Discrepancy between impedance and immunofluorescence platelet counting has implications for clinical decision making in patients with idiopathic thrombocytopenia purpura

Authors


Dr Trevor P. Baglin, Department of Haematology, Addenbrooke's NHS Trust, Cambridge CB2 2QQ, UK. E-mail: trevor.baglin@addenbrookes.nhs.uk

Summary

This study investigated whether differences occur between the impedance and immunofluorescence methods for platelet quantification in idiopathic thrombocytopenia purpura (ITP). Immunofluorescence gave a platelet count >50% higher than the impedance test in 9/35 (26%) patients, of which 4/35 (11%) were >100% higher. The clinical severity of thrombocytopenia was changed as a result of the immunofluorescence test in 14/35 (40%) patients. Neither mean platelet volume nor platelet distribution width predicted impedance/immunofluorescence method discrepancy. It is suggested that immunofluorescence platelet counts should be performed on all ITP patients when the implementation of a therapeutic or diagnostic intervention is being considered.

In patients with idiopathic thrombocytopenia purpura (ITP) clinical decisions are made based on the haemorrhagic phenotype and the degree of thrombocytopenia in the full blood count (George et al, 1996; Ancliff & Machin, 1998; Norfolk et al, 1998; Rebulla, 2001; Schiffer et al, 2001; British Committee for Standards in Haematology, 2003). To minimise the risk of poor patient outcomes it is necessary that such clinical decisions are based on an accurate and reliable measurement of the platelet count.

The impedance method used by many automated cell counters is an established routine method for determining platelet count. Immunofluorescence assays provide a rapid, accurate and reproducible test for platelet counting and could be adopted by laboratories with appropriate flow cytometry experience (Dickerhoff & Von Ruecker, 1995; Harrison et al, 2000). Discrepancies in platelet counting have previously been reported between impedance and immunofluorescence assays (Harrison et al, 2000) and are most common in patients with thrombocytopenia (Springer et al, 1998). The clinical significance of such discrepancies specifically in patients with ITP is unknown.

We investigated the frequency, severity and potential clinical consequences of discrepancies between immunofluorescence and impedance platelet counting in patients with ITP. Furthermore, the predictive value of mean platelet volume (MPV) and platelet distribution width (PDW) were assessed to prospectively identify patients at risk of inaccurate impedance platelet counting.

Methods

Samples for full blood count analysis were collected and analysed as previously described (Bowles et al, 2005). Patients with thrombocytopenia (platelet count <150 × 109/l) confirmed by blood film examination had their platelet count assayed by immunofluorescence using Trucount absolute count tubes (Becton Dickinson, Oxford, UK) as per the manufacturers instructions. Samples were analysed in duplicate and the mean of the two results was used as the immunofluorescence platelet count. Non-thrombocytopenic samples were run in parallel for every test and no discrepancy between immunofluorescence and impedance platelet count was observed in the non-thrombocytopenic controls.

Statistical evaluation was carried out by analysis of method of comparison, as described by Bland and Altman (1986). Continuous variables are summarised as mean [standard deviation (range)]. Categorical variables are expressed as the number (percentage). Groups were compared using the paired-samples t-test for paired continuous variable. Significance was defined as P < 0·05.

Results

In patients with ITP, measurement of the platelet count by the impedance method frequently under-estimated the platelet count as defined by the immunofluorescence method. In 35 patients the mean [standard deviation (range)] platelet count measured by impedance was 44 × 109/l [27·6 (3–108 × 109/l)] and by immunofluorescence was 56 × 109/l [29·9 (4–122 × 109/l)] (P < 0·001). For 9/35 (26%) patients immunofluorescence gave a platelet count that was >50% higher than the impedance test, of which 4/35 (11%) were >100% higher. The mean difference and limits of agreement in platelet count between the impedance method and immunofluorescence method are shown in Fig 1. This shows a significant under-estimation of the platelet count by the impedance method relative to the immunofluorescence method and a potential difference in an individual platelet count of up to 53 × 109/l.

Figure 1.

 Impedance testing underestimates the platelet count in idiopathic thrombocytopenia purpura when compared with the immunofluorescence method. Comparison of impedance and immunofluorescence methods for platelet counting [including the mean ± 2 standard deviations (±2 SD)] in a group of patients with idiopathic thrombocytopenia purpura. The tendency for the impedance method to underestimate the platelet count is shown by the mean difference of −13 × 109/l and a potential difference in an individual platelet count of up to 53 × 109/l.

Next we determined how frequently the clinical severity of the thrombocytopenia defined by impedance testing would be reassessed by a revised platelet count using the immunofluorescence method. Severity of thrombocytopenia was defined as severe (<20 × 109/l), moderate (20–49 × 109/l), mild 50–69 × 109/l) and very mild (70–149 × 109/l). Zero of seven (0%) patients with an impedance-measured platelet count of ≥70 × 109/l had the severity group reclassified on the basis of subsequent immunofluorescence testing. Of the remaining 28 patients with impedance-measured platelet counts <70 × 109/l, 14/28 (50%) changed severity group on the basis of the immunofluorescence platelet count. Ten patients were down-graded by one severity group and four patients were down-graded by two severity groups (Table I).

Table I.   The clinical severity of thrombocytopenia based on impedance testing can be reclassified on the basis of immunofluorescence testing in a significant proportion of patients with idiopathic thrombocytopenia purpura.
Clinical severity group changeNumber of patients (%)
  1. Clinical severity of thrombocytopenia assigned on the basis of the impedance testing was reclassified following subsequent immunofluorescence testing. Fourteen patients had the clinical severity down-graded and no patients had the clinical severity of thrombocytopenia upgraded. Severity of thrombocytopenia was defined as; severe (<20 × 109/l), moderate (20–49 × 109/l), mild 50–69 × 109/l) and very mild (70–149 × 109/l).

Severe to moderate3/35 (8·6)
Severe to mild2/35 (5·7)
Moderate to mild4/35 (11·4)
Moderate to very mild2/35 (5·7)
Mild to very mild3/35 (8·6)
No change21/35 (60·0)

Under circumstances where large platelets are common, such as thrombocytopenia as a result of reduced platelet survival (Bowles et al, 2005), we hypothesised that MPV and or PDW would predict discrepancy between the impedance and immunofluorescence methods. However, neither MPV nor PDW prospectively identified patients in whom a discrepancy would occur (data not shown).

Discussion

We found, as others have reported (Harrison et al, 2000), discrepancies in platelet count between samples simultaneously assayed by impedance and immunofluorescence methods. We extend this observation specifically to patients with ITP. In their work Harrison et al (2000) reported that in thrombocytopenia impedance testing was as likely to overestimate platelet count as underestimate platelet count when compared with immunofluorescence testing. Interestingly, in the data they provide on 17 thrombocytopenic patients with impedance counting errors [five ITP, three Bernard-Soulier disease (BS), eight leukaemia/lymphoma and one lipaemia] for all eight patients with ITP/BS the impedance method underestimated the platelet count and for 7/8 (88%) leukaemia/lymphoma patients the impedance method overestimated the platelet count when compared with the immunofluorescence method (Harrison et al, 2000). The data of Harrison et al (2000) supports our finding that the consistent common discrepancy in patients with ITP is impedance count under-estimation.

As clinical opinion on drug treatment, platelet transfusion and the risk: benefit of invasive procedures is dictated by the level of the platelet count, we investigated whether such advice might potentially change in light of more accurate platelet counting. In 50% of cases with a platelet count ≤70 × 109/l the severity risk group was changed in light of the immunofluorescence platelet count, suggesting that more accurate platelet measurement may influence decision making in individuals with thrombocytopenia.

Our results demonstrated that immunofluorescence platelet counting has clinical utility in patients with ITP when the implementation of a therapeutic or diagnostic intervention is being considered, based on the impedance measured platelet count.

Ancillary