"Platelet Variability Index As a Diagnostic Tool in Primary Immune Thrombocytopenia: Findings from a Cohort Study"

Introduction

Thrombocytopenia is a frequent cause of consultation for hematology specialists. Primary immune thrombocytopenia (ITP) has an incidence of 12/100 000 adults. It is yet today an exclusion diagnosis because any of the serologic biomarkers has a real utility. Since two decades ago it has been described that fluctuations in the platelet number are less important when the cause for thrombocytopenia is not immune. In 2021 Li et al published a platelet variability index (PVI) based on the concept of statistical volatility and demonstrated that there were differences between confirmed, suspected and discarded immune thrombocytopenia groups with statistical significance. They also obtained a sensibility of 93.8%, specificity of 89.4%, positive predictive value of 72.9% and negative predictive value of 97.7% using the cutoff point of 5 points. Therefore, our goal is to evaluate the usefulness of the PVI in patients diagnosed with ITP compared to thrombocytopenia due to other causes.

Methods

A retrospective single-center study included two groups of patients: one with the diagnosis of ITP according to the 2019 working group criteria, and the second group were patients with thrombocytopenia due to myelodysplastic syndrome (MDS) and liver disease considered as non-ITP; all with platelets ≤100x10⁹/L at the time of diagnosis. We collected the baseline characteristics of each group, the platelets at diagnosis, and the fluctuations during follow-up, as well as the platelet nadir defined as the lowest point during follow-up. With this, we evaluated the variability of platelets during follow-up and looked for differences between groups. This data was entered into a PVI calculator designed according to the one in the original study to calculate the score and subsequently perform validation.

Continuous variables were presented as medians and interquartile range (IQR), categorical variables in proportions. For comparison between groups, the X² test and Kruskal-Wallis test were used. For score validation, ROC curves were employed, and the best cutoff point for sensitivity and specificity was sought using the Youden index.

Results

We found 249 patients with thrombocytopenia, of which 131 were diagnosed with ITP (52.6%), 70 had MDS (28.11%), and 48 had liver disease (19.27%). The median platelet count at diagnosis for ITP was 32.0x10⁹/L (range;1.0-99x10x⁹/L) vs. non-ITP 46.5x10⁹/L (range;4.0-99.0x10⁹/L), p=0.047; the median platelet count at the nadir was 8.0x10⁹/L (range;1.0-98.0x10⁹/L) for ITP vs. 18x10⁹/L (range;1.0-97.0x10⁹/L) for non-ITP, p=0.058.

The PVI score was ≥5 points in 67.9% (89/131) of patients with ITP vs. 26.3% (31/118) in non-ITP, p<0.001. Furthermore, having a PVI score ≥5 points increased the likelihood of having ITP (OR 5.95, 95% CI 3.43-10.3, p<0.001).

The median PVI score was 5 points (IQR; 4-6) in ITP vs. 4 points (IQR; 3-5) in non-ITP, p<0.001. We found through ROC curves that the best score to predict ITP is 5 points with a Youden index of 0.41, Sensitivity 67.94%, Specificity 73.73%, Positive Predictive Value 74.17%, and Negative Predictive Value 67.44% with an AUC of 0.72.

There were no significant differences in the number of assessments between the ITP and non-ITP groups (median of 46, IQR: 2-100 vs. median of 52, IQR: 2-100, p = 0.265). Similarly, the follow-up duration was comparable between the groups (median of 5.7 years [IQR: 2.18 - 8.51] for ITP vs. 4.5 years [IQR: 1.5 - 8.5] for non-ITP, p = 0.307).

Conclusions

Our study found that a significant portion of patients with ITP had a PVI score of ≥5 points compared to non-ITP patients. A PVI score of ≥5 points significantly increases the probability of diagnosing ITP, with good sensitivity and specificity. The PVI can be particularly useful in cases where there is doubt in the diagnosis for differentiating between ITP and other underlying conditions. The PVI score offers an additional diagnostic tool in the management of patients with thrombocytopenia.