Article type
Year
Abstract
Background: Population-based longitudinal studies have established a continuous and positive relationship between blood pressure (BP) and mortality. Concomitantly, several reports of studies in treated hypertensive patients have described this relationship as J-shaped with an increased risk of events in patients with low BP. Possible explanations have been debated at length, including adverse effects due to over-treatment. However, these studies had generally no control group of non treated patients to investigate the role of treatment.
Objective: to determine whether the J-shaped relationship between the risk of events (cardiovascular deaths, non cardiovascular deaths and total mortality) and BP level is independent of treatment
Methods: meta-analysis of seven randomised clinical trials comparing an active antihypertensive treatment to placebo (or no treatment), using individual data of 39320 patients followed-up for 3.9 years on average. Follow-up was divided into adjacent periods of one year in which ongoing diastolic and systolic BP levels were related to the risk of events, with adjustment on major risk factors.
Results: A total of 2349 deaths were observed (52.7% cardiovascular). A J-shaped relationship was observed on total, cardiovascular and non cardiovascular mortality, in both treated and non treated patients. There was more evidence of a J-curve with DBP than with SBP. Patients with wide pulse pressure (low DBP and high SBP) did not explain this findings.
Conclusions: These results suggests that the increased risk of events observed in patients with low BP is independent of treatment and is also apparent on non-BP related events. Poor health condition related to low BP could contribute to the J-curve. The meta-analytic technique using individual patient data from randomised clinical trials offered a very powerful and appropriate method to explore the issue. It was an excellent alternative to previous epidemiological studies which could not assess the role of treatment, and to secondary analyses of clinical trial data which were limited by a lack of power.
Objective: to determine whether the J-shaped relationship between the risk of events (cardiovascular deaths, non cardiovascular deaths and total mortality) and BP level is independent of treatment
Methods: meta-analysis of seven randomised clinical trials comparing an active antihypertensive treatment to placebo (or no treatment), using individual data of 39320 patients followed-up for 3.9 years on average. Follow-up was divided into adjacent periods of one year in which ongoing diastolic and systolic BP levels were related to the risk of events, with adjustment on major risk factors.
Results: A total of 2349 deaths were observed (52.7% cardiovascular). A J-shaped relationship was observed on total, cardiovascular and non cardiovascular mortality, in both treated and non treated patients. There was more evidence of a J-curve with DBP than with SBP. Patients with wide pulse pressure (low DBP and high SBP) did not explain this findings.
Conclusions: These results suggests that the increased risk of events observed in patients with low BP is independent of treatment and is also apparent on non-BP related events. Poor health condition related to low BP could contribute to the J-curve. The meta-analytic technique using individual patient data from randomised clinical trials offered a very powerful and appropriate method to explore the issue. It was an excellent alternative to previous epidemiological studies which could not assess the role of treatment, and to secondary analyses of clinical trial data which were limited by a lack of power.