Table 2.
Subgroup analyses of the associations between NLR and modified Rankin Scale (mRS) assessed functional outcome in ischemic stroke.
| Stratified analyses | No. of patients | No. of studies | Model | Pooled ORs/RRs (95% CI) | P-value | PD-value | Heterogeneity | |
|---|---|---|---|---|---|---|---|---|
| I2 | PH-value | |||||||
| Assessment time | <0.001 | |||||||
| mRS at discharge | 1,244 | 4 | Random | 2.548 (0.954–6.805) | 0.062 | 85.1% | <0.001 | |
| mRS at 3 months | 2,891 | 9 | Random | 1.851 (1.325–2.584) | <0.001 | 91.7% | <0.001 | |
| mRS at 1 year | 371 | 1 | Random | 1.480 (0.753–2.910) | 0.256 | |||
| Demographic factors | ||||||||
| Baseline NIHSS score* | 0.010 | |||||||
| Mild | 1,429 | 3 | Random | 2.211 (0.673–7.257) | 0.191 | 86.4% | <0.001 | |
| Moderate to severe | 1,477 | 4 | Random | 1.614 (0.874–2.978) | 0.126 | 88.0% | <0.001 | |
| Age | <0.001 | |||||||
| <65 | 1,761 | 6 | Random | 1.508 (1.179–1.928) | 0.001 | 80.9% | <0.001 | |
| ≥65 | 1,899 | 5 | Random | 2.572 (1.062–6.229) | 0.036 | 90.5% | <0.001 | |
| Gender distribution | <0.001 | |||||||
| Male dominant | 3,018 | 9 | Random | 1.636 (1.307–2.047) | <0.001 | 75.6% | <0.001 | |
| Balanced | 906 | 2 | Random | 0.930 (0.829–1.043) | 0.216 | 83.2% | 0.003 | |
| Female dominant | 95 | 1 | Random | 6.400 (2.359–17.362) | <0.001 | - | - | |
| Country | <0.001 | |||||||
| Eastern | 2,719 | 10 | Random | 1.932 (1.496–2.494) | <0.001 | 81.6% | <0.001 | |
| Western | 1,416 | 3 | Random | 1.589 (0.580–4.354) | 0.367 | 71.6% | 0.030 | |
| Vascular risk factors | ||||||||
| Presence of hypertension | <0.001 | |||||||
| ≥55% and <65% | 1,935 | 3 | Random | 1.028 (0.826–1.280) | 0.806 | 89.3% | <0.001 | |
| ≥65% and <75% | 1,018 | 5 | Random | 3.187 (1.969–5.157) | <0.001 | 52.3% | 0.063 | |
| ≥75% | 971 | 3 | Fixed | 1.411 (1.254–1.588) | <0.001 | 0.0% | 0.706 | |
| Presence of diabetes mellitus | <0.001 | |||||||
| <25% | 2,450 | 5 | Random | 1.198 (0.948–1.514) | 0.130 | 88.6% | <0.001 | |
| ≥25% | 1,474 | 6 | Random | 2.258 (1.531–3.331) | <0.001 | 74.1% | 0.001 | |
| Presence of hyperlipidemia | 0.002 | |||||||
| <25% | 1,450 | 4 | Random | 2.463 (1.215–4.991) | 0.012 | 83.4% | <0.001 | |
| ≥25% | 2,271 | 5 | Random | 1.216 (0.899–1.644) | 0.204 | 91.3% | <0.001 | |
| Presence of current smoking | <0.001 | |||||||
| <35% | 1,008 | 4 | Random | 1.826 (1.178–2.831) | 0.007 | 57.2% | 0.071 | |
| ≥35% | 2,070 | 6 | Random | 1.881 (1.298–2.728) | 0.001 | 80.8% | <0.001 | |
| Methodological factors | ||||||||
| Onset time# | 0.002 | |||||||
| Within 24 h | 1,093 | 3 | Random | 1.821 (0.725–4.577) | 0.202 | 90.6% | <0.001 | |
| Within 48 h | 1,460 | 3 | Random | 1.485 (0.967–2.280) | 0.071 | 56.7% | 0.074 | |
| More than 48 h | 599 | 2 | Fixed | 1.432 (1.266–1.619) | <0.001 | 0.0% | 0.907 | |
| Sample time& | <0.001 | |||||||
| On admission | 1,075 | 5 | Random | 3.291 (1.514–7.157) | 0.003 | 78.5% | 0.001 | |
| Within 24 h | 694 | 3 | Random | 1.728 (1.157–2.580) | 0.007 | 76.6% | 0.014 | |
| Within 48 h | 825 | 2 | Random | 1.858 (1.130–3.053) | 0.015 | 18.6% | 0.293 | |
| Within 72 h | 635 | 1 | Random | 1.150 (1.046–1.264) | 0.004 | - | - | |
| Cutoff value | <0.001 | |||||||
| <4 | 2,264 | 8 | Random | 1.941 (1.472–2.559) | <0.001 | 83.0% | <0.001 | |
| ≥4 | 769 | 4 | Fixed | 3.469 (1.904–6.320) | <0.001 | 46.9% | 0.130 | |
| Definition of cutoff value | 0.267 | |||||||
| ROC curve analysis | 2,095 | 9 | Random | 2.306 (1.685–3.155) | <0.001 | 84.4% | <0.001 | |
| Continuous variable | 1,218 | 2 | Fixed | 0.913 (0.869–0.960) | <0.001 | 35.0% | 0.215 | |
| ORs/RRs calculation‡ | <0.001 | |||||||
| Multivariate | 3,038 | 9 | Random | 2.076 (1.384–3.112) | <0.001 | 86.2% | <0.001 | |
| Univariate | 1,097 | 4 | Random | 1.706 (1.200–2.426) | 0.003 | 88.1% | <0.001 | |
*
Baseline stroke severity was assessed by the NIH Stroke Scale (NIHSS) and categorized into two groups with moderate to severe (score of 5 to 20 points) and mild (0 to 4 points). Only one study (13) use Scandinavian Stroke Scale (SSS) to evaluate stroke severity.
#
Onset time was defined as time from stroke onset to recruitment/admission/diagnosis.
&
Sample time was defined as time from stroke onset to take blood sample.
‡
HRs were extracted from multivariate Cox proportional hazards models, univariate Cox proportional hazards models, or survival curve analysis.