Article, Hematology

Diagnostic value of platelet indexes for pulmonary embolism

a b s t r a c t

Aim: The aims of the study are to investigate the changes in platelet indexes, including platelet count, platelet distribution width , and Mean platelet volume , in patients with pulmonary embolism (PE) and to evaluate their diagnostic values in relation to this disease.

Methods: The study included 70 patients with PE as the observation group and 75 patients without PE as the control group. The differences in platelet count, PDW, MPV, D-dimer, and other indicators between the 2 groups were retrospectively analyzed.

Results: Platelet distribution width and MPV were significantly higher in patients with PE than in the controls (16.40% [13.70%-16.85%] vs 16.00% [11.28%-16.60%], P = .023; 9.91 +- 1.40 fL. vs 8.84 +- 1.68 fL, P b .001, respec-

tively). Multivariate logistic regression analysis showed that MPV and D-dimer were independent influencing factors for the diagnosis of PE. receiver operating characteristic curve analysis showed that MPV (with the cut-off value set at 8.45 fL) had a sensitivity of 88.7%, negative predictive value of 78.7%, specificity of 50.0%, and positive predictive value of 61.9%. D-Dimer (with the cut-off value set at 835.5 ug/L) had a sensitivity of 80.6%, negative predictive value of 77.8%, specificity of 62.1%, and positive predictive value of 66.7%. The combi- nation of D-dimer and MPV resulted in an increase in the area under the curve (0.799; 95% confidence interval, 0.724-0.874; P b .001).

Conclusion: Higher PDW and MPV levels are noticed in patients with PE. The combined application of MPV can improve the diagnostic value of D-dimer for PE.

(C) 2015


Pulmonary embolism (PE) is a clinical and pathophysiological pul- monary circulation disorder syndrome caused by partial or complete occlusion of the pulmonary artery. The existing literature has discussed the value of multiple biomarkers, such as D-dimer, brain natriuretic pep- tide (BNP), and troponin, in the diagnosis of PE and in making a progno- sis, but these biomarkers could not be used as the foundation for diagnosis [1]. Among these biomarkers, D-dimer is widely used, but its poor specificity has limited its diagnostic yield. It is important to explore other laboratory parameters, such as platelet indexes. Thrombocytosis, defined by a platelet count greater than 500 x 109/L, has been as- sociated with an increased risk of symptomatic acute PE [2]. Gunay et al

[3] reported that the mean values of mean platelet volume , platelet distribution width (PDW), and PLT were higher in PE groups than in controls (P b .05). However, Hilal et al [4] found that MPV did not correlate with the diagnosis of acute PE. Therefore, the role of plate- let indexes in diagnosing PE is uncertain. The aim of this study is to investigate the diagnostic value of platelet indexes for PE.

? The authors have no financial conflicts of interest.

* Corresponding author at: Department of cardiovascular medicine, Shantou Central Hospital, No. 114, Waima Road, Shantou City, Guangdong Province, China. Tel./fax: +86 754 88550450.

E-mail address: [email protected] (J. Huang).

Materials and methods

Patients and study design

The study retrospectively enrolled 145 consecutive patients who were admitted to our hospital between September 2009 and January 2014. Ac- cording to the European Society of Cardiology guidelines [5], PE was con- firmed by Computed tomography pulmonary angiography . Based on the CTPA results, these 145 patients were divided into 2 groups: 70 (32 males and 38 females) with PE as the observation group, aged from 20 to 86 (60.37 +- 13.75) years and 75 (43 males and 32 females) without PE as the control group, aged from 15 to 83 (56.84 +- 16.04) years. Age and sex distributions in the patients were similar between the 2 groups (P N .05). Exclusion criteria included acute coronary syndrome, hematological disorders such as thrombocytosis and idiopathic thrombocytopenic pur- pura, severe hepatic and renal diseases, chronic pulmonary hypertension,

diabetes mellitus, malignancy, and anticoagulation therapy used.

Computed tomography pulmonary angiography

All patients underwent CTPA using a 64-row multiple detector com- puted tomography scanner (LightSpeed VCT; GE Healthcare, Waukesha, WI). The imaging result of each patient was reported based on the consen- sus of 2 radiological technicians who were blinded to the plasma results.

0735-6757/(C) 2015

J. Huang et al. / American Journal of Emergency Medicine 33 (2015) 760763 761

Laboratory examination

All blood samples were drawn from the antecubital vein upon ad-

mission. For measurements of platelet indexes, blood was analyzed in


PE (n = 70)

Non-PE (n = 75)


an automated blood cell counter (SYSMEX XS-1000i; SYSMEX Corpora-

Age (y)

60.37 +- 13.75

56.84 +- 16.04


tion, Kobe, Japan) containing EDTA-K2 as an anticoagulant. The normal

Female, n (%)

38 (54.3%)

32 (42.7%)


values of PLT, PDW, and MPV are 100 to 300 x 109/L, 11.0% to 17.1%,

RBC (x1012/L)

4.03 +- 0.86

4.17 +- 0.92


Table 1 Comparison of the clinical and laboratory characteristics of the patients with PE and with- out PE

and 7.5 to 11.0 fL, respectively. Plasma D-dimer levels were measured by an immunoturbidimetric assay (ACL TOP; Instrumentation Labora- tory, Fullerton, CA). The normal values of D-dimer are 0 to 1000 ug/L. Other indicators were detected at the same time, including N-terminal pro-brain natriuretic peptide (NT-proBNP), troponin I , Red blood cell count (RBC), hemoglobin (Hgb), red cell distribution width (RDW), fibrinogen (Fib), C-reactive protein (CRP), serum creatinine (SCr), and low-density lipoprotein cholesterol (LDL-C). All assays were performed by laboratory technicians who were blinded to the CTPA results.

Data processing

Data including clinical characteristics such as age, gender, medical history, CTPA results, and plasma results were compiled into spread- sheet format (Microsoft Office Excel 2003; Microsoft Corp, Seattle, WA) for subsequent analysis.

Statistical analysis

Statistical analysis was performed using the SPSS package (version 20.0; SPSS for Windows, Chicago, IL). Descriptive statistics were repor- ted, including the mean, SD, median, interquartile range, and percen- tage. The Student t or Mann-Whitney U test was used for comparisons between 2 groups, depending on whether the data accorded with the normal distribution.

Variables with categorical data were statistically compared using the ?2 or Kruskal-Wallis test. Logistic regressions were used to perform multivariate analyses. receiver operating characteristic curves were analyzed to assess the optimal cut-off values of the influence fac- tors. Sensitivity, specificity, negative predictive value (NPV), and posi- tive predictive value (PPV) were calculated for the chosen cut-off values. P b .05 was considered statistically significant.

The ? statistic was used to evaluate the interrater reliability of CTPA results between the 2 radiological technicians.


There were no significant differences in PLT between the 2 groups (205.0 [157.5-289.8] x 109/L vs 224.0 [167.0-261.0] x 109/L, P =

.582). Platelet distribution width was higher in the patients with PE than in the patients without PE (16.40% [13.70%-16.85%] vs 16.00% [11.28%-16.60%], P = .023). Mean platelet volume was significantly higher in patients with PE than in the controls (9.91 +- 1.40 fL. vs 8.84 +- 1.68 fL, P b .001). The differences in Fib, TNI, D-dimer, and NT-proBNP between the 2 groups were also statistically significant (P b .05) (Table 1). The ? value was .812.

Univariate logistic regression analysis showed that factors in the diag- nosis of PE included PDW, MPV, TNI, D-dimer, and NT-proBNP (Table 2). Including PDW, MPV, TNI, D-dimer, and NT-proBNP as variables, multivariate logistic regression analysis showed that MPV and D-dimer

were independent factors for the diagnosis of PE (Table 3).

D-Dimer, MPV, and D-dimer + MPV were analyzed for ROC curves. The results indicated that the area under the curve (AUC) of D-dimer was larger than the AUC of MPV, and combining D-dimer and MPV re- sulted in an increase in the AUC (0.799; 95% confidence interval [CI], 0.724-0.874; P b .001) (Table 4 and Figure), suggesting that the diagnos- tic value of D-dimer for PE was increased by the combination. Taking the maximum Youden’s index (sensibility + specificity – 1) as standard, the optimal cut-off values were assessed. Sensitivity, specificity, NPV,

Hgb (g/L) RDW (%) PLT (x109/L) PDW (%) MPV (fL)

CRP (mg/L) Fib (g/L)

124.84 +- 16.12

123.60 +- 16.49


13.95 (12.98-15.70)

13.70 (12.80-14.40)


205.0 (157.5-289.8)

224.0 (167.0-261.0)


16.40 (13.70-16.85)

16.00 (11.28-16.60)


9.91 +- 1.40

8.84 +- 1.68


28.10 (14.30-62.80)

11.60 (2.75-87.50)


5.56 +- 1.73

5.12 +- 2.15


TNI (ng/mL) 0.095 (0.031-0.454)

0.016 (0.004-0.037)


D-Dimer (ug/L) 3860.0 (1038.0-6834.5)

583.0 (171.0-2583.0)


NT-proBNP (pg/mL) 787.50 (210.13-2355.50)

78.28 (22.72-251.10)


SCr (umol/L) 81.35 (64.45-101.78)

76.40 (59.30-93.20)


LDL-C (mmol/L) 1.92 +- 0.63

1.89 +- 0.71


and PPV were calculated for the chosen cut-off values (Table 5). When the cut-off value was set at 8.45 fL, MPV had a sensitivity of 88.7%, NPV of 78.7%, specificity of 50.0%, and PPV of 61.9%. D-Dimer had a sen- sitivity of 80.6%, NPV of 77.8%, specificity of 62.1%, and PPV of 66.7% when the cut-off value was set at 835.5 ug/L. The combination of MPV and D-dimer had increased specificity (90.9%) and PPV (84.6%).


Pulmonary embolism and Deep venous thrombosis are 2 manifestations of venous thromboembolism , and PE is commonly a consequence of DVT. The annual incidence of PE in the United States has been estimated at 600 000 cases. The main cause of death in the Acute stage is acute right ventricular failure, and the acute case fatality rate ranges from 7% to 11%. Pulmonary embolism is responsible for more than 100 000 deaths per year in the United States alone [1,5].

Pulmonary embolism is a severe disease endangering the whole world, with poor prognosis and high mortality. Its clinical presentation is often atypical, and the final diagnosis still depends on CTPA or pulmo- nary angiography [5], which are often difficult to perform in a timely manner because of the acuteness of the disease or lack of available equipment. Thus, the discovery of easily available and rapid diagnostic markers for acute PE is necessary. The common blood markers include D-dimer, troponin, BNP, and others [1,5]. D-Dimer is a sensitive marker of thrombosis and secondary fibrinolytic activity. The specificity of D- dimer for VTE is poor because fibrin is produced in a wide variety of con- ditions, such as cancer, inflammation, pregnancy, and infection [5]. For example, the enzyme-linked immunosorbent assay and quantitative rapid enzyme-linked immunosorbent assay dominate the rank order

Table 2

Univariate logistic regression analysis of PE


Regression coefficient




95% CI



























































































762 J. Huang et al. / American Journal of Emergency Medicine 33 (2015) 760763

Table 3

Multivariate logistic regression analysis of PE


Regression coefficient




95% CI













for the following values for PE: sensitivity, 0.95 (95% CI, 0.85-1.00); neg- ative likelihood ratio, 0.13 (95% CI, 0.03-0.58); sensitivity, 0.95 (95% CI, 0.83-1.00); and negative likelihood ratio, 0.13 (95% CI, 0.02-0.84), respectively [6]. The NPV of D-dimer is high, and the PPV is low. Our re- search was broadly consistent with this description. Current guidelines recommend that a negative D-dimer result in a highly sensitive assay safely excludes PE in patients with a low or moderate clinical probability, whereas a moderately sensitive assay excludes PE only in patients with a low clinical probability [5].

Troponins are elevated in patients with submassive or massive PE with right ventricular dysfunction and poor prognosis [7]. However, ele- vated troponin is detected in many Pathologic conditions. The common differential diagnoses include acute coronary syndrome, stroke, end- stage renal disease, acute heart failure, sepsis, acute perimyocarditis, Takotsubo, and tachycardia [8]. Brain natriuretic peptide and NT- proBNP are confirmed markers of heart failure that are associated with right ventricular dysfunction and higher all-cause inhospital or short- term mortality in patients with acute PE [9]. In our research, TNI and NT-proBNP were also significantly increased in PE. Cytoplasmic heart- type fatty acid binding protein is a new type of biomarker of myocardial damage. It was reported to be superior to Cardiac troponin T and NT- proBNP in the prediction of 30-day mortality of acute PE [10]. In recent years, RDW has been shown to be increased in patients with PE, and ele- vated RDW was an independent predictor of short-term mortality in PE [11]. However, none of these values was recommended as the standard for diagnosis of PE till now. Nordenholz et al [12] compared the diagnostic accuracy of 50 protein Biological markers of PE by measuring their AUC and found that only D-dimer, CRP, and Myeloperoxidase (the AUC was 0.93, 95% CI, 0.89-0.98; 0.78, 95% CI, 0.70-0.90; 0.78, 95% CI,

0.71-0.85, respectively) demonstrated sufficient diagnostic accuracy. So far, there are no other biomarkers playing more important roles in the diagnosis of PE than D-dimer.

Platelets are involved in multiple links of the process of thrombus formation and development. The changes in platelet parameters reflect the platelet activation. Among them, MPV represents the average volume of a single platelet and reflects the regeneration of platelets. Platelets with elevated MPV are immature, a common situation in people who have increased platelet destruction and good bone marrow compensatory function. Park et al [13] reported that platelets with large volume were more active and had a stronger coagulating effect. Platelet distribution width reflects the discreteness of the size of platelets in the blood. Increased PDW means platelet size disparity. Platelet count decreases when platelet consumption increases.

To our knowledge, the association of platelet indexes with VTE (in- cluding DVT and PE) is a relatively uninvestigated area. In particular, MPV has become interested in recent studies. Varol et al [14] revealed that MPV was significantly higher and PLT was significantly lower among patients with acute PE compared with the control group (9.6 +- 1.0 fL. vs 8.1 +- 0.8 fL, P b .001, and 227.1 +- 77.0 x 109/L vs

268.7 +- 58.4 x 109/L, P b .001, respectively). A prospective population-based study of 25 923 subjects who participated in the

Table 4

The AUC of ROC curves of D-dimer, MPV and D-dimer + MPV





95% CI











D-dimer + MPV





Figure. Comparison of ROC curves of D-dimer, MPV, and D-dimer + MPV.

Tromso Study demonstrated that increased MPV was identified as a pre- dictor for VTE, in particular VTE of unprovoked origin [15]. Mean plate- let volume might be associated with VTE, but the results are not consistent [3,4,16]. Our study showed that PLT decreased in the PE group compared with the non-PE group (but P N .05), and MPV and PDW were both higher in the PE group (P b .05). Furthermore, multivariate logistic regression analysis showed that MPV and D-dimer were indepen- dent factors for the diagnosis of PE. Thus far, the results regarding the relationship between MPV and the diagnosis of PE are inconsistent, and it remains difficult to draw a conclusion.

Currently, the clinical application of MPV alone is relatively limited. The reasons include the following: (1) MPV is a nonspecific indicator that can vary in many diseases and states such as coronary artery disease, stroke, VTE, pathologic obstetrics, inflammatory diseases, se- vere renal and hepatic impairment, type 2 diabetes mellitus, malignant tumor, and anticoagulant and antiplatelet treatments; (2) different de- tection technologies result in laboratory deviations, including the type of hematology analyzer, the anticoagulant applied, and the time from sampling to analysis [16]. Despite these issues, Canan et al [17] reported that MPV was significantly higher in patients with DVT than in patients without DVT (7.97 +- 17.8 f. vs 7.58 +- 0.87 fL, P b .01) and that D-dimer was significantly higher in patients with DVT (P b .01). They found that the specificity was elevated despite a reduction in sensitivity when combining MPV with D-dimer. Therefore, they proposed that elevated MPV might increase the specificity of D-dimer for the exclusion of DVT. Because of the poor specificity and low PPV of D-dimer for VTE, other biological markers were also combined with D-dimer to try to im- prove its diagnostic value, such as CRP and MPO. Mitchell et al [18] found that the combination of MPO and D-dimer increased the specifi- city from 59% to 73% compared with D-dimer alone. Similarly, our ROC curve analysis showed that the combination of MPV and D-dimer resul- ted in an increase in AUC (from 0.787 to 0.799) and increased the speci- ficity and PPV (from 62.1% to 90.9% and from 66.7% to 84.6%, respectively) compared with D-dimer alone. This result might suggest that MPV increases the diagnostic value of D-dimer for PE, despite its limited value alone.


Our study had certain limitations. For example, the sample size was relatively small; it was a retrospective study, and therefore, we could not observe whether an association exists between elevated MPV and the morbidity of PE and many critical patients with acute PE died before admission and their data were absent, and thus, the collected data could

J. Huang et al. / American Journal of Emergency Medicine 33 (2015) 760763 763

Table 5

Performance characteristics of D-dimer, MPV, and D-dimer + MPV


Cut-off value

The maximum Youden’s index






835.5 (ug/L)


80.6 (%)

62.1 (%)

66.7 (%)

77.8 (%)


8.45 (fL)


88.7 (%)

50.0 (%)

61.9 (%)

78.7 (%)

D-Dimer + MPV


53.2 (%)

90.9 (%)

84.6 (%)

67.4 (%)

not reflect the full panorama of PE. We look forward to larger sampling and prospective studies.


Despite the limited evidence, our present study suggests that PDW and MPV levels are higher in patients with PE and that combination with MPV can increase the specificity and PPV to improve the diagnostic value of D-dimer for PE. Because platelet indexes are convenient to detect, clinical physicians may increase their vigilance to identify suspected PE. Further prospective studies are necessary.


  1. Tapson VF. Advances in the diagnosis and treatment of acute pulmonary embolism. F1000 Med Rep 2012;4:9.
  2. Ho KM, Chavan S. Prevalence of thrombocytosis in critically ill patients and its association with symptomatic acute pulmonary embolism. A multicenter registry study. Thromb Haemost 2013;109(2):272-9.
  3. Gunay E, Sarinc Ulasli S, Kacar E, Halici B, Unlu E, Tunay K, et al. Can platelet indices predict obstruction level of pulmonary vascular bed in patients with acute pulmo- nary embolism? Clin Respir J 2014;8(1):33-40.
  4. Hilal E, Neslihan Y, Gazi G, Sinan T, Zeynep Ayfer A. Does the mean platelet volume have any importance in patients with acute pulmonary embolism? Wien Klin Wochenschr 2013;125(13-14):381-5.
  5. Torbieki A, Perrier A, Konstantinides S, Agnelli G, Galie N, Pruszczyk P, et al. ESC Com- mittee for Practice Guide1ines (CPG). Guidelines on the diagnosis and management of acute pulmonary embolism: the Task Force for the Diagnosis and Management of Acute Pulmonary Embolism of the European Society of Cardiology (ESC). Eur Heart J 2008;29:2276-315.
  6. Stein PD, Hull RD, Patel KC, Olson RE, Ghali WA, Brant R, et al. D-dimer for the exclu- sion of acute venous thrombosis and pulmonary embolism: a systematic review. Ann Intern Med 2004;140(8):589-602.
  7. Scridon T, Scridon C, Skali H, Alvarez A, Goldhaber SZ, Solomon SD. Prognostic signif- icance of Troponin elevation and right ventricular enlargement in acute pulmonary embolism. Am J Cardiol 2005;96(2):303-5.
  8. Agewall S, Giannitsis E, Jernberg T, Katus H. Troponin elevation in coronary vs. non- coronary disease. Eur Heart J 2011;32(4):404-11.
  9. Cavallazzi R, Nair A, Vasu T, Marik PE. Natriuretic peptides in acute pulmonary em- bolism: a systematic review. Intensive Care Med 2008;34(12):2147-56.
  10. Kaczynska A, Pelsers MM, Bochowicz A, Kostrubiec M, Glatz JF, Pruszczyk P. Plasma heart- type fatty acid binding protein is superior to troponin and myoglobin for rapid risk stratification in acute pulmonary embolism. Clin Chim Acta 2006;371(1-2):117-23.
  11. Ozsu S, Abul Y, Gunaydin S, Orem A, Ozlu T. Prognostic value of red cell distribution width in patients with pulmonary embolism. Clin Appl Thromb Hemost 2014;20(4): 365-70.
  12. Nordenholz KE, Mitchell AM, Kline JA. Direct comparison of the diagnostic accuracy of fifty protein biological markers of pulmonary embolism for use in the emergency department. Acad Emerg Med 2008;15(9):795-9.
  13. Park Y, Schoene N, Harris W. Mean platelet volume as an indicator of platelet activa- tion: methodological issues. Platelets 2002;13(5-6):301-6.
  14. Varol E, Icli A, Uysal BA, Ozaydin M. Platelet indices in patients with acute pulmo- nary embolism. Scand J Clin Lab Invest 2011;71(2):163-7.
  15. Braekkan SK, Mathiesen EB, Njolstad I, Wilsgaard T, Stormer J, Hansen JB. Mean platelet volume is a risk factor for venous thromboembolism: the Tromso Study, Tromso, Norway. J Thromb Haemost 2010;8(1):157-62.
  16. Leader A, Pereg D, Lishner M. Are platelet volume indices of clinical use? A multi- disciplinary review. Ann Med 2012;44(8):805-16.
  17. Canan A, Halicioglu SS, Gurel S. Mean platelet volume and D-dimer in patients with

    suspected deep venous thrombosis. J Thromb Thrombolysis 2012;34(2):283-7.

    Mitchell AM, Nordenholz KE, Kline JA. Tandem measurement of D-dimer and myeloperoxidase or C-reactive protein to effectively screen for pulmonary embo- lism in the emergency department. Acad Emerg Med 2008;15(9):800-5.