Higher Total White Blood Cell and Neutrophil Counts Are Associated With Increased Stroke Mortality Risk: The Guangzhou Biobank Cohort Study

Background: To investigate the relations of white blood cell (WBC) count and its dynamic change with future stroke mortality risk in a relatively healthy elderly population. Methods: A total of 27811 participants without stroke history at baseline were included and followed up for an average of 11.5 (SD=2.3) years. After review of available records, 399 stroke (277 ischaemic and 172 haemorrhagic) deaths were recorded among all-cause mortality. Cox proportional hazards regression was used to estimate hazard ratios (HRs) and 95% condence intervals (CIs). Results: Compared with the lowest quartile, the highest quartile of WBC count showed 53% and 67% increased mortality risk for total (adjusted HR [aHR]=1.53, 95% CI 1.16-2.02, P=0.003) and haemorrhagic (aHR=1.67, 95% CI 1.10-2.67, P=0.03) stroke, respectively; the highest neutrophil count showed 45% and 65% increased mortality risk for total (aHR=1.45, 95% CI 1.10-1.89, P=0.008) and ischaemic (aHR=1.65, 95% CI 1.10-2.47 P=0.02) stroke. The same results found for total and ischaemic stroke but not for haemorrhagic stroke were observed for both WBCs and neutrophils within the normal range level after further C-reactive protein (CRP) adjustment. Compared with the stable group, the 25% increased groups of both WBCs (aHR=1.60, 95% CI 1.07-2.40, P=0.02) and neutrophils (aHR=1.45, 95% CI 1.02-2.05, P =0.04) showed 60% and 45% increased stroke mortality risk, respectively. Conclusions: These ndings support the role of WBCs, especially neutrophils, as simple, inexpensive and readily available predictors of future stroke mortality in an elderly population. a signicant mortality risk for total stroke was shown with both increased WBCs (aHR=1.60, 95% CI 1.07-2.40, P =0.02) and increased NEUs (aHR=1.45, 95% CI 1.02-2.05, P =0.04).


Background
Stroke is mainly categorized as ischaemic and haemorrhagic [1] [, 1989 #2]. With a high prevalence of comorbidities in developed Western countries, preexisting chronic low-grade systemic in ammation has become a recognized characteristic of stroke pathophysiology [2]. Evidence now suggests that a chronic in ammatory response has been associated with an increased risk of ischaemic [3,4] and haemorrhagic [5] stroke. Total white blood cell count (WBC), a plausible marker in the pathogenesis of chronic in ammation [6], is generally conducive to stroke incidence.
Increased WBC count on admission was linked to poor outcome and increased mortality for patients with total [7,8], ischaemic [9] and haemorrhagic [10] stroke in case-control studies. However, these WBC counts may be due to the stress reaction in acute patients with stroke [2], and it is not clear that such increased WBC counts are linked directly to stroke mortality. On the other hand, a relatively high WBC count was related to the incidence, unfavourable functional outcome and mortality risk of total [11][12][13][14][15] and ischaemic stroke [13,[15][16][17][18][19][20] in prospective cohort studies, although there are still controversies regarding total [21], ischaemic [22,23] and haemorrhagic [15,19] stroke. Neutrophils (NEUs), the largest WBC subpopulation, showed similar associations with total [8,15], ischaemic [15,19,20,[24][25][26] and haemorrhagic [27] stroke. However, different types of in ammation can result in increases in not only leucocytes but also other indicators, such as C-reactive protein (CRP). CRP, a controversial independent stroke risk factor and an underlying acute in ammatory risk factor [2], was reported to be a predictor in total [13,28] and ischaemic [23,26,29,30] stroke in addition to haemorrhagic stroke. Nevertheless, no changes in WBCs or their subpopulations have been reported to be linked with stroke mortality risk to date.
In previous work, we reported that a higher WBC count was associated with all-cause, CHD and respiratory mortality [14], cardiovascular disease [31] and metabolic syndrome risk [32] in the Guangzhou Biobank Cohort Study (GBCS). Here, we aimed to systematically assess the relations of WBCs, their subpopulations and their changes with stroke mortality risk in a healthy elderly population.

Subjects
All GBCS participants were recruited from among Guangzhou permanent residents aged 50 years or older in southern China. Details of the GBCS have been reported previously [33]. The baseline (from September 2003 to February 2008) and follow-up information included a face-to-face computer-assisted interview by trained nurses on lifestyle [34], family and personal medical history and assessment of anthropometric data, blood pressure and laboratory tests. An appointment had been made for each participant in advance to ensure good health and for each of them to report in person to the designated place to sit and rest for at least half an hour before sampling and examination.

Exposure indicators
WBC and subpopulation counts were performed by a blood cell counter (KX-21, Sysmex, Japan) in Guangzhou Twelfth People's Hospital. WBCs, NEUs and lymphocytes (LYMs) were counted separately, while monocytes, eosinophils and basophils were counted as a mixture. Fasting glucose, cholesterol, triglycerides, liver and kidney function and CRP were measured by an analyser (Cobas c-311, Roche, Switzerland). The hospital laboratory performs internal and external quality control procedures according to the China Association of Laboratory Quality Control.

Study outcomes
Information on underlying causes of death up to December 2017 were mostly obtained via record linkage with the Guangzhou Centre for Disease Control and Prevention (GZCDC). Due to the lack of any other information on stroke severity, infarct volume, site of lesion and infectious complications, mortality was chosen as a primary outcome in this study. Death causes were coded according to the 10 th revision of the International Classi cation of Diseases (ICD) as follows: I60~I69 for total stroke; I60.0~I62.9 and I69.0~I69.2 for haemorrhagic stroke; and I63.0~I63.9 and I69.3 for ischaemic stroke. When the death certi cates were not issued by medical institutions, the causes were veri ed by GZCDC as part of their quality assurance programme by cross-checking past medical history and conducting a verbal autopsy by 5 senior clinicians from Guangzhou Twelfth People's Hospital and the Universities of Hong Kong, China and Birmingham, UK.

Statistical analysis
WBCs and NEUs were classi ed by quartiles. WBC and NEU change analyses were categorized into 2 models: ±10% and ±25%. Continuous variables were described by the mean ± standard deviation, and categorical variables were described by frequency and percentage. The chi-square test and Fisher's exact test were used for categorical variables, and analysis of variance (ANOVA) and the Kruskal-Wallis test were used for continuous variables. The covariates included sex, age, education, occupation, smoking habit, alcohol consumption, physical activity, body mass index (BMI, de ned as weight in kg ÷ height in m2) [35], self-rated health, hypertension, diabetes, dyslipidaemia, cancer, genitourinary disease, chest disease, platelet count (continuous) and CRP (continuous). A sensitivity analysis was conducted, including the association between leucocytes and stroke mortality in those with normal WBC levels and after excluding those with NEU levels in the top 1% and bottom 1% at baseline. All analyses were performed using STATA (Version 14.0; StataCorp LP, College Station, TX, USA). All p values were 2 sided, and statistical signi cance was de ned as p < 0.05.

Baseline characteristics
Of 30,430 participants, 286 were excluded because of previous history of stroke; 315, because of unclear stroke history; 372, because of loss to follow-up with unknown vital status; and 1646, because of incomplete information on WBCs, NEUs, lymphocytes and platelets, hypertension, diabetes, dyslipidaemia, smoking habit, alcohol consumption, physical activity, BMI, self-rated health, cancer, genitourinary disease or chest disease. Thus, 27,811 participants were enrolled in this study. After an average follow-up of 11.5 (SD=2.3) years, 399 stroke deaths (227 ischaemic and 172 haemorrhagic) were recorded. Table 1 shows basic characteristics at baseline. Compared with those in the 1 st quartile, participants from the 2 nd to the 4 th quartiles included more men, were older, had more manual and former or current smokers and drinkers. Furthermore, these subjects were more likely to have BMI ≥24 kg/m2, hypertension, diabetes and dyslipidaemia; higher levels of NEUs, LYMs, platelets and CRP; lower education levels and reduced physical activity and were less likely to have good or very good self-reported health, cancer and genitourinary disease (all P<0.001) compared with those in the 1 st WBC quartile.
Those in the 2 nd , 3 rd and 4 th WBC quartiles showed an increasing risk trend for total (P<0.001), ischaemic (P=0.01) and haemorrhagic (P=0.02) stroke. The middle of Table 2 shows NEUs in four quartiles. Signi cant associations with increased mortality risks were observed in total (aHR=1.45, 95% CI 1.10-1.89, P=0.008) and ischaemic (aHR=1.65, 95% CI 1.10-2.47, P=0.02) stroke. Unlike WBCs, neither an increased haemorrhagic mortality risk (aHR=1.14, 95% CI 0.74-1.75, P=0.56) nor an increasing risk trend (P=0.26) for NEU quartiles were observed. When excluding participants with stroke and further CRP adjustment among the 10041 participants with normal WBCs (4~10*10^9/L), a signi cant association of those in the 4 th quartile was shown only in total stroke (further adjusted HR [faHR] =1.57, 95% CI 1.02-2.42, P=0.04), and an increasing risk trend was found in both total stroke (P=0.012) and ischaemic stroke (P=0.02) (left side of Table 3). When excluding participants with NEU counts in the top 1% and bottom 1% and after further CRP adjustment, the results from 9946 participants with normal WBC levels indicated that the highest NEU quartile was related to an increased mortality risk for both total (faHR=1.55, 95% CI 1.00-2.41, P=0.05) and ischaemic (faHR=2.47, 95% CI 1.24-4.93, P=0.01) stroke and an increasing risk trend for both total (P=0.009) and ischaemic (P=0.004) stroke; however, NEUs showed neither a signi cant association with the highest quartile (P=0.18) nor an increasing risk trend (P=0.40) in haemorrhagic stroke (right side of Table 3).
Additionally, LYMs showed only a decreased risk trend for ischaemic stroke (P for crude HR [cHR] =0.03). No signi cant associations of stroke mortality risks with CRP were observed (Table 4).
WBC changes in relation to stroke mortality Table 5 shows the association between stroke mortality risk and changed leucocytes during the period of baseline (from September 2003 to February 2008) to the 1 st follow-up (from March 2008 to December 2012). Compared to those with stable disease, participants with WBC or NEU changes within 10% showed no signi cant mortality risk for total stroke. Once the change reached 25%, a signi cant mortality risk for total stroke was shown with both increased WBCs (aHR=1.60, 95% CI 1.07-2.40, P =0.02) and increased NEUs (aHR=1.45, 95% CI 1.02-2.05, P =0.04).

Discussion
In this study, we found that both WBCs and NEUs were associated with the mortality risk of total stroke, and a higher NEU count was associated with an increased mortality risk of ischaemic stroke in a relatively healthy elderly population. These associations were independent of age, sex, education, occupation, hypertension, diabetes, dyslipidaemia, smoking habit, alcohol consumption, physical activity, BMI, self-rated health, cancer, genitourinary disease, chest disease, platelets and C-reactive protein.
Growing studies on the relationship between leucocytes and stroke have focused mainly on the population at admission after stroke onset. Most of them support that an increased WBC count is related to a poor outcome or mortality [8-10, 15, 26, 27, 36, 37], except for a few cases of disagreement of initial stroke severity [10,15,30,38]. This indicates that in ammation arises together with stroke or that stroke itself leads to leucocytosis, resulting in poor outcomes. In a review [2], a series of biomarkers, such as cytokines, WBCs, CRP and interleukin 6 (IL-6), were shown to participate speci cally in stroke progression [39]. When focusing on speci c types of in ammation in mice, allergens (anaphylaxis) were found to induce IL-10 and a corresponding response, while endotoxin (lipopolysaccharide, LPS) was shown to stimulate various types of cells including leucocytes to induce the release of a series of active molecules [40]. This is evidence for the different types of in ammation involved in stroke progression.
We present a corresponding relation of stroke mortality risks with pre-existing chronic low-grade systemic in ammation. For the GBCS to collect a series of data from relatively healthy elderly individuals in South China, each appointment was made in advance to ensure good health and for each participant to be able to report in person to the designated place [32,41,42]. In addition, we excluded those with WBC levels at the top and bottom to avoid intervention in acute in ammatory reactions. Our results are consistent with some previous reports [11,13,14] showing that a higher WBC count is related to an increased total stroke mortality risk. The results were rea rmed after further CRP adjustment, similar to reports from the Japan Collaborative Cohort Study [28] and the Glasgow In ammation Outcome Study [43]. Unlike the reports of incongruent factors [15-20], we found WBC quartiles to have an increased risk trend for ischaemic stroke; such a weak association may be due to our additional adjustments. Nevertheless, a similar association with haemorrhagic stroke disappeared after further adjustments.
As the largest subpopulation, NEUs play an important role in the major processes of atherosclerosis, thrombosis and stroke [44]. Our results are consistent with a few previous publications [15,19,20] and are in contrast with a number of others [45-53] that show a higher NEU count in relation to the increased mortality risk for both total and ischaemic stroke. Taking into account WBCs and NEUs in stroke, our ndings suggest that NEUs are more relevant for predicting future stroke mortality.
CRP has been reported as an independent risk factor in clinical stroke [9,26,30]. Here, we observed no signi cant relationship between CRP and stroke mortality risk (Table 4). This is likely because our analytic data was collected from relatively healthy participants.
There are different leucocyte backgrounds in individuals, and each can reach 15% uctuation in one day [54]. Stroke events are related to chronic in ammation, while leucocytes can well explain the immediate in ammatory status. Here, we used the data from baseline to the rst follow-up and guaranteed the stability of leucocyte counts because each participant sat and rested for at least half an hour before sampling and examination. We report rst the total stroke mortality risk in relation to changes in WBCs and NEUs in healthy elderly Chinese individuals. This indicates that an increase in leucocytes, a relatively long-term or continuous chronic in ammation, promotes a higher stroke mortality risk.

Conclusions
Higher WBC and neutrophil counts were associated with increased stroke mortality risk. Neutrophils were more relevant to predicting future stroke mortality in a healthy elderly population. Abbreviations WBC: white blood cell count; HR: hazard ratio; aHR: adjusted HR; cHR: crude hazard ratio; CI: con dence interval; NEU: neutrophil; CRP: C reactive protein; LYM: lymphocyte; GBCS: Guangzhou Biobank Cohort Study; GZCDC: Guangzhou Centre for Disease Control and Prevention; ICD: International Classi cation of Diseases Declarations Ethics approval and consent to participate Our study was approved by the Guangzhou Medical Ethics Committee of the Chinese Medical Association, and all participants provided written informed consent before participation in the GBCS. The methods of this study were performed in accordance with the Declaration of Helsinki.

Consent for publication
Not applicable

Availability of data and materials
The datasets used during the current study are available from the corresponding author on reasonable request.