Mean platelet volume as a predictor of cardiovascular risk: a systematic review and meta-analysis

Summary

Aim

To determine whether an association exists between mean platelet volume (MPV) and acute myocardial infarction (AMI) and other cardiovascular events. Platelet activity is a major culprit in atherothrombotic events. MPV, which is widely available in clinical practice, is a potentially useful biomarker of platelet activity in the setting of cardiovascular disease.

Methods and Results

We performed a systematic review and meta-analysis investigating the association between MPV and AMI, all-cause mortality following myocardial infarction, and restenosis following coronary angioplasty. Results were pooled using random-effects modeling. Pooled results from 16 cross-sectional studies involving 2809 patients investigating the association of MPV and AMI indicated that MPV was significantly higher in those with AMI than those without AMI [mean difference 0.92 fL, 95% confidence interval (CI) 0.67–1.16, P < 0.001). In subgroup analyses, significant differences in MPV existed between subjects with AMI, subjects with stable coronary disease (P < 0.001), and stable controls (P < 0.001), but not vs. those with unstable angina (P = 0.24). Pooled results from three cohort studies involving 3184 patients evaluating the risk of death following AMI demonstrated that an elevated MPV increased the odds of death as compared with a normal MPV (11.5% vs. 7.1%, odds ratio 1.65, 95% CI 1.12–2.52, P = 0.012). Pooled results from five cohort studies involving 430 patients who underwent coronary angioplasty revealed that MPV was significantly higher in patients who developed restenosis than in those who did not develop restenosis (mean difference 0.98 fL, 95% CI 0.74–1.21, P < 0.001).

Conclusions

Elevated MPV is associated with AMI, mortality following myocardial infarction, and restenosis following coronary angioplasty. These data suggest that MPV is a potentially useful prognostic biomarker in patients with cardiovascular disease. Whether the relationship is causal, and whether MPV should influence practice or guide therapy, remains unknown.

Platelets play a pivotal role in atherothrombosis, the major cause of most unstable coronary syndromes [1]. Central to the pathogenesis of occlusive arterial disease is the activation of platelets at sites of vascular injury via pathologically exaggerated and deregulated versions of the protective mechanisms involved in hemostasis [1]. Platelets secrete and express a large number of substances that are crucial mediators of coagulation, inflammation, thrombosis, and atherosclerosis [2,3]. The demonstrated ability of antiplatelet drugs to reduce cardiovascular events has reinforced the major role of platelets in the atherothrombotic process [4].

Although measuring platelet activity by any of a wide variety of methods has been reported to identify individuals at increased risk for cardiovascular events, it remains a research tool that is yet to be included in routine clinical decision-making. Potential reasons include a lack of sufficient data about the optimal method of platelet testing, unknown optimal cut-off for distinguishing increased risk, and the uncertainty about the interpretation and clinical utility of results. Additionally, many methods are costly, are time-consuming, and require specialized equipment.

Within an individual, platelets are heterogeneous in size and density. Mean platelet volume (MPV), the most commonly used measure of platelet size, is a potential marker of platelet reactivity. Although there is still uncertainty about the most precise methodology for measuring MPV, it is routinely available in the inpatient and outpatient setting at a relatively low cost. Larger platelets are metabolically and enzymatically more active [5], and have greater prothombotic potential [6]. Elevated MPV is associated with other markers of platelet activity, including increased platelet aggregation, increased thromboxane synthesis and β-thromboglobulin release, and increased expression of adhesion molecules [7]. Furthermore, higher MPV is observed in patients with diabetes mellitus [8], hypertension [9], hypercholesterolemia [10], smoking [11], and obesity [12], suggesting a common mechanism by which these factors may increase the risk of cardiovascular disease.

The aim of this systematic review and meta-analysis, the first that we are aware of, is to obtain the best possible estimate of MPV and cardiovascular disease by combining data from all relevant cross-sectional and cohort studies evaluating MPV and cardiovascular disease. The emphasis of this article is to detail: (i) whether there exists an association between MPV and acute myocardial infarction (AMI); (ii) whether elevated MPV is associated with all-cause mortality following a myocardial infarction (MI); and (iii) whether MPV is associated with restenosis following percutaneous coronary intervention (PCI).

Materials and methods

Literature search and study selection

Relevant studies were identified by systematic searches of the scientific literature for all reported studies of associations between cardiovascular disease and MPV (using the terms ‘cardiovascular disease’, ‘coronary heart disease’, ‘myocardial infarction’, ‘percutaneous coronary intervention’, and ‘mean platelet volume’). We searched the MEDLINE database for articles published from January 1950 to September 2008, and identified additional studies by hand-searching references of original articles or review articles on this topic and by personal contact with investigators. All studies providing data on MPV in patients with AMI, MPV in patients undergoing coronary angioplasty or cohort studies where MPV was measured following myocardial infarction and followed longitudinally were identified and included. The search strategy had no language restrictions. We identified 161 potential studies. We excluded 25 because they were not clinical studies (reviews, editorials, or letters to the editor without data). Of the 136 clinical studies, 105 investigated MPV in clinical endpoints other than AMI and coronary restenosis, including cardiac risk factors (e.g. hypertension, diabetes, and hypercholesterolemia), cerebrovascular disease, and non-obstructive cardiac disease. Of the 31 studies that measured MPV in subjects with AMI or coronary angioplasty, two did not have non-AMI controls, three did not provide a measure of distribution of the MPV either in the original article or upon request, one cohort study did not provide quantitative data on outcomes, and one used an identical study population as another included study. These seven studies were excluded from the final selection. Figure 1 shows the study selection process.

Fig. 1.

Flow chart showing results of the search strategy and reasons for exclusion. CAD, coronary artery disease; MPV, mean platelet volume; UA, unstable angina.

Data extraction

The primary investigator (SC) abstracted data from each study and compiled summary tables. For all included studies, the following data were extracted: lead author, publication year, study population, numbers of AMI cases and non-AMI controls, numbers of restenosis cases and no restenosis controls, sex, mean age, study design, MPV assay methods, and, for cohort studies, follow-up time and primary endpoints. Mean MPVs with standard deviations and P-values between cases and controls were either extracted directly where available or calculated from the data provided. In the few studies that reported MPV ranges or medians (not reporting mean MPV), corresponding authors were contacted for detailed information.

Statistical analysis

For cross-sectional studies, the primary summary measure of association was the overall random-effects mean difference in MPV between cases and controls. For AMI studies, the comparator group was non-AMI. To assess the robustness of our findings, we evaluated the random-effects mean difference in MPV vs. each component of the non-AMI controls (e.g. patients with unstable angina, those with stable coronary disease, and those without coronary disease).

For prospective cohort studies evaluating mortality following MI, summary odds ratios (ORs) and 95% confidence intervals (CIs) were used to measure the association between MPV and incident mortality. For two studies, elevated MPV was defined as a value ≥ 10.3 fL (upper tertile) [13,14], whereas, in one study, elevated MPV was defined by a value > 9 fL (upper 40th percentile) [15]. ORs were calculated from the effect estimates of the individual studies, using both a fixed-effects model (weighting each estimate by its standard error with the Mantel–Haenszel method) and a random-effects model (estimating between-study variance in effect measures according to the method of DerSimonian and Laird) [16].

For PCI studies, the cases and controls were defined as those who did and did not develop restenosis. Summary estimates are presented with 95% CIs.

A value of P < 0.05 was considered to be statistically significant. We assessed heterogeneity among studies using I² statistics. When P < 0.05, the presence of heterogeneity was considered to be statistically significant, and when I² > 50%, the magnitude of heterogeneity was considered to be substantial [17]. However, because a lack of heterogeneity does not necessarily imply homogeneity, a summary OR was calculated, using a random-effects model, from the ORs and 95% CIs for each endpoint in each study. Funnel plots of effect size against standard error (SE) and a modified linear regression test were used to describe the presence of publication bias [18]. All statistical analyses were performed using the Comprehensive Meta-Analysis program (Biostat, Englewood, NJ, USA).

Results

MPV and MI

Sixteen cross-sectional studies involving 2809 subjects investigating a relationship between MPV and AMI were included in this analysis [19–34]. MPV, measured at the time of diagnosis, was compared in patients with and without an AMI. The non-AMI group included patients with unstable angina (seven studies), with stable coronary artery disease (seven studies), and without coronary disease (12 studies). Several studies had more than one comparator group. Table 1 summarizes these study designs.

Table 1.

Study characteristics [acute myocardial infarction (AMI) cross-sectional]

	Population		AMI
Source	Cases	Controls	Yes	No	% Men	Mean age (years)
Yilmaz [34]	NSTEMI	UA, stable CAD, sex-matched and age-matched	111	225	72	60
Boos [20]	AMI	Stable CAD, healthy	8	19	28	61
Avramakis [19]	AMI	UA, healthy	86	164	–	63
Khandekar [25]	AMI	UA, stable CAD, sex-matched and age-matched healthy	77	117	50	53.9
Kiliçli-Çamur [26]	AMI	UA, stable CAD, chest pain with no CAD	70	130	59.5	57.7
Mathur [29]	AMI	UA, healthy volunteers	15	15	52	47
Senaran [30]	AMI	UA, healthy volunteers	20	37	67	59
D’Erasmo [22]	AMI	Sex-matched and age-matched healthy	15	60	47	68
Hendra [24]	AMI	No AMI	147	150	75 (cases)	65
Erne [23]	AMI	Stable CAD, healthy	55	526	–	–
Trowbridge [31]	AMI	IHD	103	72	88	57.3
Kishk [27]	AMI	IHD, healthy young males	70	95	91	56.7 (cases)
Trowbridge [32]	AMI	UA	7	6	–	58.5
Van der Lelie [33]	AMI	Non-AMI adm to CCU	12	60	–	–
Martin [28]	AMI males	Healthy males	15	22	100	48
Cameron [21]	AMI males	Age-matched admitted healthy males	100	200	100	57.2

CAD, coronary artery disease; CCU, cardiac care unit; IHD, ischemic heart disease; UA, unstable angina.

Fourteen of the 16 studies demonstrated that patients with AMI had higher MPVs than those without AMI [19–28,30–32,34]. One study found no significant difference in MPV between the groups with and without AMI [33], and one study demonstrated a higher MPV in the non-AMI group [29]. The mean MPV in all 16 studies was 8.95 fL (95% CI 8.49–9.41). The MPV was higher in patients with AMI (9.24 fL, 95% CI 8.39–10.08) than in those without AMI (8.48 fL, 95% CI 7.77–9.19). The estimated mean difference in MPV was 0.92 fL (95% CI 0.67–1.16, P < 0.001) (Fig. 2). Because different control populations were used in each study, we performed subgroup and sensitivity analyses to further evaluate the relationship between MPV and different disease states. After removal of studies that used a population without coronary disease, studies that used a stable coronary disease population, or studies that used a population with unstable angina as the control group, the observation remained consistent – mean MPV remained higher in AMI groups than in all other groups.

Fig. 2.

Random-effects pooled mean difference of mean platelet volume (MPV) between acute myocardial infarction (AMI) cases and non-AMI controls.

Analysis of estimated mean difference in cross-sectional studies revealed no significant difference in MPV between patients with AMI and those with unstable angina (estimated mean difference 0.23 fL, 95% CI – 0.15–0.61, P = 0.237). MPV was significantly higher in the AMI group than in those with stable coronary artery disease (estimated mean difference 1.00 fL, 95% CI 0.51–1.50, P < 0.001) and in those without known coronary disease (estimated mean difference 1.11 fL, 95% CI 0.79–1.42, P < 0.001). No significant publication bias was detected among the cross-sectional analyses.

MPV and risk of death following AMI

In three cohort studies, mortality following an AMI was reported [13–15]. Table 2 summarizes the designs of these three studies. One enrolled patients with ST-elevation MI (STEMI) [14]; the other two enrolled any type of MI [13,15]. One study measured MPV 6 months after the initial event [13], whereas the other two studies measured MPV at the time of admission for the AMI [14,15]. One study reported in-hospital mortality [15], one reported mortality at 6 months [14] and one reported mortality at 2 years following the index AMI [13].

Table 2.

Study characteristics [acute myocardial infarction (AMI) cohort]

Source	Population	Sample size	% Men	Mean age (years)	Follow-up period
Huczek [14]	STEMI	388	72	60.0	6 months
Pabón Osuna [15]	AMI	1082	78	67.7	In-hospital phase of myocardial infarction
Burr [13]	AMI	1714	100	56.4	2 years

STEMI, ST-elevation myocardial infarction.

Among 3184 patients with AMI, those with an elevated MPV had a significantly higher risk of death than those with a normal MPV (11.5% vs. 7.1%, OR 1.65, 95% CI 1.12–2.52, P = 0.012) (Fig. 3). When the two studies that used a cut-off MPV of ≥ 10.3 fL [13,14] were analyzed separately from the study using a cut-off MPV of 9 fL [15], the increased risk of mortality with an elevated MPV was two-fold greater (OR 2.01, 95% CI 1.39–3.70, P < 0.001). No significant publication bias was detected among these studies.

Fig. 3.

Random-effects pooled odds ratio of mean platelet volume and mortality following myocardial infarction from prospective studies. CI, confidence interval; OR, odds ratio.

MPV and risk of restenosis

Table 3 summarizes the five cohort studies that investigated the association between MPV and risk of restenosis following PCI [35–39]. In four of the five studies, restenosis was defined as > 50% diameter narrowing [35–37,39]; in the remaining study, reocclusion defined by thrombolysis in myocardial infarction flow grade 0 or 1 was deemed to be evidence of restenosis [38]. In one study, patients who underwent stent implantation were excluded [36]; in another study, all patients underwent stent implantation [39], whereas in the remaining three studies, stent use was not described [35,37,38]. All patients received aspirin during the PCI and indefinitely thereafter. Patients in one study were also given either ticlodipine or clopidogrel during and after the procedure for at least 3 months [36]. The duration of follow-up ranged from 6 weeks to 8 months.

Table 3.

Study characteristics (restenosis cohort)

Source	Population	Sample size	% Men	Mean age (years)	Definition of restenosis	Follow-up
Yang [39]	Patients undergoing elective PTCA	174	79	60	> 50% diameter stenosis in a major vessel or major side branch	Within 6 months
Norgaz [36]	Stable angina, single-vessel disease	60	90	57.5	> 50% diameter stenosis	6 months
Avci [35]	Stable angina, single-vessel disease	102	77	56	> 50% diameter stenosis and/or > 50% late loss of acute luminal gain at follow-up coronary angiography	4–8 months
Terres [38]	Single-vessel occlusive disease	47	89	58.8	Reocclusion with TIMI flow grade of 0 or 1	8 ± 2 weeks
Smyth [37]	Single-vessel disease	47	72	57.0	> 50% diameter stenosis in two projections	4–8 months

PTCA, percutaneous transluminal coronary angioplasty; TIMI, thrombolysis in myocardial infarction.

Among the 430 subjects who underwent a PCI, 176 (40.9%) developed restenosis. In the overall cohort, the estimated mean level of baseline MPV was 8.29 fL (95% CI 8.05–8.53). Baseline mean MPV was significantly higher in patients who developed restenosis (8.67 fL, 95% CI 8.44–8.87) than in those who did not develop restenosis (7.96 fL, 95% CI 7.80–8.13). The estimated mean difference was 0.98 fL (95% CI 0.74–1.21, P < 0.001) (Fig. 4).

Fig. 4.

Random-effects pooled mean difference of mean platelet volume (MPV) between restenosis cases and no restenosis controls. AMI, acute myocardial infarction.

Discussion

This is the first systematic review and meta-analysis that we are aware of to evaluate the association between MPV and cardiovascular risk. The analysis, drawn from 24 studies of over 6000 subjects, supports the hypothesis that elevated MPV is a cardiovascular risk factor – that it is associated with adverse cardiovascular events. There were three major findings in this analysis. First, we observed significant estimated mean differences in MPV between AMI and non-AMI populations. These differences were most pronounced when those with AMI were compared with those with stable coronary disease or those without coronary disease. Second, elevated MPV was associated with increased mortality following a myocardial infarction. Third, among patients undergoing coronary angioplasty, MPV was significantly higher in patients who developed restenosis.

Platelet function and cardiovascular disease

The ex vivo measurement of platelet function is believed to provide an index of the functional capacity of platelets [40]. There is enhanced platelet activation during acute coronary syndromes [41], and measurement of platelet activation and/or aggregation may provide prognostic information in patients at risk for or following a cardiovascular event [42,43]. The relationship between residual platelet aggregability in patients on antiplatelet therapy and clinical outcome has also been investigated, with most studies reporting that high post-treatment platelet aggregability, as determined by different methods and defined in a variety of ways, is associated with adverse cardiovascular events [44].

Although a number of studies have demonstrated that platelet activity and/or aggregation is associated with the future development of adverse cardiovascular events, biomarkers of platelet activity have yet to be widely used in clinical practice. It is difficult to standardize the blood collection and processing and handling procedures for multicenter laboratory testing. The major goal of the present study was to characterize the relationship between MPV, a simple, inexpensive and widely available marker of platelet activity, and cardiovascular disease.

Elevated MPV as a risk factor and a prognostic indicator in cardiovascular disease

Increased MPV has been noted in subjects with cardiovascular risk factors, such as smoking, diabetes, obesity, hypertension, and hyperlipidemia [8–12]. In subjects with established cardiovascular disease, elevated MPV may be a marker for adverse cardiovascular events. In the largest trial to date, investigators reported that, following an MI, MPV was an independent risk factor for recurrent ischemia or death [13,45]. Importantly, MPV did not correlate with known ischemic heart disease risk factors, suggesting that an elevated MPV may contribute to adverse events through a novel mechanism. A recent study in the setting of STEMI suggested that MPV may have a role in guiding therapy [14]; only subjects with elevated MPV had a mortality reduction from abciximab. The ability to identify patients in whom targeted use of glycoprotein IIb/IIIa inhibitors (or other antiplatelet agents) is effective would be clinically helpful. Various drugs, including losartan [46] and lipid-lowering therapies [47,48], may decrease MPV, but no study to date has shown that lowering MPV lowers cardiovascular risk.

Possible mechanisms of action of MPV in cardiovascular disease

Although the precise biological pathways by which elevated MPV might influence the development or progression of cardiovascular disease and acute coronary syndromes are not completely understood, multiple mechanisms may be involved. Larger platelets are metabolically and enzymatically more active than smaller platelets, containing more prothrombotic material, with increased thromboxane A₂ and B₂ per unit volume and glycoprotein IIb–IIIa receptor expression [6,49]. They show greater aggregability in response to ADP [50] and decreased inhibition of aggregation by prostacyclin in vitro [51]. Larger platelets are denser and contain more α-granules [52], which can release prothrombotic substances, including platelet factor 4 [53], P-selectin [6], and platelet-derived growth factor [54], a chemotactic and mitogenic factor contributing to vascular neointimal proliferation [55]. Finally, larger platelets are more often reticulated, and this is an independent predictor of poor response to dual antiplatelet therapy [56]. Figure 5 illustrates some of these potential mechanisms by which larger platelets contribute to cardiovascular disease.

Fig. 5.

Potential mechanisms of action of mean platelet volume in cardiovascular thrombosis. Larger platelets contain more α-granules [52], which release prothrombotic factors such as P-selectin (P-sel), serotonin (5-HT), ADP, and β-thromboglobulin (β-TG) [6,7,53,63]. Larger platelets also have increased levels of thromboxane A₂ (TxA₂) [6,7] and express more adhesion receptors, glycoprotein (GP) Ib and GPIIb–IIIa [49]. These factors have a variety of effects on inflammation and endothelial function, promoting cell adhesion and aggregation, and vasoconstriction, and ultimately inducing thrombosis. Platelet size may be determined by megakaryocyte ploidy, which is regulated by thrombopoietin (TPO) [30,58] and interleukin-6 (IL-6) [59].

Regulation of platelet size is multifactorial. Platelet volume is determined both during megakaryopoiesis and during thrombopoiesis, and may not correlate with the age of the platelet [57]. MPV has been shown to be positively associated with levels of thrombopoietin and interleukin-6, cytokines that regulate megakaryocyte ploidy [30,58,59] and platelet number [60]. The relationship between MPV and the number of platelets is unclear [60]. Several studies have reported that increases in platelet volume are often associated with decreases in platelet count [14,24,39], perhaps as a result of small platelets being consumed in order to maintain a constant platelet functional mass [61]. Conversely, one study found that platelet count and platelet volume both increase during stimulated thrombopoiesis [62], indicating that platelet count and volume may be regulated by independent mechanisms [63]. The relationship between MPV and platelet count is not completely understood. However, most studies found no significant association between increased platelet count and incidence of AMI, restenosis, or long-term mortality [13,14,20,23–26,30,33,35–37]. Furthermore, several studies demonstrated a significant association between MPV and AMI and other cardiovascular events, even after adjusting for platelet count [14,26,34,35,38].

Analyses of the studies included in our analysis do not and cannot provide insights into whether elevated MPV actually causes cardiovascular disease and acute coronary syndromes, or is merely associated with them. However, it is likely that an increased MPV pre-dates an AMI rather than results from it. As the average platelet lifespan in blood is 7–10 days, the vast majority of platelets sampled at the time of hospital admission for an AMI would have been circulating before the event [28]. Martin et al. found that platelet size remained elevated at 6 weeks after hospital discharge, indicating that increases in MPV are chronic rather than acute [28]. Our analysis supports the hypothesis that platelet size may play a role in both the development and the consequences of cardiovascular disease. Recent genome-wide studies have identified loci involved with MPV [64], and we hope to extend this work by investigating the relationship between genetic determinants of MPV, actual MPV values, and cardiovascular phenotype. Prospective studies of MPV and cardiovascular disease among healthy populations are needed to address these important questions.

Clinical implications

MPV is simple and inexpensive to obtain, easy to interpret, and routinely measured by automated cell counters. As compared with other markers of platelet activity, which require special hemostasis laboratories, because of the stringent requirement for nearly perfect phlebotomy, blood-processing procedures, platelet isolation, and specialized equipment, MPV is a practical and prognostically important biomarker of cardiovascular disease. Formal comparison of MPV and other novel biomarkers is needed to fully characterize its relative clinical utility.

Limitations

A limitation of MPV as a marker of prognosis is uncertainty as to whether the relationship between platelet size and cardiovascular risk is continuous or categorical. A broad spectrum of MPV values was seen in our analysis, ranging from 6.5 fL to 11.75 fL in AMI groups. Given the significant but still narrow absolute difference in mean values of MPV between AMI and non-AMI groups, the clinical utility of an isolated MPV value in determining an individual’s cardiovascular risk is unclear. The MPV cut-off point that is most useful for predicting cardiovascular events in clinical practice remains unknown. Another important consideration is the variability in the timing and methods of blood sample preparation and the type and calibration of particle counters, which may have had a significant impact on the measurement of MPV. Whereas most studies measured MPV at the time of diagnosis, one study did not measure MPV until 6 months after the initial MI [13]. Furthermore, the time between collection of blood samples and measurement of MPV, although mostly standardized within each study, differed significantly between studies, from 5 min [20] to over 24 h [13]. A majority of studies used EDTA as an anticoagulant [13–15,19–26,30,32–39],which has been shown to cause platelet swelling in a time-dependent manner [7]. The four studies in which citrate was used as an anticoagulant [27,28,30,31,46] had lower mean MPV values than those using EDTA. However, as sample preparations did not vary within studies, but only between them, we do not believe that this limitation significantly altered our results or the conclusions that should be drawn from them.

Owing to the nature of this analysis, it is impossible to adjust for baseline differences among the studies. Therefore, the results should be interpreted in this context. Nevertheless, several studies that we included did adjust for differences in baseline demographics, clinical risk factors, and platelet count, and demonstrated an independent association between MPV and cardiovascular events. Finally, although MPV was ascertained at a single time point, data suggest that MPV values remain constant when assessed over a period of days [28,65] or weeks [28].

Conclusions

The present study suggests that MPV is higher in patients presenting with an AMI and in patients who develop restenosis following coronary angioplasty. Increased MPV is also associated with higher mortality following MI. These findings raise the hypothesis of the potential importance of MPV in the underlying pathophysiology of cardiovascular disease. Further studies are required to assess whether MPV provides added value in being able to identify patients at increased clinical risk and whether therapeutic modification of this marker may lead to improved cardiovascular care.