Article Text
Abstract
Objective There is ongoing controversy regarding a ‘J-curve’ phenomenon such that low and high blood pressure (BP) levels are associated with increased risks of recurrent stroke. We aimed to determine whether large treatment-related BP reductions are associated with increased risks of recurrent stroke.
Design Data are from the PROGRESS trial, where 6105 patients with cerebrovascular disease were randomly assigned to either active treatment (perindopril±indapamide) or placebo(s). There were no BP criteria for entry. BP was measured at every visit, and participant groups defined by reduction in systolic BP (SBP) from baseline were used for the analyses. Outcome was recurrent stroke.
Results During a mean follow-up of 3.9 years, 727 recurrent strokes were observed. There were clear associations between the magnitude of SBP reduction and the risk of recurrent stroke. After adjustment for cardiovascular risk factors and randomised treatment, annual incidence was 2.08%, 2.10%, 2.31% and 2.96% for participant groups defined by SBP reductions of ≥20, 10–19, 0–9 and <0 mm Hg, respectively (p=0.0006 for trend).
Conclusions The present analysis provided no evidence of an increase in recurrent stroke associated with larger reductions in SBP produced by treatment among patients with cerebrovascular disease.
- Stroke
- Epidemiology
Statistics from Altmetric.com
Introduction
A recent observational analysis of the PRoFESS trial suggested a J-shaped association, where both the highest and the lowest achieved on-treatment blood pressure (BP) levels were associated with increased risks of recurrent stroke among patients with recent ischaemic stroke.1 This finding raised concern about the possible harm of intensive BP lowering down to very low level. However, most of the previous observational studies investigated the effects of absolute on-treatment BP values but did not examine treatment-induced BP reductions. The objective of the present post hoc analysis of the PROGRESS trial was to determine whether larger reduction in systolic BP (SBP) associated with treatment increases the risks of recurrent stroke among patients with cerebrovascular disease.
Methods
The design of the PROGRESS trial has been described in detail elsewhere.2 ,3 Briefly, 6105 participants who had a history of stroke or transient ischaemic attack (TIA) within 5 years were randomly assigned to active treatment (2–4 mg perindopril±2–2.5 mg indapamide) or matching placebo(s). BP was measured in duplicate, with an interval of at least 2 min, to the nearest 2 mm Hg after 5 min of quiet rest in the seated position using a standard mercury sphygmomanometer at every visit (see online supplementary methods for details). The outcomes were recurrent stroke and its subtypes (intracerebral haemorrhage (ICH) and ischaemic stroke) (see online supplementary methods for details). The associations between outcomes and SBP reduction (baseline SBP—follow-up SBP) at each visit were investigated by Poisson linear regression models using SBP reduction (≥20, 10–19, 0–9 and <0 mm Hg) as a time-dependent variable.4
Results
Baseline characteristics are shown in online supplementary table 1. Over a mean follow-up of 3.9 years, 727 recurrent stroke events (111 ICH and 565 ischaemic stroke) occurred. For the four groupings defined by SBP reduction of ≥20, 10–19, 0–9 and <0 mm Hg, median values of SBP reduction were 29, 14, 4 and −10 mm Hg and median values for achieved on-treatment SBP were 128, 134, 138 and 149 mm Hg, respectively. In these four groupings, the numbers of stroke events/person-years were 279/7546, 127/4318, 110/3945 and 179/5821, producing adjusted annual incidence rates of recurrent stroke 2.08%, 2.10%, 2.31% and 2.96%, respectively (p=0.0006 for trend; figure 1). Comparable results were observed for subgroups defined by history of ICH, baseline SBP (below vs above mean) and randomised treatment (p homogeneity=0.13, 0.41 and 0.40, respectively). Comparable results were also observed between 2102 patients with prior lacunar infarction and 948 with prior large-artery infarction (p homogeneity=0.63). There were strong and continuous associations between SBP reduction and the risk of ICH, while the decreases in the risk of ischaemic stroke were similar in the groups achieving SBP reductions of ≥20, 10–19 and 0–9 mm Hg compared with patients who did not achieve any SBP reduction, though p for trend was significant (p=0.04; figure 1).
Discussion
The present analysis demonstrated that larger reductions in SBP up to 20 mm Hg or more were associated with lower risks of recurrent stroke. The relationship of SBP reduction and ICH was stronger than that of ischaemic stroke, for which there may be a ceiling effect of SBP reduction. These results are consistent with previous meta-regression analyses of randomised controlled trials of BP-lowering treatment, which showed that larger SBP reduction up to 10–20 mm Hg was associated with greater reduction in the risks of initial and recurrent stroke.3 ,5 The present findings are also consistent with the SPS3 trial, which demonstrated a significant reduction in recurrent ICH and a modest decrease in recurrent ischaemic stroke with intensive BP lowering targeting an SBP of <130 mm Hg among patients with recent symptomatic small subcortical stroke.6 Our previous observational analysis of PROGRESS also demonstrated continuous associations between achieved on-treatment SBP and recurrent stroke down to 112/72 mm Hg.4 Although the PRoFESS trial reported a J-shaped association,1 there are a couple of potential explanations underlying the discrepancy. PROGRESS included patients with all types of stroke and TIA while PRoFESS involved ischaemic stroke only (mainly large-artery and small-artery infarction). In PRoFESS, median time from qualifying stroke to randomisation was 15 days and the J-shaped association was prominent within 180 days of randomisation, while many PROGRESS participants were recruited beyond this time. Median achieved SBP in the largest SBP reduction group in PROGRESS did not reach 120 mm Hg, the point at which increased recurrent stroke was observed in PRoFESS.
Although the present analyses are observational and non-randomised, the comparability of the findings with those from a randomised controlled trial6 and with meta-regression analyses3 ,5 provides reassurance that they are not substantially biased. Because many participants of PROGRESS were recruited long term after the index event, the effects of SBP reduction early after stroke were not evaluated in the present analysis.
In conclusion, the present analysis provided no evidence of increase in recurrent stroke associated with larger reductions in SBP produced by treatment among patients with cerebrovascular disease participating in the PROGRESS trial.
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Collaborators PROGRESS Collaborative Group.
-
Contributors HA and JC wrote the manuscript. HA analysed the data. SM and JC obtained grant funding. CA, TO, MW, SM, GM, M-GB, CT, SH, LL and BN revised the manuscript.
-
Funding This work was supported by Servier, National Health and Medical Research Council of Australia (special grant) and Health Research Council of New Zealand (HRC98/149).
-
Competing interests TO received personal fees and research grants from Daiichi. SM and JC received personal fees and research grants from Servier. GM received personal fees from Boehringer Ingelheim, Bayer, Daiichi Sankyo, Medtronic, Novartis, Recordati, Servier and Takeda. SH received personal fees from Servier. BN received personal fees from Amgen, Astra Zeneca, GlaxoSmithKline, Pfizer, Servier and Tanabe and research support from Johnson and Johnson, Merck Schering Plough, Roche, Servier and United Healthcare Group.
-
Patient consent Obtained.
-
Ethics approval Institutional Review Board of each participating centre.
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Data sharing statement Additional unpublished data from the study—if any—are available, to Professor John Chalmers, The George Institute for Global Health, PO Box M201, Missenden Road, NSW 2050 Australia.