research paper

Platelet size deviation width, platelet large cell ratio, and mean
platelet volume have sufficient sensitivity and specificity in the
diagnosis of immune thrombocytopenia

Ken Kaito,1 Hiroko Otsubo,2 Noriko Summary

Usui,2 Miyuki Yoshida,1 Jyunko Tanno,1
Etsuko Kurihara,1 Kozue Matsumoto,1
Ryuzo Hirata,1 Kenichi Domitsu1 and
Masayuki Kobayashi2
1
Central Clinical Laboratory, Jikei University
Hospital, and 2Division of Hematology and
Oncology, Department of Internal Medicine, Jikei
University, School of Medicine, Tokyo, Japan
Received 19 November 2004; accepted for
publication 29 November 2004
Correspondence: Ken Kaito MD, Central
Clinical Laboratory, Jikei University Hospital, 3-
19-18 Nishi Shinbashi, Minato-ku, Tokyo 105-
8471, Japan. E-mail: ken.kaito@jikei.ac.jp
We investigated the significance of the platelet indices, mean platelet volume
(MPV), platelet size deviation width (PDW), and platelet-large cell ratio (P-
LCR), in the diagnosis of thrombocytopenia by comparing these levels in 40
patients with hypo-productive thrombocytopenia (aplastic anaemia; AA) and
39 patients with hyper-destructive thrombocytopenia (immune thrombo-
cytopenia; ITP). The sensitivity and specificity of platelet indices to make a
diagnosis of ITP were also compared. All platelet indices were significantly
higher in ITP than in AA, and platelet indices showed sufficient sensitivity
and specificity. The area under the curve (AUC) of the receiver operating
characteristics curve of platelet indices was large enough to enable the
diagnosis of ITP. P-LCR and PDW had the largest AUCs, which indicated
that these values were very reliable for immune thrombocytopenia. Our
results suggest that these indices provide clinical information about the
underlying conditions of thrombocytopenia. More attention should be paid
to these indices in the diagnosis of thrombocytopenia.
Keywords: platelet indices, mean platelet volume, platelet deviation width,
platelet-large cell ratio, thrombocytopenia.

In evaluating the mechanism of thrombocytopenia, it is
necessary to know which is more dominant, hypo-productive
thrombocytopenia or hyper-destructive thrombocytopenia.
For this purpose, bone marrow aspiration, platelet-associated
immunoglobulin G (PAIgG), and molecular markers for
disseminated intravascular coagulation (DIC) are often eval-
uated. Bone marrow aspiration provides information about
platelet production, such as the number of megakaryocytes
and the degree of platelet production, while PAIgG identifies
the presence of anti-platelet antibodies that lead to platelet
destruction. Bone marrow examination, which is an invasive
test, is necessary for aplastic anaemia (AA), but there is no
agreed consensus regarding its necessity for immune thromb-
ocytopenia (ITP) (George et al, 1996; Mak et al, 2000; Marsh
et al, 2003). PAIgG is often elevated in ITP, but it is not
specific to ITP and an increased PAIgG level is often found in
many other diseases (Mueller-Eckhardt et al, 1980; Von dem
Borne et al, 1986; Kelton et al, 1989). In fact, the necessity for
both bone marrow aspiration and PAIgG in ITP was not
accepted in the recent guidelines (British Committee for
Standards in Haematology General Haematology Task Force,
2003). However, these two diagnostic approaches are actually
overused in the diagnosis of ITP.
Recent advances in automated blood cell analysers have
made it possible to measure various blood cell parameters
automatically. Among these parameters, platelet indices, such
as mean platelet volume (MPV), platelet size deviation width
(PDW), and platelet large cell ratio (P-LCR), provide some
important information (Threatte, 1993; Niethammer & For-
man, 1999; Park et al, 2002), but are not accepted for routine
clinical use. If these indices really are informative regarding
platelet kinetics, they might become very useful laboratory
measures for thrombocytopenia. We investigated the signifi-
cance of these platelet indices in the diagnosis of thrombocy-
topenia by comparing the levels in hypo-productive (AA) and
hyper-destructive thrombocytopenia (ITP). The sensitivity and
specificity of platelet indices to enable the diagnosis of ITP
were also evaluated.

doi:10.1111/j.1365-2141.2004.05357.x ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 128, 698–702
 PDW, P-LCR, and MPV in Thrombocytopenia

Table I. Characteristics of the evaluated pa-

tients. No. of cases according to the platelet count
(·109/l)
Diagnosis Age, years
(n) Female/male (median) 10–20 20–40 40–60 60–80 80–100

AA (40) 27/13 26–84 (52) 2 7 10 13 8

ITP (39) 20/19 20–84 (55) 0 11 11 7 10

Patients and methods 100 %

Seventy-nine patients with thrombocytopenia (platelet coun- 12 fL
t < 100Æ0 · 109/l) were entered into this study. There were 47
women and 32 men. Their ages ranged from 20 to 84 years
(median age, 58 years). All patients gave written informed
consent for this study. The Ethics Committee for Biomedical
Research of the Jikei University School of Medicine approved
this study in July 2003. Patients were divided into two groups
according to the diagnosis of thrombocytopenia; AA and ITP. P-LCR
There were 40 patients with AA and 39 patients with ITP. The
20 %
degree of thrombocytopenia is shown in Table I. The diagnosis
and severity of AA were based on haematological, pathological,
radiological, and chromosomal analyses (International Agra-
nulocytosis and Aplastic Anemia Study, 1987), and the diagno- 2 fL 30 fL
sis of ITP was based on the guidelines previously reported PDW
(George et al, 1996). The AA group comprised two severe AA,
Fig 1. Histogram of platelet size distribution and the definition
11 moderate AA, and 27 mild AA patients. Four patients had a of platelet size deviation width (PDW), and platelet-large cell ratio
history of blood transfusion, but there were no patients who had (P-LCR). The distribution width at the level of 20% was defined as
a history of blood transfusion in the previous 5 years. The PDW, and the percentage of the platelets with a size of more than 12 fL
Sysmex-XE2100 automated blood cell analyzer (Sysmex, Kobe, was defined as P-LCR.
Japan) was used to measure platelet indices. MPV was calculated
by the following formula, MPV (fL) ¼ [(plateletcrit (%)/
platelet count (·109/l)] · 105. Plateletcrit was the ratio of the Table II. Comparison of platelet count and platelet indices between
platelet volume to the whole blood volume. PDW and P-LCR aplastic anaemia and immune thrombocytopenia.
were analysed from a histogram of platelet size distribution. The
AA P-value ITP
distribution width at the level of 20% (the peak of the histogram
is 100%) was defined as PDW, and the percentage of platelets Platelet (·109/l) 5Æ9 ± 0Æ4 ns 6Æ0 ± 0Æ4
with a size of more than 12 fL was defined as P-LCR (Fig 1). MPV (fL) 10Æ2 ± 0Æ2 <0Æ0001 12Æ2 ± 0Æ2
The sensitivity and specificity of platelet indices to make a PDW (fL) 11Æ6 ± 0Æ3 <0Æ0001 16Æ8 ± 0Æ5
diagnosis of ITP were calculated under various cut-off ranges, P-LCR (%) 25Æ7 ± 1Æ1 <0Æ0001 42Æ2 ± 1Æ5
and the receiver operating characteristic (ROC) curves were
AA, aplastic anaemia; ITP, immune thrombocytopenia; MPV, mean
drawn. Sensitivity was calculated as the ratio of the number of platelet volume; PDW, platelet size deviation width; P-LCR, platelet-
positive tests to the number of ITP (39), and specificity was large cell ratio.
calculated as the ratio of the number of negative tests to the
number of AA (40). The false-positive ratio (%) was calculated
as 100)specificity (%). similar in both groups. All platelet indices were significantly
Results are presented as the mean ± standard error (SE), higher in ITP than in AA (P < 0Æ0001). In particular, PDW
and Fisher’s protected least significant difference test and and P-LCR showed marked differences between the two types
Pearson’s correlation test were used for statistical analysis. of thrombocytopenia.
P < 0Æ05 was considered statistically significant. The correlation between the platelet count and evaluated
parameters is shown in Fig 2. In patients with AA, there were
significant inverse correlations between PDW and the platelet
Results
count and between P-LCR and the platelet count. However, no
The platelet count and platelet indices were compared between significant correlation between platelet indices and the platelet
AA and ITP and are shown in Table II. The platelet count was count was found in ITP.

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 128, 698–702 699
K. Kaito et al

Aplastic anaemia
(fL) (fL) (%)
MPV ns 24 PDW P < 0·03 P-LCR P < 0·003
14
18 50
12

10 12 30

8 6 10
0 2 4 6 8 Plt 0 2 4 6 8 Plt 0 2 4 6 8 Plt

Immune thrombocytopenia
(fL) (fL) (%)
MPV ns 24 PDW ns P-LCR ns Fig 2. Correlations between the platelet count
14
and evaluated parameters. An inverse correla-
12 18 50 tion was found between the platelet count and
both platelet size deviation width and platelet-
10 12 30 large cell ratio in aplastic anaemia. However, no
significant correlation between platelet indices
8 6 and platelet count was found in immune
10
0 2 4 6 8 Plt 0 2 4 6 8 Plt 0 2 4 6 8 Plt thrombocytopenia.

Table III. Sensitivity and specificity for the diagnosis of ITP under The sensitivity and specificity of platelet indices to make a
various cut-off ranges. diagnosis of ITP were calculated under various cut-off ranges.
Cut-off value Sensitivity (%) Specificity (%) The referential ranges at our institute were 8Æ4–12Æ0 fL for
MPV, 8Æ0–14Æ0 fL for PDW, and 10–30% for P-LCR. Under
MPV (fL) these cut-off ranges, platelet indices, especially PDW and P-
>11 87Æ2 80Æ0 LCR, showed favourable sensitivity and specificity (Table III).
>12 59Æ0 95Æ0 The ROC curves of MPV, PDW, and P-LCR are shown in
>13 11Æ1 100 Fig 3. A laboratory test with the ROC curve shifted to the
PDW (fL)
upper left indicates a better test. As apparent from these
>13 92Æ3 75Æ0
figures, ROC curves of MPV, PDW, and P-LCR showed upper
>14 76Æ9 90Æ0
>15 71Æ8 95Æ0
left-shift. The area under the curve (AUC) of the platelet
P-LCR (%) indices was very large, 9104Æ8 for MPV, 9305Æ0 for P-LCR, and
>25 100 45Æ9 9343Æ4 for PDW. Among these three indices, P-LCR and PDW
>30 91Æ4 73Æ0 were more reliable for immune thrombocytopenia.
>40 62Æ9 100

ITP, immune thrombocytopenia; MPV, mean platelet volume; PDW, Discussion

platelet size deviation width; P-LCR, platelet-large cell ratio.
It is very important to know whether thrombocytopenia is a
result of hypo-production of platelets or hyper-destruction of

100 100 100

Sensitivity(%)

Sensitivity(%)

Sensitivity(%)

75 75 75

50 50 50

25 MPV 25
PDW 25 P-LCR
AUC = 9104·8 AUC =9343·4 AUC = 9305·0
0 0 0
0 25 50 75 100 0 25 50 75 100 0 25 50 75 100
False positive (%) False positive (%) False positive (%)

Fig 3. Receiver operating characteristic (ROC) curves of mean platelet volume (MPV), platelet size deviation width (PDW), and platelet-large cell
ratio (P-LCR) to distinguish immune thrombocytopenia from thrombocytopenic patients. ROC curves of MPV, PDW, and P-LCR were shifted to the
upper left of the graph, which indicates that these parameters are sufficient to distinguish the two types of thrombocytopenia. The area under the
curve was larger in PDW and P-LCR than MPV.

700 ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 128, 698–702

platelets. ITP is diagnosed by the absence of other diseases that
cause thrombocytopenia, such as systemic lupus erythemato-
sus, malignancy, and DIC. For this purpose, PAIgG and bone
marrow aspiration are sometimes used. However, elevated
levels of PAIgG are often found in various diseases, such as
infection, autoimmune disease, and liver cirrhosis (Mueller-
Eckhardt et al, 1980; Von dem Borne et al, 1986; Kelton et al,
1989), and PAIgG was not recommended as a diagnostic
measure in recent guidelines (British Committee for Standards
in Haematology General Haematology Task Force, 2003). Bone
marrow sampling is invasive and not necessary as the first-line
diagnostic procedure; it was recommended that it should be
reserved for older patients, or patients with atypical features
(George et al, 1996; Mak et al, 2000; Marsh et al, 2003). Thus,
a new non-invasive diagnostic approach for thrombocytopenia
is needed. Recent advances in technology have made it possible
to record various platelet indices, such as MPV, PDW, and
P-LCR, with an automated haematology analyser. There have
been some reports about these platelet indices and platelet
disorders (Bessman et al, 1985; Karnad & Poskitt, 1985; Endler
et al, 2002; Henning et al, 2002). However, they are not
clinically accepted as conventional checkpoints for thrombo-
cytopenia. We evaluated the efficacy of these indices by
comparing their values between hypo-productive thrombocy-
topenia (AA) and hyper-destructive thrombocytopenia (ITP).
Mean platelet volume and PDW were elevated in ITP
(Gardner & Bessman, 1983); increasing the number of
megakaryocytes and supporting the diagnosis of ITP was
associated with larger MPV (Rajantie et al, 2004). However,
little is known about P-LCR and thrombocytopenia, and
whether platelet indices are satisfactory laboratory tests for
thrombocytopenia has not fully been discussed. In our
evaluation, not only MPV and PDW, but also P-LCR was
significantly higher in ITP than in AA. Therefore, these indices
were effective in distinguishing these two types of thrombo-
cytopenia. The sensitivity and specificity of platelet indices
were sufficient to enable a diagnosis of ITP. The ROC curve is
useful in comparing the superiority of different laboratory
tests, and a laboratory test with larger AUC is thought to be
more reliable, with less misdiagnosis than a test with smaller
AUC. Our results indicated that these platelet indices showed
favourable ROC curves and large AUC. Among the three
parameters, PDW and P-LCR were more reliable markers for
distinguishing hyper-destructive thrombocytopenia from
hypo-productive thrombocytopenia.
It is not always possible to record platelet indices. In severe
thrombocytopenia, and in the presence of red cell fragmen-
tation, a platelet histogram cannot be adequately drawn, and
the indices cannot be recorded. Thus, platelet indices have a
limited use as an indicator for thrombocytopenia. Of course,
the possibility of many other diseases that cause thrombocy-
topenia should be ruled out by routine work up. However, our
results suggested that these indices could help to distinguish
hyper-destructive thrombocytopenia and hypo-productive
thrombocytopenia very easily. Platelet indices, if reported,
ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 128, 698–702
PDW, P-LCR, and MPV in Thrombocytopenia
provide a lot of clinical information about the underlying
conditions of thrombocytopenia. Whether platelet indices are
useful in other conditions that cause thrombocytopenia, except
ITP and AA, remains unknown. More attention should be paid
to these indices for the diagnosis of thrombocytopenia.
References
Bessman, J.D., Gilmer, P.R. & Gardner, F.H. (1985) Use of mean
platelet volume improves detection of platelet disorders. Blood Cells,
11, 127–135.
British Committee for Standards in Haematology General Haematol-
ogy Task Force (2003) Guidelines for the investigation and man-
agement of idiopathic thrombocytopenic purpura in adults, children
and in pregnancy. British Journal of Haematology, 120, 574–596.
Endler, G., Klimesch, A., Sunder-Plassmann, H., Schillinger, M., Exner,
M., Mannhalter, C., Jordanova, N., Christ, G., Thalhammer, R.,
Huber, K. & Sunder-Plassmann, R. (2002) Mean platelet volume is
an independent risk factor for myocardial infarction but not for
coronary artery disease. British Journal of Haematology, 117, 399–404.
Gardner, F.H. & Bessman, J.D. (1983) Thrombocytopenia due to
defective platelet production. Haematology, 12, 23–38.
George, J.N., Woolf, S.H., Raskob, G.E., Wasser, J.S., Aledort, L.M.,
Ballen, P.J., Blanchette, V.S., Bussel, J.B., Cines, D.B., Kelton, J.G.,
Lichtin, A.E., McMillan, R., Okerbloom, J.A., Regan, D.H. &
Warrier, I. (1996) Idiopathic thrombocytopenia purpura, a practical
guideline developed by explicit methods for the American Society of
Hematology. Blood, 88, 3–40.
Henning, B.F., Zidek, W., Linder, B. & Tepel, M. (2002) Mean platelet
volume and coronary heart disease in hemodialysis patients. Kidney
and Blood Pressure Research, 25, 103–108.
International Agranulocytosis and Aplastic Anemia Study (1987)
Incidence of aplastic anemia: the relevance of diagnostic criteria.
Blood, 70, 1718–1721.
Karnad, A. & Poskitt, T.R. (1985) The automated complete blood cell
count. Use of the red blood cell volume distribution width and mean
platelet volume in evaluating anemia and thrombocytopenia.
Archives of Internal Medicine, 145, 1270–1272.
Kelton, J.G., Murphy, W.G., Lucarelli, A., Garvey-Williams, J., Santos,
A., Meyer, R. & Powers, P. (1989) A prospective comparison of four
techniques for measuring platelet-associated IgG. British Journal of
Haematology, 71, 97–105.
Mak, Y.K., Yu, P.H., Chan, C.H. & Chu, Y.C. (2000) The management
of isolated thrombocytopenia in Chinese adults: does bone marrow
examination have a role at presentation? Clinical and Laboratory
Haematology, 22, 355–358.
Marsh, J.C., Ball, S.E., Darbyshire, P., Gordon-Smith, E.C., Keidan,
A.J., Martin, A., McCann, S.R., Mercieca, J., Oscier, D., Roques,
A.W. & Yin, J.A., British Committee for Standards in Haematology
(2003) Guidelines for the diagnosis and management of acquired
aplastic anaemia. British Journal of Haematology, 123, 782–801.
Mueller-Eckhardt, C., Kayser, W., Mersch-Baumert, K., Mueller-
Eckhardt, G., Breidenbach, M., Kugel, H.G. & Graubner, M. (1980)
The clinical significance of platelet-associated IgG. A study on 298
patients with various disorders. British Journal of Haematology, 46,
123–131.
Niethammer, A.G. & Forman, E.N. (1999) Use of the platelet histo-
gram maximum in evaluating thrombocytopenia. American Journal
of Hematology, 60, 19–23.
701
K. Kaito et al

Park, Y., Schoene, N. & Harris, W. (2002) Mean platelet volume as an
indicator of platelet activation: methodological issues. Platelets, 13,
301–306.
Rajantie, J., Javela, K., Joutsi-Korhonen, L. & Kekomaki, R. (2004)
Chronic thrombocytopenia of childhood: use of non-invasive
methods in clinical evaluation. European Journal of Haematology, 72,
268–272.
Threatte, G.A. (1993) Usefulness of the mean platelet volume. Clinical
and Laboratory Medicine, 13, 937–950.
Von dem Borne, A.E., Vos, J.J., can der Lelie, J., Bossers, B. & van
Dalen, C.M. (1986) Clinical significance of positive platelet im-
munofluorescence test in thrombocytopenia. British Journal of
Haematology, 64, 767–776.

702 ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 128, 698–702