Abstract
Introduction
Previous studies have suggested that elevated neutrophils, monocytes, and neutrophil-to-lymphocyte ratio (NLR) may be associated with poor outcomes in intracerebral hemorrhage (ICH). We sought to determine if white blood cell (WBC) types were independently associated with poor outcome in ICH in a large cohort.
Methods
We performed a retrospective study of primary ICH at two academic centers. Cases were identified via ICD-9 code and verified via physician review. We included only those patients with WBC types obtained within 24 hours of ICH onset.
Results
We identified 593 patients with primary ICH and WBC differentials in the first 24 hours. Independent factors (OR, 95% CI) associated with 30-day case-fatality were age >80 (2.4 (1.4, 4.2)), p=0.0023; NIHSS greater than median (3.9 (2.4, 6.3)), p<0.0001; ICH volume quartiles (Q1:ref, Q2: 1.5 (0.7, 3.0), Q3: 3.2 (1.6, 6.6), Q4: 11.9 (5.3, 26.4)), p<0.0001; non-lobar location (3.3 (1.9, 5.9)), p≤0.0001; IVH (2.3 (1.4, 3.6)), p=0.0005, monocytes greater than median (1.6 (1.0, 2.4)), p=0.0457, and anticoagulant use (3.2 (1.8, 5.6)), p<0.0001. Elevated NLR was not associated with higher case-fatality.
Discussion
We found that elevated monocytes were independently associated with 30-day case-fatality. Future studies will investigate whether there are subgroups of ICH patients, including those with particular blood or imaging biomarkers, in which WBC types might help predict poor outcome and provide targets for intervention.
Keywords: intracerebral hemorrhage, leukocytes, outcomes, inflammation
Introduction
Secondary injury is a contributor to poor outcome in intracerebral hemorrhage (ICH).1 The initial injury, due to the mechanical effects of the hematoma, is followed by an inflammatory response which includes leukocyte infiltration. Previous studies have reported that higher peripheral monocyte count is associated with increased 30-day case-fatality2,3 and hematoma expansion,4 while other studies have suggested that an elevated neutrophil to lymphocyte ratio (NLR) is associated with worse functional outcomes5–7 and perihematomal edema growth.8 The design of future clinical trials targeting secondary brain injury might benefit from a better understanding of the relationship between WBC types in the acute setting and patient outcomes. The sterile inflammatory response in ICH is a complex and dynamic process and data stratified by early time period are needed. We therefore sought to characterize the response of WBC types in the first 24 hours following ICH and to determine if WBC types were independently associated with poor outcome in ICH.
Methods
The Indiana University Institutional Review Board, the Indiana Network for Patient Care (INPC) board of directors, and Wishard Memorial Hospital approved this study.
Cohort assembly
We identified all patients ≥18 years old with primary ICH presenting to two academic centers via a query of the INPC database (http://www.ihie.org) for ICD-9 codes 431 and 432.9. The INPC is a health information exchange that serves multiple hospital systems in Indiana.9 For inclusion in this study the index ICH had to occur between January 1, 2009 and December 31, 2011 at Wishard Memorial Hospital and between January 1, 2009 and December 31, 2013 at IU Health Methodist Hospital. A vascular neurologist (J.M.) reviewed all of the cases. We excluded patients with traumatic ICH or an aneurysm, encephalitis, or brain tumor as a cause of the hemorrhage. We also excluded patients with hemorrhagic transformation of an ischemic stroke and ICHs related to cerebral venous sinus thrombosis, carotid endarterectomy, or thrombolytic administration for ischemic stroke.
Clinical data abstraction
Under the close supervision of the vascular neurologist, abstractors ascertained demographics, vascular risk factors, and processes of care. All available referring hospital and transfer data were reviewed. The neurologist reviewed the initial imaging scan from the academic center for each patient as well as all available brain scans from the referring hospital. We used a validated method for the NIH stroke scale (NIHSS) estimation if a formal score was not reported at presentation.10 We estimated hematoma volume with the ABC/2 method11 and intraventricular (IVH) volume with a validated score.12
For all analyses in this study, we included only those patients with WBC differentials obtained within 24 hours of ICH onset. We determined onset times (or last seen normal times) via chart review. We calculated the neutrophil-to-lymphocyte ratio (NLR) as the absolute neutrophil count divided by the absolute lymphocyte count. We determined vital status via chart review and obituary query and performed a National Death Index (http://www.cdc.gov/nchs/ndi.htm) query for the vital status of all patients for whom we still could not account. All clinical data were recorded in REDCap.13
Statistical Methods
The primary purpose of this study was to assess the association of WBC types with 30-day case-fatality after adjustment for other significant factors. We first performed univariate analyses to evaluate association with 30-day case-fatality and then built a multivariate regression model from the set of factors with a p-value < 0.20. We included age in the multivariate model because of higher likelihood of 30-day case fatality with older patients; this provides an age-adjusted association for other covariates with 30-day case fatality. We used the stepwise regression approach to build the model. We performed additional univariate analyses assessing NLR and multiple outcomes as well. Our secondary outcome was time to death within one year of stroke. We applied the log-rank test to compare the survival curves between patients with monocyte values greater than the median versus those with values less than or equal to the median. We used the Kruskal Wallis test to characterize the differences in patient characteristics stratified by time from last normal to sample draw: 0–6 hours, 6–12 hours, 12–18 hours and 18–24 hours. We also assessed how the WBC types changed over time when measured at two time points in the first 24 hours in the same patient and used the Wilcoxon Signed-Rank test to analyze these differences. To assess the performance of factors on predicting outcomes we used Receiver Operating Curve (ROC) analyses along with the Youden Index, which is the maximum value for all points on the ROC curve. The maximum value corresponds to the optimum cut point of a numeric predictor and this cut point provides the best balance of sensitivity and specificity of that predictor as a diagnostic test.14,15
Results
From an overall cohort of 889 ICH patients, we excluded 128 because there was no differential obtained. We additionally excluded 86 patients with unknown onset time and 74 patients with differentials only obtained >24 hours after stroke onset. We also excluded 8 patients for whom we did not have 30-day case-fatality data, which left 593 patients for this analysis. The excluded patients presented with lower median NIHSS scores (7 vs 10, p=0.0001), lower median ICH volumes (8.4mL vs. 9.4mL, p=0.0398), and a lower proportion of IVH (42.2% vs. 49.7%, p=0.0344) than the patients included in the analysis. (Please see the Supplementary Table for the univariate analysis comparing excluded and included patients.)
30-day case-fatality analysis
The 30-day case-fatality was 191/593 (32.2%) in the overall cohort: 60/222 (27.0%) in lobar ICH, 96/272 (35.3%) in deep ICH, 20/38 (52.6%) in brainstem ICH, 13/53 (24.5%) in cerebellar ICH, and 2/8 (25%) in primary IVH. The univariate analysis of factors for 30-day case-fatality is shown in Table 1. Thirty-day case-fatality was positively associated (OR>1) with NIHSS score >10 (median), higher hematoma volume, non-lobar location, presence of IVH, presence of SAH, WBC count >9 (median), and monocyte count >0.6 (median). However, in multivariate analysis (Table 2), significant factors (OR (95% CI)) for 30-day case-fatality were age >80 (2.4 (1.4, 4.2)), p=0.0023; NIHSS greater than median (3.9 (2.4, 6.3)), p<0.0001; ICH volume quartiles (Q1:ref, Q2: 1.5 (0.7, 3.0), Q3: 3.2 (1.6, 6.6), Q4: 11.9 (5.3, 26.4)), p<0.0001; non-lobar location (3.3 (1.9, 5.9)), p≤0.0001; IVH (2.3 (1.4, 3.6)), p=0.0005, monocytes greater than median (1.6 (1.0, 2.4)), p=0.0457, and anticoagulant use (3.2 (1.8, 5.6)), p<0.0001. We performed an additional analysis excluding patients transitioned to comfort measures in the first 24 hours (n=39) and monocytes were no longer significant (OR 1.44 (95% CI 0.92, 2.26), p=0.113)).
Table 1:
Univariate association of clinical and demographic factors with 30-day case-fatality
| Dead at 30 Days |
OR (95% CI) | p-value | |
|---|---|---|---|
| N (%) | |||
| Age | |||
| ≤ 80 | 149 (31.2%) | 1.00 (– –) | 0.3048 |
| > 80 | 42 (36.2%) | 1.25 (0.82, 1.91) | |
| Sex | |||
| Male | 105 (32.6%) | 1.00 (– –) | 0.8205 |
| Female | 86 (31.7%) | 0.96 (0.68, 1.36) | |
| Race | |||
| Non-White | 38 (26.8%) | 1.00 (– –) | 0.1121 |
| White | 153 (33.9%) | 1.41 (0.92, 2.14) | |
| Initial NIHSS | |||
| ≤ 10 (median) | 47 (14.9%) | 1.00 (– –) | <.0001 |
| >10 | 144 (52.0%) | 6.20 (4.20, 9.15) | |
| Charlson | |||
| 0–1 (median) | 113 (30.5%) | 1.00 (– –) | 0.2630 |
| >1 | 78 (35.0%) | 1.22 (0.86, 1.74) | |
| HTN | |||
| Yes | 157 (31.9%) | 0.91 (0.58, 1.43) | 0.6743 |
| No | 34 (34.0%) | 1.00 (– –) | |
| Anticoagulant use | |||
| Yes | 48 (48.5%) | 2.31 (1.49, 3.59) | 0.0002 |
| No | 143 (29.0%) | 1.00 (– –) | |
| Antiplatelet use | |||
| Yes | 70 (32.3%) | 1.004 (0.70, 1.44) | 0.9845 |
| No | 121 (32.2%) | 1.00 (– –) | |
| ICH volume (mL) | |||
| Q1 (0–2.6) | 18 (11.9%) | 1.00 (– –) | <.0001 |
| Q2 (2.7–10) | 33 (21.7%) | 2.05 (1.10, 3.83) | |
| Q3 (10.1–28.9) | 48 (33.8%) | 3.77 (2.07, 6.89) | |
| Q4 (29–187.5) | 91 (62.3%) | 12.23 (6.74, 22.17) | |
| Lobar | |||
| Yes | 60 (27.0%) | 1.00 (– –) | 0.0372 |
| No | 131 (35.3%) | 1.47 (1.02, 2.12) | |
| IVH | |||
| Yes | 138 (46.8%) | 4.06 (2.79, 5.91) | <.0001 |
| No | 53 (17.8%) | 1.00 (– –) | |
| SAH | |||
| Yes | 38 (53.5%) | 2.78 (1.68, 4.59) | <.0001 |
| No | 153 (29.3%) | 1.00 (– –) | |
| WBC | |||
| ≤ 9 (median) | 81 (27.0%) | 1.00 (– –) | 0.0062 |
| > 9 | 110 (37.5%) | 1.63 (1.15, 2.30) | |
| Absolute Neutrophil | |||
| ≤ 6.1 (median) | 92 (30.3%) | 1.00 (– –) | 0.2985 |
| > 6.1 | 99 (34.3%) | 1.20 (0.85, 1.70) | |
| Absolute Lymphocyte | |||
| ≤ 1.5 (median) | 98 (32.2%) | 1.00 (– –) | 0.9882 |
| > 1.5 | 93 (32.2%) | 1.00 (0.71, 1.41) | |
| Absolute Monocyte | |||
| ≤ 0.6 (median) | 97 (27.8%) | 1.00 (– –) | 0.0061 |
| > 0.6 | 94 (38.5%) | 1.63 (1.15, 2.31) | |
| Absolute Eosinophil | |||
| ≤ 0.1 (median) | 125 (33.2%) | 1.00 (– –) | 0.4776 |
| > 0.1 | 66 (30.4%) | 0.88 (0.61, 1.26) | |
| Absolute Basophil | |||
| = 0 (median) | 121 (31.3%) | 1.00 (– –) | 0.5007 |
| > 0 | 70 (34.0%) | 1.13 (0.79, 1.62) | |
| NLR | |||
| ≤ 3.8 (median) | 94 (31.5%) | 1.00 (– –) | 0.7274 |
| > 3.8 | 97 (32.9%) | 1.06 (0.75, 1.50) | |
| Platelets | |||
| ≤ 217 (median) | 93 (31.5%) | 1.00 (– –) | 0.7032 |
| > 217 | 97 (33.0%) | 1.07 (0.76, 1.51) | |
| Onset to Presentation | |||
| 0–6 hours | 116 (31.8%) | 1.00 (– –) | 0.3606 |
| 6–12 hours | 54 (37.0%) | 1.26 (0.84, 1.88) | |
| 12–18 hours | 15 (26.3%) | 0.77 (0.41, 1.44) | |
| 18–24 hours | 6 (24.0%) | 0.68 (0.26, 1.74) | |
| Onset to First Lab Draw | |||
| 0–6 hours | 100 (33.1%) | 1.00 (– –) | 0.5479 |
| 6–12 hours | 64 (33.5%) | 1.02 (0.69, 1.50) | |
| 12–18 hours | 17 (30.4%) | 0.88 (0.48, 1.63) | |
| 18–24 hours | 10 (22.7%) | 0.59 (0.28, 1.25) | |
| Comfort Care within 24 hours | |||
| Yes | 39 (100.0%) | -- | -- |
| No | 152 (27.4%) | ||
Abbreviations:
NIHSS: NIH Stroke Scale score
HTN: Hypertension
IVH: Intraventricular hemorrhage
SAH: Subarachnoid hemorrhage
WBC: White Blood Cell count
NLR: Neutrophil to Lymphocyte Ratio
Table 2:
Multivariate association of clinical and demographic factors with 30-day case-fatality
| Model | OR (95% CI) | p-value |
|---|---|---|
| Age | ||
| ≤ 80 | 1.00 (– –) | 0.0023 |
| > 80 | 2.40 (1.37, 4.22) | |
| Initial NIHSS | ||
| ≤ 10 (median) | 1.00 (– –) | <.0001 |
| >10 | 3.89 (2.40, 6.31) | |
| ICH volume (mL) | ||
| Q1 (0–2.6) | 1.00 (– –) | <.0001 |
| Q2 (2.7–10) | 1.48 (0.74, 2.96) | |
| Q3 (10.1–28.9) | 3.22 (1.58, 6.57) | |
| Q4 (29–187.5) | 11.87 (5.34, 26.37) | |
| Lobar | ||
| Yes | 1.00 (– –) | ≤0.0001 |
| No | 3.31 (1.86, 5.89) | |
| IVH | ||
| Yes | 2.25 (1.42, 3.55) | 0.0005 |
| No | 1.00 (– –) | |
| Absolute Monocyte | ||
| ≤ 0.6 (median) | 1.00 (– –) | 0.0457 |
| > 0.6 | 1.55 (1.01, 2.38) | |
| Anticoagulant use | ||
| Yes | 3.20 (1.83, 5.60) | <.0001 |
| No | 1.00 (--) | |
Abbreviations:
NIHSS: NIH Stroke Scale score
IVH: Intraventricular Hemorrhage
NLR analysis
We did not identify an independent association between NLR and case-fatality at any time point. In univariate logistic regression we did not find an association between values greater than our median NLR (3.8) or a previously cited optimal cut point of 4.585 and discharge mRS or case-fatality at any time point out to 1 year. The Youden cut point was 9.29 (above the 75th percentile) and in univariate was associated with a higher case-fatality at 30 days and 90 days. In addition, when we accounted for time from presentation to blood draw in our model, we observed additional significant associations of NLR with discharge Rankin score of ≥3 (p=0.034) and one-year case-fatality (p=0.026). After adjustment for baseline NIHSS values, however, these associations were no longer present (data shown in Table 3).
Table 3:
Univariate logistic regressions for NLR and monocyte cut points
| Discharge Rankin Score ≥3 | 30-Day case-fatality | 90-Day case-fatality | 1-Year case-fatality | |
|---|---|---|---|---|
| NLR ≥4.585 | 1.18 (0.76, 1.84) AUC=0.5198 p=0.4698 | 1.03 (0.73, 1.45) AUC=0.5032 p=0.8822 | 1.00 (0.71, 1.40) AUC=0.5006 p=0.9785 | 0.95 (0.68, 1.32) AUC=0.5070 p=0.7395 |
| Our Median: NLR >3.8 | 1.17 (0.76, 1.82) AUC=0.5201 p=0.4702 | 1.06 (0.75, 1.50) AUC=0.5077 p=0.7274 | 1.02 (0.72, 1.42) AUC=0.5019 p=0.9308 | 0.94 (0.68, 1.31) AUC=0.5076 p=0.7202 |
| Youden Cut Point*: NLR ≥9.286 | 1.84 (0.99, 3.43) AUC=0.5439 p=0.0541 | 1.53 (1.01, 2.31) AUC=0.5362 p=0.0428^ | 1.60 (1.06, 2.41) AUC=0.5397 p=0.0251^ | 1.49 (0.99, 2.24) AUC=0.5328 p=0.0569 |
| Abs. monocyte (log) 2 | 1.27 (0.82, 1.96) AUC=0.5330 p=0.2725 | 1.48 (1.03, 2.12) AUC=0.5607 p=0.0324^^ | 1.32 (0.93, 1.87) AUC=0.5410 p=0.1193 | 1.37 (0.97, 1.91) AUC=0.5395 p=0.0704 |
| Our Median: Abs. monocyte > 0.6 | 1.16 (0.74, 1.82) AUC=0.5179 p=0.5114 | 1.63 (1.15, 2.31) AUC=0.5595 p=0.0061 | 1.36 (0.96, 1.92) AUC=0.5375 p=0.0808 | 1.28 (0.92, 1.80) AUC=0.5305 p=0.1447 |
| Youden Cut Point: Abs. monocyte ≥ ** | 1.02 (0.63, 1.67) AUC=0.5021 p=0.9335 | 1.04 (0.65, 1.66) AUC=0.5023 p=0.8864 | 1.04 (0.65, 1.65) AUC=0.5023 p=0.8847 | 1.14 (0.79, 1.66) AUC=0.5133 p=0.4818 |
Values reported as OR (95%CI)
Abbreviations:
NLR: Neutrophil to Lymphocyte Ratio
AUC: Area Under the Curve
NLR Youden cut point was the same for 30-day, 90-day, and 1-year outcomes.
Monocyte Youden cut points vary by outcome–0.713, 0.331, 0.372, and 0.495 respectively
NLR effect disappears when adjusted for baseline NIHSS.
Monocyte effect disappears when adjusted for baseline NIHSS.
Monocyte analysis
We performed additional analyses evaluating monocytes and outcomes. In univariate logistic regression we found an association between values greater than our median monocyte count (0.6) and 30-day case-fatality, which remained significant after adjustment for other variables. We also found an association between log absolute monocytes2 and 30-day case-fatality; after adjustment for baseline NIHSS, however, this association was no longer present. The Youden cut points varied by outcome but were not significant. (Data shown in Table 3.)
Survival analysis within one year of stroke
The Figure shows survival curves within one year of stroke stratified by median value of monocyte. (We do not have one-year case-fatality data for 29 of the 593 patients.) There was a significant difference in survival history between the two groups (p-value = .033). For example, 25% of the patients in the lower than the median group died within 20 days (95% CI: 8, 61) of ICH compared with 5 days (95% CI: 3, 11) of ICH in the other group. In an exploratory analysis we then assessed how these two groups were different in terms of patient characteristics (Table 4). The only variable associated with higher monocyte values with a p-value <0.05 was anticoagulant use (p=.0382).
Table 4:
Patient characteristics stratified by monocyte value
| Monocytes greater than median value of 0.6 at baseline | ||||
|---|---|---|---|---|
| Variable | Overall N=593 | No N=349 | Yes N=244 | P-value |
| Age >80 | 116 (19.6%) | 71 (20.3%) | 45 (18.4%) | 0.5657 |
| Women | 271 (45.7%) | 171 (49.0%) | 100 (41.0%) | 0.0539 |
| White | 451 (76.1%) | 266 (76.2%) | 185 (75.8%) | 0.9110 |
| ICH volume >30mL | 143 (24.2%) | 80 (23.1%) | 63 (25.8%) | 0.4397 |
| NIHSS >10 | 277 (46.7%) | 153 (43.8%) | 124 (50.8%) | 0.0936 |
| Charlson Index | 1 (0, 2) | 1 (0, 2) | 1 (0, 2) | 0.5466 |
| Hypertension | 493 (83.1%) | 283 (81.1%) | 210 (86.1%) | 0.1112 |
| Antiplatelet use | 217 (36.6%) | 118 (33.8%) | 99 (40.6%) | 0.0925 |
| Anticoagulant use | 99 (16.7%) | 49 (14.0%) | 50 (20.5%) | 0.0382 |
| Baseline Rankin 0–1 | 398 (67.1%) | 236 (67.6%) | 162 (66.4%) | 0.7540 |
| Prior ICH | 62 (10.5%) | 31 (8.9%) | 31 (12.7%) | 0.1375 |
| Prior ischemic stroke | 132 (22.3%) | 75 (21.6%) | 57 (23.4%) | 0.6027 |
| First SBP | 183.6 ± 38.7 | 182.8 ± 38.3 | 184.8 ± 39.2 | 0.5375 |
| First DBP | 99.8 ± 26.1 | 98.9 ± 25.6 | 101.1 ± 26.9 | 0.3107 |
| Lobar | 222 (38.0%) | 132 (37.8%) | 90 (36.9%) | 0.8165 |
| IVH | 295 (49.8%) | 167 (47.9%) | 128 (52.5%) | 0.2694 |
| SAH | 71 (12.0%) | 45 (12.9%) | 26 (10.7%) | 0.4087 |
| NLR | 3.8 (2.2, 8.1) | 3.9 (2.2, 7.5) | 3.6 (2.1, 9.0) | 0.7488 |
Abbreviations:
NIHSS: NIH Stroke Scale score
SBP: Systolic Blood Pressure
DBP: Diastolic Blood Pressure
IVH: Intraventricular hemorrhage
SAH: Subarachnoid hemorrhage
NLR: Neutrophil to Lymphocyte Ratio
Time interval analysis of sample draw
To explore the relationship of time with onset to WBC differential, we stratified the cohort into 6-hour increments. Of the 593 patients included, 302 (50.9%) had their first samples drawn in 0–6 hours after stroke onset, 191 (32.2%) in the 6–12 hour window, 56 (9.4%) in the 12–18 hour window, and 44 (7.4%) in the 18–24 hour window. Demographic, risk factor, WBC type, and imaging data stratified by time window are shown in Table 5. The major findings were that more severely affected patients presented earlier and that there were significant differences in WBC count, absolute neutrophils, absolute lymphocytes, absolute eosinophils, and NLR. The monocyte count was not different across time windows.
Table 5:
Patient characteristics stratified by time from onset to sample draw
| 0–6 hours (n=302) | 6–12 hours (n=191) | 12–18 hours (n=56) | 18–24 hours (n=44) | p-value | |
|---|---|---|---|---|---|
| Age (mean, SD) | 65 (56, 78) | 67 (56, 78) | 68.5 (61.5, 78) | 68 (57.5, 80.5) | 0.4824 |
| Women (%) | 128 (42.4%) | 95 (49.7%) | 27 (48.2%) | 21 (47.7%) | 0.4221 |
| White (%) | 223 (73.8%) | 151 (79.1%) | 45 (80.4%) | 32 (72.7%) | 0.4584 |
| NIHSS (median, IQR) | 10 (4, 23) | 12 (4, 21) | 6.5 (3, 14) | 5.5 (3, 14) | 0.0117 |
| GCS (median, IQR) | 13 (8.5, 15) | 14 (9, 15) | 15 (13, 15) | 15 (12.5, 15) | 0.0005 |
| Charlson (median, IQR) | 1 (0, 2) | 1 (0, 2) | 1 (1, 2) | 1 (0, 3.5) | 0.2010 |
| HTN (%) | 248 (82.1%) | 161 (84.3%) | 44 (78.6%) | 40 (90.9%) | 0.3715 |
| Antiplatelet use | 108 (35.8%) | 61 (31.9%) | 25 (44.6%) | 23 (52.3%) | 0.0440 |
| Anticoagulant use | 44 (14.6%) | 39 (20.4%) | 11 (19.6%) | 5 (11.4%) | 0.2473 |
| ICH volume, mL (median, IQR) | 11.6 (2.3, 34.1) | 9.4 (3.0, 26.3) | 8.4 (4, 19.3) | 7.3 (2.1, 19.1) | 0.6097 |
| Lobar (%) | 109 (36.1%) | 65 (34.0%) | 29 (51.8%) | 19 (43.2%) | 0.0813 |
| IVH (%) | 147 (48.7%) | 102 (53.4%) | 24 (42.9%) | 22 (50.0%) | 0.5271 |
| SAH (%) | 35 (11.6%) | 24 (12.6%) | 8 (14.3%) | 4 (9.1%) | 0.8645 |
| WBC (median, IQR) | 8.5 (6.6, 11.2) | 9.7 (7.7, 12.1) | 8.9 (7.0, 11.7) | 9.4 (7.6, 12.4) | 0.0328 |
| Abs. neutrophil (median, IQR) | 5.3 (3.8, 8.1) | 7.1 (4.9, 9.5) | 6.3 (4.4, 9.5) | 7.3 (5.5, 10) | <0.0001 |
| Abs. lymphocyte (median, IQR) | 1.7 (1.2, 2.5) | 1.4 (0.9, 2.2) | 1.3 (1.0, 1.8) | 1.1 (0.9, 1.6) | <0.0001 |
| Abs. monocyte (median, IQR) | 0.6 (0.4, 0.8) | 0.6 (0.4, 0.8) | 0.6 (0.5, 0.8) | 0.6 (0.3, 0.9) | 0.8255 |
| Abs. eosinophil (median, IQR) | 0.1 (0.0, 0.2) | 0.1 (0, 0.2) | 0.1 (0, 0.2) | 0.1 (0, 0.1) | 0.0001 |
| Abs. basophil (median, IQR) | 0 (0, 0.1) | 0 (0, 0.1) | 0 (0, 0.1) | 0 (0, 0.1) | 0.4346 |
| NLR (median, IQR) | 3.1 (1.9, 5.5) | 4.7 (2.5, 9.8) | 4.7 (3.2, 9.9) | 6.8 (3.9, 10.9) | <0.0001 |
| Platelets | 215.5 (173.5, 270) | 223.5 (182, 275) | 211.5 (175, 264) | 206 (186, 295) | 0.7264 |
| Comfort care within 24 hours | 25 (8.3%) | 9 (4.7%) | 3 (5.4%) | 2 (4.6%) | 0.4018 |
Abbreviations:
NIHSS: NIH Stroke Scale score
GCS: Glasgow Coma Score
HTN: Hypertension
IVH: Intraventricular hemorrhage
SAH: Subarachnoid hemorrhage
WBC: White Blood Cell count
NLR: Neutrophil to Lymphocyte Ratio
Within-patient subgroup analysis of sample draw
To explore WBC type values longitudinally, we analyzed within-patient variability in a subgroup analysis. We identified 198 patients with multiple WBC differentials in the first 24 hours following stroke: 172 had two differentials, 25 had three differentials, and 1 had five differentials. The medians of the first and last measurements within the first 24 hours for each patient are shown in Table 6. There were statistically significant differences for each type, though the most striking differences were increases in WBCs, neutrophils, and NLR, and a decrease in lymphocytes. The proportions of patients with increasing values in WBC, in neutrophils, and in NLR were 59%, 71% and 78% respectively, while the proportion of patients with increasing values in lymphocytes was only 19% and in monocytes only 38%. The overall 30-day case-fatality in this subgroup with multiple WBC differentials in the first 24 hours following stroke was 68 out of 198 (34.3%). We did not find statistically significant differences in case-fatality stratified by decreasing, same, or increasing values in any of the WBC types (data not shown). In a separate analysis, we evaluated a subgroup of 90 patients with a first test within 6 hours of ICH and a second test between 6 and 24 hours. We found no difference in the changes of monocyte values, but non-survivors at 30 days had a significantly (p=0.034) higher increase in NLR (median increase of 6.6) than the survivors (median increase of 2.5).
Table 6:
Within-patient changes in WBC types in the first 24 hours after ICH (n=198)
| Measure | Time 1 Median, (IQR) | Time 2 Median, (IQR) | p-value (Wilcoxon Sign-Rank) |
|---|---|---|---|
| WBC | 9.4 (6.8, 11.9) | 10.2 (7.4, 13.3) | <0.0001 |
| Abs. neutrophil | 6.1 (4.4, 8.8) | 8.4 (5.8, 11.1) | <0.0001 |
| Abs. lymphocyte | 1.6 (1.0, 2.5) | 1.1 (0.7, 1.5) | <0.0001 |
| Abs. monocyte | 0.6 (0.5, 0.8) | 0.6 (0.4, 0.8) | 0.0383 |
| Abs. eosinophil | 0.1 (0, 0.2) | 0 (0, 0.1) | <0.0001 |
| Abs. basophil | 0 (0, 0.1) | 0 (0, 0) | 0.0012 |
| NLR | 3.57 (2.12, 7.60) | 8.10 (5.09, 14.22) | <0.0001 |
Abbreviations:
WBC: White Blood Cell count
NLR: Neutrophil to Lymphocyte Ratio
Discussion
In this study we found that higher initial peripheral monocyte count was independently associated with higher 30-day case-fatality in ICH and further found a survival difference that persisted at one year. We also found that the general trend in the first 24 hours after ICH was increased neutrophils, decreased lymphocytes, and increased NLR. We did not identify an independent association between elevated WBC count, neutrophils, or NLR and worse outcome.
The primary injury in ICH is mechanical and results in tissue disruption and mass effect. Secondary injury follows and includes a sterile inflammatory response. Leukocyte infiltration has attracted attention as a possible contributor to poor outcome. Several previous studies have evaluated the association between elevated WBC count and poor outcome in ICH with mixed results.16–19 An analysis of the INTERACT trial data found that higher WBC was associated with poor outcome in univariate but was not independently associated with poor outcome after adjustment for other variables.20
Recent attention has focused on the role of individual WBC types in sterile inflammation in ICH. Neutrophils are the first WBC type to enter the hematoma bed after ICH1 and the relative increase in neutrophils is commonly reported as the ratio of absolute neutrophils to absolute lymphocytes (NLR). NLR is an easily obtained and inexpensive way to measure inflammation and has been associated with adverse outcomes in stroke patients.21–23 A recent study in ICH patients in Italy found that higher neutrophils, lower lymphocytes, and a higher NLR was associated with poor 3 month outcome.5 A Chinese group found a similar independent association with higher NLR and poor functional outcome at 3 months,7 while conversely a German group found an independent association with inpatient mortality but not poor functional outcome at 3 months.6 A group in New York found that NLR was associated with perihematomal edema growth.8 Stratifying by time from onset to sample draw in our overall cohort, we found that NLR appears to increase in the first 24 hours after ICH and is driven by an increase in neutrophils. These findings were consistent with our subset of patients with multiple draws in the first 24 hours as well, as more than 70% were noted to have increasing neutrophils and NLRs in that time frame. Our median NLR of 3.8 is lower than that reported by the groups noted above (with baseline NLRs ranging from 4.66 to 7.3). More than half of our patients (302 of 593) had their blood drawn within 6 hours of symptom onset and the median NLR for that subset was 3.1 (Table 5). The higher NLR in the Italian study might be related to a mean onset-to-sample time of 17.3 hours (15.8–19.0). But the Chinese group’s onset-to-sample time was 4 hours (3–7) and the New York group’s was 5.4 hours (3.4–11.4), and their NLRs were 7.3 and 5.4, respectively. The reasons for NLRs higher than ours in these groups also with early samples are unclear. Higher NLR was not independently associated with poor outcome at any time point in our study, however. The reasons for the differences between our study and those finding an association with poor outcome at 3 months are unclear but might be related to size of and differences in cohorts. In contrast to the Italian group our cohort has more HTN (83% vs 66%, fewer deep ICHs (46% vs 64%), and more intraventricular extension (50% vs 23%), for example. The Chinese cohort was younger (50 vs 66), more likely men (74% vs 54%), had larger ICHs (17mL vs 9mL), and much lower case-fatality rates (11.7% at 90 days vs 32.2% at 30 days in our cohort). It is also possible that obtaining a differential in the first 6 hours might be too early to see significant alterations in WBC types and not predictive of overall trajectory or outcome. In a subanalysis (n=90) we found that the median NLR increase in the first 24 hours in patients who died by 30 days was higher than that of those who survived.
A few clinical studies have assessed monocytes in ICH. An analysis of the GERFHS cohort found that monocytes were independently associated with 30-day case-fatality2 and a larger cohort from the ERICH study reported similar findings.3 This study confirms and extends those findings by including a large cohort with stratification by early time windows, within-patient analyses over the first 24 hours, an exploratory analysis comparing higher than median monocyte counts vs lower than median counts, and case-fatality data out to one year post-ICH. In another large cohort the Boston group found that a higher neutrophil count was associated with a lower risk of hematoma expansion and that elevated monocyte count was associated with higher risk of hematoma expansion.4 Higher monocyte counts have been associated with worse outcomes in sterile inflammation in several conditions.24 The reasons for higher case-fatality in our cohort are unclear. We did not find striking baseline differences in the usual predictors of poor ICH outcome or significant differences in 30-day case-fatality in patients with increasing or decreasing monocyte counts in the first 24 hours. Interestingly, when patients who were transitioned to comfort care within 24 hours of ICH were excluded from the analysis, monocytes were no longer independently associated with 30-day case-fatality. One possibility is that patients with higher monocyte counts were more likely to have hematoma expansion and in turn more likely to transition to early comfort care. We cannot comment on whether elevated monocytes were associated with hematoma expansion as we do not have serial head CT data.
Monocyte response is complex because monocytes are heterogeneous with differing functions. Some monocytes activate inflammatory pathways while others are associated with healing and repair. Conventional laboratory testing in the acute hospital setting does not differentiate between the two, so whether there are differences in the proportions of monocyte subsets in patients with high overall monocyte counts or poor outcomes is unclear. Mechanistically, inflammatory monocyte infiltration could lead to a higher likelihood of hematoma expansion, more blood-brain barrier breakdown, and more cerebral edema following the initial insult. One study found that inflammatory monocytes were the most prominent cells in the peri-ICH period in mice.25 This group further found that elevated serum CCL2, which is the main chemokine associated with the inflammatory pathway, was associated with worse functional outcome at 7 days in a human cohort. An experimental neutrophil depletion in mice also led to decreased monocyte infiltration and improved functional outcome in another study.26 These lines of evidence support a link between an inflammatory monocyte response and poor clinical outcome.
Strengths of this study include a large and well-characterized cohort, extensive review of referring hospital data, and case-fatality data out to one year. We obtained all WBC differential data in the first 24 hours following ICH, which yielded a sizable cohort of patients with multiple measurements. There are several limitations to this work. Retrospective studies have well-known inherent limitations. We are unable to comment on hematoma expansion or perihematomal edema. We are also unable to account for the effects of infection (including infections both preceding and resulting from ICH), inflammatory conditions, or lab error on the WBC data. It is possible that differential cell counts were altered in some of our patients because of infection and/or that counts might predispose particular patients to infection and lead to worse outcomes. We are therefore unable to comment on whether counts are associated with poor outcome independent of infection. Elevated monocytes were only associated with 30-day case-fatality (and not the later time points) in this study, and early infection might have played a role in that finding. ICHs in the posterior limb of the internal capsule and thalamus have been associated with worse outcomes,27 but we do not have specific deep locations abstracted in our cohort. We also did not evaluate WBC types beyond 24 hours. It is likely that the trajectory of WBC types (including monocytes) changes in the days following ICH and possible that these changes could have an effect on outcomes. Finally, we do not have computerized volumetric analyses of the hematomas and instead used the ABC/2 score, which has been shown to overestimate hematoma volumes in some studies.28,29
The inflammatory response following ICH is complex. An improved understanding of how infiltrating leukocytes affect hematoma expansion, edema formation, neuronal damage, and blood-brain barrier breakdown could help improve future clinical trial design. Future studies will investigate whether there are subgroups of ICH patients, including those with particular blood or imaging biomarkers at multiple time points, in which WBC types might help predict clinical outcomes and provide targets for intervention.
Supplementary Material
Figure 1:
Kaplan-Meier survival curve for one-year case-fatality by monocytes greater than median versus less than median
Acknowledgements
The authors would like to thank Peter Castelluccio for his assistance.
Sources of Funding
This work was supported by awards from the IU Health Values Fund (IUH VFR365), the IU CTSI PDT (ICTSI NIH/NCRR RR025761), the IUH/IUSM Strategic Research Initiative, and an IU CTSI KL2 award (NIH, UL1TR001108, Shekhar PI).
Conflict of Interest/Disclosures
Dr. Mackey is funded by Research Grant; Significant; IUH-VFR-365, IUH/IUSM Strategic Research Initiative, and CTSI PDT. NIH LRP recipient. Indiana University CTSI KL2 award recipient.
A.D. Blatsioris is funded by Research Grant; Significant; IUH-VFR-365, IUH/IUSM Strategic Research Initiative.
C. Saha is funded by Research Grant; Significant; IUH-VFR-365, IUH/IUSM Strategic Research Initiative.
E.A.S. Moser is funded by Research Grant; Significant; IUH-VFR-365, IUH/IUSM Strategic Research Initiative.
R.J.L. Carter is funded by Research Grant; Significant; IUH-VFR-365, IUH/IUSM Strategic Research Initiative.
Dr. Cohen-Gadol reports no disclosures.
Dr. Leipzig reports no disclosures.
Dr. Williams reports no disclosures.
Footnotes
Details page
We confirm that this manuscript complies with all instructions to authors. All authorship requirements have been met. All authors approved the final manuscript. This manuscript has not been published elsewhere and is not under consideration at another journal. We have adhered to strict ethical guidelines and have received IRB approval for our work as outlined in the Methods section. We have used the STROBE reporting checklist.
Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.
References:
- 1.Wang J, Doré S. Inflammation after intracerebral hemorrhage. J Cereb Blood Flow Metab 2007;27:894–908. [DOI] [PubMed] [Google Scholar]
- 2.Adeoye O, Walsh K, Woo JG, et al. Peripheral monocyte count is associated with case fatality after intracerebral hemorrhage. Journal of stroke and cerebrovascular diseases : the official journal of National Stroke Association 2014;23:e107–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Walsh KB, Sekar P, Langefeld CD, et al. Monocyte Count and 30-Day Case Fatality in Intracerebral Hemorrhage. Stroke 2015;46:2302–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Morotti A, Phuah CL, Anderson CD, et al. Leukocyte Count and Intracerebral Hemorrhage Expansion. Stroke 2016;47:1473–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lattanzi S, Cagnetti C, Provinciali L, Silvestrini M. Neutrophil-to-Lymphocyte Ratio Predicts the Outcome of Acute Intracerebral Hemorrhage. Stroke 2016. [DOI] [PubMed] [Google Scholar]
- 6.Giede-Jeppe A, Bobinger T, Gerner ST, et al. Neutrophil-to-Lymphocyte Ratio Is an Independent Predictor for In-Hospital Mortality in Spontaneous Intracerebral Hemorrhage. Cerebrovasc Dis 2017;44:26–34. [DOI] [PubMed] [Google Scholar]
- 7.Qin J, Li Z, Gong G, et al. Early increased neutrophil-to-lymphocyte ratio is associated with poor 3-month outcomes in spontaneous intracerebral hemorrhage. PloS one 2019;14:e0211833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gusdon AM, Gialdini G, Kone G, et al. Neutrophil-Lymphocyte Ratio and Perihematomal Edema Growth in Intracerebral Hemorrhage. Stroke 2017;48:2589–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.McDonald CJ, Overhage JM, Barnes M, et al. The Indiana Network For Patient Care: A Working Local Health Information Infrastructure. Health Affairs 2005;24:1214–20. [DOI] [PubMed] [Google Scholar]
- 10.Williams LS, Yilmaz EY, Lopez-Yunez AM. Retrospective Assessment of Initial Stroke Severity With the NIH Stroke Scale. Stroke 2000;31:858–62. [DOI] [PubMed] [Google Scholar]
- 11.Kothari RU, Brott T, Broderick JP, et al. The ABCs of Measuring Intracerebral Hemorrhage Volumes. Stroke 1996;27:1304–5. [DOI] [PubMed] [Google Scholar]
- 12.Hallevi H, Dar NS, Barreto AD, et al. The IVH score: a novel tool for estimating intraventricular hemorrhage volume: clinical and research implications. Crit Care Med 2009;37:969–74, e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Schisterman EF, Perkins NJ, Liu A, Bondell H. Optimal cut-point and its corresponding Youden Index to discriminate individuals using pooled blood samples. Epidemiology 2005;16:73–81. [DOI] [PubMed] [Google Scholar]
- 15.YOUDEN WJ. Index for rating diagnostic tests. Cancer 1950;3:32–5. [DOI] [PubMed] [Google Scholar]
- 16.Suzuki S, Kelley RE, Dandapani BK, Reyes-Iglesias Y, Dietrich WD, Duncan RC. Acute leukocyte and temperature response in hypertensive intracerebral hemorrhage. Stroke 1995;26:1020–3. [DOI] [PubMed] [Google Scholar]
- 17.Castellanos M, Leira R, Tejada J, et al. Predictors of good outcome in medium to large spontaneous supratentorial intracerebral haemorrhages. J Neurol Neurosurg Psychiatry 2005;76:691–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Leira R, Dávalos A, Silva Y, et al. Early neurologic deterioration in intracerebral hemorrhage: predictors and associated factors. Neurology 2004;63:461–7. [DOI] [PubMed] [Google Scholar]
- 19.Sun W, Peacock A, Becker J, Phillips-Bute B, Laskowitz DT, James ML. Correlation of leukocytosis with early neurological deterioration following supratentorial intracerebral hemorrhage. J Clin Neurosci 2012;19:1096–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yu S, Arima H, Heeley E, et al. White blood cell count and clinical outcomes after intracerebral hemorrhage: The INTERACT2 trial. J Neurol Sci 2016;361:112–6. [DOI] [PubMed] [Google Scholar]
- 21.Zhu B, Pan Y, Jing J, et al. Neutrophil counts, neutrophil ratio, and new stroke in minor ischemic stroke or TIA. Neurology 2018;90:e1870–e8. [DOI] [PubMed] [Google Scholar]
- 22.Gökhan S, Ozhasenekler A, Mansur Durgun H, Akil E, Ustündag M, Orak M. Neutrophil lymphocyte ratios in stroke subtypes and transient ischemic attack. Eur Rev Med Pharmacol Sci 2013;17:653–7. [PubMed] [Google Scholar]
- 23.Nam KW, Kim TJ, Lee JS, et al. High Neutrophil-to-Lymphocyte Ratio Predicts Stroke-Associated Pneumonia. Stroke 2018. [DOI] [PubMed] [Google Scholar]
- 24.Spahn JH, Kreisel D. Monocytes in sterile inflammation: recruitment and functional consequences. Arch Immunol Ther Exp (Warsz) 2014;62:187–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hammond MD, Taylor RA, Mullen MT, et al. CCR2+ Ly6C(hi) inflammatory monocyte recruitment exacerbates acute disability following intracerebral hemorrhage. J Neurosci 2014;34:3901–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Sansing LH, Harris TH, Kasner SE, Hunter CA, Kariko K. Neutrophil depletion diminishes monocyte infiltration and improves functional outcome after experimental intracerebral hemorrhage. Acta Neurochir Suppl 2011;111:173–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Delcourt C, Sato S, Zhang S, et al. Intracerebral hemorrhage location and outcome among INTERACT2 participants. Neurology 2017;88:1408–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Tilling EJ, El Tawil S, Muir KW. Do Clinicians Overestimate the Severity of Intracerebral Hemorrhage? Stroke 2019;50:344–8. [DOI] [PubMed] [Google Scholar]
- 29.Xu X, Chen X, Zhang J, et al. Comparison of the Tada formula with software slicer: precise and low-cost method for volume assessment of intracerebral hematoma. Stroke 2014;45:3433–5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.