Mortality rate and biomarker expression within COVID-19 patients who develop acute ischemic stroke: a systematic review and meta-analysis

Objective: To describe the mortality difference between acute ischemic stroke (AIS) and non-AIS groups within COVID-19 patients. Materials & methods: We included observational studies through September 2020 that categorized COVID-19 patients into two groups (with and without AIS). Results: Eight studies with a total sample size of 19,399 COVID-19 patients were included. The pooled risk difference showed that patients with COVID-19 who developed AIS had significantly higher mortality than those without AIS by a risk difference of 24% (95% CI: 0.10–0.39; p = 0.001). In two studies, the COVID-19+AIS group had significantly higher lymphocytes, procalcitonin and creatinine levels. Conclusion: Developing AIS significantly adds to the mortality of COVID-19. Timely interventions to manage those patients are strongly recommended.

inflammation [9], and the more recent concept of cerebral vasculitis [10]. Antiphospholipid antibodies have also been positive in severely infected patients with AIS, regardless of their genetic predisposition [5,6].
Although healthcare resources including personnel, hospital and intensive care unit beds and physicians should be directed toward caring for COVID-19 patients, this should not compromise the care of patients presenting with AIS [11]. A German nationwide cohort study using all hospitalized patients' administrative database showed that the absolute number of stroke admissions declined during the pandemic. However, patients presenting with AIS continued to receive tissue plasminogen activator and mechanical thrombectomy according to standards [12].
Patients with AIS are usually at higher risk of mortality and morbidity than healthy controls [13][14][15][16]. However, the mortality rates of COVID-19 patients who develop AIS compared with those without AIS remain unanswered. This systematic review and meta-analysis aimed to compare the mortality rates of COVID-19 patients who developed AIS with those who did not.

Materials & methods
The study was conducted following the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist for systematic reviews [17]. The review underwent critical appraisal following the A MeaSurement Tool to Assess systematic Reviews 2 (AMSTAR 2) checklist [18].

Search Strategy
We systematically searched PubMed, Scopus and Web of Science from database inception to August 2020. The search strategy was: ("brain" OR "cerebral" OR "cerebrovascular" OR "CNS" OR "large vessel" OR "central nervous system") AND ("infarction" OR "occlusion" OR "thrombosis" OR "thrombotic" OR "cerebrovascular accident" OR "stroke") AND ("coronavirus" OR "COVID" OR "SARS-COV 2" OR "severe respiratory distress syndrome"). A manual search was performed to minimize results bias by searching for relevant missed publications in included articles' references. The reference lists of the studies included in our study were searched to trace potential further studies. The search was updated in September 2020 to include articles published after the first search.

Inclusion & exclusion criteria
Studies satisfying the following criteria were included: • Study design: studies that were described as observational studies (cohort or case-control studies); • Population: studies on COVID-19 patients categorized into two groups, with and without AIS, with confirmed temporality between COVID-19 and stroke; • Outcome: studies reporting the mortality rate and demographic, clinical, and laboratory findings; and • Language: only studies written in the English language were included. We excluded in vitro or animal studies; reviews, thesis, books, conference papers or articles without available full texts; duplicated, overlapping, unreliably extracted or incomplete data; and preprints that were not peer-reviewed.
Screening & study selection process Two independent reviewers screened all records for eligibility. Eligibility screening was performed in two steps: in the first step, titles and abstracts were screened, and in the second step, full-text articles of the selected abstracts were retrieved and assessed for eligibility. Disagreements were resolved by discussion with a third reviewer.

Data extraction
We extracted data on patients' demographic characteristics, past medical history, clinical presentation, laboratory values, treatments and clinical outcomes. Two independent reviewers extracted data to a uniform Microsoft Excel sheet. A third independent reviewer performed a further check on the retrieved data to confirm data accuracy. All disagreements were resolved through discussion.
Assessing the risk of bias in the included studies We used the Newcastle Ottawa Scale (NOS) to assess the risk of bias in the included observational studies. The NOS is a star-based score in which studies are judged on three domains: the selection of the study groups, the comparability of the groups and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies, respectively.

Data analysis
The differences between COVID-19 with AIS and COVID-19 without AIS groups were pooled as risk difference (RD) with the corresponding 95% CIs in the random effects meta-analysis model. For continuous outcomes, the standardized mean difference (SMD) between the study groups was pooled with the corresponding 95% CI in the random effect meta-analysis model. All analysis was done by the RevMan software for Windows (Review Manager Version 5.3. Copenhagen: The Nordic Cochrane Centre, 2014). A p-value <0.05 was considered for statistical significance.

Heterogeneity
Heterogeneity was assessed by visual inspection of the forest plots to check the pooled estimates' overlapping 95% CIs. Heterogeneity was tested by the chi-square test and quantified by the I 2 test. The outcomes were considered heterogeneous when the p-value was <0.1 and I 2 >50%.

Literature search results
Our literature search yielded 5100 unique records. Following the title and abstract screening, 31 studies remained eligible for full-text screening. Finally, eight cohort studies (n = 19,399 patients) were eligible for inclusion in this meta-analysis [19][20][21][22][23][24][25][26]. The flow diagram of the study selection process is shown in Figure 1.

Baseline characteristics of included studies
We included eight studies with a total population of 19,399 COVID-19 patients, including 288 with AIS versus 19,111 without AIS. Three studies were conducted in China [20][21][22], one in Spain [19], one in Italy [23], two in the USA [24,25] and one was multinational [26]. The summary of the included studies and their population characteristics are shown in Table 1.

Quality of the included studies
As assessed by the NOS, the included studies' overall quality was acceptable ( Table 2). Regarding selection bias, all studies had good quality in the exposed cohort's representativeness, selection of the nonexposed cohort and ascertainment of exposure. Four studies did not clarify whether the outcome was presented initially or not [22,[24][25][26]. As for comparability, all studies presented comparable cohorts except Annie et al. [24]. All studies showed good quality in outcome assessment, sufficiently long follow-up and the cohort's adequacy.

Discussion
Overall, 19,399 COVID-19 patients were included in the qualitative and quantitative evidence synthesis. This study shows that COVID-19 patients who developed AIS were at a 24% higher risk of mortality than non-AIS COVID-19 patients. In addition, specific laboratory values as lymphocyte count, procalcitonin level and creatinine level were higher in the COVID-19 with AIS group. There was no significant difference in the remaining laboratory values, which could be explained by the limited number of available studies. The high mortality rate in COVID-19 with AIS group in our meta-analysis (29.6%) is concordant with recent studies on COVID-19 [27][28][29]. A recent systematic review showed that the proportion of patients with COVID-19 who developed stroke was 1.8% (95% CI: 0.9-3.7%), and the in-hospital mortality rate was 34.4% (95% CI: 27.2-42.4%) [27]. However, this study is different from our meta-analysis in that it included patients with all types of stroke and not only ischemic stroke. It also did not directly compare the mortality of COVID-19 patients with stroke versus those without stroke [27]. In a recent systematic review by Tan et al. investigating the clinical features, neuroimaging findings and outcomes of AIS in COVID-19 patients, the pooled incidence of AIS in COVID-19 patients from observational studies was 1.2%, with a high mortality rate of 38% [28]. This study did not include a direct comparison between COVID-19 with AIS against those without AIS. Dmytriw et al. investigated the effects of race differences among COVID-19 with AIS patients and found that the mortality rate was significantly higher in African-Americans than other races (55.6% vs 28.6%; p = 0.05) [29]. Another study by Yaghi et al. found that 0.9% of all COVID-19 hospitalized patients experienced an ischemic stroke, and these patients had a higher inpatient mortality rate than contemporary and historical ischemic stroke controls before the COVID-19 pandemic (63.6% vs 9.3%; p < 0.001) [30]. Even patients with a history of stroke before their COVID-19 infection had higher mortality, as evidenced by a recent study in Wuhan, China, on 1875 patients, which found that patients  with a history of stroke had worse outcomes compared with those without a history of stroke, with lower rates of discharge and higher mortality risks (24.0 vs 43.3%, 14.0 vs 8.3%, respectively) [31]. Compared with the previously reported mortality rates of ischemic stroke patients in the general population, COVID-19 patients with AIS in our meta-analysis had a higher mortality rate (29.6%) [15,32,33]. Saposnik et al. reported the mortality rates for 12,262 AIS patients ranged from 12.2 to 12.6% at 30 days and 22.5 to 22.9% at 1 year [15]. A previous systematic review and meta-analysis showed that from 2000 to 2008, early (21 days to 1 month) case fatality of ischemic stroke ranged from 13 to 23% in high-income countries and from 13 to 19% in low to middle-income countries [32]. Ganesh et al. conducted a retrospective study on 319,972 stroke/TIA hospitalized patients and found that the crude 30-day mortality rate decreased from 15.8% in 2003/2004 to 12.7% in 2012/2013 in provinces with stroke systems while remained 14.5% in those without such systems [33]. By 8 April 2020, the total number of confirmed COVID-19 cases globally was 133.9 million, with 2.9 million deaths (mortality rate of 2.2%) [34]. A systematic review and meta-analysis of published evidence on COVID-19 until July 2020 showed that the COVID-19 infection fatality rate across populations was 0.68% (0.53-0.82%) [35]. Thus, this meta-analysis shows that the mortality rate of COVID-19 patients with AIS is greater than the mortality rate of either COVID-19 patients in general or patients with acute ischemic stroke alone, suggesting a compounded mortality effect of stroke and COVID-19 when they co-occur.
Regarding laboratory findings, the difference between COVID-19 with AIS and COVID-19 without AIS was only significant for lymphocyte count, procalcitonin and creatinine, which were significantly higher in AIS patients. Creatinine level is one of the most critical indicators that correlates with ischemic and hemorrhagic strokes [36]. Snarska et al. have investigated the impact of kidney functions on in-hospital patients with ischemic and hemorrhagic strokes. They have found that in both, creatinine was a predictor for stroke outcome [36]. Our analysis of the available literature has shown that two studies [22,26] have assessed creatinine levels for COVID-19 stroke patients. The meta-analysis pooled results show that creatinine levels among COVID-19 with AIS patients  Redrawn from [49].
Our meta-analysis is limited by the relatively small number of available studies, the small sample size in some reports and a limited number of available parameters compared between groups. This is because studies did not consistently report AIS outcomes in the literature, and only a few studies provided head-to-head comparisons between COVID-19 with AIS and COVID-19 without AIS populations. Pooled studies were not homogenous, Acute cardiac injury since ACE2 is highly expressed on cardiac myocytes, therefore susceptibility to embolic events. which was resolved by doing sensitivity analysis, excluding small studies. Although there might be a suggestion that mortality in COVID-19 with AIS was not necessarily higher due to AIS per se but rather the severity or co-occurring complications of COVID-19 such as renal failure and pneumonia, the included studies' data did not consistently provide such manifestations, complications or severity of COVID-19 for their patients, so including such variables in the meta-analysis was not possible. However, because this study compares groups of patients (COVID-19 with AIS vs those without AIS), the aim was not to establish direct causation between AIS and mortality. Another limitation is that only two studies investigated laboratory findings associated with AIS in COVID-19 patients [22,26]. Although Fan et al.'s study had a substantially small sample size of six AIS patients compared to 80 COVID-19 patients without AIS [22], this was compensated by Shahjouei et al. large sample size with a relatively higher number of COVID-19 with AIS patients (n = 123) [26].

Secondary processes
Nonetheless, this meta-analysis expands the literature by providing evidence on the increased mortality in COVID-19 patients with AIS, which requires medical attention and prompt management to reduce the mortality.

Conclusion
COVID-19 patients with AIS have higher rates of mortality compared with COVID-19 patients without AIS. There is a suggestion that COVID-19 with AIS patients have higher levels of lymphocytes, procalcitonin and creatinine than those without AIS. Future prospective, well-designed studies should provide more data on the predictors of mortality among COVID with AIS patients. Also, future clinical trials are needed to evaluate interventions to reduce mortality in this population.

Future perspective
On the basis of this study's findings, the management of COVID-19 patients should follow a risk stratification model. The healthcare for COVID-19 patients who developed AIS or at risk of developing AIS should be maximized to lower their mortality. Specific biomarkers such as lymphocyte count and levels of procalcitonin and creatinine should be monitored closely as they can predict the occurrence of AIS in COVID-19 patients. Future prospective, well-designed studies should provide more data on the predictors of mortality among this population. Future clinical trials are needed to evaluate interventions to reduce mortality in COVID-19 with AIS patients.

Summary points
• COVID-19 is an evolving disease with varying clinical courses, ranging from asymptomatic infection to severe complications, including death. • Previous reports indicated an association between COVID-19 and inflammatory coagulopathy that causes disseminated vascular obstructions, including but not limited to acute ischemic stroke (AIS). • This systematic review and meta-analysis aimed to compare the mortality rates of COVID-19 patients who developed AIS versus those who did not and determine the significant demographic, clinical and laboratory characteristics of COVID-19 patients who developed AIS. • A total of eight studies with 19,399 COVID-19 patients, including 288 with AIS versus 19,111 without AIS, were included in this study. • Six studies reported the mortality rates in both COVID-19 with AIS and COVID-19 without AIS patients. Two studies compared biomarkers' levels between COVID-19 with AIS (n = 129) and COVID-19 without AIS patients (n = 6280). • The meta-analysis showed higher mortality in the COVID-19 with AIS group compared with the COVID-19 without AIS group (RD: 0.24; 95% CI: 0.10-0.39; p = 0.001). • Higher lymphocytes, procalcitonin and creatinine levels were seen in COVID-19 patients with AIS compared with those without AIS. • The overall standardized mean difference between the two groups did not favor either of the two groups in terms of the white blood cells, hemoglobin, platelets, C-reactive protein, blood urea nitrogen, alanine aminotransferase, D-dimer, lactate dehydrogenase or creatine kinase. • COVID-19 patients with AIS in our meta-analysis had a higher mortality rate than previously reported mortality rates of AIS patients in the general population. • In conclusion, COVID-19 patients who develop AIS have higher mortality than COVID-19 patients without AIS; thus, they need timely interventions to lower their mortality. The biomarkers mentioned herein should be monitored closely in COVID-19 patients as they are associated with AIS development.

Author contributions
All authors have seen and approved the content, fulfilled the authorship criteria and contributed significantly to work. All authors presented substantial contributions to the article and participated in the version's final approval to be submitted.