Prevalence of Triggering Factors in Acute Stroke - Journal of Stroke ...

2 downloads 0 Views 136KB Size Report
Manjari Tripathi, DM,* Rohit Bhatia, DM,* Mamta Bhusan Singh, DM,* and Anupriya ... Address correspondence to Ashish Sharma, DM, Department of. Dentistry ...
Prevalence of Triggering Factors in Acute Stroke: Hospital-based Observational Cross-sectional Study Ashish Sharma, DM,* Kameshwar Prasad, DM,* M. V. Padma, DM,* Manjari Tripathi, DM,* Rohit Bhatia, DM,* Mamta Bhusan Singh, DM,* and Anupriya Sharma, MDS†

Background: Although chronic risk factors for stroke are reasonably well understood, the acute precipitants, or triggers, of stroke relatively remain understudied. Identification of particular time periods during which stroke risk is elevated could prove a valuable strategy to reduce stroke incidence through the introduction of appropriate prevention strategies during a period of vulnerability. The aim of this study was to determine the prevalence of trigger factors in acute stroke patients and to investigate the association of the presence of trigger factors with initial stroke severity at presentation (National Institutes of Health Stroke Scale (NIHSS) score in ischemic stroke patients and volume of hematoma in hemorrhagic stroke patients). Methods: This was a hospital-based observational cross-sectional study. All consecutive patients of recent stroke (reporting within 1 week of stroke onset) were included in the study. This study examined the prevalence of 11 predefined triggers (including both well-established and potential triggers) in predefined hazard periods. Results: In total, 290 patients participated in the study. Presence of any trigger factor out of 11 trigger factors studied was seen in 128 (44.2%) of 290 patients, 104 (46.4%) of 224 ischemic stroke patients and 24 (36.4%) of 66 hemorrhagic stroke patients. Psychological stress was present in 51 (17.6%) patients, among psychological stress: stressful life event in 34 (11.7%), negative affect in 17 (5.9%), acute alcohol abuse in 31 (10.7%), clinical infections in 24 (8.3%), and anger and coffee intake in 12 (4.1%) each. Sexual activity, trauma, and surgery were present in 5 (1.7%), 4 (1.4%), and 5 (1.7%) patients, respectively. None of the patients reported exposure to recreational drug abuse, startling event, and unusual vigorous physical exertion in hazard periods. Two or more trigger factors were present in 16 (5.5%) patients. Clinical variables independently associated with the presence of trigger factors in acute stroke after multivariate analysis were younger age (,60 years) and stroke severity at initial presentation (ie, higher NIHSS score and higher hematoma volume). Conclusions: Trigger factors were present in 44.2% of acute stroke patients. Psychological stress (17.6%), acute alcohol abuse (10.7%), and clinical infections (8.3%) were the most common triggers. Younger age (,60 years) and stroke severity at initial presentation were independently associated with the presence of trigger factors in acute stroke patients. However, these associations need to be further explored in community-based studies. Key Words: Prevalence—triggering factors—acute stroke—hospital based. Ó 2015 by National Stroke Association

From the *Department of Neurology, All India Institute of Medical Sciences, New Delhi; and †Department of Dentistry, Dr RPGMC, Kangra, Himachal Pradesh, India. Received July 19, 2014; revision received August 20, 2014; accepted August 27, 2014. Financial disclosure/conflict of interest: None declared.

Address correspondence to Ashish Sharma, DM, Department of Dentistry, G-38, Type 4, Dr RPGMC, Kangra, HP 176001, India. E-mail: [email protected]. 1052-3057/$ - see front matter Ó 2015 by National Stroke Association http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2014.08.033

Journal of Stroke and Cerebrovascular Diseases, Vol. 24, No. 2 (February), 2015: pp 337-347

337

A. SHARMA ET AL.

338

Introduction Although chronic risk factors for stroke are reasonably well understood, the acute precipitants, or triggers, of stroke relatively remain understudied. Vascular events could be precipitated by acute factors, called triggers, which increase the risk of the disease over a relatively short period of time and may directly lead to its onset.1 Some triggers may exert a single, sharp, and short transient effect on the pathophysiological process, whereas others may exert more varied and pervasive effects, probably amplifying risk at multiple points and over a longer period.2 Thus, the period of time associated with an increased risk, called hazard period, starts more or less quickly after trigger initiation, and its duration may vary according to the type of trigger.3 With all that is, known about stroke epidemiology, however, it remains extremely difficult, if not impossible, to predict when a stroke will occur, even among those with a heavy burden of risk factors. Our knowledge of stroke precipitants, or triggers, to be contrasted with risk factors remains relatively primitive. Nonetheless, recent evidence suggests that predicting not just who is at risk of stroke, but when stroke is most likely to occur, may be increasingly possible.4 Why does a given patient, perhaps one with a history of hypertension and diabetes mellitus for decades, have a stroke today? Is this shortterm risk of stroke a predictable or stochastic event? Can the stroke-prone state be measured in some way? If this state can be measured, and is therefore potentially predictable, is there something that can be done to prevent the consequent stroke? There are several potential triggers for stroke. However, to date, research has been mainly focused on acute alcohol abuse and clinical infection.5 Anger, negative or positive emotions, birthday, and psychological distress were significantly associated with stroke onset; however, these associations were identified in single studies. More studies are needed on factors such as physical exertion, acute stress, sexual activity, or anger. Identification of a short-term state of elevated stroke risk could have several therapeutic implications. Patients with a history of stroke, or with a significant burden of risk factors for stroke, might be targeted for more intensive stroke prevention therapy during the period of increased risk.

Aims and Objectives Objectives The objectives of this study were (1) to determine the prevalence of triggering factors in acute stroke patients and (2) to investigate the association of the presence of triggering factors in acute stroke patients with initial stroke severity at presentation (National Institutes of Health Stroke Scale [NIHSS] score in Ischemic stroke

patients and volume of hematoma in hemorrhagic stroke patients).

Hypothesis The acute stroke patients with the presence of triggering factors have more severe stroke compared with acute stroke patients without the presence of triggering factors.

Methodology This was a hospital-based observational cross-sectional study conducted in the Department of Neurology, AIIMS, New Delhi, India, over a period of 15 months from March 2012 to May 2013. All consecutive patients of recent stroke (reporting within 1 week of stroke onset) attending neurology OPD or admitted in neurology ward were considered for inclusion in the study. The study was approved by the ethics committee of the institute. In all, 290 patients were included in the study (fulfilling inclusion and exclusion criteria).

Inclusion Criteria All adult patients ($18 years) with recent stroke (reporting within 1 week of stroke onset): both ischemic and hemorrhagic.

Exclusion Criteria (1) Patients not consenting for study, (2) patients with dementia (Mini-Mental State Examination # 26) and aphasia, (3) patients not able to identify time of onset of stroke symptoms, (4) too ill to complete interview, (5) poor memory around time of stroke, (6) comatose/stuperpose/intubated patients, and (7) patients with other forms of stroke (venous thrombosis and aneurysmal bleed). The following demographic and clinical variables were recorded in a predesigned Performa: age, gender, date and time of stroke onset, risk factor profile (presence or absence of hypertension/diabetes mellitus/dyslipidemia/smoking/heart disease/alcohol intake), type of stroke, ischemic stroke (arterial territory, NIHSS score6 at initial presentation, etiology as per Trial of Org 10172 in acute stroke treatment classification7), and hemorrhagic stroke (site of bleed and hematoma volume). Date of onset was classified into one of the 4 seasons: winter season (December, January, and February), premonsoon season/summer season (March, April, and May), monsoon/summer monsoon (June, July, and August), and postmonsoon season (September, October, and November). The classification was in accordance with the Indian Meteorological Department guidelines.8

TRIGGERING FACTORS IN ACUTE STROKE

The time of onset was distributed into four 6-hour intervals: 00:00-05:59, 06:00-11:59, 12:00-17:59, and 18:0023:59 hours. In hemorrhagic stroke, to calculate hematoma volume, the formula ABC/2 was used,9 where A is the greatest hemorrhage diameter by computed tomography, B is the diameter 90 to A, and C is the approximate number of computed tomographic slices with hemorrhage multiplied by the slice thickness. All measurements for A and B were made using the centimeter scale on the computed tomographic scan to the nearest .5 cm. A, B, and C were then multiplied and the product was divided by 2, which yielded the volume of hemorrhage in cubic centimeters. This study examined the prevalence of 11 trigger factors (including both well-established and potential triggers) in predefined hazard periods. For all trigger factors, patients were asked about usual frequency of exposure and the presence of exposure in the ‘‘hazard period.’’ The hazard periods were predefined according to the estimated duration of effect of each potential trigger factor. Patients were interviewed within 1 week after the event using a questionnaire especially designed for this purpose. The questionnaire included details on possible exposure to triggers and included recent psychological stress, anger, acute alcohol abuse, clinical infection, coffee consumption, recreational drug abuse, sexual intercourse, sudden posture change because of startling event, trauma, surgery, and unusual vigorous physical exertion. Patients’ cognitive state was assessed using the Mini Mental State test. No surrogates were used for information about potential triggers, taking into account subjective perceptions of potential triggers; data were collected from the patients.

List of Triggers with Definitions and Hazard Period Duration

339

such events, using a 5-point scale with ‘‘0’’ indicating none, ‘‘2’’ indicating moderately upset, and ‘‘4’’ indicating ‘‘upset very much.’’ An elevated stressful life events score was defined as the presence of at least 1 stressful experience within the preceding month subjectively perceived as greater than moderately upsetting (score $3 points).12 Anger Anger was assessed using the onset anger scale, a 7level self-report anger scale.13 Patients were defined as exposed to anger if they reported a peak level of anger of 5 and more (‘‘very angry,’’ ‘‘furious,’’ or ‘‘enraged’’) in the 7-level onset anger scale during 2 hours preceding the onset of stroke symptoms. Acute Alcohol Abuse Acute alcohol abuse was defined as alcohol intake of more than 40 g (equivalent to 4 standard drinks) within the 24 hours preceding stroke or more than 150 g (equivalent to 15 standard drinks) within the week preceding stroke. Clinical Infection The diagnosis of clinical infections was based on the presence of fever alone, typical symptoms alone, or fever with typical symptoms. Subjects were classified as having antecedent infection present within 7 days, beyond 7 days but within the prior month, or no infection within 30 days. Coffee Consumption Exposure to coffee consumption as triggering risk factor for stroke was defined as consumption of coffee during 2 hours preceding the onset of stroke symptoms in infrequent drinker (,1 cup/d).

Recent Psychological Stress Recent psychological stress was measured with scales of stressful life events (within the prior 1 month) and negative affect (within the prior 1 week). The negativeaffect scale, modified from the list of negative emotions of Zevon and Tellegen,10 includes 5 categories: (1) nervous, distressed, scared; (2) sad, depressed; (3) angry, upset, irritated; (4) angry at self, guilty, dissatisfied with self; and (5) calm, content (scored negatively). Subjects were instructed to indicate the intensity for each emotion category during the preceding week using a 5-point scale ranging from 0 (none) to 4 (very much).11 An elevated negative-affect score of $10 points was defined as exposed to psychological stress. The questionnaire regarding stressful life events included 5 descriptions of potentially major stressful experiences, and subjects were asked to answer whether they had such experiences within the past month. Those answering ‘‘yes’’ were instructed to indicate the intensity they had been upset by

Recreational Drug Abuse (Prior 1 Week) Patients were enquired regarding use of recreational drugs (cocaine, amphetamine, heroin, and marijuana) and over the counter medications (propylphenylamine and ephedra) in 1 week before stroke onset. Sexual Intercourse Sexual activity within 2 hours preceding stroke onset was enquired. Sudden Posture Change Because of Startling Event Patients were asked to recall any sudden change in posture during the day preceding stroke onset. For each reported sudden change in posture, the exact time and reason were specified. Only patients reporting sudden changes in posture in response to a startling event, such as getting up suddenly from bed in

A. SHARMA ET AL.

340

Table 1. Demographic and clinical characteristics of all study subjects and of subgroups on basis of type of stroke

Characteristics Age, y, mean (SD) Age groups, n (%) #40 y 41-60 y .60 y Gender, n (%) Male Female Time quadrant of stroke onset, n (%) 00:00-05:59 h 06:00-11:59 h 12:00-17:59 h 18:00-23:59 h Season of stroke onset, n (%) Winters (December/January/February) Pre-monsoon (March/April/May) Monsoon (June/July/August) Postmonsoon (September/October/November) Hypertension, n (%) Diabetes mellitus, n (%) Dyslipidemia, n (%) Smoking, n (%) Alcohol intake, n (%) Heart disease, n (%) RHD: 13 (4.5) IHD: 5 (1.7) Lone AF: 6 (2.1) No known risk factors, n (%)

Total N 5 290

Ischemic stroke 5 224 (77.2%)

Hemorrhagic stroke 5 66 (22.8%)

54.1 (13.5)

54.6 (13.7)

52.2 (12.4)

55 (19) 132 (45.5) 103 (35.5)

43 (19.2) 97 (43.3) 84 (37.5)

12 (18.2) 35 (53) 19 (22.8)

210 (72.4) 80 (27.6)

159 (71) 65 (29)

51 (77.3) 15 (22.7)

37 (12.8) 145 (50) 69 (23.8) 39 (13.4)

29 (12.9) 113 (50.4) 53 (23.7) 29 (12.9)

8 (12.1) 32 (48.5) 16 (24.2) 10 (15.2)

43 (14.8) 115 (39.7) 69 (23.8) 63 (21.7) 184 (63.4) 87 (30) 68 (23.4) 78 (26.9) 75 (25.9)

32 (14.3) 87 (38.8) 55 (24.6) 50 (22.3) 126 (56.2) 83 (37.1) 62 (27.7) 67 (29.9) 63 (28.1)

11 (16.7) 28 (42.4) 14 (21.2) 13 (19.7) 58 (87.9) 4 (6.1) 6 (9.1) 11 (16.7) 12 (18.2)

24 (8.3)

24 (10.7)

38 (13)

34 (15.2)



4 (6.1)

Abbreviations: AF, atrial fibrillation; IHD, ischemic heart disease; RHD, rheumatic heart disease.

response to the patient’s child/grandchild falling down and crying or getting up suddenly in response to an unexpected very loud noise, were considered exposed. Trauma/Surgery (Prior 1 Month) Subjects were enquired for motor vehicle accidents, sports injury, and spinal manipulative therapy (especially chiropractic neck manipulation). Subjects were also enquired for any major surgery (general and cardiac). Unusual Vigorous Physical Exertion (within 1 Hour Preceding Stroke Onset) Subjects were asked about the frequency and timing of moderate exertion, with deep breathing, and the frequency and timing of vigorous exertion, with panting and overheating. Exposure to unusual vigorous physical exertion as triggering risk factor for stroke was defined as vigorous physical exertion within 1 hour preceding stroke onset in subjects who exercise less than 3 times per week.

Statistical Analysis Sample size of 300 patients was calculated to allow prevalence estimation of triggering factors in acute stroke patients with 5% precision, assuming prevalence estimate of up to 20%. In addition, this sample size would be enough for detection of odds ratio (OR) of 2.5 or more when analyzing factors related to presence of trigger factors in acute stroke. Data were entered into an electronic database for statistical analysis (SPSS, version 20.0). Data were presented as number (%) or mean (standard deviation [SD]) as appropriate. Categorical variables were compared using the chi-square test or Fisher exact test and continuous variables using Student t test for independent samples. Variables with a possible (P , .1) association with trigger factors in stroke on bivariate analysis were included in a multivariable logistic regression model, examining for associations of the presence of trigger factors in acute stroke as a binary dependent variable. The results were reported as OR (95% confidence interval [CI]). The P value less than .05 was considered as statistically significant.

TRIGGERING FACTORS IN ACUTE STROKE

341

Table 2. Prevalence of various triggering factors* in all subjects and in subgroups on basis of type of stroke

Type of trigger Psychological stress Stressful life eventsy (previous 1 mo) Negative affect (previous 1 wk) Anger (during 2 h preceding onset) Acute alcohol abuse Within 24 h preceding stroke (.40 g) Within last 1 wk (.150 g) Clinical infection Within 7 d of stroke onset 8 d to 1 mo of stroke onset Coffee intake (within 2 h preceding stroke onset in infrequent consumers) Sexual activity (within 2 h preceding stroke onset) Trauma (within previous 1 mo preceding stroke onset) Surgery (within previous 1 mo preceding stroke onset)

Total N 5 290, n (% total pt.; 95% CI): 128 (44.2%; 38.4%-50%)

Ischemic stroke 5 224, n (% ischemic stroke pt.): 104 (46.4%)

Hemorrhagic stroke 5 66, n (% hemorrhagic stroke pt.): 24 (36.4%)

51 (17.6; 13.4%-22.4%) 34 (11.7) 17 (5.9) 12 (4.1; 2.1%-7.1%) 31 (10.7; 7.4%-14.8%) 20 (6.9) 11 (3.8) 24 (8.3; 5.4%-12%) 14 (4.8) 10 (3.4) 12 (4.1; 2.2%-7.1%)

37 (16.5) 23 (10.3) 14 (6.2) 9 (4) 26 (11.6) 17 (7.6) 9 (4) 21 (9.3) 14 (6.2) 7 (3.1) 8 (3.5)

14 (21.2) 11 (16.7) 3 (4.5) 3 (4.5) 5 (7.5) 3 (4.5) 2 (3) 3 (4.5) — 3 (4.5) 4 (6)

.38

5 (1.7; .6%-4%) 4 (1.4; .4%-3.5%)

4 (1.8) 4 (1.8)

1 (1.5) —

.70

5 (1.7; .6%-4%)

4 (1.8)

1 (1.5)

.70

P value

.87 .35

.21

.59

Abbreviations: CI, confidence interval; pt., patients. *None of the patients reported recreational drug abuse, startling event, and unusual vigorous physical exertion in hazard periods. yDeath/major health issue in the family/close relatives or friends, 8 (2.8%); financial issue, 6 (2.1%); serious argument/problem in family, 9 (3.1%); serious argument/problem at workplace, 5 (1.7%); legal problems, 2 (.7%); others, 4 (1.4%).

Results In total, 290 patients participated in the study (Table 1). The mean age of entire study population was 54.1 years (SD 13.5). Stroke in young (age # 40 years) included 55 (19%) patients; 210 (72.4%) patients were men. Regarding the type of stroke, 224 (77.2%) patients had ischemic stroke and 66 (22.8%) patients had hemorrhagic stroke; 145 (50%) patients had stroke onset in time quadrant between 06:00 and 11:59. Maximum patients, n 5 115 (39.7%), were recruited in pre-monsoon season (March, April, and May) as these months came twice in study period of 15 months. Regarding risk factor profile, hypertension was present in 184 (63.4%), diabetes mellitus in 87 (30%), dyslipidemia in 68 (23.4%), smoking in 78 (26.9%), alcohol use in 75 (25.9%), and heart disease in 24 (8.3%). Presence of any trigger factor out of 11 trigger factors studied was seen in 128 (44.2%) of total 290 patients, 104 (46.4%) of 224 ischemic stroke patients, and 24 (36.4%) of 66 hemorrhagic stroke patients (Table 2). Psychological stress was present in 51 (17.6%) patients, among psychological stress: stressful life event in 34 (11.7%), negative affect in 17 (5.9%), acute alcohol abuse in 31 (10.7%), clinical infections in 24 (8.3%), and anger and coffee intake in 12 (4.1%) each. Sexual activity, trauma, and surgery were present in 5 (1.7%), 4 (1.4%), and 5 (1.7%) patients, respectively. Stressful life events

included death/major health issue in the family/close relatives or friends, 8 (2.8%); financial issue, 6 (2.1%); serious argument/problem in family, 9 (3.1%); serious argument/problem at workplace, 5 (1.7%); legal problems, 2 (.7%); and others, 4 (1.4%). None of the patients reported exposure to recreational drug abuse, startling event, and unusual vigorous physical exertion in hazard periods. Two or more trigger factors were present in 16 (5.5%) patients. The difference in prevalence of triggers in 2 stroke subtypes (ischemic and hemorrhagic) was not significant. However, definite conclusion could not be made because of relatively small number of hemorrhagic stroke patients. On bivariate analysis (Tables 3-5), comparing acute stroke (including both ischemic and hemorrhagic stroke) patients without trigger factors, those with the presence of trigger factors were younger: in patients younger than 60 years, triggers were present in 50.3%, whereas in patients older than 60 years, triggers were present in 32% (P value 5 .008). Triggers were present in 47.6% men and 35% women (P value 5 .05) and triggers were present in 55.1% smokers in comparison with 44.9% nonsmokers (P value 5 .02). Further, comparing acute ischemic stroke patients without trigger factors, those with the presence of trigger factors were younger: in patients younger than 60 years,

A. SHARMA ET AL.

342

Table 3. Bivariate analysis to examine clinical correlates of presence of trigger factors in entire study subjects: comparison of clinical characteristics in subgroups on the basis of presence or absence of trigger factor

Variables Age, y, mean (SD) Age groups, n (%) #40 y 40-60 y $60 y* Gender, n (%) Male Female* Time quadrant of stroke onset, n (%) 00:00-05:59 h 06:00-11:59 h 12:00-17:59 h 18:00-23:59 h* Season of stroke onset, n (%) Winters* (December/January/February) Pre-monsoon (March/April/May) Monsoon (June/July/August) Postmonsoon (September/October/November) Hypertension, n (%) Diabetes mellitus, n (%) Dyslipidemia, n (%) Smoking, n (%) Alcohol intake, n (%) Heart disease, n (%) No known risk factor, n (%) Type of stroke, n (%) Ischemic stroke Hemorrhagic stroke*

Total N 5 290 54.1 (13.5)

Trigger present, 128 (44.1%) 51.6 (12)

Trigger absent, 162 (55.9%)

P value, OR (95% CI)

56 (14.2)

.005 .008 .03, 2.04 (1.04-4) .003, 2.25 (1.31-3.85)

55 (19) 132 (45.5) 103 (35.5)

27 (49.1) 68 (51.5) 33 (32)

28 (50.9) 64 (48.5) 70 (68)

210 (72.4) 80 (27.6)

100 (47.6) 28 (35)

110 (52.4) 52 (65)

37 (12.8) 145 (50) 69 (23.8) 39 (13.4)

14 (37.8) 63 (43.4) 38 (55.1) 13 (33.3)

23 (62.2) 82 (56.6) 31 (44.9) 26 (66.7)

43 (14.8) 115 (39.7) 69 (23.8) 63 (21.7) 184 (63.4) 87 (30) 68 (23.4) 78 (26.9) 75 (25.9) 24 (8.3) 38 (13.1)

17 (39.5) 52 (45.2) 33 (47.8) 26 (41.3) 82 (44.6) 40 (46) 29 (42.6) 43 (55.1) 37 (49.3) 7 (29.2) 19 (50)

26 (60.5) 63 (54.8) 36 (52.2) 37 (58.7) 102 (55.4) 47 (54) 39 (57.4) 35 (44.9) 38 (50.7) 17 (70.8) 19 (50)

224 (77.2) 66 (22.8)

104 (46.4) 24 (36.4)

120 (53.6) 42 (63.6)

.05, 1.6 (.9-2.8) .12 .68, 1.21 (.47-3.11) .25, 1.53 (.73-3.22) .03, 2.45 (1.08-5.55) .80 .52, 1.26 (.61-2.57) .39, 1.40 (.64-3.03) .85, 1.07 (.48-2.37) .84, 1 (.6-1.6) .68, 1.10 (.6-1.8) .77, .9 (.5-1.5) .02, 1.8 (1-3.1) .29, 1.3 (.7-2.2) .12, .4 (.2-1.2) .43, 1.31 (.66-2.6) .14, 1.51 (.86-2.67)

Abbreviations: OR, odds ratio; CI, confidence interval. *Reference category for calculating odds.

triggers were present in 55.7%, whereas in patients older than 60 years, triggers were present in 37.5% (P value 5 .004). Triggers were present in 50.3% men and 36.9% women (P value 5 .07). The mean (SD) NIHSS score in patients with triggers was 8.2 (3.4) in comparison with mean NIHSS score of 6.2 (3.3) in patients without triggers (P value , .001). Similarly, in hemorrhagic stroke patients, the mean (SD) hematoma volume in patients with triggers was 11.9 (4.2) in comparison with mean hematoma volume of 7.7 (2.8) in patients without triggers (P value , .001). There was no association of age with the presence of triggers in hemorrhagic stroke patients. However, definite conclusion could not be made because of relatively small number of hemorrhagic stroke patients. Note is made of wide CIs in statistical estimates of hemorrhagic stroke patients, again because of small number (n 5 66) of these patients in the study. Examining the entire study population, clinical variables independently associated with the presence of trigger factors in acute stroke patients on multivariate

analysis were younger age (,60 years) and higher NIHSS score in ischemic stroke patients and higher hematoma volume in hemorrhagic stroke patients (Tables 6 and 7). Those variables with P value less than .10 on bivariate analysis were entered into this multivariate analysis. Note is made of wide CIs in statistical estimates of hemorrhagic stroke patients because of small number (n 5 66) of these patients in the study. In ischemic stroke patients, trigger factors such as psychological stress, acute alcohol abuse, and anger (in descending order) were associated with higher NIHSS score compared with clinical infection and coffee intake (Table 8). Whereas in hemorrhagic stroke patients, trigger factors such as anger, psychological stress, and acute alcohol abuse (in descending order) were associated with higher hematoma volume compared with clinical infection and coffee intake (Table 8). But these are preliminary observations as numbers of patients in individual groups were small especially in hemorrhagic stroke patients.

TRIGGERING FACTORS IN ACUTE STROKE

343

Table 4. Bivariate analysis to examine clinical correlates of presence of trigger factors in ischemic stroke subjects: comparison of clinical characteristics in subgroups on basis of presence or absence of trigger factor

Variables Age, y, mean (SD) Age groups, n (%) #40 y 40-60 y $60 y* Gender, n (%) Male Female* Time quadrant of stroke onset, n (%) 00:00-05:59 h 06:00-11:59 h 12:00-17:59 h 18:00-23:59 h* Season of stroke onset, n (%) Winters* (December/January/February) Pre-monsoon (March/April/May) Monsoon (June/July/August) Postmonsoon (September/October/ November) Hypertension, n (%) Diabetes mellitus, n (%) Dyslipidemia, n (%) Smoking, n (%) Alcohol intake, n (%) Heart disease, n (%) No known risk factor, n (%) Severity: NIHSS score, mean (SD) NIHSS score, n (%) 0-4* 5-9 .9 Territory, n (%) ACA* MCA PCA Etiology, n (%) Large-artery* atherosclerotic Cardioembolic Small-vessel disease Other known causes Unknown

Total N 5 224, n (%)

Trigger present, 104 (46.4%)

Trigger absent, 120 (53.6%)

P value, OR (95% CI)

54.6 (13.7)

52 (12)

57 (15)

22 (51.2) 56 (57.7) 26 (31)

21 (48.8) 41 (42.3) 58 (69)

.004 .001 .02, 2.33 (1.09-4.97) ,.001, 3.04 (1.65-5.62)

159 (71) 65 (29)

80 (50.3) 24 (36.9)

79 (49.7) 41 (63.1)

29 (12.9) 113 (50.4) 53 (23.7) 29 (12.9)

12 (41.4) 53 (46.9) 30 (56.6) 9 (31)

17 (58.6) 60 (53.1) 23 (43.4) 20 (69)

32 (14.3) 87 (38.8) 55 (24.6) 50 (22.3)

14 (43.8) 41 (47.1) 27 (49.1) 22 (44)

18 (56.2) 46 (52.9) 28 (50.9) 28 (56)

126 (56.2) 83 (37.1) 62 (27.7) 67 (29.9) 63 (28.1) 24 (10.7) 34 (15.2) 7.1 (3.5)

63 (50) 39 (47) 29 (46.8) 37 (55.2) 32 (50.8) 7 (29.2) 16 (47.1) 8.2 (3.4)

63 (50) 44 (53) 33 (53.2) 30 (44.8) 31 (49.2) 17 (70.8) 18 (52.9) 6.2 (3.3)

65 (29) 101 (45) 58 (26)

18 (27.7) 44 (43.6) 42 (72.4)

47 (72.3) 57 (56.4) 16 (27.6)

11 (4.9) 177 (78.7) 37 (16.4)

3 (27.3) 89 (50.3) 12 (32.4)

8 (72.7) 87 (49.7) 25 (67.6)

67 (29.9) 23 (10.3) 71 (31.7) 12 (5.4) 51 (22.8)

26 (38.8) 7 (30.3) 36 (50.7) 9 (75) 26 (51)

41 (61.2) 16 (69.6) 35 (49.3) 3 (25) 25 (49)

43 (19.2) 97 (43.3) 84 (37.5)

.07, 1.73 (.95-3.12) .16 .41, 1.56 (.53-4.61) .12, 1.96 (.82-4.68) .02, 2.89 (1.11-7.54) .94 .74, 1.14 (.5-2.6) .63, 1.24 (.5-3) .98, 1.01 (.4-2.5) .22, 1.39 (.81-2.36) .89, 1.03 (.6-1.78) .94, 1.01 (.56-1.83) .08, 1.65 (.93-2.94) .41, 1.27 (.71-2.28) .08, .43 (.17-1.10) .93, 1.03 (.49-2.14) ,.001 ,.001 .04, 2.01 (1.03-3.94) ,.001, 6.85 (3.1-15.12) .06 .14, 2.72 (.7-10.62) .74, 1.28 (.28-5.7) .06 .47, .69 (.25-1.9) .16, 1.62 (.82-3.19) .02, 4.73 (1.17-19.1) .18, 1.64 (.78-3.42)

Abbreviations: ACA, anterior cerebral artery; CI, confidence interval; MCA, middle cerebral artery; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; PCA, posterior cerebral artery. *Reference category for calculating odds.

Discussion There are 2 broad aspects of research on stroke triggers: First, to establish whether potential triggers are associated with stroke onset, to know the duration of triggering effect, and to understand pathophysiological basis of triggering mechanism. Second, to determine the prevalence of triggers in acute stroke, to look for demographic/clin-

ical correlates of triggers, and to know the interaction with chronic risk factors and stroke outcomes. The bulk of research on stroke triggers till now has focused on first aspect. This study focusing on second aspect examined the prevalence of 11 predefined trigger factors (including both well-established and potential triggers) in predefined varying hazard periods for each individual trigger and looked for possible association with various clinical

A. SHARMA ET AL.

344

Table 5. Bivariate analysis to examine clinical correlates of presence of trigger factors in hemorrhagic stroke subjects: comparison of clinical characteristics in subgroups on the basis of presence or absence of trigger factor

Variables Age, y, mean (SD) Age groups, n (%) #40 y 40-60 y $60 y* Gender, n (%) Male Female* Time quadrant of stroke onset, n (%) 00:00-05:59 h* 06:00-11:59 h 12:00-17:59 h 18:00-23:59 h Season of stroke onset, n (%) Winters* (December/January/February) Pre-monsoon (March/April/May) Monsoon (June/July/August) Post Monsoon (September/October/ November) Hypertension, n (%) Diabetes mellitus, n (%) Dyslipidemia, n (%) Smoking, n (%) Alcohol intake, n (%) No known risk factor, n (%) Severity: hematoma volume, ml, mean (SD) Hematoma volume, ml, n (%) #6* 7-10 .10 Site, n (%) Putamen* Thalamus Cerebellum Lobar

Total N 5 66

Trigger present, 24 (36.4%)

Trigger absent, 42 (63.6%)

52.2 (12.4)

51 (12)

53 (13)

12 (18.2) 35 (53) 19 (28.8)

5 (41.7) 12 (34.3) 7 (36.8)

7 (58.3) 23 (66.7) 12 (63.2)

51 (77.3) 15 (22.7)

20 (39.2) 4 (26.7)

31 (60.8) 11 (73.3)

8 (12.1) 32 (48.5) 16 (24.2) 10 (15.2)

2 (25) 10 (31.2) 8 (50) 4 (40)

6 (75) 22 (68.8) 8 (50) 6 (60)

11 (16.7) 28 (42.4) 14 (21.2) 13 (19.7)

3 (27.3) 11 (39.3) 6 (42.9) 4 (30.8)

8 (72.7) 17 (60.7) 8 (57.1) 9 (69.2)

58 (87.9) 4 (6.1) 6 (9.1) 11 (16.7) 12 (18.2) 4 (6.1) 9.2 (3.9)

19 (32.8) 1 (25) 0 (0) 6 (54.5) 5 (41.7) 3 (75) 11.9 (4.2)

39 (67.2) 3 (75) 6 (100) 5 (45.5) 7 (58.3) 1 (25) 7.7 (2.8)

19 (29) 26 (39) 21 (32)

2 (10.5) 8 (30.8) 14 (66.6)

17 (89.5) 18 (69.2) 7 (33.3)

44 (66.7) 12 (18.2) 6 (9.1) 4 (6.1)

17 (38.6) 3 (25) 0 (0) 4 (100)

27 (61.4) 9 (75) 6 (100) 0 (0)

P value, OR (95% CI) .45 .89 .78, 1.22 (.27-5.37) .85, .89 (.27-2.86)

.37, 1.77 (.49-6.34) .55 .73, 1.36 (.23-7.97) .25, 3.00 (.45-19.59) .50, 2.00 (.26-15.38) .82 .48, 1.72 (.37-7.95) .42, 2.00 (.36-10.91) .85, 1.18 (.20-6.98) .11, .29 (.06-1.35) .62, .56 (.05-5.75) — .17, 2.46 (.66-9.17) .67, 1.31 (.36-4.71) .13, 5.85 (.57-59.08) ,.001 .001 .12, 3.77 (.70-20.37) .001, 17 (3.03-95.25) .01 .38, .52 (.12-2.23) — —

Abbreviations: CI, confidence interval; OR, odds ratio. *Reference category for calculating odds.

variables (eg, age, gender, risk factor profile, and stroke severity). This study does not aim to establish the potential triggers as risk factors for stroke onset as it does not have nonstroke controls. It aims to determine the prevalence of potential triggers already reported in literature and to determine whether there is association between the presence of triggers and stroke severity—a question that has not been studied previously. Despite our best efforts, we could not find any previous study in literature with similar aims and objectives as in our study. Most of the studies published in literature till now focus on establishing individual potential trigger factor’s association with stroke onset. This is true as we first need to establish association of various individual trigger

factors with stroke onset before proceeding to determine the prevalence and possible associations. In our chosen list of trigger factors, acute alcohol abuse and clinical infections are the most well-established trigger factors for stroke.5 Regarding psychological stress, a recent study by Guiraud et al14 suggested it as possible trigger factor for stroke onset.14 Using a case-crossover approach, they found patients were exposed to 1 or more life events more often during the first month preceding stroke onset than during the 5 control periods (OR 5 2.96; 95% CI 2.19-4.00). There is limitation on other triggers, such as anger, coffee intake, startling event, unusual vigorous physical exertion, and sexual activity.5 Most of these triggers have a single study in stroke patients. Hence, there is a need to conduct further studies.

TRIGGERING FACTORS IN ACUTE STROKE

Table 6. Multivariate binary logistic regression analysis examining associations of presence of trigger factors in acute ischemic stroke patients

Variables* Age (y) #40 40-60 $60y NIHSS score 0-4y 5-9 .9

Presence of trigger factor in acute ischemic stroke patients, P value, OR (95% CI) .001 .01, 2.84 (1.25-6.42) ,.001, 3.54 (1.82-6.90) ,.001 .02, 2.28 (1.13-4.59) ,.001, 8.12 (3.52-18.69)

Abbreviations: CI, confidence interval; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio. *Those variables with P value ,.10 in bivariate analysis were entered into this multivariate analysis; only data for significant results are presented. yReference category for calculating odds.

Direct comparison with previous studies regarding prevalence of triggers is difficult because of varying definitions and hazard periods. Triggers were observed in 44% (presence of any trigger out of list of 11 triggers) of stroke patients. The most frequent in descending order include psychological stress (17.6%), acute alcohol abuse (10.7%), clinical infection (8.3%), anger, and coffee intake (4.1% each). The difference in prevalence of triggers in 2 stroke subtypes (ischemic and hemorrhagic) was not significant. However, definite conclusion could not be made because of relatively small number of hemorrhagic stroke patients. Partially similar to our study, Fradis and Iacobescu15 studied in 300 acute stroke patients (100 carotid ischemic, Table 7. Multivariate binary logistic regression analysis examining associations of presence of trigger factors in acute hemorrhagic stroke patients

Variables* Hematoma volume (ml) #6z 7-10 .10

Presence of trigger factor in acute hemorrhagic stroke patients, P value, OR (95% CI)y .003 .12, 3.77 (.70-20.37) .001, 17 (3.03-95.25)

Abbreviations: CI, confidence interval; OR, odds ratio. *Those variables with P value ,.10 in bivariate analysis were entered into this multivariate analysis; only data for significant results are presented. yWide confidence intervals are because of small number (ie, 66) of hemorrhagic stroke patients. zReference category for calculating odds.

345

100 vertebrobasilar ischemic, and 100 hemorrhagic) the precipitating factors (PF) of stroke. Precipitating factors were observed in 63.7% of cases. Of the 407 PF found in the whole sample, 60% occurred 24 hours before stroke and 40% 2-3 days before that. The most frequent PF were physical effort (32.3%), extreme temperature (29%), psychic trauma (23.3%), and head and body positions that may unfavorably influence cerebral circulation (23%). These PF were usually observed in associations. The differences regarding the PF prevalence in the 3 stroke subtypes were not statistically significant. But the main limiting issues of this study were that the PF of stroke noted in study were not predefined and also no individual hazard periods were defined. In other case-crossover study by Koton et al,16 200 consecutive acute stroke patients were interviewed 14 days after the event using a validated questionnaire, partially similar to our study. Reported exposure to potential triggers including negative and positive emotions, anger, sudden posture changes as response to a startling event, heavy physical exertion, heavy eating, and sudden temperature changes during a 2-hour hazard period before stroke onset were compared with the same period during the preceding day and to average exposures in the last year. Of 200 patients interviewed, 76 (38%) reported exposure to at least 1 trigger. Independent triggers included negative emotions, anger, and sudden changes in posture. Regarding comparison of individual trigger prevalence with previous studies, relevant studies regarding acute alcohol abuse and clinical infection are listed in Table 9. These studies have similar trigger definition and hazard period duration as of our study. Regarding psychological stress, recent study by Guiraud et al14 has shown prevalence of 39.3% for stressful life events in hazard period of 1 month; 4.1% prevalence of anger found in our study was lower compared with case-crossover study by Koton et al,16 according to which 7% of the patients had an anger outburst during the 2 hours before stroke onset. Further independent association (on multivariate analysis) of younger age (,60 years) and higher initial stroke severity at presentation with the presence of trigger factors in acute stroke suggest these triggers to be possible risk factors for stroke onset. But in comparison with conventional chronic risk factors, which increase the stroke risk over a long period of time, trigger factors increase stroke risk for relatively short period of time. However, these associations need to be further explored in community-based studies. A more complete characterization of triggering factors with the understanding of triggering mechanisms in relation to how they interact with the conventional cardiovascular risk factor scenario requires further investigations. The association observed in this study between the presence of triggering factor and stroke severity needs explanation. Although the exact explanation remains

A. SHARMA ET AL.

346

Table 8. Comparison of stroke severity on basis of type of triggering risk factors Measure of stroke severity/type of triggering risk factor

Psychological Acute alcohol Clinical infection, stress, n abuse, n n

NIHSS score (ischemic stroke); 37; 10.4 (3.1) mean (SD) Hematoma volume (hemorrhagic 14; 13.5 (3.4) stroke); mean (SD)

26; 9.1 (3.8)

21; 6.4 (3.1)

5; 13.2 (3.6)

3; 11.3 (3)

Anger, n

Coffee intake, n P value*

9; 8.89 (2.4)

8; 6.4 (2.5)

,.001

3; 14 (2.6)

4; 7.5 (1.3)

.03

Abbreviation: NIHSS, National Institutes of Health Stroke Scale. *P value calculated by 1-way analysis of variance.

unknown, we speculate that triggers (in context of ischemic stroke) may be activating plaques in the large arteries and contributing to train of events leading to their occlusion, and thus, a larger proportion of patients with triggers have large artery occlusions and have greater severity. In context of hemorrhagic stroke patients, triggers may contribute to transient more rise in blood pressure and, hence, increase the volume of hematoma. Our study, however, did not have the power to explore these

potential explanations, and more studies with larger sample size are required. Regarding methodological issues, patients were not blinded to study objectives and interviewer was not blinded to patient status, particularly stroke severity. This study was hospital based and confined to a single center; this factor potentially limits the generalizability of our results. This was a hospital-based cross-sectional study. The possibility that a reversal of the cause-and-

Table 9. Summary of previous studies relevant* in context of our study regarding prevalence of acute alcohol abuse and clinical infection in acute stroke patients in hazard periods Study reference Acute alcohol abuse Haapaniemi (1997) Finland(37) Hillbom (1999) Finland (38) Present study

Clinical infections Garu (1995) Germany (39) Garu (1998) Germany (40) Paganini-Hill (2003) United States (41) Present study

Brief study details

Results

Case–control, hospital based 506 cases (72% males) Mean age: 49 y Case–control, hospital based 212 cases (87% males) Mean age: 44 y Observational cross-sectional study, hospital based No nonstroke controls 290 cases (72% males) Mean age: 54 y

Prevalencey of acute alcohol abuse (.40 g in 24 h/.150 g in 1 wk preceding stroke onset): 15%

Case–control, hospital based 197 cases (58% males) Mean age: 63 y Case–control, hospital based 166 cases (66% males) Mean age: 61 y Case–control, hospital based 233 cases (56% males) Mean age: 57 y Observational cross-sectional study, hospital based No nonstroke controls 290 cases (72% males) Mean age: 54 y

Prevalencey of acute alcohol abuse (.40 g in 24 h/.150 g in 1 wk preceding stroke onset): 17% Prevalence of acute alcohol abuse (.40 g in 24 h/.150 g in 1 wk preceding stroke onset): 10.7%

Prevalencey of any infection (within 1 mo preceding stroke onset): 24.4% Prevalencey of any infection (within 1 mo preceding stroke onset): 27.1% Prevalencey of any infection/inflammation (within 1 mo preceding stroke onset): 18.7% Prevalence of any infection (within 1 mo preceding stroke onset): 8.3%

*These studies are relevant to our study in context to similar definition and duration of hazard periods; otherwise, study design and objectives were different. yPrevalence in cases (ie, acute stroke patients).

TRIGGERING FACTORS IN ACUTE STROKE

effect relationship occurs is inevitable. No attempt was made to establish association with stroke onset. The sample size was calculated to detect factors strongly related to the presence of stroke triggers (ie, ORs . 2.5), but the power of the study was insufficient to detect milder correlations. Our sample size did not allow assessment of stroke type, ischemic stroke territory and etiology, and hemorrhagic stroke site as possible associations. Also, our sample size did not allow for detailing associations of each type of individual trigger factors. Furthermore, our study did not have the power to explore mechanism of higher stroke severity in those with the presence of triggers. To summarize, to the best of our knowledge, our study is the first in India to describe the prevalence of potential triggers in acute stroke and to demonstrate an association between the presence of trigger and stroke severity. Whether these are risk factors for stroke onset or what is the mechanism by which triggers increase stroke severity remains to be studied in further studies. The practical implications of research on triggers are still hypothetical. The main potential implication is a better knowledge of pathophysiological mechanisms leading to stroke onset. There is also potential for preventive strategies directed against the short-term risk posed by triggering activities as a complement of the long-term risk factor reduction approach. For instance, we could contemplate avoiding trigger exposure as much as possible in high-risk patients (eg, aggressive treatment of infections) or transient reinforcement of traditional prevention in high-risk patients exposed to a trigger (eg, use of antithrombotic drugs). However, this remains to be demonstrated and would not apply to all triggers.

References 1. Muller JE, Abela GS, Nesto RW, et al. Triggers, acute risk factors and vulnerable plaques: the lexicon of a new frontier. J Am Coll Cardiol 1994;23:809-813.

347 2. Stone PH. Triggering myocardial infarction. N Engl J Med 2004;35:1716-1718. 3. Tofler GH, Muller JE. Triggering of acute cardiovascular disease and potential preventive strategies. Circulation 2006;114:1863-1872. 4. Elkind MSV. Why now? Moving from stroke risk factors to stroke triggers. Curr Opin Neurol 2007;20:51-57. 5. Guiraud V, Amor MB, Mas J-L, et al. Triggers of ischemic stroke: a systematic review. Stroke J Cereb Circ 2010; 41:2669-2677. 6. Goldstein LB, Bertels C, Davis JN. Interrater reliability of the NIH stroke scale. Arch Neurol 1989;46:660-662. 7. Adams HP Jr, Bendixen BH, Kappelle LJ, et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke J Cereb Circ 1993;24:35-41. 8. Attri SD, Tyagi A. Microsoft Word—NAT-preface 8.11.doc—climate_profile.pdf [Internet]. India Meterological Department. Available at: http://www.imd.gov.in/ doc/climate_profile.pdf. Accessed August 30, 2013. 9. Kothari RU, Brott T, Broderick JP, et al. The ABCs of measuring intracerebral hemorrhage volumes. Stroke 1996;27:1304-1305. 10. Zevon MA, Tellegen A. The structure of mood change: an idiographic/nomothetic analysis. J Pers Soc Psychol 1982;43:111-122. 11. Cohen S, Tyrrell DA, Smith AP. Psychological stress and susceptibility to the common cold. N Engl J Med 1991; 325:606-612. 12. Ruberman W, Weinblatt E, Goldberg JD, et al. Psychosocial influences on mortality after myocardial infarction. N Engl J Med 1984;311:552-559. 13. Mittleman MA, Maclure M, Sherwood JB, et al. Triggering of acute myocardial infarction onset by episodes of anger. Determinants of myocardial infarction onset study investigators. Circulation 1995;92:1720-1725. 14. Guiraud V, Touze E, Rouillon F, et al. Stressful life events as triggers of ischemic stroke: a case-crossover study. Int J Stroke Off J Int Stroke Soc 2013;8:300-307. 15. Fradis A, Iacobescu E. Exogenous precipitating factors in stroke. Neurol Psychiatr (Bucur) 1989;27:35-43. 16. Koton S, Tanne D, Bornstein NM, et al. Triggering risk factors for ischemic stroke: a case-crossover study. Neurology 2004;63:2006-2010.