Blood Pressure Threshold Violations in the First 24 Hours After Admission for Acute Stroke
Frequency, Timing, Predictors, and Impact on Clinical Outcome
Background and Purpose— Admission blood pressure (BP) and significant decreases in BP after acute stroke have been correlated with outcome. Few data are available on the impact of extreme values at any time point within the first 24 hours.
Methods— BP was measured hourly for 24 hours in 325 consecutive patients with acute ischemic stroke. Predefined endpoints were systolic BP ≥200, diastolic BP ≥110, or systolic BP <100 mm Hg during the first 24 hours, and significant systolic BP decreases by >26 mm Hg within 4 hours after admission. Multiple logistic regression analysis identified independent predictors of each end point and determined the impact on dependency at 3 months defined as modified Rankin scale score ≥3.
Results— Upper threshold violations occurred in 70% of cases during the admission process, and more frequently in patients arriving early after stroke; 30% of cases exhibited such values at a later time point. History of hypertension (P<0.01) and higher NIHSS on admission (P<0.05) were independent predictors. Systolic BP <100 mm Hg occurred at random and was associated with younger age (P<0.05). Night time admission was the strongest independent predictor of systolic BP decreases >26 mm Hg (P<0.0001). Diabetes, NIHSS on admission, and age were associated with adverse outcome at 3 months, whereas threshold violations and decreases were not. There was a trend for administration of antihypertensives being associated with poor outcome (P<0.1).
Conclusion— External stimuli, premorbid risk factors, diurnal BP variations, and disease-immanent mechanisms all influence the course of BP after acute stroke. Monitoring should precede any medical BP treatment.
Medical treatment of strongly increased or critically reduced blood pressure (BP) in the acute phase of stroke is much debated.1–4 Most patients with acute stroke have increased BP values in the acute stage.5–11 Extreme hypertension and hypotension on admission have both been correlated with adverse outcome in acute stroke patients.8–10,12–15 However, there is no consensus of the “optimal range” of BP.16,11 In daily practice, BP management is even more complicated as all studies mentioned focused on admission BP to predict outcome. Few data are available on the impact of any extreme BP values documented within the first 24 hours after admission.
Significant decreases of BP during the first hours after admission “should be avoided” according to current guidelines,3,4 because this may compromise cerebral perfusion.17 Castillo et al11 showed that decreases of systolic BP by >20 mm Hg within the first 24 hours after admission are correlated with poor outcome at 3 months. However, Chamorro et al18 reported that moderate BP lowering is associated with favorable outcome. Mattle et al5 showed that BP decreases occurred significantly more frequent after recanalization of the brain arteries in patients treated with thrombolysis, but it is unclear whether this finding can be transferred to the entire stroke population.
We therefore examined the evolution of systolic BP and diastolic BP of stroke patients admitted within 72 hours of symptom onset and investigated which factors were associated with predefined threshold violations during the first 24 hours after admission. Furthermore, we analyzed whether any of the given thresholds was associated with adverse outcome after 3 months.
Materials and Methods
We included consecutive patients admitted to our certified comprehensive stroke unit between December 2004 and December 2005. Only patients with ischemic stroke were included. Patients with transient neurological deficits were included in case diffusion-weighted brain MRI identified a new ischemic lesion consistent with clinical presentation. Patients with sinus thrombosis, primary or secondary intracerebral hemorrhage, and patients with subarachnoid hemorrhage were excluded, as were patients who spent <24 hours on the stroke unit or who had to be intubated within 24 hours. All patients or their proxies gave informed consent for data collection and for a telephone interview after 3 months. Because the study was purely observational, no ethics committee was involved.
Systolic BP ≥200 mm Hg or diastolic BP ≥110 mm Hg within the first 24 hours was defined as upper threshold violation. We also documented the frequency and time of systolic BP ≥220 mm Hg to compare the clinical impact of this particular threshold recommended in the official guidelines of the European Stroke Initiative.4
Systolic BP <100 mm Hg was defined as the lower threshold. In case values decreased below this limit, medical treatment by volume expansion was recommended but left to the discretion of the treating physician.
Because the term “abrupt decrease in blood pressure”4 is ill-defined, we chose the upper quartile of differences between systolic BP on admission and after 4 hours as end points and declared this as “significant” decrease.
BP values were repetitively recorded during the admission process and the first 24 hours on the stroke unit. Patients’ first BP was recorded either by emergency medical personnel or in the emergency department with varying devices and methods. All BP values obtained before the patient arrived on the stroke unit were collected and the highest or lowest value were used for the evaluation of upper or lower threshold violations before admission to the stroke unit. The respective value was termed “preadmission” BP.
The first value recorded on the stroke unit was declared as “admission” BP. Plausibility checks of BP values on the stroke unit were performed by registered nurses instructed about the study. On arrival in the stroke unit, patients were connected to an automated multimodal monitoring system (Siemens SC9000XL; Siemens) using an oscillometric BP measuring method. Measurements were recorded at least every hour or even more frequently if indicated in single cases. In those cases the maximum and minimum values of the 1-hour period were taken for upper or lower threshold comparison. Violations of the values prespecified here were confirmed by a second measurement after 5 minutes.
Age, sex, and other risk factors were recorded. History data were obtained from the patient, relatives, or general practitioners (GPs). History of chronic heart failure (New York Heart Association [NYHA] I or more as known by patient, proxies, or GP), diabetes, and other risk factors was also assessed.
Time from symptom onset was recorded in full hours. In patients with undeclared time of symptom onset, the last time the patient had been seen without symptoms was used. Patients were then subdivided in 4 groups, namely <2 hours from symptom onset, 2 to <6 hours, 6 to <24 hours, and >24 to 72 hours. These times were selected to separate candidates for intravenous thrombolysis and intra-arterial thrombolysis from patients arriving within 24 hours but beyond a time-window for specific treatment. We also included patients who were admitted between 24 and 72 hours after stroke onset to better-define effects on BP associated mainly with the admission procedure and only to a lesser extent with the disease itself.
To check for diurnal modification of BP, patients were grouped in day-time admissions (6:00 am to 5:59 pm) and night-time admission (6:00 pm to 5:59 am).
Stroke severity on admission was determined using the NIHSS.19 Stroke etiology was classified according the TOAST criteria. We noted whether patients received thrombolytic therapy or antihypertensive agents during the study period.
Patients were followed-up by validated standardized telephone interviews 3 months after symptom onset to determine clinical outcome.19 Dependency was defined as modified Rankin scale score ≥3.
Frequency and timing of upper threshold violations (systolic BP ≥200 mm Hg, systolic BP ≥220 mm Hg, and diastolic BP ≥110 mm Hg diastolic) and of lower threshold violations (systolic BP <100 mm Hg) within the first 24 hours were graphically displayed. The total number of patients violating the threshold after 24 hours was set to 100%. We chose “time from admission” as reference time point for threshold violations instead of “time from symptom onset,” because it was impossible to determine whether patients arriving late after symptom onset did not already have violated the threshold before the first BP measurement. However, to account for the more physiological time point of symptom onset, a second diagram displaying threshold violations systolic BP ≥200 mm Hg with separate graphs for each time from symptom onset to admission group was drawn. The scale was set to cumulative percent within the group.
The single factor analysis for the respective endpoints was performed using Fisher exact test for dichotomized items and χ2 test for trend analysis for ordinal items. Mann-Whitney U tests were performed for continuous variables. In a further step, multiple logistic regression analysis was used to determine independent predictors of: (1) systolic BP decreases by >26 mm Hg within 4 hours of admission; (2) systolic BP ≥200 mm Hg at any time within 24 hours of admission; (3) diastolic BP ≥110 mm Hg at any time within 24 hours of admission; and (4) systolic BP <100 mm Hg at any time within 24 hours of admission.
Systolic BP ≥220 mm Hg at any time within 24 hours of admission was not tested in the regression model because single factor analysis yielded no significances.
Multiple logistic regression analysis was also used to determine whether any of these 4 end points predicted adverse outcome at 3 months. For this analysis the respective end points were included as predictors, as well as the parameters NIHSS, age, history of diabetes, history of hypertension, history of chronic heart failure, thrombolytic therapy, antihypertensive therapy, sex, and stroke etiology.
Three hundred twenty-five of 509 patients admitted to our stroke unit during the study period were included in the study. Ninety-seven patients had a different disorder, subarachnoid hemorrhage, sinus thrombosis, or secondary intracerebral hemorrhage. Fifty-two patients left the stroke unit earlier than 24 hours after admission, and 36 patients were mechanically ventilated within 24 hours. Demographics and baseline characteristics are given in Table 1.
Frequency and Timing of BP Threshold Violations
Seventy-seven of 325 patients (24%) had systolic BP ≥200 mm Hg during the observation period. Fifty-eight patients (75%) had systolic BP ≥200 mm Hg within the first 4 hours of admission and 71 (92%) had systolic BP ≥200 mm Hg within 12 hours. Only 6 patients (6%) had systolic BP ≥200 mm Hg >12 hours after admission.
Eighteen of 325 patients (6%) had systolic BP values ≥220 mm Hg, and 74 patients (24%) had diastolic BP values ≥110 mm Hg during the 24-hour study period. Figure 1 illustrates that the time course of all upper threshold violations was very similar.
Figure 2 shows that the pattern of threshold violations systolic BP ≥200 mm Hg was very similar for all subgroups of time from symptom onset. Most threshold violations were evident on the first measurement. However, with shorter times from symptom onset to admission there was an increasing absolute likelihood that the patient would have systolic BP ≥200 mm Hg (29%, 25%, 19%, and 13%; P=0.011).
Thirty-six out of 325 patients (11%) had systolic BP <100 mm Hg during the study. In contrast to the clustering of upper threshold violations within the first hours, the rate of lower BP violations revealed a constant frequency over time (Figure 1).
Predictors of BP Threshold Violations
Table 2 displays in detail data for the predefined end points. Systolic BP ≥200 mm Hg was associated with history of hypertension, thrombolysis, high NIHSS scores, and older age. As said, time from symptom onset to admission was also associated with systolic BP ≥200 mm Hg. “Other determined” and “undeclared” stroke etiology were negatively associated with systolic BP ≥200 mm Hg (Table 2). Factors associated with diastolic BP ≥110 mm Hg were similar to those associated with systolic BP ≥200 mm Hg (Table 2).
After multivariate logistic regression, history of hypertension (OR, 2.2; 95% CI, 1.0–4.7; P<0.05) and NIHSS on admission (OR, 1.06; 95% CI, 1.0–1.13) were independent predictors of systolic BP ≥200 mm Hg. Only history of hypertension (OR, 3.2; 95% CI, 1.5–7.2; P<0.01) was a predictor of diastolic BP ≥110 mm Hg (Table 3). There were no factors associated with systolic BP ≥220 mm Hg.
Low systolic BP values were associated by history of chronic heart failure, “other defined etiology,” and younger age. Administration of antihypertensives was inversely associated with systolic BP <100 mm Hg, because patients receiving antihypertensives did usually continue to have high BP values despite treatment. There was significant interaction between etiology and age, because “other defined etiology” applied in our data set mainly to young patients with, eg, arterial dissections. After multiple logistic regression, only young age remained a significant predictor of systolic BP <100 mm Hg (Table 3).
Significant Systolic BP Decrease
The median decrease of the systolic BP within the first 4 hours was 10 mm Hg (interquartile range, 4 mm Hg increase to 26 mm Hg decrease). A decrease by >26 mm Hg within the first 4 hours after admission was therefore considered to be “significant” (n=81).
History of chronic heart failure, longer time from symptom onset to admission, lower NIHSS scores on admission, absence of thrombolysis, and night-time admission were associated with systolic BP decreases (Table 2). After logistic regression analysis, only night-time admission remained an independent predictor of significant systolic BP decreases (OR, 2.8; 95% CI, 1.6–4.6; P<0.0001; Table 3).
In single factor analysis, a number of parameters were associated with poor outcome (modified Rankin score ≥3). Details are displayed in Table 4.
Administration of antihypertensives was more strongly associated with poor outcome than any BP threshold violation. In multiple regression analysis (Table 5) there was a trend for administration of antihypertensives to be associated with adverse outcome (P<0.1). BP threshold violations were not associated with outcome. Strong predictors of adverse outcome were higher NIHSS on admission (OR, 1.2; 95% CI, 1.13–1.3 per point; P<0.0001) and higher age (OR, 1.04; 95% CI, 1.02–1.07 per year; P<0.0001). Diabetes was the only classical risk factor associated with poor outcome (P<0.05).
Twenty-four percent of patients with acute ischemic stroke experienced systolic BP ≥200 or diastolic BP ≥ 110 mm Hg. Only 6% of patients had extreme systolic BP values ≥220 mm Hg during the study, which is in line with some previous observations.11 However, other authors reported frequencies of 8% for systolic BP ≥200 mm Hg13 and of 0.1% for systolic BP ≥220 mm Hg.7 The difference is best explained by the very short median time from symptom onset to admission in our study and because other studies did not consider a time-interval of 24 hours. We identified an additional 30% of patients with upper threshold violations beyond admission. However, only ≈10% of patients experienced extremely high BP after >12 hours.
Patients were more likely to have high BP in case of more severe strokes and if previously hypertensive. It also occurred more frequently in patients who were admitted soon after the event. This finding could be explained by maintenance of high systolic BP in case of occluded arteries. Mattle et al5 found that BP decreases significantly after successful thrombolysis, whereas it remained elevated in case of persisting vessel occlusion. Yong et al9 concluded similarly from their analysis of the European Cooperative Acute Stroke Study II (ECASS-II) trial data. The lower likelihood of patients arriving very early to experience significant BP decreases also points in this direction. Patients with more severe strokes on admission and those who arrive early are much more likely to have occluded arteries than others and thus may exhibit threshold violations more frequently. Because higher clinical severity is known to be closely correlated with shorter times to admission, only clinical severity was an independent predictor of systolic BP ≥200 in our study.
Systolic BP ≥200 mm Hg was associated with poor outcome in single factor analysis, which is inline with others,9,10,12,15 but not independently. This might be explained because we did not use absolute BP values as predictors but only dichotomous threshold violations. However, this approach was chosen because clinical practice frequently relies on threshold violations to initiate management. Furthermore, data from larger trials such as the Tinzaparin in Acute Ischaemic Stroke Trial (TAIST) are likely to include first BP values well beyond the very acute phase as we did in our study,15 limiting comparison of the data.
Low systolic BP values occurred at a quite constant rate over time in our study and the only independent predictor of systolic BP <100 mm Hg was younger age. In single factor analysis, systolic BP <100 mm Hg was associated with chronic heart failure, an observation that has been made before by Vemmos et al.16 In our study systolic BP <100 mm Hg was not independently associated with poor outcome, probably because systolic BP <100 mm Hg was frequently found in young patients with a good long-term prognosis.
The association of low systolic BP with adverse long-term outcome may only be valid for cases with low systolic BP on admission.16 This subgroup was too small to be tested in our cohort. Low systolic BP in the course of the disease are, however, probably not associated with hemodynamic compromise and instead an indicator of recanalization.5
Significant systolic BP decreases were particularly frequent in patients admitted during the evening and night hours in our study. This new finding is of major interest for the management of BP in the acute phase after stroke because it should increase caution to actively treat BP in patients admitted during night-time. Furthermore, extreme decreases very likely occur as a reaction to the admission process. After admission to the unit, the stress relieves and BP decreases. If this relaxation coincides with the normal circadian decrease of the BP during the night, then extreme decreases can occur.
We did not find that decreases were associated with unfavorable outcome, as Castillo et al11 found in their study. They suggested that their finding could explain the poor outcome of patients treated with antihypertensive drugs in the Intravenous Nimodipine West European Stroke Trial (INWEST).14 We also found a trend in our study that antihypertensive treatment was associated with poor outcome, but we found a trend for favorable outcome in patients with more pronounced decreases. Furthermore, administration of antihypertensive agents was inversely associated with systolic BP <100 mm Hg in our study. Antihypertensive medication does thus not induce a significant decrease of BP on a regular basis and decreases are not all caused by medication. Chamorro18 found that mild BP reduction improves outcome, and this may again be explained because BP only decreases significantly in patients who have recanalized vessels. Outcome may be shifted by treatment to the worse end if BP decreases are induced by treatment in patients who still have compromised brain perfusion.5,9 Antihypertensives therefore should be used only with great caution in the first hours after stroke.
Our study has several limitations, mainly the observational design. Data should be interpreted cautiously, because factors not controlled for in our protocol might influence the results. For example, we did not adjust for the obviously meaningful patency of cerebral vessels. Furthermore, our patients may not be representative of all stroke patients. For example, 18% of our patients received thrombolysis, a very high proportion compared with other stroke services. Our cut-off values represented the local policy of our hospital, which are different from official guidelines.3,4
Some conclusions can be drawn from our data nevertheless. First, BP evolution after stroke is dependent on disease-specific mechanisms such as vessel status, as well as on medical history and the patient’s individual stress reaction. Second, patients without significant BP elevations during the first 12 hours on the stroke unit will only rarely experience upper threshold violations thereafter. This may be of major importance for patient allocation to a monitoring unit or to a regular ward in case of shortage of resources and triage requirements. Low systolic BP is much less predictable. It is probably not indicated to treat systolic BP even above suggested thresholds in the first 4 to 6 hours after admission unless conditions such as thrombolysis or cardiac failure demand BP lowering. Significant decreases of BP are a frequent finding, especially in patients admitted during night-time.
Follow-up data were provided by the “Qualitätssicherungsprojekt Schlaganfall Nordwestdeutschland” (Quality control project “Stroke” North-Western Germany).
- Received April 1, 2008.
- Revision received May 21, 2008.
- Accepted June 17, 2008.
Adams HP, Jr, Adams RJ, Brott T, del Zoppo GJ, Furlan A, Goldstein LB, Grubb RL, Higashida R, Kidwell C, Kwiatkowski TG, Marler JR, Hademenos GJ, Stroke Council of the American Stroke Association. Guidelines for the early management of patients with ischemic stroke: A scientific statement from the Stroke Council of the American Stroke Association. Stroke. 2003; 34: 1056–1083.
Hack W, Kaste M, Bogousslavsky J, Brainin M, Chamorro A, Lees K, Leys D, Kwiecinski H, Toni P, Langhorne P, Diener C, Hennerici M, Ferro J, Sivenius J, Gunnar N, Bath P, Olsen TS, Gugging M, European Stroke Initiative Executive Committee and the EUSI Writing Committee. European Stroke Initiative Recommendations for Stroke Management-update 2003. Cerebrovasc Dis. 2003; 16: 311–337.
Mattle HP, Kappeler L, Arnold M, Fischer U, Nedeltchev K, Remonda L, Jakob SM, Schroth G. Blood pressure and vessel recanalization in the first hours after ischemic stroke. Stroke. 2005; 36: 264–268.
Carlberg B, Asplund K, Hagg E. Factors influencing admission blood pressure levels in patients with acute stroke. Stroke. 1991; 22: 527–530.
Stead LG, Gilmore RM, Decker WW, Weaver AL, Brown RD Jr. Initial emergency department blood pressure as predictor of survival after acute ischemic stroke. Neurology. 2005; 65: 1179–1183.
Yong M, Diener HC, Kaste M, Mau J. Characteristics of blood pressure profiles as predictors of long-term outcome after acute ischemic stroke. Stroke. 2005; 36: 2619–2625.
Castillo J, Leira R, Garcia MM, Serena J, Blanco M, Davalos A. Blood pressure decrease during the acute phase of ischemic stroke is associated with brain injury and poor stroke outcome. Stroke. 2004; 35: 520–526.
Leonardi-Bee J, Bath PM, Phillips SJ, Sandercock PA, IST Collaborative Group. Blood pressure and clinical outcomes in the International Stroke Trial. Stroke. 2002; 33: 1315–1320.
Ahmed N, Nasman P, Wahlgren NG. Effect of intravenous nimodipine on blood pressure and outcome after acute stroke. Stroke. 2000; 31: 1250–1255.
Sprigg N, Gray LJ, Bath PM, Boysen G, De Deyn PP, Friis P, Leys D, Marttila R, Olsson JE, O'Neill D, Ringelstein B, van der Sande JJ, Lindenstrom E, TAIST Investigators. Relationship between outcome and baseline blood pressure and other haemodynamic measures in acute ischaemic stroke: data from the TAIST trial. J Hypertens. 2006; 24: 1413–1417.
Eames PJ, Blake MJ, Dawson SL, Panerai RB, Potter JF. Dynamic cerebral autoregulation and beat to beat blood pressure control are impaired in acute ischaemic stroke. J Neurol Neurosurg Psychiatry. 2002; 72: 467–472.
Chamorro A, Vila N, Ascaso C, Elices E, Schonewille W, Blanc R. Blood pressure and functional recovery in acute ischemic stroke. Stroke. 1998; 29: 1850–1853.