in about half of the patients within the first two weeks of treatment, and about one third needed hospitalisation or drug treatment. Dyspnoea symptoms occurred in conjunction with other symptoms from the airways or skin in 23 out of the 36 cases. In the WHO database there were 318 reports of asthma or bronchospasm, 516 reports of dyspnoea, and 7260 reports of cough in relation to 11 different angiotensin converting enzyme inhibitors.

CONCLUSION

Symptoms of airway obstruction in relation to treatment with angiotensin converting enzyme inhibitors seem to be a rare but potentially serious reaction generally occurring within the first few weeks of treatment.

Adverse reactions in healthy and immunocompromised children under six years of age vaccinated with the Danish BCG vaccine, strain Copenhagen 1331: implications for the vaccination policy in Sweden

A retrospective analysis of the adverse reactions reported between 1979 and 1991, in the 139,000 children under six years of age vaccinated in Sweden with the Danish BCG vaccine, strain Copenhagen 1331, showed an incidence of 1.9 per 1000 vaccinated children. Regional lymphoglandular swellings and/or abscesses were most commonly reported in 1.4 per 1000. Serious, disseminated, BCG infections developed in four infants vaccinated neonatally. Three of the infants suffered from severe, combined, immunodeficiency syndrome, undiagnosed at the time of vaccination. The incidence of severe, combined, immunodeficiency syndrome was higher in the BCG-vaccinated population (4 per 100,000 infants vaccinated within a year of their births), compared with all newborns in Sweden (1 per 100,000). The mean age at the onset of symptoms was 2.4 months for the seven non-BCG-vaccinated infants versus 1.3 months for the four BCG-vaccinated ones, while the immunodeficiency syndrome was diagnosed at an average age of 7.6 months in those who were not vaccinated versus 5.3 months in those BCG-vaccinated. It is recommended that the selective BCG vaccination of infants at high risk of exposure to tuberculosis should be postponed to six months of age to reduce the risk of inoculating infants suffering from immunodeficiency syndromes.

Metformin-associated lactic acidosis in Sweden 1977-1991

Since the withdrawal of phenformin in 1978, the use of metformin has increased from 13,500 to 22,000 patient years/year. During the period 1977-91 a total of 18 cases of metformin-associated acidosis was reported, of which 16 had lactic acidosis. The incidence of reported acidosis and lactic acidosis decreased from 1.50 cases per 10,000 patient years in 1977-81 to 0.24 cases per 10,000 patient years 1987-91, probably due to lower doses doses and reduced usage in the very old. All the reports described patients with several other concomitant diseases, mainly cardiovascular and renal, when the acidosis was diagnosed. It is important continuously to re-evaluate metformin therapy and to stop treatment at the onset of impaired renal or cardiovascular function.

Nonsteroidal anti-inflammatory drug use in relation to major upper gastrointestinal bleeding

In a study in the United States, Sweden, and Hungary, 335 cases of gastric bleeding without predisposing factors were compared with 670 control subjects, and 239 cases of duodenal bleeding were compared with 489 control subjects. For aspirin taken at least every other day during the week before the onset of bleeding (regular use), the relative risk of gastric bleeding was 4.4 (95% confidence interval [CI], 2.9 to 6.7); for occasional use, it was 3.3 (95% CI, 2.1 to 5.0). For ibuprofen, the corresponding estimates were 1.0 (95% CI, 0.4 to 2.6) and 1.1 (95% CI, 0.5 to 2.4). For naproxen, the estimate for regular use was 4.0 (95% CI, 1.5 to 11). The estimates for any use of piroxicam (crude estimate), indomethacin, and diclofenac during the week before onset were 18 (95% CI, 4.1 to 83), 1.6 (95% CI, 0.4 to 5.9), and 0.9 (95% CI, 0.2 to 4.2), respectively. The corresponding relative risks of duodenal bleeding were 7.1 (95% CI, 4.2 to 12) and 2.2 (95% CI, 1.3 to 3.7) for the regular and occasional use of aspirin, 2.4 (95% CI, 0.5 to 11) and 0.8 (95% CI, 0.3 to 2.0) for ibuprofen, 12 (95% CI, 2.8 to 54) and 9.9 (95% CI, 2.3 to 44) for naproxen, 17 (95% CI, 3.6 to 79) for any use of piroxicam (crude estimate), and 1.7 (95% CI, 0.2 to 14) for any use of indomethacin. There was a significant trend in the risk of gastric bleeding with increasing dose of regular aspirin use (p = 0.002). The relative risk estimates for the regular use of 325 mg or less were significantly elevated for both gastric and duodenal bleeding at 3.1 and 6.4, respectively.

An international comparison of case definition of severe adverse cutaneous reactions to medicines

There is substantial intercountry variation in the proportion of cases of toxic epidermal necrolysis (TEN) and Stevens-Johnson syndrome (SJS) which are attributed to specific drugs. This study was undertaken to determine whether these differences might reflect biases in diagnosis of these conditions. A total of 138 reactions in 5 countries originally diagnosed as TEN or SJS were coded on to standardised forms. A single observer blind to the original diagnosis assessed each case according to specified criteria. This observer's diagnoses were compared with the original diagnoses. Overall, 111 of the 138 cases had information adequate for assessment. The blinded observer agreed with the diagnosis for 61% of cases where the original diagnosis was TEN and 58% of cases where the original diagnosis was SJS. There was no significant difference in rates of agreement when reactions attributed to sulphonamide antibiotics were compared with reactions attributed to other drugs. There were substantial and significant differences in percentage agreement between the blinded observer's diagnosis and the original diagnoses between countries. The lowest rates of agreement between the blinded observer and the original reports occurred in the US. Our results illustrate the difficulty in comparing reaction rates based on spontaneous reports between countries where the systems for gathering such reports vary. This illustrates the need for a minimum quantity of standard data and precise definitions of reactions if spontaneous reports of adverse reactions are to provide useful information about severe adverse skin reactions associated with drugs.

Effects of exposure to benzodiazepine during fetal life

Dysmorphism and mental retardation have been reported in 7 Swedish children born of mothers who had taken high doses of benzodiazepines regularly during pregnancy. To explore this association further, we examined benzodiazepine use during pregnancy in 104,000 women whose deliveries were registered by the US public health insurance system, Medicaid, during 1980-83. Fetal outcomes were assessed from the health claims profiles of their offspring, up to 6-9 years after delivery. 80 pregnant women had received 10 or more benzodiazepine prescriptions during the 4 years. Their records showed heavy general use of health care and frequent alcohol and substance abuse, and other disorders that could confound any effect of the benzodiazepines. For the 80 pregnancies, 3 intrauterine deaths were identified as well as 2 infants with congenital abnormalities whose curtailed records suggested neonatal death. Records of 64 surviving children could be linked to these 80 pregnancies whilst records for 11 apparent survivors could not be located. 6 of the 64 survivors had diagnoses consistent with teratogenic abnormalities. The high rate of teratogenicity after heavy maternal benzodiazepine use occurs with multiple alcohol and substance exposure and thus may not be due to benzodiazepine exposure.

Comparing risk estimates of sulphonamide-induced agranulocytosis from the Swedish Drug Monitoring System and a case-control study

A comparison has been made of risk estimates for agranulocytosis connected with sulphasalazine and trimethoprim-sulphamethoxazole (T-SM) therapy calculated from data in the Swedish Drug Monitoring System ("spontaneous" reports, sales and prescription information) and a population based case-control study (the IAAAS). The relative risk for agranulocytosis during sulphasalazine treatment was calculated to be 107 and 123 by the spontaneous reporting system and the case-control study, respectively. The corresponding excess risk in both systems was 1.8. For T-SM the relative risk was 17 in the spontaneous reporting system and 12 in the case-control study, while the excess risk calculated for 3 days of treatment was 0.9 in the spontaneous reporting system, and 1.6 for 3 or more days of treatment in the case-control study. It is concluded that the Swedish Drug Monitoring System gives an appropriate estimate of the risk of developing agranulocytosis in association with the two drugs studied.

Liver damage from flucloxacillin, cloxacillin and dicloxacillin

We describe a case of severe cholestatic liver disease, which persisted for 7 years, and was probably induced by flucloxacillin. We also report a survey of 77 liver reactions which were probably or possibly induced by penicillinase-resistant penicillins and spontaneously reported to the Swedish Adverse Drug Reactions Advisory Committee. The reactions were usually cholestatic with a tendency to a protracted course. There is some evidence for an immunoallergic idiosyncratic reaction. The incidence of reported liver reactions was estimated from sales data to be 1.6-2.9 per million DDD (defined daily doses) or, for flucloxacillin 1:11,000-1:30,000 prescriptions. Female sex, age and high daily doses seemed to be associated with higher risk of liver reactions from flucloxacillin.

Angio-oedema in relation to treatment with angiotensin converting enzyme inhibitors

OBJECTIVE

To evaluate and describe the clinical course of angio-oedema reactions induced by angiotensin converting enzyme inhibitors.

DESIGN AND METHODS

All reports of angio-oedema reactions associated with angiotensin converting enzyme inhibitors submitted to Swedish Adverse Reactions Advisory Committee were reviewed and the clinical courses summarised. Numbers of cases judged to be induced by angiotensin converting enzyme inhibitors were related to their annual usage, estimated from total sales of defined daily doses, as well as to the estimated number of new patients. All cases of angio-oedema associated with angiotensin converting enzyme inhibitors reported to the World Health Organisation's international drug information system were also summarised.

RESULTS

36 of the 38 reported cases in Sweden between 1981 and 1990 were judged to be related to angiotensin converting enzyme inhibitors. During 1981 through 1990, altogether 1309 cases of angio-oedema associated with angiotensin converting enzyme inhibitors were registered with the international drug information system. The incidence of reported cases of angio-oedema increased largely in parallel with the increased sales (usage) of angiotensin converting enzyme inhibitors. Of the 36 Swedish patients, 77% experienced the reaction within the first three weeks after starting treatment. 10 patients needed hospitalisation, two of whom had life threatening laryngeal obstruction. With one exception all 36 patients were free of symptoms within one week after discontinuing the drug.

CONCLUSIONS

Angio-oedema induced by angiotensin converting enzyme inhibitors is a rare but potentially life threatening reaction, which in most instances occurs shortly after the start of treatment. Any patient in whom the reaction is suspected should have the treatment interrupted and, if necessary, be admitted for observation.

Severe hypoglycaemia during treatment with glipizide

In the years 1980-87, 19 cases of severe hypoglycaemia during treatment of Type 2 diabetes with glipizide were reported to the Swedish Adverse Drug Reactions Advisory Committee. Patient age was 75 +/- 9 years (mean +/- SD) (significantly higher than in all patients on glipizide in Sweden), and the duration of glipizide treatment ranged from 1 day to 4 years. Eleven patients presented with coma, three with reduced consciousness and five with other symptoms. An uneventful early recovery occurred in 14 patients. The remaining five patients had prolonged or recurrent hypoglycaemia for up to 60 h. Two of the patients, both with complicating disorders, died. In a case-control substudy, patients with glipizide-associated hypoglycaemia were found to have renal impairment more often than age- and sex-matched controls treated with glipizide (odds ratio 4.0). The median dose of glipizide (10 mg per day) was identical to that in controls. Other drugs, notably diuretics and benzodiazepines, were more commonly used by hypoglycaemic patients (median 5 vs 2 concomitant drugs, p less than 0.001). We conclude that the use of glipizide is associated with a risk of developing severe hypoglycaemia with a clinical course that is not always benign.

Evaluation of a prescription based record-linkage model for epidemiological studies of long-term adverse effects of drugs--with special regard to combined oral contraceptives

The Department of Drugs of the Swedish National Board of Health and Welfare undertook a study of the possibilities of a new scheme for post-marketing surveillance by means of prescription and register based epidemiological studies, primarily of combined oral contraceptives (COC). Based on available data on COC usage patterns and incidence rates of the disease at study, it was estimated that study periods, including the necessary time periods for disease development and generation of a sufficient number of cases, would amount to at least 1 to 13 years for cardiovascular outcomes and 8 to 17 years for reproductive cancers. Prospective and unbiased exposure ascertainment would be the most important advantage. However, delay in follow up, the need for extensive individual questionnaire probing and fear of violation of personal integrity could adversely affect the feasibility of the scheme. Chiefly on the grounds of the extended study periods and magnitude of the necessary infrastructure, it was not judged cost-effective to pursue such a scheme for COC exposure only. It was, however, suggested that it would be considered for epidemiological surveillance of other drugs that are commonly used and for which short term and frequent serious side effects are expected, as for instance lipid lowering compounds, beta-blockers, bensodiazepines and other psychotropic drugs.

Acetazolamide-associated aplastic anaemia

Eleven cases of acetazolamide-associated aplastic anaemia were reported in Sweden during a 17-year period. There were six women and five men with a median age of 71 years (range 63-85 years). The median dose of acetazolamide was 500 mg, and the median duration of treatment was 3 months (range 2-71 months). Ten of the eleven patients died, all within 8 weeks after detection of their aplastic anaemia. The relative risk of developing aplastic anaemia when taking acetazolamide was 13.3 (95% confidence limits (CL); 6.8-25.3). The estimated incidence of reported acetazolamide-associated aplastic anaemia is approximately one in 18,000 patient years. The results strongly indicate that acetazolamide treatment is associated with a substantial increase in the risk of developing aplastic anaemia.

Trimethoprim-sulphamethoxazole-associated blood dyscrasias. Ten years' experience of the Swedish spontaneous reporting system

During the 10-year period 1976-1985, a total of 154 cases of blood dyscrasia were reported in Sweden which were evaluated as having a probable or possible causal relationship with trimethoprim-sulphamethoxazole (T-SM). There were 61 cases of leucopenia (of which 16 had agranulocytosis), 28 cases of thrombocytopenia, and two of non-haemolytic anaemia. There were also 32 cases of bicytopenia and 31 cases of tricytopenia. The median age varied from 38 years in the leucopenia group to 81 years in those with tricytopenia. The overall fatality rate was 17%, ranging from 2% in the group with mild leucopenia to 52% in the group with tricytopenia. In relation to sales and prescription data, the overall incidence of reported T-SM blood dyscrasias was 5.3 per million defined daily doses, and among out-patients the incidence was one case per 18,000 prescriptions. Thus the overall incidence of any blood reaction to T-SM appears to be low. In relation to prescription data, elderly people were overrepresented among the serious reactions.

Quality criteria for early signals of possible adverse drug reactions

The main function of the World Health Organisation's International Collaborative Programme on Drug Monitoring is to provide a reliable early warning of possible health hazards caused by medicines. Described here is an attempt to devise criteria that would produce a well-founded early signal of an adverse reaction on the basis of reports sent in by national collaborating centres and combined in the WHO database. To reduce the frequency of spurious associations (false-positive signals) it is suggested that publication be delayed until a few case-histories meeting the suggested criteria have been sent in. The criteria were tested retrospectively against early published case-reports on drug-associated agranulocytosis. 19 suspected associations were examined and a signal in the database was defined by there being three or more cases containing stipulated information about the patient and the treatment. The WHO database had reports on all the associations, suggested criteria for a signal being met in 15 instances. This signal was present when the first case was published in 7 instances and within three months of first publication in 1. Moreover, in 3 instances where publication came first the cases presented had been collected by a national drug monitoring centre. The WHO databank has the potential to provide doctors and scientists with signals which then should be evaluated in detail.