Characterization of the severe phenotype of pyruvate kinase deficiency.

Abstract

To the Editor: Pyruvate kinase (PK) deficiency is the most common cause of hereditary non-spherocytic hemolytic anemia and is characterized by considerable genotypic heterogeneity, with over 350 documented pathogenic mutations in the PKLR gene.1, 2 Clinical manifestations range from a mild, asymptomatic well-compensated anemia to a severe transfusion-dependent hemolytic anemia from birth.3, 4 Other complications of PK deficiency include iron overload, pulmonary hypertension, endocrinopathies, liver failure, biliary disease, and extramedullary hematopoiesis, among others.3, 4 Splenectomy, a common supportive treatment, may partially ameliorate the anemia and reduce transfusion requirements.5 Hemoglobin concentrations correlate poorly with symptoms in PK deficiency6; therefore, transfusion requirements are typically used to classify disease severity, with those who are regularly transfused (often labeled “transfusion-dependent”) despite splenectomy assumed to be the most severely-affected subgroup. Our understanding of the clinical characteristics of this most severe subgroup is quite limited. Therefore, in this study, we aimed to describe the differences in clinical characteristics and disease complications between regularly transfused splenectomized patients with PK deficiency, and those who were also splenectomized but did not require regular transfusions. As is observed in other hereditary hemolytic anemias, we hypothesized that patients with pyruvate kinase deficiency requiring regular transfusions would have higher rates of disease complications. Using data collected from the Pyruvate Kinase Deficiency Natural History Study (PKD NHS),3 a prospective, international 30-site observational study, we aimed to compare demographics, complications, and laboratory results between two groups of patients with PK deficiency, defined by their clinical severity. All participants in the PKD NHS had molecularly confirmed PK deficiency and only splenectomized patients were included in this analysis. Regular transfusions were defined as ≥6 discrete red cell transfusion episodes per year. Transfusion frequency was observed over a 3-year period (divided into three 1-year periods) post-splenectomy. Patients were categorized into two groups based on transfusion frequency and splenectomy status. The most severe PK deficiency phenotype group was defined as one that received regular transfusions post-splenectomy. The comparison PK deficiency group did not receive regular transfusions post-splenectomy. Phenotype stability over the 3-year period was also assessed. Continuous parameters were compared between the two groups using the Wilcoxon rank-sum test and binary parameters were compared using the Fisher's exact text. Different genotype classifications between the two groups were compared using the Cochran-Armitage test. Of the 255 patients enrolled in the PKD NHS, 154 splenectomized patients were included in this analysis (30 patients in the most severe PK deficiency phenotype group and 124 patients in the comparison PK deficiency group). Patients in the most severe PK deficiency phenotype group were followed in 21 of the 30 participating centers. Results of the analysis comparing the two groups are described in Table 1. Given our definition of disease-severity, the most severely affected patients were more likely to have iron overload (93% vs 51%, P < .001), have received chelation therapy (90% vs 42%, P < .001), and have had more lifetime transfusions (median: 77 vs 15, P < .001) than the comparison PK deficiency group. The most severe patients were more likely to be female (77% vs 51%, P = .013) and older at the time of splenectomy (median age: 5 vs 3.6, P = .011). Rates of other PK deficiency complications including pulmonary hypertension, extramedullary hematopoiesis, liver cirrhosis, endocrinopathy, and bone fracture appeared similar between the two groups. Laboratory values, including hemoglobin, total bilirubin, normalized PK enzyme activity, and median absolute reticulocyte count appeared similar between the two groups. In a second analysis comparing the two groups, but with exclusion of the Amish population (N = 52), findings were unchanged except that the age of splenectomy was no longer significantly different (median: 5.1 vs 4.9 years, P = .96) between the most severe vs comparator PK deficiency group. The underlying genetic mutation patterns (missense mutations vs non-missense, such as frameshift variants or deletions) also appeared similar between the groups. Phenotype over time was fairly stable for patients who met the criteria for the comparator group at the time of enrollment, but variable for patients who met the criteria for most severe at enrollment, year 1, or year 2 (Table S1). Many of the latter patients alternated severity level from year-to-year; just six (27%) of the 22 patients who were most severe at any time point met the definition of most severe at all three time points. In summary, splenectomized patients with PK deficiency who are not regularly transfused appear to have similar rates of PK deficiency-associated complications (except for iron overload), and similar relevant laboratory values and genotypes when compared to those who are regularly transfused. If severe hemolysis results in higher rates of PK deficiency complications and hemolysis severity drives transfusion frequency, the findings of this study are counterintuitive, in that iron overload (and associated chelation therapy, an expected complication of transfusions) was the only PK deficiency complication more common in the most severe phenotype group. The similarity observed between the most severe phenotype patients and comparison PK deficiency patients could result from a protective effect of transfusion, as has been described in thalassemia intermedia.7 It could also suggest transfusion-dependence does not truly reflect disease severity, but is reflective of varying provider practices and/or patient symptoms. The latter hypothesis is supported by the fact that many patients classified as severe at baseline no longer met criteria for this classification one and 2 years later, demonstrating that transfusion requirements in patients with PK deficiency fluctuate significantly over time. Transfusion-dependence, therefore, does not necessarily imply a worse clinical outcome in PK deficiency. Furthermore, there are no transfusion guidelines for PK deficiency to help determine a universal transfusion threshold. Early in life, transfusion requirements often decrease with age likely due to fewer infection-associated hemolytic episodes, timing of splenectomy, and age-differences in tolerance of anemia. The opposite is sometimes seen later in life, with transfusion requirements increasing during middle-age, possibly due to physiologic age-related decline in cardiopulmonary fitness. Physiologic determinants of severity of disease in PK deficiency likely significantly relate to intrinsic red cell factors, such as PK protein levels8 or non-PKLR red cell gene variants, of which transfusion requirements may be a poor surrogate. Our analysis is limited by the lack of a standard threshold in patients with PK deficiency for the number of transfusions per year that meet criteria for regularly transfused. Use of a stricter definition (≥7 transfusions per year) would have significantly diminished the small sample size making comparisons between the two groups difficult, and likely diminished the relevance of the analysis as it would be defining a rather niche group in an already rare disease. Furthermore, the indication and goals for initiating (and discontinuing) regular transfusions, for example to treat a complication or improve subjective symptoms, were not collected. In conclusion, the most severely affected patients with PK deficiency defined as a requirement for regular RBC transfusions following splenectomy have a similarly high rate of disease complications and laboratory abnormalities as splenectomized patients not requiring regular transfusions, with the exception of higher rates of iron overload. Transfusion requirements seem to fluctuate considerably over time and are not a valid marker of disease severity in PK deficiency. A prospective study of a universal transfusion protocol may be useful in understanding the role of transfusions in the management of patients with PK deficiency. H. Al-Samkari is the recipient of the National Hemophilia Foundation-Shire Clinical Fellowship Award, the Harvard KL2/Catalyst Medical Research Investigator Training Award, and the American Society of Hematology Scholar Award. H. Al-Samkari wrote the first draft of the manuscript and contributed to data analysis, creation of tables and figures, critical revision of the manuscript, and final approval; H. Al.-Sayegh contributed to data analysis, critical revision of the manuscript, and final approval; W.L. contributed to data analysis, critical revision of the manuscript, and final approval; R.G. contributed to study design, data collection, data analysis, creation of tables and figures, critical revision of the manuscript, and final approval; and the following authors contributed to data collection, critical revision of the manuscript, and final approval: E.v.B., D.M., W.B., S.E., B.G., H.Y., S.C., K.K., N.K., J.D., D.P., C.K., J.K., Y.P., A.T., M.W., Y.R., J.R., H.W., S.H., V.B., M.V., J.K., S.S., M.S., M.R., H.B., M.M., K.A. Al-Samkari: Consultancy (Agios, Dova, Moderna), Research funding (Agios, Dova, Amgen). Van Beers: Consultancy and Research Funding (Agios). Barcellini: Consultancy (Agios, Alexion, Apellis, Biocryst, Bioverativ, Incyte, Momenta, Novartis). Eber: Consultancy (Agios). Kollmar: None. Despotovic: None. Pospíšilová: None. Knoll: None. Thompson: Consultancy (Bluebirdbio, Celgene, Novartis), Research Funding (Bluebirdbio, Celgene, Novartis, Baxalta). Wlodarski: None. Glader: Scientific Advisor and Consultancy (Agios), Research Funding (Agios). Morton: None. Chonat: Advisory board (Alexion, Agios). Kuo: Honoraria (Alexion, Novartis); Consultancy (Agios, Alexion, Bluebirdbio, Celgene, Novartis, Pfizer); Data Safety Monitoring Board Chair (Bioverativ). Holzhauer: Consultancy (Pfizer, Agios, Novartis), Research Funding (Pfizer). Verhovsek: None. Kunz: Consultancy (Novartis, Bluebird Bio). Sheth: None. Rose: None. Bradeen: None. McNaull: None. Rothman: Consultancy and Research Funding (Agios). Kwiatkowski: Consultancy (Celgene, Bluebirdbio, Imara, Agios). Research Funding (Apopharma, Novartis, Terumo, Bluebirdbio). London: Consultancy (ArQule, Jubliant Draximage) and Research Funding (Agios, Bluebird Bio, Rocket Pharma). Grace: Consultancy (Agios, Dova), and Research Funding (Agios, Novartis, Pfizer). The remaining authors have no conflicts of interest. Table S1. Severity status of splenectomized patients with PK deficiency at three time-points (enrollment, first follow-up year, second follow-up year), for patients with known severity status at all three time-points (N = 116) Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.