RESEARCH LETTERS Although its risk/benefit ratio has been questioned, nifedipine is still used in the treatment of hypertension and angina. We suggest that nifedipine may be associated with a risk of aplastic anaemia. Since 1980, we have done case-control surveillance on agranulocytosis and aplastic anaemia; initially part of the International Study on Agranulocytosis and Aplastic Anaemia (ISAAA). 1 With the collaboration of haematology services in the Metropolitan Area of Barcelona (4·12–4·30 10 6 inhabitants), all cases of aplastic anaemia (white blood cell count 3·5 10 9 /L platelets 50 10 9 /L, haemoglobin <100 g/L, or haematocrit of <30%, with a compatible bone-marrow biopsy sample) were identified. Controls paired for age, sex, and hospital were selected according to a list of diagnoses judged to be independent of the reason for use of most drugs. Cases and controls were interviewed during hospital admission with a structured questionnaire by trained interviewers. Exposure window was the 5 months ending 1 month before admission. Confounding was controlled by a multiple regression model, including age, sex, exposure to acetylsalicylic acid, paracetamol, analgesic drugs, indometacin, butylpyrazolidines, other non-steroidal anti- inflammatory drugs, chlorphenamine, nifedipine, other calcium-channel blockers, penicillamine, gold salts, or both drugs associated with a known risk of aplastic anaemia (including acetazolamide, allopurinol, carba- mazepine, chloramphenicol, sulfasalazine, and ticlopidine), rheumatoid arthritis, and occupational exposure to benzene, other solvents, and pesticides. Since the number of exposed patients was low, and with the aim of confirming the estimates obtained with the multiple regression model, we did a stratified analysis with two strata, defined according to the presence or absence of simultaneous exposure to the drugs known to induce aplastic anaemia. To find out the incidence of aplastic anaemia among users of nifedipine, the consumption of nifedipine in the study area was investigated and the absolute risk (and 95% CI according to Poisson distribution) was estimated, assuming that all exposed cases were due to nifedipine. Estimates on consumption of nifedipine were provided by the National Health System Prescription Pricing Authority, General Directorate of Drugs, Ministry of Health. During a follow-up of 74·5 10 6 person-years, 178 cases of aplastic anaemia were identified, giving an incidence of 2·39 per 10 6 inhabitants per year. Of these, 147 could be interviewed and were compared with 1295 controls. Six cases (4·1%, table) and 11 controls (0·8%) had been exposed to nifedipine during the window period. Multi- variate odds ratio was 4·7 (1·5–14·6). The stratified estimate was 4·6 (1·7–12·8). All six cases died within 5 months of the diagnosis. All of them had taken nifedipine for at least 7 months. Unfortunately, we do not know whether nifedipine was taken by any of the cases after aplastic anaemia had been diagnosed. Only one case had taken another drug known to be associated with aplastic anaemia (case 2, allopurinol). None had been exposed to benzene, other solvents, or pesticides. Other calcium-channel blockers had been used by three cases and 17 controls (1·51 [0·4–5·7]). During the study period, 187·6 million daily defined doses of nifedipine were consumed in the study area. Assuming that the six exposed cases can be attributed to nifedipine, the incidence would be 11·7 cases per million users per year (95% CI 4·2–25·4), which gives a risk-ratio estimate of 4·9 (1·8–10·6) compared with the incidence of aplastic anaemia in the general population. THE LANCET • Vol 352 • August 22, 1998 619 Research letters Fatal aplastic anaemia associated with nifedipine Joan-Ramon Laporte, Luisa Ibáñez, Elena Ballarín, Eulàlia Pérez, Xavier Vidal Case Age Sex Date of Daily dose (mg), Induction Other pharmacological and occupational exposures* ( yrs) diagnosis formulation period 1 62 M June 1988 20, conventional 1 year Nitroglycerin (occasionally, 1 year), acetylsalicylic acid (occasionally, several years), clorazepate (4 days) 2 79 M March 1989 30, conventional 1 year Allopurinol (6 months), digoxin (1 year), potassium chloride (7 months), furosemide (7 months), isosorbide mononitrate (1 year), topical nitroglycerine (2 years), paracetamol (occasional, 1 year), temazepam (variable, several months) 3 70 M May, 1991 30, conventional 2 years Paracetamol (7 days), cold medication† (7 months), cloxacillin (6 months) 4 82 F Jan, 1991 30, conventional 7 months Tenoxicam (3 months), diclofenac (years), paracetamol (3 months), acetylsalicylic acid (2 years), digoxin (2 years), dipyridamole (2 years), methaqualone+dipenhydramine (6 months) 5 67 M Feb, 1993 10, conventional 5 years Atenolol (5 years), acetylsalicylic acid (occasionally) 6 74 M April 1994 40, retard 8 months Furosemide (8 months) *During 5-month period ending 1 month before diagnosis. †Including paracetamol+chlorphenamine+dextromethorphan. People with aplastic anaemia exposed to nifedipine