1 Introduction

Intracerebral haemorrhage (ICH) accounts for approximately 15% of all stroke, and is considered the least treatable form of stroke [1]. Spontaneous ICH is associated with high mortality and morbidity with a case fatality of 40% at 1 month [2, 3]. Due to an ageing population, incidence of ICH is likely to rise in the future, so finding effective management strategies is imperative [4].

Despite the significant burden of ICH, limited effective management options exist. Elevated blood pressure (BP) is a common feature after the onset of ICH [5] and is associated with expansion of the underlying haematoma [6] and poor outcome [7]. Therefore, BP control remains one of the mainstays of acute ICH management. Two large BP lowering trials in acute ICH, the second Intensive Blood Pressure Reduction in Acute Cerebral Haemorrhage Trial and the second Antihypertensive Treatment of Acute Cerebral Haemorrhage trial, showed discordant results [8, 9]. However, pooled analysis from these trials suggest that early and sustained BP control may improve outcomes [10], and intensive BP lowering as part of an “ABC care bundle” in ICH, that additionally consisted of rapid anticoagulant reversal and prompt neurosurgical referral in suitable patients, was associated with lower case fatality[11].

The relationship between systemic BP and cerebral physiology is complex, and may be adversely affected in ICH. First, cerebral autoregulation (CA) maintains a constant cerebral blood flow (CBF) despite variations in cerebral perfusion pressure (CPP). There is already evidence of impaired CA and lower CBF velocity in acute ICH [12], which raises concern that intensive BP lowering may result in ischaemia. Secondly, the haematoma and its growth can cause surrounding vasogenic oedema, further contributing to raised intracranial pressure (ICP), mass effect and midline shift [13], which may be associated with serious adverse effects [14]. Intracranial hypertension is common in spontaneous ICH with Kamel and colleagues reporting it in up to 70% of ICH patients undergoing ICP monitoring [15]. However, there is a relative paucity of data on the effect of systemic BP lowering on ICP and CPP. We therefore conducted a systematic review investigating the effects of BP lowering on ICP in acute ICH.

2 Methods

The protocol for this systematic review was registered on PROSPERO (CRD42019134470), and was in accordance with the Preferred Reporting Items for Systematic Reviews Meta-analysis [17]. The checklist results are presented in Supplementary material Figure 1.

2.1 Search Strategy

Studies were identified with a search strategy across three databases (MEDLINE, EMBASE and CENTRAL), between 1974 and April 2019. Appropriate subject headings or subcategories for each database were used (Supplementary material Figure 2). The reference lists of included papers were searched for any other titles that were relevant, as were reference lists of relevant papers.

2.2 Inclusion and Exclusion Criteria

Studies with BP measurements, and ICP or surrogate measures in acute ICH were included. Eligibility was assessed by reading abstracts, and, if necessary, whole articles. The effects of ICP changes on clinical outcome were assessed. Excluded were non‐English language articles, non-human studies, healthy population, participants under 18 years old and non-ICP or surrogate measure outcomes. Studies were screened initially by title and abstract by two reviewers (MK and JSM). TGR was asked to adjudicate any disagreements. Included studies were evaluated as full papers by both reviewers against the inclusion and exclusion criteria (MK and JSM).

2.3 Data Extraction

The following data were extracted: (1) number of patients and controls (where applicable), (2) sex, (3) age, (4) previous hypertension, (5) previous history of stroke, (6) systolic and diastolic BP, (7) heart rate, (8) ICP, (9) CPP, (10) central venous pressure, (11) clinical outcome, (12) neurosurgical procedures, and (13) main conclusions. The methodological quality of the selected studies was assessed by the Newcastle-Ottawa scale (NOS) for observational studies. This scoring system evaluates the quality of an article based on 3 broad perspectives: the cohort selection (0–4 points), comparability (0–2 points), and assessment of outcomes (0–3 points). A score of ≥7 points was suggestive of a high-quality study. Both reviewers (MK and JSM) undertook the methodological quality screening and any discrepancies were settled by consensus.

3 Results

3.1 Summary of Included Studies

A detailed flow diagram of study selection is shown in Fig. 1. 1226 publications met the search criteria and were evaluated. After removing duplicates, 1089 articles were screened. 978 articles were removed after title and abstract screening. 7 further articles were found through other sources. 117 articles were screened at full text, with 99 articles excluded at this stage; the main reasons being that abstracts had not been available and full text review of the articles showed them to be irrelevant on full text review, and no ICP or surrogate measures were recorded. Overall, eligibility criteria were met by 18 articles (Table 1).

Fig. 1
figure 1

PRISMA flow diagram. From: Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(7): e1000097. doi:10.1371/journal.pmed1000097

Full size image
Table 1 Characteristics of included studies
Full size table

3.2 Risk of Bias in Included Studies

According to the results of the NOS, two out of the eighteen studies scored ≥ 7, indicating high quality studies, eleven studies scored ≥ 4 indicating moderate quality and five studies scored < 4 indicating low quality. Supplementary material Figure 3 provides the risk of bias indicators of the included studies.

3.3 Demographics of Included Studies

Patient numbers ranged from 1 to 385. Nine studies included patients with ICH only [18,19,20, 22,23,24,25,26,27]; the remaining nine studies included other brain injury patients as well as ICH patients. These studies included patients with subarachnoid haemorrhage (SAH), traumatic brain injury, acute ischaemic stroke, hypoxic brain injury, anoxia, encephalitis and arteriovenous malformation [16, 21, 28,29,30,31,32,33,34]. The mean age of patients varied from 43 years to 70 years. Two studies randomized patients to a control or placebo group [16, 24], whilst one study randomized patients to different blood pressure targets [20].

3.4 BP and ICP Changes in Included Studies

The main findings of the included studies are presented in Table 1. Overall, sixteen studies included ICP measures [18, 19, 21,22,23,24,25,26,27,28,29,30,31,32,33,34], eleven studies included ICP and CPP measures [19, 22,23,24, 26,27,28, 30,31,32,33], and two studies included CPP measures only [16, 20]. Dihydropyridine calcium channel blockers (CCBs) were the main agent used to lower BP, with six studies using these agents and including ICP measures [18, 22, 27, 28, 31, 32]. Dihydropyridine CCBs used included nicardipine [18, 22, 27, 31], nimodipine [18], clevidipine [28] and nifedipine [32]; with one study using two CCBs [18]. In two of these studies [22, 32], it was not possible to determine exact numbers, thus a trend in behaviour of BP and ICP was determined from the data available and the studies were not included in Fig. 2. Dihydropyridine CCB administration was associated with BP reduction in all studies. However, ICP response was variable, with three studies showing a reduction with nicardipine administration [18, 22, 31], and one study an increase [27]. ICP also increased with the administration of nifedipine [32] and nimodipine [18], but did not significantly change with clevidipine [28].

Fig. 2
figure 2

Effect of dihydropyridine calcium channel blockers on BP and ICP. BP Blood pressure, ICP intracranial pressure

Full size image

Five studies used other agents to lower BP and included ICP measures [19, 27, 30, 32, 33], including glyceryl trinitrate (GTN) [27] and diltiazem (a non-dihydropyridine CCB) [27]. One study was a case report using labetalol and nitroprusside from which it was not possible to ascertain exact values, instead a trend was determine [19]. Hayashi et al assessed the effects of chlorpromazine, reserpine, furosemide and thiopental on arterial BP, ICP and CPP, from which trends were determined [32]. Suarez et al used 23.4% saline for intracranial hypertension, with only mean arterial pressure (MAP) values reported, so this study was not included in Fig. 3 [33]. One study used paracetamol to lower temperature, but because this was associated with BP reduction it was included [30]. Hirayama et al used nitroglycerin and diltiazem to lower blood pressure [27]. BP reduction was seen in all studies, with the exception of furosemide administration [32]. ICP decreased with administration of labetalol and nitroprusside [19], furosemide [32], thiopental [32] and 23.4% saline [33], but increased with administration of nitroglycerin [27], diltiazem [27], chlorpromazine [32] and reserpine [32]. There was no significant change in ICP with paracetamol administration [30].

Fig. 3
figure 3

Effect of other BP lowering agents on BP and ICP. BP Blood pressure, ICP intracranial pressure

Full size image

3.5 BP and CPP Changes in Included Studies

Ten studies included BP lowering and CPP measures [16, 19, 20, 22, 27, 28, 30,31,32,33]. However, only three studies have been included in Fig. 4 [28, 30, 31] because either it was not possible to determine the exact degree of BP lowering [16, 20], MAP only was reported [33], the exact degree of CPP lowering was not possible to determine [27] or it was not possible to determine exact degree of both BP and CCP lowering [19, 22, 32]. CPP was reduced with administration of labetalol and nitroprusside [19], nicardipine [22, 27, 31], nitroglycerin [27], diltiazem [27], clevidipine [28], paracetamol [30], nifedipine [32], chlorpromazine [32] and reserpine [32]. However, CPP increased significantly with administration of 23.4% saline [33], and increased with GTN, though this was not statistically significant [16]. There was no change in CPP with administration of furosemide [32] or thiopental [32]. There was no significant difference in CPP between the group randomized to a BP target <150 mmHg compared to <180 mmHg when using labetalol, hydralazine and enalapril to lower BP [20].

Fig. 4
figure 4

Effect of all BP lowering agents on BP and CPP. BP Blood pressure, ICP intracranial pressure

Full size image

3.6 Quantitative Analysis

Meta-analysis of data was not deemed to be possible due to the lack of outcome measures. Where data were available, figures were created to demonstrate the effect of BP lowering on BP and ICP or CPP.

4 Discussion

This systematic review highlights the therapeutic variation employed in blood pressure lowering, and demonstrates the complex relationship between BP and ICP in acute ICH.

4.1 Dihydropyridine Calcium Channel Blockers and ICP

Dihydropyridine CCBs were the most common agent used to lower BP. Administration reduced BP significantly in all studies, but produced varying results on ICP.

Hirayama et al included patients with low initial BP, which further decreased with administration of nicardipine, whilst ICP significantly increased [27]. These patients may represent a subset of patients who are more unwell and have a poor outcome. Hayashi et al found that administration of nifedipine decreased arterial BP, but increased ICP [32]. Interestingly, the increase in ICP was more significant in patients who already had a higher initial ICP, and again may represent a more unwell patient subgroup.

There were several differences between the studies themselves, which may have resulted in the differential effect of BP lowering on ICP. Many of the articles did not provide information on when BP lowering medication was given in the time course of acute ICH which may be relevant in any effect on ICP. Furthermore, studies may also have included several other interventions with potential impact on BP and ICP, for instance Nishayama et al reported that glycerin fructose solution was also given to lower ICP, and thus elucidating the relationship between BP lowering and ICP was difficult [22].

4.2 GTN and ICP/CPP

Recently, there has been significant interest in using GTN to lower BP in acute ICH, though trials have shown conflicting results. The RIGHT trial [35] and subgroup analysis of the ENOS trial [36] found that GTN improved functional outcome in stroke. However, the RIGHT-2 trial [37] showed that GTN did not affect functional outcome, moreover, GTN administration resulted in worse functional outcome in ICH patients [38].

In a study investigating the effects of GTN on patients with recent stroke, BP decreased with administration of transdermal GTN, but CPP increased, although this was not significant [16]. Cerebral vasodilators are usually thought to reduce CPP through dilatation of cerebral vasculature. Willmot et al also measured zero-flow pressure, (ZFP), as a measure of cerebral downstream pressure [16]. GTN did not significantly alter ZFP; the authors postulating that venodilatation increased blood flow out of the cranium, maintaining CPP.

In contrast, Hirayama et al found that nitroglycerin reduced BP, increased ICP, and decreased CPP [27]. These two studies included different patient populations: Willmot et al [16] mild-moderate ischaemic and haemorrhagic stroke patients, and Hirayama et al [27] patients undergoing surgical evacuation of the haematoma. In addition, Willmot et al [16] measured CPP non-invasively using measures of middle cerebral artery blood flow, whereas Hirayama et al [27] measured CPP by subtracting mean ICP from mean systolic blood pressure. Therefore, differences in patient population and study methodology may account for any differences in the effect of BP lowering using GTN on ICP.

4.3 BP, ICP, CPP and Clinical Outcomes

Studies included limited information of clinical outcomes following acute ICH. Two studies investigated the difference in characteristics between survivors and non-survivors, and between those improving and deteriorating [23, 26]. Both studies found that increased ICP and decreased CPP were associated with poorer outcome [23, 26]. Several studies noted that raised ICP was associated with poor outcome [23,24,25,26, 34]. Increased BP and ICP variability were also associated with poorer outcome [26, 29]. These studies suggest that not only is it important to control ICP, but also to reduce ICP variability.

4.4 Strengths and Limitations

To our knowledge, this is the first systematic review investigating the effect of BP lowering on ICP in acute ICH, which has included a variety of BP lowering agents. Overall, the review highlights the need for further research.

There are several limitations with the studies included in this systematic review, in particular related to methodological heterogeneity, lack of a control population and small study size. The review was limited to data reported in the selected articles and did not examine individual patient data. Although the majority of the studies included in the review were of moderate quality, some studies were of low methodological quality. Only English language articles were included. Due to the lack of outcome measures, meta-analysis was not possible.

It is also important to note that this review included a specific population of patients with acute ICH. Patients who undergo ICP monitoring are likely to have moderate ICH and less co-morbidities, and this may not reflect the whole ICH population. Furthermore, it is possible that agents administered in the studies did not alter ICP, and the natural history of ICP in ICH was seen instead. However, there were temporal associations between administration of BP-lowering agents and ICP.

4.5 Future Work

Considering the methodological variation and the small patient numbers, further work is needed to develop a better understanding of the effect of BP lowering on ICP in acute ICH. A large prospective observational or randomized intervention study is needed to further elucidate the relationship between BP lowering on ICP in acute ICH. The agent used to lower BP and the timing of administration should be considered carefully. Understanding the relationship between BP lowering and ICP in acute ICH may help to develop more effective management strategies, which is much needed, considering the significant burden of acute ICH.

5 Conclusions

Administration of a CCB reduced BP, but had a varying effect on ICP in patients with ICH, other agents were also used to lower BP, and similarly had a varying effect on ICP. This review found considerable variation in methodology of included studies and small patients numbers, suggesting further work is needed to understand the effect of BP lowering on ICP in acute ICH such as large prospective observational or randomized intervention study.