Preeclampsia has an association with both platelet count and mean platelet volume: A systematic review and meta-analysis

Background Preeclampsia (PE) is a pregnancy-specific disorder characterized by endothelial dysfunction, and activation of the coagulation system. Alteration of PLT parameters is the common hematological abnormality observed in women with PE. The main aim of this study was to systematically review previous studies from around the world to generate evidence about the relationship between platelet count (PC) and PE, as well as mean platelet volume (MPV) and PE, by calculating the pooled weighted mean difference (WMD) of PC and MPV between PE and normotensive (NT) groups. Methods Relevant articles which were published in the English language from January 10, 2011, to January 10, 2021, were systematically searched through PubMed, Web of Science, and African journals online. In addition, reference probing of published articles searching was employed through Google Scholar and Google for searching grey literature. The methodological qualities of articles were assessed using Joana Brigg’s institute critical appraisal checklist. A random-effects model was used to estimate pooled WMD of PLT parameters between the two groups with the respective 95% confidence intervals (CI) using Stata version 11.0. The I2 statistics and Egger’s regression test were used to assess heterogeneity and publication bias among included studies, respectively. Results A total of 25 articles were included in this systematic review and meta-analysis. Of which, 23 studies were used in each PC and MPV analysis. The overall pooled WMD of PC and MPV between PE and NT groups were -41.45 × 109/L [95% CI; -51.8, -31.0] and 0.98 fl [95% CI; 0.8, 1.1], respectively. The pooled WMD revealed that PC decreased significantly in the PE group compared to the NT group while MPV increased significantly in the PE group. Conclusions This systematic review and meta-analysis indicated that there is a significant decrease in PC and a significant increase in MPV during PE development among pregnant women. As a result, a change in these parameters among pregnant women may indicate the development of PE.

Introduction Preeclampsia (PE) is a pregnancy-specific medical disorder that is characterized by activation of the coagulation system, and endothelial cell dysfunction [1]. It is diagnosed by elevated blood pressure (BP), a systolic BP (SBP) of at least 140 mmHg, or a diastolic BP (DBP) of 90 mmHg after 20 th GW with proteinuria � 1+ in urine dipstick [2]. Women with PE present diverse signs and symptoms associated with multiple organ systems, including headache, thrombocytopenia, severe hypertension, chest pain, pulmonary edema, low oxygen saturation, and abnormal liver and kidney function [3].This disorder is a major cause of maternal and fetal morbidity and mortality [4]. Around 5% of pregnancies are complicated by PE worldwide [5]. World Health Organization (WHO) estimated that the incidence of PE is higher in developing countries than in developed countries [6].
The placenta has always been a central figure in the etiology of PE [4].A two-stage model was developed to easily explain its complex pathogenesis [7]. First, deficient spiral artery remodeling in the uterus is caused by inadequate fetal trophoblast invasion of uterine tissue. The resulting abnormal implantation reduces maternal blood flow to the placenta and fetus, ultimately leading to the development of placental ischemia and an increase in oxidative stress [8,9]. The second stage of PE starts when the placenta responds to progressive ischemia or hypoxia; the placenta may secret and release reactive oxygen species, chemokines, pro-inflammatory cytokines, and anti-angiogenic factors into the maternal circulation that contribute to endothelial damage [7,10,11].
Among the potential mediators, the balance between pro-angiogenic factors and antiangiogenic factors is particularly clinically important [12].The disproportionate levels of proangiogenic factors like vascular endothelial growth factor (VEGF), placental growth factor (PlGF) and transforming growth factor-β (TGFβ), and anti-angiogenic factors such as soluble endoglin (sEng) and soluble fms-like tyrosine kinase-1 (sFlt-1) are believed to cause generalized maternal endothelial dysfunctions [4]. For instance, the elevated levels of sFlt are thought to bind and reduce the bioavailability of VEGF, impairing their endogenous production of nitric oxide and causing vasoconstriction. The production of nitric oxide is induced by VEGF which neutralizes reactive oxygen species and vasoconstrictor signaling [13].
Furthermore, the release of Chemokines and pro-inflammatory cytokines induces inflammation in the maternal circulation, and the release of reactive oxygen species contributes to further placental oxidative stress and endothelial dysfunction [14].
The contact of platelet (PLT) with the injured endothelium activates the coagulation system which leads to an increase in PLT consumption and production [15]. Activation of the coagulation system with increased PLTs aggregation leads to multisystem dysfunction in PE [16,17]. The elevated consumption of PLTs due to the abnormal coagulation system and PLT activation leads to thrombocytopenia which can be used as an important sign of PE [15,18]. Mean platelet volume (MPV) is one of the groups of PLT parameters that is mainly related to PLT morphology and proliferation kinetics [19]. It is a marker of PLT size, function, and activation as the number and size of pseudopodia increases during PLT activation [20]. The increased consumption of PLTs obligated the bone marrow to produce and release young and large PLTs [15,21] leading to an increase in MPV in PE patients [22,23]. There were studies conducted on the association of PLT parameters with PE, however, the findings were controversial. For instance, several studies [17,24,25] found a significant association between PE and platelet count (PC), as well as PE and MPV; on the other hand, some studies [26-28] found non-significant differences in these parameters between PE and NT groups. Therefore, the primary goal of this systematic review and meta-analysis was to determine the pooled weighted mean difference (WMD) of PC and MPV between PE and normotensive (NT) groups and generate evidence about the association between PC and PE, and MPV and PE.

Study design
This systematic review and meta-analysis was performed based on an updated preferred reporting item for systematic review and meta-Analysis (PRISMA) guideline 2020 [29]. This systematic review and meta-analysis analyzed findings from published articles to evaluate the pooled WMD of some PLT parameters between preeclamptic and NT pregnant women globally.

Eligibility criteria
Studies that met the following criteria were considered eligible for inclusion in this study.
Types of studies. We include all published original studies with cross-sectional, case-control, and cohort study designs. There were language restrictions; studies published only in the English language were included. Moreover, we have included studies that were published from January 10, 2011, to January 10, 2021.
Types of participants. We included studies that were carried out on the pregnant woman with a primary clinical diagnosis of PE and also having a control NT group.
Study area. Those studies that were conducted all over the world. Types of outcome measures. The primary outcome of the study was determining the WMD of PC and MPV between PE and NT groups. Studies reported the outcome of interest (PC and MPV) and express the results as mean and standard deviation (SD) or median and interquartile range (IQR) for both PE and NT groups were included in this study.
Publication condition. Studies that meet the eligibility criteria were included regardless of their publication status (published, unpublished and grey literature, etc.).
After a thorough screening of the abstracts and the full texts of the studies, articles having methodological problems were excluded. Studies that did not report the value of PC and MPV in both groups were excluded. Studies whose entire text was not available for free were also excluded. Furthermore, studies conducted among pregnant women, but who had co-morbidities like; HIV/AIDS, malaria, hypertension, and coagulation disorder were excluded from the study.

Information source
Data were collected through searching previous literature on electronic databases such as PubMed, Science direct, and African journals online using search terms. In addition, reference probing of published articles searching was employed through Google Scholar and Google for searching grey literature.

Search strategy
This was done independently by three reviewers (MW, ZG, and YG). A systematic search of published articles was performed on electronic databases using search terms. The search terms were developed following the Medical Subject Headings (Mesh) thesaurus in combination with free text key terms, and they were used in combination using Boolean operators like "OR" or "AND". The searching terms used in electronic databases were "platelet parameter", "platelet indices", "platelet count", "thrombocytes count", "mean platelet volume", "mean platelet size", "MPV", and "preeclampsia". During the literature search, the results were limited by study population, language, availability of full text, and date of publication. Relevant articles which were written in the English language and published from January 10, 2011, to January 10, 2021, were systematically searched.

Selection process
After studies are identified through extensive searching, the search results were imported into EndNote X9 (Thomson Reuters, New York, USA) to organize and remove duplicated articles. After the removal of the duplicates, two authors (MW and YG) independently and meticulously screened the title and abstract of each of the retrieved articles for eligibility. In case of disagreement, consensus on which articles to screen full-text was reached by discussion. If necessary, the third reviewer (ZG) was consulted to make the final decision. Next, the two reviewers independently screened full-text articles for inclusion. Again, in case of disagreement, the consensus was reached on inclusion or exclusion by a discussion with the third reviewer.

Data collection process
We designed a data extraction form, which two review authors (MW and FA) used to extract data from eligible studies. Reviewers worked independently to extract study details. A third reviewer (ZG) reviewed data extraction, and resolve conflicts. The mean±SD values of PC and MPV in each group were extracted. The results of these parameters that were expressed as median (IQR) were also extracted and changed to mean and SD using a formula that was recommended by Wan et al. [30]. Additionally, the first author's name, publication year, study country, study design, the sample size of the PE group, the sample size of the NT group, and PE severity status among PE groups were extracted. Moreover, the standard error (SE) of WMD was calculated to do Egger's test and funnel plot. Extracted data were compared, with any discrepancies being resolved through discussion. Then, MW entered the extracted data into a Microsoft Excel spreadsheet.
Data item. • PC: is a laboratory test to measure the number of PLTs per microliter of blood.
• MPV: is a laboratory test used to measure a PLT volume that reflects the bone marrow activity [31].
• PE severity: PE cases are divided into two categories based on the severity of the condition: mild PE and severe PE. Preeclamptic patients can be considered with mild PE when the BP ranged from 140/90-160/110 mmHg, and proteinuria � 1 on a urine dipstick while severe PE when BP is � 160/110 mmHg with proteinuria >3+ on a urine dipstick and edema and other major symptoms [2]. We extracted this data as mild or severe directly from the included studies. If the studies did not classify them as mild and severe, they were taken as undefined PE.

Risk of bias assessment
The methodological qualities of included studies were appraised in detail using the Joana Brigg's institute (JBI) tool [32] by two reviewers (FA and FG). The tool consists of 8 items to assess the internal and external validity for cross-sectional studies, 10 items for case-control studies, and 11 items for cohort studies. Each item was assessed as 0 for "not done", 1 for "done", UC for "unclear", and NA for "not applicable". Any disagreements during the review process were resolved by consensus and when the disagreements continue after discussion, a third person (ZG) was consulted to finally settle the discrepancy. Finally, the overall qualities of the study were scored according to the number of done items per study quality assessment items. Articles with an average score of 50% and above were included in this meta-analysis study.

Effect measures
We planned to analyze the pooled WMD of PC and MPV and their 95%confidence interval (CI) between PE and NT groups in pair-wise meta-analyses by extracting the value of these parameters in each group from the report of previous studies.

Statistical analysis and interpretation
After the relevant data had been extracted from the studies using Microsoft excel format, the authors then analyzed the results by using STATA version 11.0 (STATA Corporation, College Station, TX, USA), and it was performed by MW and FG together. The original studies were summarized and presented by using a table and the forest plot. The random-effects model was used for pooling WMD analysis of PLT parameters between groups with the respective 95% CIs. Heterogeneity between included studies was determined using Higgin's I-squared statistics. The I 2 values of 25%, 50%, and 75% were considered as low, medium, and high heterogeneity, respectively [33]. To explore the potential source of heterogeneity, subgroup analysis was carried out by study area, study design, and PE severity. Sensitivity analysis was performed to evaluate the influence of each study on the pooled measures by omitting one single study in each turn and recalculating the pooled WMD for the remainders. Publication bias was evaluated by inspection of funnel plots and Egger's test (a statistical analog for funnel plots) [34]. All analyses were performed using Stata 11.0 (StataCorp LP, College Station, TX, USA), and P < 0 05 was considered to be statistically significant.

Search and study selection
A total of 1, 496 studies were identified through database searching and Google search which were done on pregnant women and published from January 10, 2011, to January 10, 2021, in the English language. After removing duplicates, a total of 675 studies remained for screening. Out of which, 650 articles were excluded after reading their title (626 articles) and full-text (24 articles).Among 24 studies that were excluded after reading the full text, 19of them did not report the outcome variable (PLT parameters) in both groups clearly [1,10,18,22,[35][36][37][38][39][40][41][42][43][44][45][46][47][48][49], and three studies did not report the criteria to diagnose PE [50][51][52], and the remaining two studies had not clearly described the study design [53,54]. Finally, after excluding non-relevant articles, a total of 25 studies that reported the outcome of interest in terms of mean±SD or median (IQR) were included and used for the final analysis (Fig 1).
A total of 27 studies were included for pooled WMD analysis of PC (Table 1). Of them, eight studies were included by splitting four studies [24, 25, 60, 67] into two based on PE severity. The included studies for PC analysis comprised a total of6, 250 pregnant women (1,943 PE and 4, 307NT). Of preeclamptic women, 319 women were with mild PE, 480 with severe PE  Table 2).The included studies comprised a total of 6, 609 pregnant women (2,034PE and 4,575NT).

The association of mean platelet volume with preeclampsia
The random-effects meta-analysis for MPV showed that the estimated pooled mean of MPV in PE patients and NT pregnant women was 9.8fl [95% CI; 9.6fl, 10 (Fig 3).

Heterogeneity and publication bias analysis
The existence of heterogeneity and publication bias was determined within included studies.
The heterogeneity of the included study was high according to Higgin's I-squared statistics

PLOS ONE
(I 2 = 99.7%; p<001 for PC and I 2 = = 97.7%; p<001 for MPV). The heterogeneity indicated that there was a high variation of studies which lead us to subgroup analysis. The subgroup analyses were performed based on the year of publication, PE severity, and region where the studies were conducted. However, there was no significant reduction in heterogeneity. The Egger's test was used to determine the presence of publication bias for the analysis of pooled WMD of PLT parameters between PE and NT pregnant women. The P-value in Egger's test showed that the publication bias was marginally insignificant in all PLT parameter analyses [P = 0.565 for PC and P = 0.811 for MPV]. All the values were greater than 5%; indicating no evidence of publication bias within included studies (Tables 3 & 4) and no need for trim and fill analysis. Besides, a funnel plot was also depicted to illustrate the presence/absence of publication bias in all PC (Fig 4) and MPV (Fig 5) analyses.

PLOS ONE
Preeclampsia has an association with both platelet count and mean platelet volume: Meta-analysis subgroup analysis by study design showed that there was a significant decrement of PC in the PE group compared to the NT group in case-control and cross-sectional studies with a pooled WMD of -43.5×10 9 /L [95% CI; -60.3×10 9 /L, -26.7×10 9 /L] and -53.1×10 9 /L [95% CI; -70.1× 10 9 /L,-36.0× 10 9 /L], respectively. However, there was no significant difference among the groups in cohort studies with WMD -20.1×10 9 /L [95% CI; -42.1×10 9 /L, 1.9×10 9 /L] in the retrospective cohort and -15.3×10 9 /L [95% CI; -34.5× 10 9 /L, 4.0× 10 9 /L] in prospective cohort studies (Fig 8). Subgroup analysis in estimated pooled weighted mean difference of mean platelet volume between preeclampsia and normotensive groups. The subgroup analysis of MPV by PE severity revealed that the pooled WMD of MPV between severe PE group and NT group was smaller [WMD: 1.7fl; 95%CI; 0.9fl, 2.6fl] compared to the WMD between mild PE and NT groups [WMD: 1.2fl; 95%CI; 0.6fl, 1.8fl], the difference was significant in both subgroups (Fig  9). The subgroup analysis by continent showed that there was a significant difference of MPV between groups in Asia, Europe, and Africa with a pooled WMD of 0.6fl [95% CI; 0.5fl, 0.8fl],

Sensitivity analysis
Sensitivity analysis was carried out on the pooled WMD of PC and MPV among PE and NT pregnant women. The analysis was done to evaluate the effect of each study on the estimated pooled WMD. The result showed that omitted studies did not show a significant effect on the pooled estimated WMD of PC analysis among PE and NT pregnant women (Fig 12). However, the sensitivity statistics on the pooled WMD of MPV analysis indicated that one study [70] has a determinant effect on the pooled estimated WMD of MPV with an estimate WMD of 1.13 which was out of the overall pooled estimated range [WMD: 0.98; 95% CI; 0.84, 1.124] (Fig 13).

Discussions
During a normal pregnancy, the level of PC drops physiologically due to hemodilution, increased PLT consumption in peripheral tissue, and increased PLT aggregation as a result of

PLOS ONE
increased levels of thromboxane A2 [71]. This Pregnancy-induced thrombocytopenia is mild and has no negative consequences for the mother or fetus, however, significant thrombocytopenia is associated with medical conditions and can have serious maternal-fetal consequences [72]. These changes may aggravate pregnancy-related disorders like PE. In established PE, changes in the coagulation system lead to a decrease in PC [24], which suggests an early sign of the disease [72]. Moreover, the progression of PE to the severe stage leads to increased turnover of PLTs [48]. The main objective of this systematic review and meta-analysis was to assess the association between PE and PC, PE and MPV. The association was determined by the pooled weighted mean difference. Accordingly, the majority of the included studies showed that PC was significantly lower in PE pregnant women than NT pregnant women [16,17,24,25,50,57,60,61,63,65,66,[68][69][70]. However, some studies showed that PC had not a significant difference between PE and NT pregnant women [26-28, 55, 56, 58, 62, 64]. Moreover, in the majority of the included studies, MPV was significantly higher in PE pregnant women than NT pregnant women [16, 17, 24, 25, 38, 50, 55, 56, 58, 60, 63-66, 68, 70]. Indeed, some studies showed MPV had no significant difference between PE and NT pregnant women [26,28,57,59,62,69]. The reason for the discrepancy might be the difference in sample size, study design, diagnostic method, geographical location, and/or variation in gestational week between cases and controls.

PLOS ONE
Nevertheless, in this systematic review and meta-analysis, there was a significant decrement of pooled PC in the PE group compared to the NT group [WMD:-41.45 × 10 9 /L; 95% CI; -51.8, -31.0]. This could be due to increased consumption of PLT due to an abnormal coagulation system along with PLT activation [15,18]. In preeclamptic women, plasma PLTs activation markers like β-thrombomodulin and PLT factor-4 significantly increased [69]. The increment of plasma PLTs activation markers like β-thrombomodulin and PLT factor-4 [73,74]and the expression of activation markers on the PLTs surface confirm PLT activation [75,76]. The activation of PLTs leads to PLT consumption. Besides, impaired endothelial synthesis of Prostacyclin and nitric oxide has been related to PE [69,77]. Both Prostacyclin and nitric oxide relax blood vessels and inhibit PLT activation [78]. Therefore the impairment of the synthesis of these molecules in PE may cause blood vessel constriction and leads to PLT activation and consumption [15,24].
Moreover, this systematic review and meta-analysis revealed that the overall pooled WMD of MPV values was significantly increased in the PE group when compared with the NT group [WMD: 0.98fl; 95%CI; 0.8, 1.1]. This demonstrated that PE is associated with a significant increase in MPV. The finding was in agreement with a systematic review and meta-analysis

PLOS ONE
done by Bellos et al. which claimed that MPV was significantly higher in preeclamptic patients than the NT pregnant women [79]. This could be due to the increased PLT synthesis in the bone marrow and release of large PLTs as a result of increased PLT consumption and destruction, increasing MPV in PE [80].
In the subgroup analysis by PE severity, the pooled WMD of all PLT parameters was smaller among women with mild PE and NT pregnant women compared to the difference between severe PE and NT pregnant women. This might be due to the increased activation of the coagulation system as the disease progresses from mild to severe stage [1]. So, it makes the WMD of PLT parameters in severe PE higher and more significant. The subgroup analysis by study design showed that the pooled WMD of PC was not significant in both retrospective and prospective cohort studies while MPV was not significant only in retrospective cohort studies. Moreover, Subgroup analysis based on continent showed the pooled WMD of MPV was significant in all regions while the pooled WMD of PC between PE and NT groups was significant in Asia and Africa regions.

Strengths and limitations of the study
The present review summarizes the current literature regarding the association of PC and MPV with PE which may give a clue to use these PLT parameters in the diagnosis of PE. It is

PLOS ONE
control design; thus, selection bias cannot be excluded and studies that were published other than in English languages were not included. Even though subgroup analysis was performed the heterogeneity was still observed in all analyses.

Conclusions
The pooled data from our systematic review and meta-analysis suggest that PE is associated with a lowering PC and increasing MPV. Thus, the measurement of PLT parameters, like PC and MPV, among pregnant women can be used as easily available, economical, and inexpensive clinical indicators in the assessment of PE. As a result, combining PLT Parameters with BP is more important than utilizing BP alone. However, the predictive performances like cutoff value, sensitivity, and specificity of these parameters have to be explored for the early diagnosis of PE. Further, multicenter longitudinal studies are required to evaluate their role at various GW of pregnancy.