Predicting the probability of Gaucher disease in subjects with splenomegaly and thrombocytopenia

Hematologists are frequently involved in the diagnostic pathway of Gaucher disease type 1 (GD1) patients since they present several hematological signs. However, GD1 is mainly underdiagnosed because of a lack of awareness. In this multicenter study, we combine the use of a diagnostic algorithm with a simple test (β-glucosidase activity on Dried Blood Spot) in order to facilitate the diagnosis in a population presenting to the hematologist with splenomegaly and/or thrombocytopenia associated with other hematological signs. In this high-risk population, the prevalence of GD1 is 3.3%. We propose an equation that predicts the probability of having GD1 according to three parameters that are routinely evaluated: platelet count, ferritin, and transferrin saturation.


Scientific Reports
| (2021) 11:2594 | https://doi.org/10.1038/s41598-021-82296-z www.nature.com/scientificreports/ of GD1 in a high-risk population presenting to the hematologist with splenomegaly and/or thrombocytopenia associated with other hematological signs or symptoms suggestive of GD1. Preliminary results of this study on the first 196 patients have been previously published, showing a GD1 prevalence of 3.6% in a high-risk population 7 .

Materials and methods
Study design. We designed a multicenter study among 35 hematology centers in Italy. According to the feasibility questionnaire, we expected to enroll 500 subjects. The enrolment started in September 2010 and closed in December 2018. Inclusion and exclusion criteria were based on the published algorithm for the non-Ashkenazi population 5 : • Inclusion criteria: splenomegaly and/or thrombocytopenia and at least one sign or symptom among bone pain history, anemia, MGUS, polyclonal gammopathy in subjects under 30 years of age, splenectomy; • Exclusion criteria: onco-hematological diseases, portal hypertension due to liver diseases, hemoglobinopathies, or chronic hemolytic anemias.
Demographic, clinical, and laboratory data were collected at enrolment, gathered in a specific case report form, and collected by the coordinating center at Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy.
Beta-glucosidase activity. The beta-glucosidase activity on Dried Blood Spot (DBS) was centralized at Ospedale Gaslini, Genoa, Italy 8,9 . Normal values range from 4.4 to 17.7 pmol/punch/h. Subjects showing activity below 4.4 pmol/punch/h were recalled to be assessed with the gold standard assay on nucleated cell homogenates (leucocytes, EBV-lymphoblasts, or fibroblasts). If the enzymatic defect was confirmed, the diagnosis was completed with the molecular GBA analysis.
Statistical analysis. The prevalence of GD1 and its 95% confidence interval (CI) based on the exact method were calculated. Demographic, clinical, and laboratory variables of patients affected by GD1 were compared to those of unaffected patients using Wilcoxon rank-sum (Mann-Whitney U-test) test (for continuous variables) or Fisher's exact test (for categorical variables). Using univariate and multiple logistic regression models, we analyzed the predictive role of platelets (thousands/mm 3  Ethical aspects. The study was approved by the ethical review committee of the coordinating center "Comitato Etico Milano Area 2" (Protocol number 714/10) and by all participating Centers, and was carried out in compliance with the principles established in the Helsinki Declaration. Informed consent was obtained from all individual participants included in the study.

Results
The prevalence of GD1 in high risk predominantly Caucasian population is above 3%. Five hundred subjects have been enrolled in the study. Forty-five have been excluded because they did not fulfill the inclusion and exclusion criteria ( Fig. 1). Demographic, clinical, and laboratory characteristics are presented in Table 1. Ninety-one percent (91%) of the subjects were Caucasian. The mean age at enrolment was 46.9 years, and 31.9% (145/455) were females. The majority had splenomegaly (89.7%), and approximately half (47.9%) thrombocytopenia associated with other signs/symptoms. Anemia was the most common adjunctive sign (23.1%) ( Table 1).
DBS showed normal values in 379 subjects, while 76 (16.7%) had a reduced β-glucosidase activity. These 76 patients and a patient with a family history of GD1 presenting with a β-glucosidase activity slightly above the lower normal range were recalled to test the conventional enzymatic activity. Among the 65 patients tested with β-glucosidase activity on nucleated cell homogenates (12 did not answer), 15 were diagnosed with Gaucher disease type 1, with a prevalence of 3.3% (15/455, 95% CI 1.9-5.4%). In 14, the molecular analysis of the GBA gene identified the mutations. In one patient, no mutations of GBA gene nor PSAP, encoding for saposin C, were identified. Among GD1 patients, 7/15 (46.7%) were female, 14 of Caucasian origin, and the mean age at diagnosis was 43.5 years. They showed a lower platelet count compared to non-GD1 patients (84.000/mm 3 vs 131.000/ mm 3 , p = 0.0006), a higher serum ferritin level (551 ng/dL vs 139 ng/dL, p = 0.0002) which was associated to a lower transferrin saturation (20.8% vs 25.7%, p = 0.03).
Platelet count, ferritin, and transferrin saturation predict the probability of having GD1 in this high-risk population. Considering 159 subjects (13 with GD1) with complete data on ferritin, platelets, and TSAT, the best discrimination between GD1 and non-GD1 subjects was provided by platelets (AUC = 0.79), while ferritin and transferrin saturation showed lower AUCs (Fig. 2).
The addition of a history of bone pain to this model did not lead to an AUC increase, and thus, this variable was omitted.

Discussion
This study shows that in a predominantly Caucasian high-risk population presenting to the hematologist with splenomegaly and/or thrombocytopenia associated with other hematological signs, including anemia and MGUS, 3.3% of patients have Gaucher disease. These data confirm our previously published preliminary data 7 . Similar studies have been reproduced in different regions worldwide, with different prevalence, ranging from no cases detected in a Canadian study on 221 subjects 10 to 7.0% in an adult cohort in China 11 . A similar approach is under evaluation in the Italian pediatric population 12 , and promising preliminary results have been presented in a Chinese study 13 . Altogether, these data support the use of the previously published algorithm by Mistry et al. 5 associated with a simple first-level diagnostic test to screen high-risk populations. Diagnostic confirmation with β-glucosidase activity on nucleated cell homogenates is necessary to confirm the diagnosis. Of note, given that false negatives in the DBS test may arise due to methodological differences in blood spot drying, transport and storage [14][15][16] testing with the gold standard diagnostic exam is warranted when there is any clinical suspicion of Gaucher disease, even in the presence of normal DBS values.
Since lysosomal storage disorders, including GD1, are underdiagnosed, other approaches have been proposed to increase the diagnostic rate. Namely, the new-born screening has been experimented in several regions with different results according to the ethnicity of the tested population 6,[17][18][19] . However, the new-born screening raises unique issues that are primarily related to the inevitable detection of a disease with late-onset phenotypes.
Among the enrolled subjects, the only clearly different parameters between GD1 and non-GD1 patients were platelet count and serum ferritin. Hyperferritinemia with normal transferrin saturation is a common finding in naïve patients with GD1 20 , with prevalence ranging between 63 and 81% 21,22 . Recently, together with other Italian www.nature.com/scientificreports/ referral groups for iron disorders, we have proposed a new diagnostic flow-chart 23 , which enhances hyperferritinemia role when associated with splenomegaly and thrombocytopenia.
Here we propose an equation that predicts the probability of having GD1 according to platelet count and ferritin and TSAT levels and thus may support hematologists when evaluating a subject with splenomegaly and/ or thrombocytopenia.

Conclusion
High-risk population testing is effective in identifying Gaucher disease patients who present to the hematologist with splenomegaly and/or thrombocytopenia. The evaluation of the probability of having GD1 according to an equation and the use of DBS as a first-level test are potentially useful tools that can facilitate the diagnostic process.