Pharmacoepidemiologic studies on cardiovascular drugs - with special reference to the effectiveness and safety of blood pressure lowering drugs

Abstract: Background: Ischaemic heart disease (IHD) and stroke are the main causes of morbidity and mortality in developed countries and hence constitute a large public health problem. Cardiovascular drugs are used in an effort to prevent (e. g. blood pressure lowering drugs (BPLD) and lipid lowering drugs) and palliate (e. g. nitrates) these disorders. The efficacy and safety of these drugs have been evaluated in randomised clinical trials. However, conclusions from these trials are not representative of patients in general. This necessitates epidemiologic studies to evaluate the long-term effectiveness and risk of cardiovascular drugs in routine care. Methods: 283 Swedish municipalities were studied during 1989 - 1993 by multivariate Poisson regression at the ecologic level. Drug utilisation was categorised into four equal-size groups by quartiles and compared with mortality from IHD and stroke. The relation between stroke mortality and utilisation of BPLD was also analysed in 49 municipalities by correlation of direct age standardised (equivalent average rate method) rates. At the individual level, the incidence of ischaemic cardiac events and stroke in relation to the use of BPLD and blood pressure level was analysed in a prospective population based cohort of 500 men born 1914 and living in Malmö in 1982-83 by multivariate 10-year survival analysis. Results: Compared with the municipalities within the lowest fourth of drug utilisation, mortality from IHD and stroke was reduced in the three other fourths of lipid-lowering drugs utilisation. Only the municipalities within the uppermost BPLD utilisation had a lower IHD (-5%) and stroke mortality (-10%). Utilisation of nitrates was associated with a gradually increased risk of mortality from stroke (+3%, +7%, +11%) and an increased risk of mortality from IHD (+10%, +9%, +13%). At the individual level, use of BPLD was associated with increased incidence of ischaemic cardiac events both in those with diastolic blood pressure above, and in those at or below, 90 mmHg. However, after adjustment for confounders the two-fold risk increase among the former disappeared, whereas the four-fold risk increase among the latter remained. Systolic blood pressure <160 mmHg conveyed a lower risk of stroke, but this benefit was only seen in untreated men. In men being treated with BPLD the opposite was true, the risk of stroke being increased almost threefold. Main conclusions: Both BPLD and lipid lowering drugs seemed to be effective in the general population, but the effectiveness of routine treatment with BPLD appeared to be low; reduced IHD and stroke mortality was reached in only one fourth of the municipalities. This need not indicate a suboptimal utilisation level of BPLD but rather that various factors may counteract their effects in routine care. The safety of nitrates should be further analysed as they might enhance the risk of IHD and stroke mortality. A plausible explanation of the increased risk in BPLD users with low blood pressure is that it was confined to a subgroup of subjects not represented in randomised clinical trials in whom lowering of blood pressure could be detrimental (the J-curve phenomenon).