Hereditary stomatocytosis in the general population: A genetically based prevalence estimate from a 109 039 individual Danish cohort

Hereditary spherocytosis (HS) is caused by mutations in genes such as ANK1, EPB42, SLC4A1, SPTA1, or SPTB,¹ leading to altered red blood cell (RBC) membrane proteins, reduced deformability, and decreased RBC lifespan. Dehydrated hereditary stomatocytosis (xerocytosis) is caused by variants in the PIEZO1 gene and, less commonly, KCNN4 variants, affecting RBC hydration and stability.²

The prevalence of HS is generally reported around 1:2000, while dehydrated hereditary stomatocytosis estimates range from 1:10 000 to 1:50 000, though mostly based on sporadic cases or older studies lacking comprehensive diagnostic methods.^{1, 2} Current prevalence estimates may be inaccurate due to underdiagnosis, referral bias, and the lack of systematic population-based testing. A study analyzing complete blood counts of 48 million North American patients suggested that up to 1:8000 individuals could potentially have dehydrated hereditary stomatocytosis based on specific biochemical markers, although definitive confirmation through genetic or specialized testing was not feasible.³ Interestingly, a study identified a gain-of-function PIEZO1 allele in one-third of African Americans, which, in a mouse model, induced a phenotype resembling dehydrated hereditary stomatocytosis and provided significant resistance to malaria.⁴

We tested the hypothesis that both hereditary spherocytosis and dehydrated hereditary stomatocytosis are more frequent in the white general population than previously estimated. For this purpose, we studied 109 039 white individuals of Danish descent from the Copenhagen General Population Study(H-KF 01-144/01) examined between 2003 and 2015.⁵ All individuals had hemoglobin, red cell distribution width (RDW), and MCHC measured, and all individuals had DNA obtained for further genetic analysis. Individuals with biochemical signs of hemolysis were genetically tested for the most frequent causes of hereditary hemolysis. All individuals aged 40–100 years and 25% of inhabitants aged 20–39 years living in a suburban part of the Capital Region of Denmark were invited, of these 43% participated. All participating individuals underwent a physical examination, had blood samples drawn, and completed a questionnaire on lifestyle and health on the day of enrollment. Blood samples for hemoglobin, mean corpuscular volume, and MCHC were drawn at enrollment and analyzed fresh on an ADVIA 120 Hematology System. RDW was calculated as standard deviation of mean corpuscular volume divided by mean corpuscular volume multiplied by 100. As previously proposed by Kaufman et al.³ potential hemolysis was considered in individuals with RBC indices suggesting hemolysis: increased MCHC and high RDW (as a sign of reticulocytosis), or increased MCHC and low hemoglobin. Increased MCHC was defined as MCHC >95th percentile, equal to MCHC >36.3 g/dL for women and MCHC >35.3 g/dL for men. As reticulocyte count was not measured directly, we used RDW as measure of reticulocytosis since high RDW is well correlated with a high reticulocyte count.⁶ High RDW was defined as RDW >95th percentile equal to RDW >14.5% for women and RDW >14.3% for men. Low hemoglobin was defined as hemoglobin ≤10th percentile equal to hemoglobin ≤12.4 g/dL for women and hemoglobin ≤13.5 g/dL for men. DNA from peripheral blood leucocytes was subjected to a broad targeted panel of genes potentially housing variants causing hereditary hemolysis in individuals of Scandinavian descent (Supplementary Methods and Table S1).

Table S2 displays the baseline characteristics of the 109 039 individuals in the study, categorized by RBC indices of hemolysis. Among these, 187 individuals had high MCHC and high RDW and 107 had high MCHC and low hemoglobin. Notably, 14 individuals had high MCHC and both low hemoglobin and high RDW (Figure S1). In total, 280 individuals had potential hemolysis, on whom next-generation sequencing was performed. After variant filtering, 48 individuals harbored a variant in genes that can cause hereditary spherocytosis. Of these, 7/48 individuals had definite hereditary spherocytosis (likely pathogenic and pathogenic variants) and 5/48 individuals had probable hereditary spherocytosis (hot variants of uncertain significance [VUS]). Forty-five individuals harbored a variant in genes that can cause dehydrated hereditary stomatocytosis. Of these, 2/45 individuals had definite dehydrated hereditary stomatocytosis (likely pathogenic and pathogenic variants) and 14/45 individuals had probable dehydrated hereditary stomatocytosis (hot VUS). Of the individuals with probable or definite dehydrated hereditary stomatocytosis, 11/16 harbored one of two combinations of variants in PIEZO1, either c.2423G > A/2344G > A or c.2815C > A/c.7374C > G. These variants have previously been reported in cis in the literature and likely constitutes relatively common haplotypes in the European population (Table S2). Among the 19 individuals with probable hereditary spherocytosis or probable dehydrated hereditary stomatocytosis, 2 individuals overlapped (Figure S1). None of the 280 individuals with RBC indices of hemolysis had genetic evidence of other causes of hereditary hemolysis than hereditary spherocytosis or dehydrated hereditary stomatocytosis.

Among the 12 individuals with definite or probable spherocytosis, 3 had been diagnosed with hereditary spherocytosis and 1 had been diagnosed with unspecified anemia. None of the 16 individuals with either definite or probable dehydrated hereditary stomatocytosis had been diagnosed with dehydrated hereditary stomatocytosis, but 2 had been diagnosed with unspecified anemia, 1 had been diagnosed with iron deficiency anemia, and notably 1 had been diagnosed with hereditary spherocytosis (Table 1). None of the 280 sequenced individuals were diagnosed with hereditary lymphedema, which can result from biallelic PIEZO1 loss-of-function variants.

When estimating a conservative prevalence of hereditary spherocytosis in the general population defined as only individuals with variants definitely causing hereditary spherocytosis, the prevalence would be 0.0064% (95% confidence interval [95% CI]: 0.0026%–0.013%) equal to 1:16000. For dehydrated hereditary stomatocytosis, the corresponding conservative prevalence estimate would be 0.0018% (95% CI: 0.00022%–0.0066%) equal to 1:55000 (Figure S2). If combining individuals with either definite hereditary spherocytosis or probable hereditary spherocytosis (the latter based on hot VUS), the prevalence of hereditary spherocytosis would be 0.011% (95% CI: 0.0057%–0.019%) equal to 1:9000. Likewise, when combining individuals with either definite dehydrated hereditary stomatocytosis or probable dehydrated stomatocytosis, the prevalence of dehydrated hereditary stomatocytosis would be 0.015% (95% CI: 0.0084–0.024) equal to 1:7000 (Figure S2). However, it is important to note that including hot VUS in prevalence calculations is speculative and should be interpreted with caution.

Our study is limited by not being able to ascertain with certainty whether some of the individuals sharing potential hemolysis-causing variants were related. We assessed relatedness between carriers of the shared gene variants using a custom implementation of the Somalier relatedness score (https://github.com/brentp/somalier), which we applied to estimate relatedness based on variants with allele frequencies 25%–75% in our overall population (103 variant positions). Due to the limited size of our NGS panel, we cannot calculate a standardized relatedness score, but our custom relatedness rating indicated that relatedness is not likely to be a major determining factor in the individuals we found had shared hemolysis variants. Importantly, all shared variants were classified as hot VUS and thus did not influence the conservative prevalence estimates.

To our knowledge, this is the first study to screen for hereditary spherocytosis and dehydrated hereditary stomatocytosis in the general population using RBC indices and genetic testing in combination.

As only 25% of individuals with genetic evidence of hereditary spherocytosis was correctly diagnosed in our study, previous studies using hospital diagnoses might be inaccurate in assessing prevalence of hereditary spherocytosis in the general population.

Our finding is consistent with Kaufman et al.,³ who analyzed hematological indices from 48 403 254 patients undergoing complete blood counts, predominantly outpatients, with a smaller proportion of hospitalized individuals. This study suggested a prevalence of dehydrated hereditary stomatocytosis of approximately 1:8000 based on specific biochemical indices. However, this study did not have access to DNA on the studied patients, which made genetic confirmation of these biochemical indices impossible.

Treatment for dehydrated hereditary stomatocytosis differs substantially from that of hereditary spherocytosis.² In dehydrated hereditary stomatocytosis, iron overload is common and may necessitate monitoring and treatment by either iron chelation or phlebotomy. Splenectomy is the cornerstone treatment for individuals with symptomatic hereditary spherocytosis but contraindicated for individuals with dehydrated hereditary stomatocytosis.² Importantly, in our study, one individual had genetic evidence of dehydrated hereditary stomatocytosis but was diagnosed with hereditary spherocytosis at a hospital, potentially leading to mistreatment.

Strengths of our study include the large study population of 109 039 individuals all of whom had RBC indices measured and DNA available for genetic analyses. Our study was conducted using data on individuals from the general population; therefore, findings may be more generalizable compared with studies using either individuals with a pre-existing condition or blood donors.

The prevalence estimates in this study are weakened by only having studied prevalence of hereditary hemolysis according to RBC indices, likely leading to underestimation, as some individuals with hereditary hemolysis might only be detected through other biochemical markers of hemolysis, like high bilirubin, lactate dehydrogenase, or ferritin. Functional validation, for example, by osmotic gradient ektacytometry, was not possible. Additionally, our next-generation sequencing panel is not well-suited for detecting copy number variations or assessing consanguinity, both of which may affect the accuracy of prevalence estimates.

In conclusion, this study estimates the conservative prevalence of hereditary spherocytosis to be at least 1:16 000, potentially rising to 1:9000 when including probable cases based on hot VUS. Similarly, for dehydrated hereditary stomatocytosis, the prevalence is at least 1:55 000 but may increase to 1:7000 when considering hot VUS. While prevalence estimates incorporating hot VUS carry uncertainty, this aligns with previous data by Kaufmann et al.,³ indicating that dehydrated hereditary stomatocytosis may be several-fold more prevalent than previously reported.

Research grants from the Capital Region of Denmark, the Independent Research Fund Denmark, and Saniona AB made this study possible. The Copenhagen General Population Study is supported by the Danish Heart Foundation and Copenhagen University Hospital—Herlev and Gentofte. The funders had no input in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. The researchers acted independently from the study sponsors in all aspects of this study.

All authors declare: Saniona AB provided an unrestricted grant to AG for this specific project. No other support was received from commercial organizations for the submitted work; outside of the submitted work, JH received research funding from Sanofi A/S and AG has done consultancy/advisory board for Agios, Bristol Myers Squibb, Novartis, Novo Nordisk, Pharmacosmos, and Vertex Pharmaceuticals and received research funding from Agois, Bristol Myers Squibb, Novo Nordisk, Saniona, and Sanofi. All other authors declare no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

All participants provided written, informed consent.