obesity - prevalence - stroke, epidemiology - cross-sectional studies
obesidade - prevalência - acidente vascular cerebral, epidemiologia - estudos transversais
Obesity has reached epidemic proportions globally, with at least 2.8 million people
dying each year as a result of being overweight or obese[1]. Once associated with high-income countries, obesity is now also prevalent in low-
and middle-income countries[1]. In Brazil, a study by the Ministry of Health in 2011, showed that 48% of the population
was either overweight (33%) or obese (15%)[2]. In 2006, these figures were 22.7% and 11.4%[2],[3].
Because obesity is believed to cause a number of established risk factors for cardiovascular
diseases such as hypertension, dyslipidemia and diabetes, the growing prevalence of
obesity is assumed to increase the global cardiovascular disease burden[4]. For instance, in adults, the risk for ischemic stroke increases almost linearly
starting at a body mass index (BMI) of ≈20, and adults with a BMI of > 30 kg/m2 have about a 70% increased risk for ischemic stroke compared with patients with a
BMI < 25 kg/m2,
[5]. As stroke is currently the second largest cause of mortality in Brazil, we aimed
to measure the prevalence of overweight and obese patients with incident stroke from
five distinct Brazilian macro-regions.
METHODS
Study population
This was a cross-sectional study. We extracted data from the Joinville Stroke Registry,
which has data of patients with first-ever stroke, from five Brazilian cities: 1.
Sobral (Ceará state/northeast region); 2. Sertãozinho (São Paulo state/southeast region);
3. Campo Grande (Mato Grosso do Sul state/midwest region); 4. Joinville (Santa Catarina
state/southern region); 5. Canoas (Rio Grande do Sul state/southern region). We used
the 2016 intercensus data from the Ministry of Health[6]. All the cities have hospitals with computed tomography (CT) services available
on a 24-hour basis. The data from cities outside Joinville were extracted from an
ongoing study sponsored by the Ministry of Health and the Brazilian National Council
for Scientific and Technological Development (grant 402396/2013-8).
Period of data collection and ascertainment strategy
This study sample was extracted between January and December 2016. Staff in all the
cities were trained to follow the “ideal” population-based study according to the
World Health Organization steps criteria[7]. These methods have been extensively published elsewhere[8]. In brief, using multiple overlapping sources, we identified all inpatients and
outpatients with first-ever stroke. To evaluate hospitalized patients on a daily basis,
previously-trained research nurses registered all hospitalized stroke patients confirmed
by a neurologist. A neuroradiologist, unaware of the patients’ symptoms, analyzed
the brain CT scans and digital angiographies daily. We did not monitor patients potentially
at risk (i.e., “hot pursuit”), such as those submitted to aortic, carotid, coronary,
or peripheral investigation or intervention.
Baseline data
We ascertained demographic, socioeconomic data, cardiovascular risk factors and stroke
severity. Cardiovascular risk factors were: hypertension (a history of treated hypertension
or current use of antihypertensive medication); diabetes (a history of diabetes or
current treatment for diabetes mellitus); coronary artery disease (history of angina,
acute myocardial infarction, or coronary revascularization); atrial fibrillation (known
previous atrial fibrillation or current treatment for atrial fibrillation); hypercholesterolemia
(current treatment for hypercholesterolemia); smoking (current smoker). Stroke severity
was measured with the National Institutes of Health Stroke Scales for ischemic stroke[9]. Minor events were defined as National Institutes of Health Stroke Scales scores
≤ 3, moderate events as scores from 4–10, and severe events as scores > 10. The physical
activity levels were stratified in low, medium, high activity and inactive. The following
criteria were used to stratify physical activity levels: 1) Inactive: no physical
activities at home or work; without paid labor or recreational physical activities;
2) Low: less than 30 minutes of physical activity per day; sitting job without physical
activity; 3) Moderate: sedentary work with 30 minutes to one hour physical activity
per day or 30 minutes to one hour of paid labor or recreational physical activity
per day; 4) Active: sedentary work with ≥ one hour physical activity per day; heavy
manual work[10]. Social class was defined according to the Brazilian Criteria of Economic Classification
based on the year 2013 National Household Sample Survey[11].
Diagnostic criteria
We included all patients with a first-ever ischemic stroke, hemorrhagic stroke or
“primary” intracerebral hemorrhage, who were residents of each of the five cities.
We defined stroke as the presence of signs of sudden focal or global cerebral dysfunction
lasting longer than 24 hours without any apparent nonvascular cause[12]. The ischemic stroke patients were classified according to the modified Trial of
Org 10172 in Acute Stroke Treatment criteria[13]. The stroke investigation routine followed the guidelines of the Brazilian Society
of Cerebrovascular Diseases[14].
Outcomes
We used the World Health Organization criteria for overweight and obesity definitions[15]. The BMI (weight/height[2]) was obtained by nurse interview[16]. We compared the proportion of normal weight, overweight (BMI > 25) and obesity
(BMI > 30) among stroke patients from five cities. The prevalence of overweight and
obese patients were stratified by age, education, social class, stroke subtype, stroke
severity and previous level of physical activity.
Statistical analysis
We performed descriptive statistical analyses to compare baseline characteristics
using the Chi-square test, the Student’s t test, or the Mann-Whitney test according
to distribution. To compare proportions, we calculated binomial confidence intervals
at 95%. This study was approved by the Ethics in Research Committee.
RESULTS
We registered 1,255 first-ever stroke patients from five Brazilian cities. All cities
had a medium-size country population, which means between 100,000 to 500,000 inhabitants.
The demographic and socioeconomic baseline data are shown in [Table 1]. Except for Joinville city, the human development index was within the national
cut-off point. The overall mean age was 66 years old. The proportion of patients with
less than four years of education, or who were illiterate, ranged from 23% (95%CI,
17–29) in Canoas to 59% (95%CI, 47–67) in Sobral. The proportion of socio-economically
poor patients (class D/E) ranged from 13% (95%CI, 9–19) in Canoas to 48% (95%CI, 38–58)
in Sobral.
Table 1
Demographics and socioeconomic data of 1,255 patients with first-ever stroke.
|
Variable
|
Campo Grande (n = 372)
|
Canoas (n = 205)
|
Joinville (n = 505)
|
Sertãozinho (n = 66)
|
Sobral (n = 107)
|
All (n = 1,255)
|
|
Demographics
|
|
City population
|
863,982
|
342,634
|
569,545
|
121,412
|
203,682
|
2,101,255
|
|
HDI (2010)
|
0.784
|
0.750
|
0.809
|
0.761
|
0.714
|
0.754*
|
|
Age (SD)
|
65.6 (14.7)
|
64.5 (13.6)
|
65.6 (14)
|
65.7 (16.5)
|
65.5 (16)
|
73.1 (16.1)
|
|
Men (%)
|
185 (49.7)
|
104 (50.7)
|
270 (53.4)
|
33 (50)
|
55 (51.4)
|
647 (51.5)
|
|
Education (years)
|
|
≤ 4
|
148 (40.1)[5]
|
47 (23.0)[1]
|
128 (25.3)
|
25 (39)[2]
|
61 (59.2)[3]
|
408 (32.8)[4]
|
|
4–8
|
142 (38.5)
|
125 (61.2)
|
256 (50.8)
|
27 (42.2)
|
31 (30.1)
|
581 (46.7)
|
|
9–11
|
63 (17.1)
|
28 (13.7)
|
90 (17.8)
|
8 (12.5)
|
9 (8.8)
|
198 (15.9)
|
|
≥ 11
|
16 (4.3)
|
4 (2.0)
|
31 (6.1)
|
4 (6.3)
|
2 (1.9)
|
57 (4.6)
|
|
Socio-economic class
|
|
A1
|
0
|
0
|
1 (0.2)
|
0
|
0
|
1 (0.1)
|
|
A2
|
4 (1.1)
|
2 (1.0)
|
5 (1.0)
|
2 (3.0)
|
0
|
13 (1.0)
|
|
B1
|
7 (1.9)
|
4 (2.0)
|
13 (2.6)
|
3 (4.6)
|
4 (3.7)
|
31 (2.5)
|
|
B2
|
59 (15.8)
|
56 (27.4)
|
108 (21.4)
|
21 (31.8)
|
10 (9.4)
|
254 (20.2)
|
|
C1
|
88 (23.6)
|
59 (28.7)
|
163 (32.3)
|
17 (25.8)
|
13 (12.1)
|
340 (27.2)
|
|
C2
|
128(34.5)
|
57 (27.8)
|
126 (24.9)
|
9 (13.6)
|
28 (26.2)
|
348 (27.7)
|
|
D
|
86 (23.1)
|
26 (12.6)
|
87 (17.2)
|
13 (19.7)
|
48 (44.9)
|
260 (20.7)
|
|
E
|
0
|
1 (0.5)
|
2 (0.4)
|
1 (1.5)
|
4 (3.7)
|
8 (0.6)
|
Data are number of patients (%) unless otherwise indicated 2016 national intercensus;
HDI: human development index. *Brazil HDI; Unavailable in: one patient1; two patients2;
four patients3; 10 patients4.
The sample distribution by stroke types, clinical severity and cardiovascular risk
factors are shown in [Table 2]. The proportion of hemorrhagic stroke was significantly higher in Sertãozinho city
(23%, 95% CI, 13–35; p = 0.001). The most common type of ischemic stroke was undetermined,
ranging from 41% (95% CI, 36–45) in Joinville to 65% (95% CI, 57–72) in Canoas. At
hospital admission, almost half of the patients arrived with severe strokes in Sertãozinho
and Sobral (47%, 95% CI, 35–60 and 47%, 95% CI, 38–56, respectively) as opposed to
20% (95% CI, 16–24) in Joinville. As expected, hypertension was the most common cardiovascular
risk factor. Physical inactivity ranged from 53% (95% CI, 43-63) in Sobral to 80%
(95% CI, 73–85) in Canoas.
Table 2
Stroke types, cardiovascular risk factors and severity of patients with first-ever
stroke.
|
Variable
|
Campo Grande (n = 372)
|
Canoas (n = 205)
|
Joinville (n = 505)
|
Sertãozinho (n = 66)
|
Sobral (n = 107)
|
All (n = 1,255)
|
|
Stroke types
|
|
Hemorrhagic stroke
|
53 (14.2)
|
29 (14.1)
|
44 (8.7)
|
15 (22.7)
|
19 (17.8)
|
160 (12.7)
|
|
Ischemic stroke
|
308 (85.8)
|
176 (85.8)
|
461 (91.3)
|
51 (77.3)
|
88 (82.2)
|
1,084 (86.3)
|
|
Small artery occlusion
|
40 (13)
|
21 (12.0)
|
87 (18.8)
|
5 (9.8)
|
24 (27.3)
|
177 (16.3)
|
|
Large artery occlusion
|
16 (5.2)
|
23 (13.0)
|
77 (16.7)
|
5 (9.8)
|
15 (17.0)
|
136 (12.6)
|
|
Cardioembolic
|
34 (11)
|
18 (10.3)
|
110 (23.8)
|
7 (13.7)
|
8 (9.1)
|
177 (16.3)
|
|
Undetermined
|
218 (70.8)
|
114 (64.7)
|
187 (40.5)
|
34 (66.7)
|
41(46.6)
|
594 (54.8)
|
|
National Institute of Health Stroke Scale
|
|
0-3
|
123 (33.1)
|
93 (45.3)
|
246 (48.7)
|
12 (18.7)[1]
|
23 (22.8)[2]
|
497 (39.9)[3]
|
|
4–10
|
97 (26.1)
|
40 (19.5)
|
159 (31.5)
|
21 (32.8)
|
27 (26.7)
|
344 (27.6)
|
|
> 10
|
152 (40.8)
|
72 (35.2)
|
100 (19.8)
|
31 (48.5)
|
51 (50.5)
|
406 (32.5)
|
|
Risk factors
|
|
Hypertension
|
269 (72.3)
|
137 (66.8)
|
341 (67.5)
|
39 (59.0)
|
65 (60.7)
|
851 (67.8)
|
|
Diabetes
|
103 (27.7)
|
49 (23.9)
|
142 (28.1)
|
17 (25.7)
|
17 (15.9)
|
328 (26.1)
|
|
Dyslipidemia
|
UN
|
60 (29.2)
|
146 (28.9)
|
8 (12.1)
|
21 (19.6)
|
235 (18.7)
|
|
Atrial fibrillation
|
9 (2.4)
|
1 (0.4)
|
46 (9.1)
|
5 (7.6)
|
0
|
61 (4.8)
|
|
Myocardial infarction
|
16 (4.3)
|
11 (5.3)
|
36 (7.1)
|
5 (7.6)
|
3 (2.8)
|
71 (5.6)
|
|
Smoking
|
56 (15)
|
68 (33.1)
|
109 (21.6)
|
13 (19.7)
|
25 (23.3)
|
271 (21.6)
|
|
Physical activity
|
|
Inactive
|
248 (66.7)
|
163 (79.5)
|
353 (69.9)
|
43 (65.2)
|
57 (53.3)
|
864 (68.8)
|
|
Low activity
|
89 (23.9)
|
35 (17.1)
|
110 (21.8)
|
9 (13.6)
|
40 (37.4)
|
283 (22.6)
|
|
Medium
|
21 (5.7)
|
6 (2.9)
|
35 (6.9)
|
9 (13.6)
|
6 (5.6)
|
77 (6.1)
|
|
High activity
|
14 (3.7)
|
1 (0.5)
|
7 (1.4)
|
5 (7.6)
|
4 (3.7)
|
21 (2.5)
|
Data are number of patients (%) unless otherwise indicated; National Institute of
Health stroke scale only for ischemic strokes: Unavailable in: 2 patients1; 6 patients2;
8 patients3; 9 patients4; 68 patients5; 7 patients6; 2 patients7; one patients8; 17
patients9; 94 patients10
From 1,255 patients with first-ever ischemic stroke and hemorrhagic stroke, 64% (95%CI,
62–67) were overweight and 26% (95%CI, 24–29) were obese ([Tables 3] and [4]). More than double the overweight and obese patients were between 55 and 74 years
old. In patients with less than eight years of education, the prevalence of being
overweight was 78% (95%CI, 75–81) and the prevalence of obesity was 69% (95%CI, 63–74).
Despite social class C having more overweight patients than patients with a normal
BMI [(56% (95%CI, 53–60) vs 52% (95%CI, 48–57)], this difference was not significant
(OR 1.17, 95%CI, 0.9-1.5). This difference was also not significant when we compared
obese patients [(58% (95%CI, 52–63) with patients with normal BMI (OR 1.2, 95%CI,
0.9–1.7). As expected, comparing patients with normal weight versus overweight patients
who were both physically inactive, those who were overweight had a 40% higher risk
of stroke (OR 1.4, 95%CI, 1.09–1.8), reaching 75% higher odds of risk when compared
with obese patients (OR 1.75, 95%CI, 1.27–2.42).
Table 3
Prevalence of overweight and obesity status in stroke patients stratified by cities.
|
Variable
|
Campo Grande (n = 372)
|
Canoas (n = 205)
|
Joinville (n = 505)
|
Sertãozinho (n = 66)
|
Sobral (n = 107)
|
All (n = 1,255)
|
|
Overweight
|
250 (67,2)
|
141 (68,8)
|
323 (63.9)
|
33 (63.4)
|
60 (56.1)
|
807 (65.1)
|
|
Obesity
|
90 (24,2)
|
59 (28,8)
|
145 (28.8)
|
16 (30.7)
|
16 (15)
|
326 (26.3)
|
Data are number of patients (%) unless otherwise indicated.
Table 4
Normal BMI, overweight and obesity prevalence stratified by age, education, social
class, stroke type and severity and physical activity of patients with stroke.
|
Variable
|
BMI < 25 (n = 432)
|
BMI ≥ 25 (n = 807)
|
BMI ≥ 30 (n = 326)
|
|
Age (years)[1]
|
|
< 15
|
1 (0,2)
|
0
|
0
|
|
15–34
|
19 (4,4)
|
18 (2,2)
|
7 (2,1)
|
|
35–54
|
62 (14,4)
|
164 (20,3)
|
80 (24,6)
|
|
55–74
|
199 (46,1)
|
422 (52,4)
|
168 (51,5)
|
|
75–94
|
144 (33,3)
|
202 (25,0)
|
71 (21,8)
|
|
> 94
|
7 (1,6)
|
1 (0,1)
|
0
|
|
All
|
432 (34,9)
|
807 (65,1)
|
326 (26,3)
|
|
Education (years)[2]
|
|
< 4
|
149 (34,9)[6]
|
253 (31,5)[5]
|
88 (27,2)[3]
|
|
4–8
|
197 (46,0)
|
376 (46,9)
|
163 (50,5)
|
|
9–11
|
61 (14,2)
|
137 (17,1)
|
59 (18,3)
|
|
> 11
|
21 (4,9)
|
36 (4,5)
|
13 (4,0)
|
|
Social class[3]
|
|
A1
|
0
|
1 (0,1)
|
0
|
|
A2
|
5 (1,1)
|
8 (1,0)
|
3 (0,9)
|
|
B1
|
13 (3,0)
|
17 (2,1)
|
5 (1,5)
|
|
B2
|
73 (16,9)
|
178 (22,1)
|
82 (25,2)
|
|
C1
|
98 (22,8)
|
238 (29,5)
|
98 (30,1)
|
|
C2
|
128 (29,7)
|
217 (26,9)
|
90 (27,6)
|
|
D
|
110 (25,4)
|
146 (18,1)
|
48 (14,7)
|
|
E
|
5 (1,1)
|
2 (0,2)
|
0
|
|
Stroke type[4]
|
|
Ischemic
|
366 (84,7)
|
706 (87,5)
|
285 (87,4)
|
|
Small artery occlusion
|
64 (17,5)
|
113 (16,0)
|
46 (16,1)
|
|
Cardioembolic
|
62 (16,9)
|
114 (16,1)
|
53 (18,6)
|
|
Large artery occlusion
|
61 (16,7)
|
74 (10,5)
|
29 (10,2)
|
|
Undetermined
|
179 (48,9)
|
405 (57,4)
|
157 (55,1)
|
|
Hemorrhagic
|
66 (15,3)
|
101 (12,5)
|
41 (12,6)
|
|
National Institute of Health stroke scale[5]
|
|
< 4
|
162 (37,9)[5]
|
334 (41,5)[3]
|
128 (39,6)[3]
|
|
4–10
|
109 (25,6)
|
229 (28,5)
|
109 (33,7)
|
|
> 10
|
156 (36,5)
|
241 (30)
|
86 (26,7)
|
|
Physical activity[6]
|
|
Inactive
|
275 (63,6)
|
574 (71,2)
|
246 (75,5)
|
|
Low
|
119 (27,6)
|
164 (20,3)
|
58 (17,8)
|
|
Medium
|
24 (5,6)
|
52 (6,4)
|
17 (5,2)
|
|
High
|
14 (3,2)
|
17 (2,1)
|
5 (1,5)
|
BMI: body mass index; Data are number of patients (%) unless otherwise indicated BMI:
body mass index. Unavailable in: 16 patients1; 22 patients2; 3 patients3; 2 patients4;
5 patients5; 4 patients6; Social class according to the Research Companies Brazilian
Association.
DISCUSSION
In this cross-sectional study, with 1,255 patients from five distinct Brazilian macro-regions,
we found that the number of overweight patients with first-ever stroke were greater
than the number with a normal BMI and stroke. Most of the former had less than eight
years of education, belonged to social class C and were significantly more physically
inactive.
Obesity is a worldwide pandemic[1],[17],[18]. In 2010, the Global Burden of Disease Study reported that being overweight or obese
was estimated to cause 3.4 million deaths, 3.9% of years of life lost, and 3.8% of
disability-adjusted life-years worldwide[19]. According with this study, 7% of Brazilian men and 21% of women were obese in 2013[19].
As expected, these findings have an impact on the stroke burden. A recent survey compared
362,339 stroke hospitalizations in 2003–2004 to 421,815 hospitalizations in 2011–2012
in the EUA, and found an absolute increase from 4% to 9% in the prevalence of obesity
in among adults aged 18-64 years[20]. In a previous population-based study conducted in Joinville, we found that 16%
(95%CI 14–19) of 601 patients with first-ever ischemic stroke were obese in the 2005-2006
period. Six years later, (2012–2013 period), the obesity prevalence of 786 ischemic
stroke patients jumped significantly to 23% (95%CI, 20–27)[8]. This proportion is similar to our finding of 26% (95%CI, 24–29) in this study from
five Brazilian cities. [Table 5] shows the prevalence of patients with stroke being overweight or obese, from different
countries and periods[21],[22],[23],[24],[25],[26],[27]. Our findings were similar to a Finland cohort[22] but higher than those from Germany and China[23],[24],[25],[26],[27].
Table 5
Prevalence of overweight and obesity status among cohorts of stroke patients from
1951 to 2016.
|
Study
|
Year
|
Sample
|
Overweight prevalence
|
Obesity prevalence
|
|
Aparicio et al, USA[21]
|
1951–2011
|
782
|
68.5 (65.1–71.8)
|
24.3 (21.3–27.5)
|
|
Hu et al, Finland [22]
|
1972–2004
|
3,228
|
68.9 (67.3–70.5)
|
24.6 (23.1–26.1)
|
|
Wang et al, China[24]
|
1987–1998
|
1,108
|
32.4 (29.6–35.2)
|
4.7 (3.6–6.2)
|
|
Mitchell et al, EUA[25]
|
1992–2008
|
1,201
|
70.7 (68.1–75.9)
|
39.5 (36.7–42.3)
|
|
Park et al, USA/CA/SCO[26]
|
1997–2001
|
852
|
73 (69.9–75.9)
|
31.6 (28.5–34.8)
|
|
TEMPiS, Germany[27]
|
2003–2005
|
1,521
|
58 (55.5–60.5)
|
19.4 (17.4–21.5)
|
|
Joinville, Brazil[8]
|
2005–2006
|
610
|
––––
|
16.2 (13.4–15.4)
|
|
Kailuan, China[28]
|
2006–2007
|
1,547
|
55.7 (53.2–58.2)
|
9.9 (8.5–11.5)
|
|
Joinville, Brazil[8]
|
2012–2013
|
786
|
––––
|
23.4 (20.5–26.5)
|
|
5 Brazilian cities
|
2016
|
1,255
|
64.3 (61.6–66.9)
|
25.9 (23.9–28.5)
|
The causality between obesity and stroke is debatable[18],[19],[20]. A meta-analysis of 21 cohort studies reported that risk of ischemic stroke was
22% in patients who were overweight and 64% among those who were obese. For “primary”
intracerebral hemorrhage, the risk was not significant[28]. Nevertheless, other studies have shown that the association with obesity was substantially
attenuated after the control of hypertension and diabetes variables for obesity[24].
This is study has some limitations. As a cross-sectional analysis, we cannot assess
the causality of the associations. The BMI data were obtained by nurse interview and
have the possibility of information bias. Unfortunately, abdominal circumference measurement
and diet habits were not extracted. The strengths include the large sample extracted
from a prospective, population-based study, which showed, to the best of our knowledge,
a prevalence of higher BMI among stroke patients from a large middle-income country.
In conclusion, we found that the number of patients who were overweight and had a
stroke was greater than the number of patients with stroke and normal BMI. Furthermore,
by extracting data from cities from northeast, southeast and southern Brazil, our
findings endorse that we need strong educational campaigns focusing on prevention
of this condition.
