BP is one of the most important issues to be managed for IS patients in the acute phase, and whether an acute anti-hypertensive treatment aiming to reduce the elevated BP in such patients is beneficial has been a major question in the clinic. Controversial opinions and supportive evidence on both sides exist at the same time; thus, a systematic review of the available data was needed. In this study, with a systematic search for RCTs in IS patients, in which an enhanced BP decrease after the anti-hypertensive intervention within the acute phase was shown, a neutral (or near-harmful) effect of such therapy was found in the outcomes we analyzed. The results revealed that anti-hypertensive therapy has no evidence of providing a benefit to the prognosis of patients and could be an excessive medical treatment in clinical practice. In other words, unless otherwise indicated, anti-hypertensive treatment in the acute phase is not recommended for IS patients, in accordance with previous opinions in this field .
The neutral effect on functional outcome, i.e., death or dependency and mRS distribution, was also found in previous studies and meta-analyses [20, 35, 41–43]. Death or disability after a stroke event has led to plenty of harm and placed burdens on stroke patients and their family members, and this was the most important prognosis . For the medical team, functional outcome is one of the most important indexes to evaluate the prognosis of stroke patients. Thus, for the crucial clinical questions of functional outcome improvement, the content of a treatment and the timepoint to apply it for their patients is the most important consideration for physicians, and strong supportive evidence for these treatments is needed. Previous studies, including multiple meta-analyses [35, 41–43], tried to uncover the impact of anti-hypertensive therapies on functional outcome. All of these studies obtained neutral results, and they concluded that anti-hypertensive therapy has neither beneficial nor harmful effects on functional outcome and prognosis. This result has also been found in the current meta-analysis, which could be due to multiple reasons. From a pathophysiological view, the beneficial and harmful effects of BP-lowering therapy indeed exist simultaneously, which could neutralize each other and lead to unwanted functional outcomes and adverse event results. Of course, when more clinical trials related to this issue are available, preferably with a higher number of included patients and a longer follow-up period, it is possible that eventually a small but significant harmful effect can be found, but the current results indicate that such a therapy lacks evidence to be beneficial for IS patients.
Lowered BP by anti-hypertensive treatment for AIS patients revealed a slight but not significant harmful effect on mortality. BP was shown to be an important factor in the prognosis of ischemic stroke [9, 51, 52]. On the one hand, increased BP helps maintain cerebral perfusion via the direct induction of CBF and progressive vasoconstriction of arterioles . On the other hand, less BP reduction is related to a higher perihematomal edema ratio and hematoma expansion , along with blood-brain barrier dysfunction . The mortality results of this meta-analysis supported that these two opposite effects from BP reduction counteract each other in the rapidly changing pathophysiology during the acute phase when anti-hypertensive treatments were used. As the mortality for AIS patients showed a U-shaped curve in observational studies [8, 10, 13], it is possible that these anti-hypertensive treatments have reduced BP to the “flat bottom” of the mortality curve, and neither beneficial nor unfavorable effects could be seen in the current study. Most previous studies showed a neutral effect on the mortality of these anti-hypertensive treatments, including most original RCTs [17, 20–22, 25, 27] and several systematic reviews [35, 42, 43]. The current meta-analysis excluded those studies that did not effectively decrease BP after anti-hypertensive treatment; thus, a lower variation and a higher absolute effect on mortality was achieved, but this effect was still not significant. However, with a nonsignificant result, a trend favoring the control group drew attention, especially in the sensitivity analysis, that a significant harmful effect of the lowered BP by anti-hypertensive therapy could be seen after CHHIPS 2009  was excluded. Compared to the other studies, CHHIPS 2009  had a relatively lower number of enrolled participants, higher average age of patients, and more severe baseline stroke severity (Table 2). Moreover, this study was premature because a significant beneficial effect was found from their anti-hypertensive treatment. Thus, the possibility of this study being biased cannot be excluded. Since the result shown here supported the harmfulness of effective anti-hypertensive treatment, it could at least partially reveal the possibility when more data are available in the future, and a stronger conclusion could be made. Careful consideration is needed when applying this result in clinical practice, but the potential harmful effect to increase mortality by such therapies needs to be taken into consideration. In addition, one previous meta-analysis  of both ischemic and hemorrhagic stroke patients showed that anti-hypertensive treatments increased short-term mortality, an effect that was not significant in the current study. In their study, a large portion of the effect was contributed by the BEST trial , which included both ischemic and hemorrhagic stroke patients who had mild symptoms on stroke onset compared with patients in other studies; thus, nonnegligible clinical heterogeneity existed between the BEST trial  and the studies enrolled in the current meta-analysis. The short-term mortality in our study showed a trend preferring the control group, which is related but not fully equivalent to the result of this meta-analysis. In conclusion, a trend of harmful effects could be found in our study, which is worth taking into consideration, and a stronger and clearer conclusion in the future will need more large-scale RCTs that achieve significant BP differences between the intervention and control groups.
High BP is one of the most important and widely accepted risk factors for stroke events , and SBP level was shown to be related to recurrent stroke . It is also reasonable that BP after IS onset, as one of the central cardiovascular indexes, has impacts on vascular events, including recurrent ischemic stroke and hemorrhagic stroke, and myocardial infarction events. In addition, disruption of the central nerve BP regulation reflex axis by the index stroke could lead to unfavorable influences, including these outcomes. However, in this study, the SAEs analyzed reflected neutral effects as well. One possible explanation could be that the available data were not enough to draw a clear conclusion, such as ischemic stroke events, for which only 3 RCTs were included, and for the relatively rare SAEs discussed here, including myocardial infarction events and hemorrhagic stroke events, only 41 and 51 events were reported in total, respectively. Small amounts of data from original studies were unable to reveal small effects, especially for those relatively rare events. Aside from our study, the previous meta-analyses also did not report positive results on these terms, which showed that these SAEs were of interest, but more data are needed. In the future, RCTs that report these SAEs are necessary for this field.
This study has several advantages. First, to our knowledge, this is the first meta-analysis that only included studies with a significant BP difference between groups; thus, the efficiency of the conclusions in this study was increased. Previous meta-analyses often mixed the results of all related RCTs together, which decreased the efficiency of revealing the real effect of enhanced BP decrease. On the one hand, the potential non-BP-lowering effects of these anti-hypertensive drugs could contribute to the prognosis of AIS patients. On the other hand, since the BP did not decrease significantly compared with that in the control group in these studies, its effect on prognosis could be buried by comparable amounts of effects other than BP lowering. In contrast, in the current meta-analysis, the studies that failed to see a significant difference in BP between treatment and control groups were excluded; thus, the influence of potential non-BP-lowering effects of these anti-hypertensive drugs was carefully decreased, and the effect of BP decrease was enlarged. Second, only the patients in the IS subgroup were included when the study population of the RCT was all stroke patients; thus, more detailed and precise results were made by this study. Hemorrhagic and ischemic stroke have different pathophysiological processes and different treatments in clinical practice, but many previous studies and meta-analyses did not differentiate between them, which led to a decreased power of their results. Finally, as the dynamic change in the brain infarcted area and hemodynamics within 72 h has a nonnegligible influence on prognosis and such an effect could change after the acute phase, the purification of effects from acute phase BP management also increased the significance of this study.
However, some limitations of this study still need to be mentioned. First, although heterogeneity analysis revealed that for most outcomes, the statistical heterogeneity was not considerable, and a random effects model was used for other outcomes, the clinical heterogeneity still calls for attention, since it can be potentially increased by multiple factors among different RCTs included in this study. The included RCTs have different interventions, including ACEIs, ARBs, BBs, CCBs, diuretics, etc., which would increase the heterogeneity and potentially decrease the effectiveness because of the different pharmacological characteristics of these anti-hypertensive drugs. The patients included in this study had different initial severities of stroke. In addition, the extent of the significant BP difference and the timepoints it was observed among these studies were different. It is reasonable that a relationship exists between the extent of BP differences and outcomes, although such a relationship was hardly measured in our study because of the different baseline BPs among the included RCTs. In addition, if the BP has a larger effect in the superacute phase than later, or vice versa, which is possible, then a higher variability of the results would be observed. The factors mentioned above, including anti-hypertensive therapies, baseline stroke severity, and the extent and timepoint of significant BP difference, increased clinical heterogeneity and therefore decreased efficiency in searching for the influence on the outcomes. Aside from that, except for the significantly increased mortality in the sensitivity analysis after excluding one study, we did not find any statistically significant effects from enhanced BP decrease. It is possible that such decreased BP is neither beneficial nor harmful for patients, but it is also possible that the number of included patients in currently available studies may not be enough to overcome the clinical heterogeneity among the RCTs and that the real effects were masked. In addition, the unavailability of specific individual data in the original RCTs included in the current meta-analysis limited further subgroup analyses and additional searches for the source of heterogeneity. Finally, although we tried to eliminate the effects of anti-hypertensive treatments used in RCTs other than BP lowering, these nonrelevant effects were minimized but not eliminated in the current study; thus, the results still need further discussion when applied in clinical practice.