Summary. Background: Distinguishing inherited thrombocytopenias from immune thrombocytopenia (ITP) can be difficult, and patients are therefore at risk of misdiagnosis and inappropriate treatments. Although it is known that the most common inherited forms of thrombocytopenia are characterized by increased platelet size, the diagnostic power of this feature has never been investigated. Objectives: The aim of this study was to test the hypothesis that platelet size can be used to differentiate ITP from inherited macrothrombocytopenias. Patients/methods: We measured mean platelet volume (MPV) and mean platelet diameter (MPD), within 2 h of blood sampling, in 35 patients with inherited macrothrombocytopenias [15 MYH9-related disease (MYH9-RD), three biallelic and 17 monoallelic Bernard–Soulier syndrome (BSS)], and 56 with ITP. Using receiving operating characteristic analysis, we searched for the best cut-off values to differentiate between these conditions. Results: As expected, platelets were larger in inherited macrothrombocytopenias than in ITP. An MPD larger than 3.3 μm differentiated MYH9-RD and BSS from ITP with 0.89 sensitivity and 0.88 specificity, and an MPV larger than 12.4 fL had 0.83 sensitivity and 0.89 specificity. Combining MPD with MPV increased sensitivity and specificity to 0.97 and 0.89, respectively. Conclusion: Platelet size evaluation by both an appropriate cell counter and blood film examination is useful for differentiating inherited macrothrombocytopenias from ITP.
All guidelines for diagnosis and treatment of immune thrombocytopenia (ITP) agree that diagnosis of this disorder relies on exclusion of alternative causes of thrombocytopenia [1,2]. In most cases, history, physical examination and a complete blood count are sufficient to exclude secondary thrombocytopenias (e.g. multisystem autoimmune diseases, lymphoproliferative diseases, drug-induced thrombocytopenias, infections, and myelodysplastic syndromes); however, differential diagnosis is sometimes difficult. In particular, the differential diagnosis between ITP and inherited thrombocytopenias can challenge the clinician. The low platelet count of patients with constitutional forms of thrombocytopenia is recognized with increasing frequency in asymptomatic subjects during routine blood counts. Although the genetic origin of thrombocytopenia can be suspected in the presence of a family history that reveals one or more relatives with a low platelet count, a negative family history does not exclude the possibility of genetic thrombocytopenia, as some disorders are transmitted recessively. Moreover, genetic thrombocytopenias are often sporadic, as shown by the observation that eight of 19 consecutive patients with MYH9-related disease (MYH9-RD) had unaffected parents . Diagnostic difficulties are well illustrated by two retrospective studies that evaluated 170 cases with a diagnosis of ITP. These studies revealed that 11 patients were actually affected by inherited thrombocytopenias [4,5]. Misdiagnosis was not without consequences, as all 11 patients underwent needless splenectomy, administration of steroids and/or intravenous immunoglobulins. Moreover, another study showed that seven of 46 consecutive subjects with inherited thrombocytopenias had been previously splenectomized because they were misdiagnosed with ITP . Finally, a survey of the literature allowed us to identify a further 17 papers published in the last 10 years reporting 50 patients with inherited thrombocytopenias, in most cases Bernard–Soulier syndrome (BSS) and MYH9-RD, who were misdiagnosed with ITP and had been treated accordingly, 34 of them having been splenectomized (references available upon request).
In light of these studies, being misdiagnosed with ITP represents an actual risk for patients with inherited thrombocytopenias, and only the identification of simple, reliable tests for the differentiation of these conditions might reduce the number of such diagnostic errors.
As ITP is an autoimmune disorder, the search for platelet autoantibodies might represent the most attractive possibility. However, the predictive value of the tests available at this time remains unsatisfactory. The measurement of platelet-associated immunoglobulins is considered to be a sensitive tool, but its specificity is low , whereas the more recent assay for specific autoantibodies by means of antigen capture techniques has high specificity, but very poor sensitivity .
For a long time, there has been a general consensus that the evaluation of platelet size is a useful tool for differential diagnosis, as the less rare genetic forms (i.e. MYH9-RD, and monoallelic and biallelic BSS) have very large platelets and are therefore classified as ‘macrothrombocytopenias’, whereas platelet size is only modestly increased in ITP . However, this concept has not been supported by experimental data, as no comparative study has been performed. To this end, we applied simple techniques for measuring platelet size in a large case series of patients, and evaluated the ability of these parameters to discriminate between inherited macrothrombocytopenias and ITP.
Patients and methods
Thirty-five patients with inherited macrothombocytopenias (15 subjects with MYH9-RD, three with biallelic BSS, and 17 with monoallelic BSS) and 56 subjects with well-documented ITP referred to the IRCCS San Matteo Hospital Foundation of Pavia (Departments of Internal Medicine, Haematology and Paediatric Haematology) for diagnostic or follow-up purposes were entered into the study between January 2007 and October 2008. Forty healthy subjects were enrolled as controls over the same period. The mean age of the investigated subjects was 26.2 years for controls, 20.5 years for those with ITP, and 32.6 years for those with inherited thrombocytopenias. Two patients with ITP and four with inherited thrombocytopenias were asplenic because of a previous splenectomy.
The diagnosis of inherited thrombocytopenias was performed according to the diagnostic algorithm proposed by the Italian Gruppo di Studio delle Piastrine [6,10] and was always confirmed by molecular analysis. All 15 patients with MYH9-RD had monoallelic mutations of MYH9; two patients with biallelic BSS had mutations in the gene for glycoprotein (GP)Ibα and one in the gene for GPIX; all 17 subjects with monoallelic BSS had the C > T transition at position 515 of the gene for GPIbα (‘Bolzano’ mutation), resulting in the Ala156→Val substitution, which prevents GPIbα–von Willebrand factor interaction (‘variant’ BSS) .
The diagnosis of ITP was made according to guidelines of the American Society of Hematology and the British Society for Haematology [1,2], and was confirmed by the evaluation of subsequent clinical evolution and response to therapy.
This study was approved by the Ethics Committee of the IRCCS San Matteo Hospital Foundation, and all investigated subjects or their parents or legal guardian gave written informed consent.
Manual platelet counting
Whole blood platelet counts were performed on EDTA-anticoagulated blood. After blood dilution in ammonium oxalate solution, the counting procedure was performed by optical microscopy in a Neubauer chamber, as indicated by the International Committee for Standardization in Hematology .
Platelet count and determination of mean platelet volume (MPV) using cell counters
EDTA-anticoagulated blood samples were analyzed to determine platelet count and MPV within 2 h of sampling by two different automated blood cell analyzers: the ADVIA 120 (Bayer, Tarrytown, NY, USA) and the Sysmex XE-2100 (Sysmex Corporation, Kobe, Japan). The former instrument uses two-dimensional laser light scatter to identify and analyze platelets, and the latter is based on the impedance method. Instrument settings for this investigation were those routinely used for blood cell counts according to the manufacturer’s instructions.
Platelet diameters were measured by optical microscopy on May–Grünwald–Giemsa-stained peripheral blood films and software-assisted image analysis (Axio-vision 4.5; Carl Zeiss, Göttingen, Germany). Blood smears were prepared without anticoagulation. The largest diameter of each platelet was measured (Fig. S3). The mean platelet diameter (MPD) was the mean value obtained in 200 cell measurements.
Continuous data were presented as median and 25th to 75th percentiles, and categorical variables as counts and percentages. They were compared between groups by means of the Mann–Whitney U-test or the Kruskall–Wallis test and by the Fisher exact test, respectively. The graphical method of Bland and Altman and Lin’s concordance correlation were used to assess agreement between methods. Receiving operating characteristic (ROC) analysis was used to identify the optimal cut-off of diameters and volumes for discriminating inherited thrombocytopenias from ITP; sensitivity and specificity, together with their 95% confidence intervals, were reported.
Stata 10 (Stata Corporation, College Station, TX, USA) was used for computation. A two-sided P-value < 0.05 was considered to be statistically significant. No multiple test adjustment was performed for the explorative post hoc comparisons.
Figure S1 compares platelet counts obtained using the three counting techniques according to both the Bland and Altman graphical method and Lin’s plot. Although this shows a good average agreement, platelet counts obtained with the manual method were slightly higher than those obtained with the optical method, which, in turn, were higher than those obtained with the impedance method. The analysis of platelet counts obtained in different categories of subjects (Table 1) reveals the origin of these differences. The counting methods gave similar results in controls and ITP patients, but gave discordant results in MYH9-RD and BSS patients, with the highest values being obtained by the manual method and the lowest by impedance counting. As MYH9-RD and BSS are macrothrombocytopenic disorders, it seems reasonable to hypothesize that electronic counters failed to recognize the altered proportions of very large platelets, with the impedance instrument being more prone to this error than the optical one.
Table 1. Platelet counts measured by three methods in patients and controls
Platelet count, × 109 L−1 (median (25th to 75th percentiles)
Inherited thrombocytopenias with platelet macrocytosis
MYH9-RD (n = 15)
Biallelic BSS (n = 3)
Monoallelic BSS (n = 17)
Table 2 compares the MPVs obtained using the electronic instruments and MPDs measured by computer-assisted blood film examination in different categories of subjects. Whereas the optical counter measured MPV in all investigated subjects, the impedance instrument did not report this value in 50% of patients with ITP and in those subjects with inherited macrothrombocytopenias, owing to the abnormalities in their platelet volume distribution curves (Fig. S2). MPVs obtained with the optical counter and MPDs measured on blood smears were much higher in MYH9-RD and both monoallelic and biallelic BSS samples than in control samples, thus confirming the definition of these disorders as macrothrombocytopenias. Also in ITP, MPVs and MPDs were higher than in controls, but with a large degree of overlap. The platelet size data placed ITP between genetic macrothrombocytopenias and healthy subjects.
Table 2. Platelet size measured by electronic counters and microscopic examination of peripheral blood films
MPV (fL) (median (25th to 75th percentiles)*
MPV (fL) (median (25th to 75th percentiles)*
MPD (mm) (median (25th to 75th percentiles)*
BSS, Bernard–Soulier syndrome; ITP, immune thrombocytopenia; MPD, mean platelet diameter; MPV, mean platelet volume; MYH9-RD, MYH9-related disease. *Number of patients in which the counter measured MPV. †P < 0.001 vs. healthy subjects. ‡P < 0.05 vs. ITP. §P < 0.05 vs. healthy subjects. ¶P < 0.001 vs. ITP.
Healthy subjects (n = 40)
10.5 (10.2–11.2) 
8.2 (7.8–8.7) 
ITP (n = 56)
11.8 (11.2–12.4) 
9.7† (8.5–10.7) 
Inherited thrombocytopenias with platelet macrocytosis
MYH9-RD (n = 15)
– (–) 
17†,‡ (11–17.8) 
Biallelic BSS (n = 3)
– (–) 
16.7 (7–20.9) 
Monoallelic BSS (n = 17)
– (–) 
14.8†,‡ (14.3–15.5) 
As shown in Fig. 1, a good correlation was found between MPVs and MPDs, although a few notable exceptions were observed. Indeed, five patients with inherited thrombocytopenias (four with MYH9-RD and one with biallelic BSS) had MPDs larger than 5 μm, which is at least twice as high as the mean of controls, whereas their MPVs were reduced or normal. Figure S3 shows representative images of platelets from these cases as compared with those of ITP and healthy subjects. Interestingly, platelet counts obtained in these patients using the optical counter were 50–70% lower than those measured by manual counting (data not shown). Altogether, these findings suggest that in these subjects the optical instrument did not recognize a high percentage of platelets, probably because of their very large sizes, and therefore dramatically underestimated both platelet counts and volumes.
Platelet size for differentiating inherited macrothrombocytopenias from ITP
As platelet size was larger in inherited macrothrombocytopenias than in ITP, we examined the ‘optimal’ cut-off levels of MPV and MPD, using ROC analysis to discriminate between these two conditions (Fig. S4). With regard to MPV, the best cut-off value for distinguishing MYH9-RD/BSS from ITP was 12.4 fL (sensitivity of 0.83; specificity of 0.89), whereas the best cut-off value for MPD was 3.3 μm (sensitivity of 0.89; specificity of 0.88). The positive and negative predictive values for MPV were 0.83 and 0.90, whereas those for MPD were 0.81 and 0.93 (Table 3). Therefore, MPD was a better parameter for distinguishing between inherited thrombocytopenias and ITP. However, the best sensitivity for discriminating inherited macrothrombocytopenias from ITP was obtained by combining MPV and MPD data. Indeed, blood film evaluation easily identified the five patients with very large platelets (MPD larger than 5 μm) whose MPV was greatly underestimated by the electronic counter (see above and Fig. S3). In this way, sensitivity was increased to 0.97, while specificity remained unchanged (0.89), as no ITP subject had an MPD larger than 5 μm.
Table 3. Sensitivity, specificity and predictive values of the cut-off levels of mean platelet volume (MPV) and mean platelet diameter (MPD) identified by receiving operating characteristic analysis in discriminating inherited macrothrombocytopenias from immune thrombocytopenia (ITP)
Positive predictive value
Negative predictive value
Ninety-five per cent confidence intervals are in parentheses. The positive and negative predictive values were calculated on the basis of the 38% incidence of inherited macrothrombocytopenias at our institution.
ITP vs. inherited macrothrombocytopenias
MPV > 12.4 fL
MPD >3.3 μm
Although guidelines may suggest that platelet size is a useful parameter for discriminating between inherited macrothrombocytopenias and ITP, no study has yet formally supported this statement. The major difficulty in performing such a study was associated with the rarity of inherited thrombocytopenias and the low number of patients that are encountered in single centers. Moreover, multicenter studies have been hampered by differences in the cell counters used at different centers, making MPV values difficult to compare. As our institution is a reference center for both inherited and acquired thrombocytopenias, we were able, for the first time, to directly compare MPVs measured using the same instruments in a large case series of patients with inherited macrothrombocytopenias and in a corresponding number of subjects with ITP. We also measured platelet counts and MPDs by microscopic techniques, as these ‘old’ methods allowed identification of platelets with absolute certainty on the basis of their morphology.
Our study confirmed the limitation of electronic counters in recognizing very large platelets, as both impedance and optical instruments overestimated the degree of thrombocytopenia in MYH9-RD and BSS, when compared with manual counting. Discrepancies between the results from impedance and manual counting were the most relevant, and some patients with giant platelets were classified as having profound thrombocytopenia by the former method, whereas they had only a moderate decrease in platelet count by the latter.
Limitations of cell counters in identifying large platelets resulted in underestimation of MPV in subjects with macrothrombocytopenias. This was particularly evident in some patients with MYH9-RD or biallelic BSS, in whom the optical counter measured MPVs below the normal range although their MPDs were greatly increased. Moreover, the impedance counter did not provide MPV data for subjects with inherited macrothrombocytopenias or for half of the ITP patients, owing to the abnormalities of their MPV distribution curves. This prevented the use of the impedance counter for comparison of platelet volumes in ITP and inherited thrombocytopenias.
On the basis of the data presented in Table 2, we noticed that patients with MYH9-RD and BSS (both biallelic and monoallelic) have platelets that are usually larger than those of ITP patients. Thereafter, using ROC analysis, we identified the optimal cut-off values for MPV and MPD, and tested both their diagnostic sensitivity and specificity (Table 3 and Fig. S4). As expected on the basis of the limitations of the electronic counters discussed above, the best results were obtained for the cut-off values of MPD, which distinguished genetic macrothrombocytopenias from ITP with a sensitivity of 0.89 and a specificity of 0.88. Moreover, by combining MPV data obtained by the optical counter with MPD data measured by blood film examination, we obtained a sensitivity of 0.97 and specificity of 0.89. Thus, combining the ability of microscopy to identify giant platelets with the accuracy and precision of the optical instrument in analyzing platelets without marked volume abnormalities resulted in a powerful diagnostic test.
We are aware that our study did not include all inherited macrothrombocytopenias described so far  and that our results cannot, therefore, be fully extrapolated. However, the disorders that we investigated represent, at least in Italy, the most common forms and those more frequently misdiagnosed as ITP . We acknowledge that the cut-off values for MPV and MPD have been tailored to our case series of patients and that further studies in other groups of patients are required to confirm their diagnostic efficacy. Naturally, different cell counters have different normal values, and the reported cut-off values for MPV are strictly dependent on the optical instrument that generated them. Thus, other centers will need to optimize the cut-off value according to the results obtained in control subjects with the instrument that they routinely employ.
In conclusion, we believe that combining microscope evaluation of MPDs with the value of MPVs obtained using an optical counter represents a powerful tool for distinguishing the most frequent forms of inherited macrothrombocytopenias from ITP.
P. Norris enrolled patients, analyzed results and wrote the paper; C. Klersy performed statistical analysis and contributed to paper preparation; M. Zecca, A. Pecci, L. Arcaini, F. Passamonti and F. Locatelli enrolled patients and contributed to data analysis and paper preparation; F. Melazzini performed experiments and made the figures; V. Bozzi, C. Ambaglio and V. Terulla performed experiments; C.L. Balduini designed the study and wrote the paper.
This work was supported by grants from Telethon Foundation (Grant GGP06177) and the Italian Institute of Health (Italy–USA Project on Rare Diseases).
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.