Journal of Neurology

, Volume 253, Issue 11, pp 1478–1483

Use of calcium channel blockers after stroke is not associated with poor outcome

A cohort from the registry of the Canadian stroke network

Authors

    • Division of NeurologyOttawa Hospital
  • Jiming Fang
    • Institute for Clinical Evaluative Sciences
  • Marc Kawaja
    • Dept. of Clinical EpidemiologyMemorial University of Newfoundland
  • Antoine Hakim
    • Ottawa Health Research InstituteUniversity of Ottawa, Canadian Stroke Network
ORIGINAL COMMUNICATION

DOI: 10.1007/s00415-006-0249-1

Cite this article as:
Dowlatshahi, D., Fang, J., Kawaja, M. et al. J Neurol (2006) 253: 1478. doi:10.1007/s00415-006-0249-1

Abstract

Background

Control of hypertension is essential for the secondary prevention of stroke. Although several trials have assessed the role of calcium channel blockers (CCBs) in the acute stroke, few address their safety and efficacy in secondary prevention. The recovery process after stroke requires neurite outgrowth, which may be dependent on activation of calcium channels and NMDA receptors. We asked whether treatment of hypertension using CCBs is safe during the recovery of patients following stroke and whether it contributes to their functional outcome.

Methods

The Registry of the Canadian Stroke Network provided access to information from 1545 patients with ischemic stroke. Primary outcome variables were mortality and functional outcome, which was assessed using the Stroke Impact Scale-16.

Results

Patients discharged on CCB had a 2.5% 6-month mortality rate compared with 5.5% in those who were not on CCB at discharge (OR 0.38, 95% CI 0.17−0.88). There was no change in 6-month mortality with respect to treatment with ACE-I, B-blockers or diuretics at discharge. Patients that were admitted on CCB had improved SIS-16 at 6-months if they were also discharged on CCB, as compared with patients who had their CCB discontinued (73.8 veas 66.8, p = 0.032).

Conclusions

CCB treatment at the time of discharge did not impede functional recovery, and was associated with reduced mortality and improved SIS-16 at 6 months.

Keywords

strokecalcium channel blockersmortalityrecoverytreatment

Background

Outside the acute period, control of hypertension is essential for the prevention of stroke [4, 15, 16, 18, 19]. Calcium channel blockers (CCB), along with angiotensin converting enzyme inhibitors (ACE-I), angiotensin receptor blockers (ARB), beta-blockers and diuretics, are frequently prescribed in the inpatient setting as well as at discharge. Multiple trials have assessed treatment of hypertension in stroke patients with ACE-Is, ARBs and diuretics [10]. Although several trials have assessed the role of CCBs in the acute stroke setting [1, 13], few addressed their safety and efficacy in secondary prevention.

It is widely accepted that NMDA-mediated calcium influx contributes to cell death following cerebrovascular ischemia. Despite the experimental evidence, several clinical trials attempting to antagonize these receptors, or otherwise interfere with calcium conduction, were unsuccessful [5, 6]. This discrepancy may be due to the non-specific calcium blockade of these agents particularly since NMDA receptors and calcium channels have differential effects on cell survival, based on their regional expression on neurons. For example, synaptic NMDA receptors appear to play a neuroprotective role, whereas extra synaptic NMDA receptor stimulation contributes to cell death [11].

More recent data suggest that neurite outgrowth requires activation of the calcium channels [3]. Since the recovery process after stroke requires neurite outgrowth, blocking calcium entry into the cell may impede recovery after stroke. By extrapolating from the clinical and animal literature, we asked whether generalized non-specific calcium blockade might be detrimental to the recovery of patients with stroke [12].

Despite the advancing knowledge of calcium signalling in neurite outgrowth and recovery, little is known about the clinical effect of calcium antagonism during the recovery period following stroke. Given the widespread use of CCBs in this population, we proposed a cohort study to assess the effect of CCB treatment following stroke on functional recovery and mortality. Specifically, based on recent experimental studies we hypothesised that non-selective calcium channel blockade may impede recovery after stroke.

Methods

The Registry of the Canadian Stroke Network (RCSN) was established by the Canadian Stroke Network, which is funded by the Canadian Networks of Centres of Excellence. The registry has been described in detail elsewhere [23]. Briefly, the RCSN is partnered with the Institute for Clinical Evaluative Sciences (ICES) and is supported in Ontario by the Ministry of Health and Long-term Care. Collaborators from 21 Canadian hospitals in 8 provinces collected clinical stroke data characterizing the onset of symptoms, and were followed until six months post-stroke. Potential sites were selected based on applications from stroke neurologists and anticipated volumes, with attempts made to ensure representation from most Canadian provinces. Phase one of the RCSN consisted of 21 tertiary care institutions with specific stroke care resources representing 20% of all admitted stroke patients in the country between July 2001 and February 2002. Phase two of the RCSN collected patient information between June 2002 and December 2002, and included the original 21 institutions, as well as 4 large community hospitals. Approval was obtained from the research ethics board at each participating institution, thereby conforming to the ethical standards laid down in the 1964 Declaration of Helsinki.

The RCSN provided access to clinical information from 1545 patients with acute ischemic stroke; no patients with hemorrhagic stroke were included in the study. All patients provided informed consent and data included demographics, medical history, stroke severity, and pre-hospital, emergency and in-hospital medications, as well as 6-month mortality and functional status (table 1). We did not have access to rates of recurrent strokes in this sample. Data were entered electronically, and the aggregate database was managed at ICES in Toronto, Ontario.
Table 1

Patient demographics by group

Group

 

All patients

A

B

C

D

 

CCB on admission

  

Yes

Yes

No

No

 

CCB on discharge

  

Yes

No

Yes

No

 

Sample size

 

1545

236

106

78

1125

 
       

P

Gender - n(%)

Female

681 (44.1%)

102 (43.2%)

46 (43.4%)

40 (51.3%)

493 (43.8%)

0.6231

Male

864 (55.9%)

134 (56.8%)

60 (56.6%)

38 (48.7%)

632 (56.2%)

 

Age (year) - (n)

Mean

69.9 (1545)

72.5 (236)

73.8 (106)

71.3 (78)

68.9 (1125)

0.0001

Median

72 (1545)

73 (236)

76 (106)

74 (78)

71 (1125)

 

CNS at Admission - (n)

Mean

8.7 (1253)

8.9 (197)

8.2 (87)

7.9 (61)

8.7 (908)

0.2268

Median

9 (1253)

9.5 (197)

9 (87)

8.5 (61)

9 (908)

 

Level of Consciousness - n(%)

Alert

1352 (88%)

219 (93.6%)

90 (86.5%)

65 (85.5%)

978 (87.2%)

0.1134

Confused

48 (3.1%)

≤5 (1.3%)

≤5 (3.8%)

≤5 (2.6%)

39 (3.5%)

 

Drowsy

52 (3.4%)

9 (3.8%)

≤5 (2.9%)

≤5 (5.3%)

36 (3.2%)

 

Unconscious

84 (5.5%)

≤5 (1.3%)

7 (6.7%)

≤5 (6.6%)

69 (6.1%)

 

Rankin Score at Discharge - (n)

Mean

2.2 (816)

2.4 (143)

2.4 (56)

2.4 (47)

2.1 (570)

0.1242

Median

2 (816)

2 (143)

3 (56)

2 (47)

2 (570)

 

Rankin Score at Discharge - n(%)

0

146 (17.9%)

18 (12.6%)

7 (12.5%)

8 (17%)

113 (19.8%)

0.1552

1

162 (19.9%)

22 (15.4%)

13 (23.2%)

9 (19.1%)

118 (20.7%)

 

2

145 (17.8%)

33 (23.1%)

≤5 (8.9%)

7 (14.9%)

100 (17.5%)

 

3

141 (17.3%)

28 (19.6%)

12 (21.4%)

6 (12.8%)

95 (16.7%)

 

4

175 (21.4%)

37 (25.9%)

17 (30.4%)

13 (27.7%)

108 (18.9%)

 

5

43 (5.3%)

≤5 (3.5%)

≤5 (3.6%)

≤5 (6.4%)

33 (5.8%)

 

6

≤5 (0.5%)

0 (0%)

0 (0%)

≤5 (2.1%)

≤5 (0.5%)

 

TPA given - n(%)

 

162 (10.6%)

18 (7.7%)

15 (14.3%)

10 (12.8%)

119 (10.6%)

0.2663

Medical History:

Smoking - n(%)

 

735 (47.8%)

109 (46.4%)

44 (41.9%)

34 (43.6%)

548 (48.9%)

0.1335

Diabetes - n(%)

 

182 (29%)

38 (31.1%)

11 (25%)

13 (38.2%)

120 (28%)

0.8501

Hypertension - n(%)

 

1016 (66%)

211 (89.4%)

91 (86.7%)

61 (79.2%)

653 (58.3%)

0.0001

Hyperlipidemia - n(%)

 

552 (35.8%)

105 (44.5%)

40 (37.7%)

26 (33.3%)

381 (33.9%)

0.0207

TIA - n(%)

 

313 (20.3%)

53 (22.6%)

24 (22.6%)

20 (25.6%)

216 (19.2%)

0.3532

Stroke - n(%)

 

370 (24%)

69 (29.4%)

24 (22.6%)

17 (21.8%)

260 (23.2%)

0.2157

Atrial Fibrillations - n(%)

 

234 (15.2%)

35 (14.9%)

23 (21.7%)

11 (14.1%)

165 (14.7%)

0.3533

CHF - n(%)

 

94 (6.1%)

19 (8.1%)

6 (5.7%)

≤5 (6.4%)

64 (5.7%)

0.6851

Previous MI - n(%)

 

259 (16.8%)

55 (23.5%)

26 (24.5%)

11 (14.1%)

167 (14.9%)

0.0109

Valvular - n(%)

 

55 (3.6%)

10 (4.3%)

≤5 (3.8%)

≤5 (2.6%)

39 (3.5%)

0.1652

Medications on Admission:

ACE Inhibitor- n(%)

 

460 (29.8%)

79 (33.5%)

40 (37.7%)

29 (37.2%)

312 (27.7%)

0.0287

Diuretic - n(%)

 

326 (21.1%)

58 (24.6%)

30 (28.3%)

23 (29.5%)

215 (19.1%)

0.0118

Beta Blocker - n(%)

 

440 (28.5%)

83 (35.2%)

34 (32.1%)

24 (30.8%)

299 (26.6%)

0.0449

ASA - n(%)

 

338 (21.9%)

64 (27.1%)

22 (20.8%)

13 (16.7%)

239 (21.2%)

0.1464

Clopidogrel - n(%)

 

0 (0%)

0 (0)

0 (0)

0 (0)

0 (0)

 

Aggrenox - n(%)

 

0 (0%)

0 (0)

0 (0)

0 (0)

0 (0)

 

Statin - n(%)

 

39 (2.5%)

7 (3%)

≤5 (1.9%)

≤5 (2.6%)

28 (2.5%)

0.9467

CCB = clacium channel blocker, CNS = Canadian Neurological Score, TIA =transient ischemic attack, CHF = congestive heart failure

Functional recovery at 6 months was assessed using the Stroke Impact Scale-16 (SIS-16), which is a psychometric assessment tool for physical function in patients with stroke. The questionnaire assesses 16 categories of physical function including 7 activities of daily living & instrumental activities of daily living items, 8 mobility items, and 1 hand function item, providing a total score ranging from 0 to 100 (highest function). Due to a lower ceiling effect, SIS-16 can differentiate more subtle levels of disability as compared to the Barthel Index [6] and has been validated in recent literature [7, 8, 14].

Primary outcome variables were 6-month mortality and SIS-16 scores. A research nurse conducted follow-up telephone interviews with patients or their proxies. Out of 1545 patients initially included in the study, 35 refused follow-up, and 184 were lost to follow-up (figure 1). Mortality data were available for 1510 patients, whereas SIS-16 scores at 6 months were available for 1128 patients. The cohort was divided into 4 groups based on CCB status on admission and at discharge (table 1). For measuring differences between the groups, one-way ANOVA was used in contiguous variables and χ2 analysis was used for categorical variables. General linear models were used to assess the effect of CCB use on SIS-16; these models included all variables listed in table 1 as covariates. Given the sample sizes, the power to detect an SIS-16 difference of 5 at α=0.05 was greater than 0.8. χ2 analysis was also used to assess the effect of CCB status on 6-month mortality. Logistic regression was used to determine the odds of CCB treatment at discharge with respect to 6-month mortality, and was adjusted for age, stroke severity as well as all variables listed in table 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs00415-006-0249-1/MediaObjects/415_2006_249_f1.jpg
Fig. 1

Final sample sizes used for mortality and SIS-16 analyses

Results

Table 1 shows the cohort of patients divided by their CCB status on admission and discharge. There were significant differences between the groups in age, history of hypertension, hyperlipidemia, MI, and antihypertensive medications at admission. All subsequent analyses added these variables as covariates.

Multivariate analyses showed that patients discharged on CCB had a lower 6-month mortality rate compared to those who were not on CCB at discharge (Odds Ratio {OR} 0.38, 95% Confidence Interval {CI} 0.17–0.88); patients discharged on CCB had a mortality rate of 2.5% whereas those not on CCB at discharge had a mortality rate of 5.5% (table 2). This effect appears independent of CCB status on admission (table 3); patients admitted on CCB had a 5.7% lower 6-month mortality if discharged on CCB, and patients not on CCB at admission had a 5.2% lower 6-month mortality rate if discharged on CCB. There was no statistically significant association between 6-month mortality and treatment with ACE-I, B-blockers or diuretics at discharge.
Table 2

6-month mortality of patients discharged with CCBs compared to those not on CCB at discharge

Group

 

All patients

“On”

“Off”

  

CCB on discharge

  

Yes

No

  

Sample size

 

1510

311

1199

  
     

p

OR (95% CI)

Lost to follow-up

 

184

36

148

  

Patients with follow-up

 

1326

275

1051

  

6-month follow-up - n(%)

Alive

1261 (95.1%)

268 (97.5%)

993 (94.5%)

  

Dead

65 (4.9%)

7 (2.5%)

58 (5.5%)

0.0421

0.38 (0.17-0.88)

Table 3

6-month mortality by CCB status at admission and at discharge

Group

 

All patients

A

B

C

D

 

CCB on admission

  

Yes

Yes

No

No

 

CCB on discharge

  

Yes

No

Yes

No

 

Sample size

 

1510

234

104

77

1095

 
       

p

Lost to follow-up

 

184

27

15

12

133

 

Patients with follow-up

 

1326

207

89

65

962

 

6-month follow-up- n(%)

Alive

1261 (95.1%)

203 (96.7%)

81 (91%)

65 (100%)

912 (94.8%)

 

Dead

65 (4.9%)

7 (3.3%)

8 (9%)

0 (0%)

50 (5.2%)

0.0497

In all patients combined, general linear models revealed CCB status at discharge did not have a significant effect on SIS-16 scores at 6 months. However, in patients who were admitted on CCB, there was an increase in SIS-16 score if they were also discharged on CCB, as compared with patients who had their CCB discontinued (mean 73.8 vs 66.8, p = 0.032; figure 2). In this same group, discharge with diuretics was also associated with a small but statistically significant increase in SIS-16 (71.02 to 75.39, p = 0.015). Furthermore, discharge on B-blockers in this same group was associated with a statistically significant decrease in SIS-16, although the magnitude was clinically insignificant (1.7 difference, p = 0.016).
https://static-content.springer.com/image/art%3A10.1007%2Fs00415-006-0249-1/MediaObjects/415_2006_249_f2.jpg
Fig. 2

Patients admitted on CCB had a significantly higher SIS-16 score if also discharged on CCB (top panel). Conversely, CCB treatment at the time of discharge was not associated with a significant difference in SIS-16 if patients were not on CCB at the time of admission (bottom panel)

Patients who were not on CCB on admission did not show statistically significant effect of CCB treatment at discharge on SIS-16 (figure 2). There was no effect of ACE-I, B-blockers or diuretics at discharge on SIS-16 in this group.

Discussion

Calcium channel blockers bind to “slow” L-type voltage gated channels found predominately in cardiac and smooth muscle tissue. They are lipophilic and readily cross the blood brain barrier. Early laboratory experiments suggested a role for L-type calcium channel antagonism in reduction of focal cerebral ischemia [17, 24] and improved recovery [20]. However, clinical trials failed to show improved outcome in stroke patients receiving CCBs [21, 22].

This study shows a clinically and statistically significant decreased 6-month mortality rate in non-hemorrhagic stroke patients discharged on CCBs. This mortality difference appears independent of prior CCB use, as determined by medication history at the time of admission.

Contrary to our hypothesis, in the subset of patients that were already using CCBs prior to their stroke, continuation of CCB use at discharge was associated with an improved 6-month functional recovery as measured by a 7-point improvement on the SIS-16. This increase of SIS-16 from 66.8 to 73.8 is equivalent to a Modified Rankin Score improvement from 3 (moderate disability) to 2 (slight disability) [7].

In the CCB-naïve group, treatment with CCB at discharge resulted in a non-significant trend towards worsened functional recovery at 6-months (7 point difference on SIS-16; figure 2). Although our sample was powered at 0.8 to detect a difference of 5 points on the SIS-16, this difference warrants re-evaluation using a larger sample size.

Despite these findings, it is difficult to attribute mortality and recovery benefits to CCBs. The increased functional recovery seen in patients already taking CCBs at admission may reflect more rigorous medical care prior to their stroke. Treatment with this class of drug may also reflect more aggressive hypertension control, as suggested by the similar albeit smaller improvement seen in patients discharged with diuretics. Better blood pressure control may have led to a lower rate of recurrent stroke, which we were unable to measure in this study. This can be further elucidated using the patient sample from the upcoming Phase III of the Registry of the Canadian Stroke Network, where information regarding to recurrent vascular events following discharge was collected. Interestingly, a marginally worse recovery was seen in patients discharged on B-blockers, an effect that may be due to comorbid cardiac disease, which further complicates the interpretation of these data.

Nevertheless, this study provides observational evidence that treatment with CCBs following non-hemorrhagic stroke does not impede the process of stroke recovery, and may in fact help it. Benefits obtained from treatment with CCBs in terms of hypertension control may outweigh potential risks. This is supported by a recent meta-analysis of patients with hypertension showing a superiority of CCBs for primary stroke prevention relative to other agents [2].

There are several important limitations to this study. First, the RCSN has an inherent selection bias: patients able to consent to the registry had lower overall mortality rates, and may not be representative of a general stroke population [23]. Furthermore, 12% of this consenting cohort was lost to follow-up. Second, medication compliance information was not available for the 6-month period following discharge. This study, therefore, assumes that patients were taking these medications throughout the follow-up period, which was not verified. Third, blood pressure information was not available for the 6–month period following discharge, thereby obscuring the role of aggressive hypertension control on observed mortality and recovery differences. Fourth, medications were assessed by class, which may mask any drug-specific effects. This is particularly important with CCBs, which are commonly divided into central and peripherally acting categories. Finally, the cohort design introduces the possibility of unknown confounding variables, including unknown or pleiotropic effects of other antihypertensive medications, as well as ecological bias due to treatment differences between source institutions. These issues could be addressed using a prospective design in a randomized control trial.

In summary, CCB treatment at the time of discharge did not impede functional recovery in this cohort. The finding that CCB treatment was associated with reduced mortality and improved SIS-16 at 6 months should be interpreted with caution; our cohort design is susceptible to a variety of confounders and a controlled experimental study is needed to properly test this hypothesis.

Acknowledgments & Funding

We thank the Canadian Stroke Network, Registry of the Canadian Stroke Network, Institute for Clinical Evaluative Sciences, Kevin Willis and Moira Kapral for access to the Registry cohort, as well as assistance with the design of the study and review of the manuscript.

Copyright information

© Steinkopff Verlag Darmstadt 2006