Introduction
Methods
Study definitions
Grouping and outcome definitions
Statistical analysis
Results
Baseline characteristics
Characteristics | BMI groups, kg/m2 | P-value | P for linear trend | |||
---|---|---|---|---|---|---|
< 18.5 | 18.5- < 23 | 23- < 27.5 | ≥ 27.5 | |||
N, % | 309, 2.18% | 3855, 27.25% | 7491, 52.95% | 2491, 17.61% | ||
Age (year), median (IQR) | 72 (61–78) | 65 (57–73) | 62 (54–69) | 60 (52–67) | < 0.0001 | < 0.0001 |
Male, N (%) | 177 (57.28%) | 2599 (67.42%) | 5303 (70.79%) | 1641 (65.88%) | < 0.0001 | 0.2082 |
Ethnicity(Han), N (%) | 299 (96.76%) | 3776 (97.95%) | 7268 (97.02%) | 2387 (95.82%) | < 0.0001 | < 0.0001 |
Stroke history, N (%) | 80 (25.89%) | 806 (20.91%) | 1663 (22.20%) | 585 (23.48%) | 0.0361 | 0.1016 |
Medical comorbidities, N ( %) | ||||||
Diabetes mellitus | 41 (13.27%) | 760 (19.71%) | 1829 (24.42%) | 680 (27.30%) | < 0.0001 | < 0.0001 |
Atrial fibrillation | 42 (13.59%) | 303 (7.86%) | 489 (6.53%) | 152 (6.10%) | < 0.0001 | < 0.0001 |
Myocardial infarction | 10 (3.24%) | 69 (1.79%) | 144 (1.92%) | 55 (2.21%) | 0.2572 | 0.6764 |
Hypertension | 159 (51.46%) | 2085 (54.09%) | 4835 (64.54%) | 1808 (72.58%) | < 0.0001 | < 0.0001 |
Lipid metabolism disorders | 16 (5.18%) | 226 (5.86%) | 572 (7.64%) | 261 (10.48%) | < 0.0001 | < 0.0001 |
Heavy drinkinga, N ( %) | 38 (12.30%) | 552 (14.32%) | 1075 (14.35%) | 345 (13.85%) | 0.7171 | 0.9632 |
Current smoker, N ( %) | 99 (32.04%) | 1210 (31.39%) | 2411 (32.19%) | 783 (31.43%) | 0.8085 | 0.8819 |
Waist circumference (cm), median (IQR)b | 72 (69–79) | 80 (75–86) | 87 (80–94) | 96.5 (88–105) | 0.0001 | < 0.0001 |
TOAST subtypes, N (%) | ||||||
Large-artery atherosclerosis | 78 (25.24%) | 992 (25.73%) | 1936 (25.84%) | 661 (26.54%) | 0.0016 | |
Small-vessel occlusion | 52 (16.83%) | 790 (20.49%) | 1723 (23.00%) | 572 (22.96%) | ||
Cardioembolism | 28 (9.06%) | 282 (7.32%) | 442 (5.90%) | 129 (5.18%) | ||
Stroke of other determined etiology | 5 (1.62%) | 50 (1.30%) | 86 (1.15%) | 30 (1.20%) | ||
Stroke of undetermined etiology | 146 (47.25%) | 1741 (45.16%) | 3304 (44.11%) | 1099 (44.12%) | ||
NIHSS at admission, median (IQR) | 5 (2–9) | 4 (2–6) | 3 (2–6) | 3 (2–6) | < 0.0001 | < 0.0001 |
Intravenous thrombolysis, N (%) | 55 (17.80%) | 465 (12.06%) | 751 (10.03%) | 249 (10.00%) | < 0.0001 | < 0.0001 |
Arterial thrombolysis or mechanical thrombectomy, N (%) | 3 (0.97%) | 25 (0.65%) | 32 (0.43%) | 11 (0.44%) | 0.2573 | 0.096 |
Characteristics | WHtR groups | P-value | P for linear trend | |||
---|---|---|---|---|---|---|
< 0.47 | 0.47- < 0.52 | 0.52- < 0.57 | ≥ 0.57 | |||
N | 1169 | 1316 | 1190 | 1130 | ||
Age (years), median (IQR) | 65 (57–74) | 64 (56–72) | 63 (55–70) | 63 (55–71) | < 0.0001 | < 0.0001 |
Male, N (%) | 824 (70.49%) | 943 (71.66%) | 790 (66.39%) | 657 (58.14%) | < 0.0001 | < 0.0001 |
Ethnicity(Han), N (%) | 1140 (97.52%) | 1264 (96.05%) | 1130 (94.96%) | 1049 (92.83%) | < 0.0001 | < 0.0001 |
Stroke history, N (%) | 192 (16.42%) | 269 (20.44%) | 237 (19.92%) | 234 (20.71%) | 0.0297 | 0.0191 |
Medical comorbidities, N (%) | ||||||
Diabetes mellitus | 279 (23.87%) | 274 (20.82%) | 268 (22.52%) | 286 (25.31%) | 0.0558 | 0.2682 |
Atrial fibrillation | 111 (9.50%) | 108 (8.21%) | 102 (8.57%) | 63 (5.58%) | 0.0042 | 0.0016 |
Myocardial infarction | 19 (1.63%) | 14 (1.06%) | 22 (1.85%) | 18 (1.59%) | 0.4208 | 0.6407 |
Hypertension | 711 (60.82%) | 819 (62.23%) | 776 (65.21%) | 801 (70.88%) | < 0.0001 | < 0.0001 |
Lipid metabolism disorders | 64 (5.47%) | 71 (5.40%) | 94 (7.90%) | 101 (8.94%) | 0.0005 | < 0.0001 |
Heavy drinkinga, N (%) | 134 (11.46%) | 188 (14.29%) | 140 (11.76%) | 130 (11.50%) | 0.0882 | 0.5481 |
Current smoker, N (%) | 315 (26.95%) | 419 (31.84%) | 359 (30.17%) | 296 (26.19%) | 0.0054 | 0.4982 |
Body mass index, median (IQR) | 22.84 (20.76–24.80) | 23.44 (21.72–25.12) | 24.68 (23.05–26.36) | 26.71 (24.49–29.05) | < 0.0001 | < 0.0001 |
TOAST subtypes, N (%) | ||||||
Large-artery atherosclerosis | 281 (24.04%) | 347 (26.37%) | 298 (25.04%) | 298 (26.37%) | 0.0575 | |
Small-vessel occlusion | 233 (19.93%) | 270 (20.52%) | 269 (22.61%) | 276 (24.42%) | ||
Cardioembolism | 97 (8.30%) | 91 (6.91%) | 94 (7.90%) | 59 (5.22%) | ||
Stroke of other determined etiology | 21 (1.80%) | 21 (1.60%) | 20 (1.68%) | 15 (1.33%) | ||
Stroke of undetermined etiology | 537 (45.94%) | 587 (44.60%) | 509 (42.77%) | 482 (42.65%) | ||
NIHSS at admission, median (IQR) | 3 (2–6) | 3 (2–6) | 3 (2–6) | 3 (2–6) | 0.0832 | 0.0609 |
Intravenous thrombolysis, N (%) | 143 (12.23%) | 171 (12.99%) | 128 (10.76%) | 94 (8.32%) | 0.0016 | 0.0007 |
Arterial thrombolysis or mechanical thrombectomy, N (%) | 12 (1.03%) | 10 (0.76%) | 4 (0.34%) | 2 (0.18%) | 0.0274 | 0.0028 |
Clinical outcomes
Number at risk | Number of events, N (%) | Unadjusted Hazard Ratio(95%CIa | P-value | Adjusted Hazard Ratio(95%CI) b | P-value | Pinteraction with stroke subtypec | |
---|---|---|---|---|---|---|---|
mortality | |||||||
BMI groups, kg/m2 | 0.6751 | ||||||
< 18.5 | 309 | 32 (10.36%) | 2.55 (1.75–3.73) | < 0.0001 | 1.70 (1.16–2.49) | 0.007 | |
18.5- < 23 | 3855 | 162 (4.20%) | 1.00 (Ref) | 1.00 (Ref) | |||
23- < 27.5 | 7491 | 235 (3.14%) | 0.74 (0.61–0.91) | 0.0035 | 0.94 (0.77–1.16) | 0.5779 | |
≥ 27.5 | 2491 | 57 (2.29%) | 0.54 (0.40–0.73) | < 0.0001 | 0.76 (0.56–1.03) | 0.0784 | |
WHtR groups | 0.6714 | ||||||
≥ 0.57 | 1130 | 35 (3.10%) | 0.64 (0.43–0.97) | 0.0357 | 0.70 (0.46–1.07) | 0.0977 | |
0.52- < 0.57 | 1190 | 25 (2.10%) | 0.43 (0.27–0.68) | 0.0004 | 0.42 (0.26–0.67) | 0.0003 | |
0.47- < 0.52 | 1316 | 63 (4.79%) | 1.00 (Ref) | 1.00 (Ref) | |||
< 0.47 | 1169 | 67 (5.73%) | 1.21 (0.86–1.71) | 0.2785 | 0.91 (0.64–1.29) | 0.5937 | |
Recurrence of stroke | |||||||
BMI groups, kg/m2 | 0.6263 | ||||||
< 18.5 | 309 | 36 (11.65%) | 1.22 (0.87–1.72) | 0.2525 | 1.13 (0.80–1.59) | 0.5005 | |
18.5- < 23 | 3855 | 382 (9.91%) | 1.00 (Ref) | 1.00 (Ref) | |||
23- < 27.5 | 7491 | 741 (9.89%) | 1.00 (0.88–1.13) | 0.9329 | 1.03 (0.91–1.17) | 0.6484 | |
≥ 27.5 | 2491 | 265 (10.64%) | 1.07 (0.92–1.26) | 0.3709 | 1.13 (0.96–1.32) | 0.1499 | |
WHtR groups | 0.9101 | ||||||
≥ 0.57 | 1130 | 132 (11.68%) | 1.07 (0.84–1.35) | 0.6012 | 1.10 (0.87–1.40) | 0.4345 | |
0.52- < 0.57 | 1190 | 101 (8.49%) | 0.76 (0.59–0.98) | 0.032 | 0.77 (0.60–0.99) | 0.0440 | |
0.47- < 0.52 | 1316 | 145 (11.02%) | 1.00 (Ref) | 1.00 (Ref) | |||
< 0.47 | 1169 | 111 (9.50%) | 0.86 (0.67–1.10) | 0.2339 | 0.83 (0.65–1.07) | 0.1440 | |
Combined vascular event | |||||||
BMI groups, kg/m2 | 0.6467 | ||||||
< 18.5 | 309 | 39 (12.62%) | 1.24 (0.90–1.73) | 0.1934 | 1.14 (0.82–1.58) | 0.4469 | |
18.5- < 23 | 3855 | 307 (10.56%) | 1.00 (Ref) | 1.00 (Ref) | |||
23- < 27.5 | 7491 | 784 (10.47%) | 0.99 (0.88–1.11) | 0.8354 | 1.02 (0.91–1.15) | 0.7288 | |
≥ 27.5 | 2491 | 275 (11.04%) | 1.05 (0.90–1.22) | 0.5664 | 1.10 (0.94–1.28) | 0.2474 | |
WHtR groups | 0.8802 | ||||||
≥ 0.57 | 1130 | 140 (12.39%) | 1.04 (0.83–1.30) | 0.7576 | 1.07 (0.85–1.35) | 0.5679 | |
0.52- < 0.57 | 1190 | 107 (8.99%) | 0.74 (0.58–0.94) | 0.0194 | 0.75 (0.58–0.95) | 0.0198 | |
0.47- < 0.52 | 1316 | 158 (12.01%) | 1.00 (Ref) | 1.00 (Ref) | |||
< 0.47 | 1169 | 115 (9.84%) | 0.82 (0.64–1.04) | 0.1005 | 0.79 (0.62–1.00) | 0.0511 |
Discussion
References | Design | Number of participants | Main Results |
---|---|---|---|
Liu Z et al. 2021 [18] | Placebo-controlled, randomized clinical trial | 1033 | BMI is inversely associated with short-term mortality, and U-shaped or J-shaped with short-term functional outcomes |
Oesch L et al. 2017 [19] | Twenty-five studies, non-randomized studies | 299,750 | One study showed the association between WHtR and mortality was U-shaped BMI and mortality after stroke: ten of twelve studies support the existence of an obesity paradox BMI and non-fatal outcome: seven of nine studies reported the association between higher body weight and improved non-fatal outcomes |
Rozen G et al. 2022 [20] | Real-world national cohort | 84,185 | Inverse association between BMI and in-hospital mortality |
Aparicio HJ et al. 2017 [14] | Nested case–control study | 782 (stroke patients) | Overweight and low obese patients but not high obese patients had reduced 10-year mortality |
Akyea RK et al. 2021 [21] | A prospective cohort study | 30,702 | Overweight or obesity was associated with better long-term outcomes, including lower risk of major adverse cardiovascular events and mortality |
Jang SY et al. 2015 [22] | A nested case study within a prospective nationwide cohort | 2057 | Extreme obesity (BMI > 30 kg/m2) is associated with short-term good functional outcomes, especially for the young |
Pirson FAV et al. 2019 [23] | A post hoc analysis of a national randomized trial for acute ischemic stroke | 366 | Higher BMI was associated with better short-term prognosis, including improved functional outcome, reduced mortality for large vessel occlusion patients BMI did not affect endovascular treatment effect |
Freeman C et al. 2022 [24] | Retrospective cohort study | 392 | The association between higher BMI and functional gains was affected by age, motor function on admission and diabetes |
Scherbakov N et al. 2011 [25] | a multicenter randomized trial and six non-randomized studies | 218,826 | Higher mortality in undernourished patients (a randomized trial) Inverse association of BMI and mortality (three studies, follow-up time 5–10 years) Increased BMI associated with high mortality (two Asian studies) Weight loss > 3 kg associated with increased mortality (a population-based study) |
Xu J et al. 2019 [26] | a nationwide prospective cohort | 1227 | A Chinese study showed that the BMI paradox existed in insulin-resistant patients but does not in insulin-sensitive ischemic stroke patients |