Importance of Immediate Thiamine Therapy in Children with Suspected Thiamine-Responsive Megaloblastic Anemia—Report on Two Patients Carrying a Novel SLC19A2 Gene Mutation

Abstract

Thiamine-responsive megaloblastic anemia (TRMA) is an autosomal recessive disorder characterized by the development of megaloblastic anemia, diabetes mellitus, and sensorineural deafness. We report on the first two Croatian patients with TRMA, compound heterozygotes for nonsense, c.373C > T; p.(Gln125Ter) and novel missense variant, c.1214C > G; p.(Thr405Arg) in SLC19A2 gene. The first was diagnosed at 4 months with diabetes mellitus and severe anemia requiring transfusions. As TRMA was suspected, thiamine therapy was immediately started to prevent further transfusions and insulin therapy. His brother developed extreme anemia at 3 weeks of age while waiting for the results of the genetic test. Severe anemia in this sibling may have been prevented if thiamine had been initiated earlier.

Introduction

Thiamine-responsive megaloblastic anemia (TRMA, OMIM # 249270) syndrome is a rare autosomal recessive condition characterized by a clinical triad of megaloblastic anemia, diabetes mellitus, and sensory deafness. ¹ It is also known as Rogers' syndrome, ¹ which Rogers described for the first time in 1969. TRMA syndrome is caused by the intracellular deficiency of thiamine, the essential vitamin resulting from the defective thiamine transporter protein 1, THTR1. ² Unlike other tissues that have two high-affinity thiamine transporters, pancreatic β cells, bone marrow, and a subset of cochlear cells have a single transporter, THTR1, explaining the clinical findings in patients with TRMA. ^{3 4} However, at high concentrations, thiamine can cross the cell membrane via passive diffusion, ³ allowing treatment of such condition with pharmacological doses of thiamine, 50 to 100 mg/d of oral thiamine regardless of age. The exact dose of thiamine needed to achieve clinical benefit is unknown and there is no consensus on the starting dose of thiamine or the titration/adjustment of thiamine therapy. However, no additional clinical benefits were found with thiamine doses > 150 mg/d and no reported side effects up to 300 mg/d. ⁵

TRMA is a rare condition identified in individual cases and small series of patients, and thus, the frequency and prevalence of TRMA could not be determined. In addition, there are scarce data on clinical characteristics and long-term follow-up in these patients. Other early childhood clinical findings, such as thrombocytopenia, short stature, optic atrophy, retinal changes, severe neurological disorders including stroke and focal or generalized epilepsy, as well as cardiovascular anomalies with sudden death, high-output heart failure, cardiac arrhythmia, and congenital heart defects such as atrial septal defect, are recorded. ^{2 5} Clinical and genetic studies have not shown a genotype–phenotype correlation, as the number of different mutations in the SLC19A2 gene is almost equal to the number of pedigrees reported, and different clinical presentations are often found within the same family.

We report on two Croatian patients diagnosed with TRMA, compound heterozygotes for nonsense, c.373C > T; p.(Gln125Ter) and novel missense type, c.1214C > G; p.(Thr405Arg) in SLC19A2 gene. The clinical course in both of our patients suggested the need for early introduction of thiamine therapy, even prior to the genetic confirmation of the disease.

Written consent for the publication of patient information was received by the parents of the patients.

Case Reports

Patient 1 is the second child of unrelated, healthy parents. He was born at term from an uneventful pregnancy, a birth weight of 3,680 g. He showed pallor and low weight gain at 4 months of age (weight at 2 months of age: 5,800 g, +0.74 standard deviation [SD]; weight at diagnosis: 5,925 g, −1.11 SD). Laboratory examination indicated severe megaloblastic anemia: hemoglobin (Hgb) 5.8 g/dL (3.60 mmol/L; normal range of ages 6.21–8.50), hematocrit 17% (0.17; normal range for ages 0.28–0.39), and mean corpuscular volume (MCV) 92.4 fL (normal range for ages 73.8–89.4), requiring erythrocyte concentrate transfusion. Peripheral blood smear showed pronounced anisopoikilocytosis with dacrocytes, schizocytes, and hypochromic erythrocytes. He had a normal leukocyte count and morphology and a low platelet count. Additional work-ups led to the diagnosis of diabetes mellitus (blood glucose 24.4 mmol/L), HbA1c 54 mmol/mol (7.1%). The insulin level was 39.0 pmol/L (normal range: 15.6–149.4), and C-peptide was 0.41 nmol/L (normal range: 0.13–0.72). There have been no clinical symptoms or no laboratory signs of diabetic ketoacidosis. The antibodies to islet cell antibody, glutamic acid decarboxylase, and islet antigen-2 were negative. Insulin therapy was started immediately (only one dose of 0.5 IU aspart insulin subcutaneous (SC) followed by detemir 2 IU SC twice daily (0.66 IU/kg/d) and blood glucose levels were normalized without hypoglycemia or significant hyperglycemia. The serum was lipemic at admission with elevated triglyceride levels (9.69 mmol/L, normal: up to 1.70 mmol/L), but normalized after insulin was applied and glucose levels stabilized (after 7 days, the triglyceride level was 1.77 mmol/L, after 6 months 0.96 mmol/L). Laboratory work-ups excluded iron, B12 and folate deficiency, as well as autoimmune hemolytic anemia. Bone marrow aspiration showed moderate myelodysplastic characteristics (limited megaloblastoid forms and no sideroblasts) with regular morphology of other bone marrow cells, possibly due to micronutrient deficiency. The combination of neonatal diabetes mellitus and nonmicrocytic anemia has led to a presumptive diagnosis of TRMA. With parental consent, 20 days after presentation, we began empiric therapy with oral thiamine (100 mg once daily PO) and observed a rapid improvement in glycemic control enabling insulin to be stopped after 2 days (the first day of treatment with thiamine received 1 IU of insulin detemir SC twice, the second day just 1 IU of insulin detemir in the morning, and the third day insulin was discontinued). His Hgb level stabilized, preventing further transfusions of erythrocytes. Analysis of all coding regions and the exon/intron boundary of the SLC19A2 gene by Sanger sequencing was performed at the Exeter Molecular Genetics Laboratory, United Kingdom. Patient-1 was found to be a compound heterozygote for a nonsense, c.373C > T; p.(Gln125Ter) and a novel missense variant, c.1214C > G; p.(Thr405Arg). Both variants are expected to be pathogenic to support a genetic diagnosis of TRMA syndrome. Patient's mother and brother are carriers of a novel missense, and father is the carrier of a nonsense mutation of the SLC19A2 gene. As heterozygous single mutation carriers, they are asymptomatic. The diagnostic evaluation aimed to search for possible additional clinical findings including audiological assessment with auditory brain stem response audiometry, electroencephalography and neurologic examination, electrocardiogram, and cardiac ultrasound gave normal findings. However, fundoscopy revealed retinal pigmentary changes thus necessitating further evaluation. Visual evoked potential gave normal result indicating proper binocular vision, while retinography (flash electroretinography) showed dysfunction of retinal cone cells. Cone-rod dystrophy has been previously reported in patients suffering from TRMA syndrome. ^{6 7}

At the 28 months follow-up, aged 32 months, Patient-1 remained insulin free with a blood glucose level of 3.8 mmol/L, C-peptide 0.31 nmol/L, insulin 19.8 pmol/L, HbA1c 34 mmol/mol (5.3%), stable Hgb levels: Hgb 10.3 g/dL (6.39 mmol/L, normal range for age > 12 months 6.76–8.65), Hct 29% (0.29, normal range for age > 12 months 0.32–0.40), MCV 92.1 fL (normal range for age > 12 mmol 73.8–89.4), no need for iron supplementation. He has normal physical development with catch-up weight (at 28 months: weight 15.3 kg, +1.26 SD; height 93.1 cm, +0.64 SD) as well as normal neuromotor development. Apart from retinal cone cells dysfunction, there were no additional findings on regular screening for other clinical features of TRMA.

At the second day of life, blood count revealed Patient-2 is the newborn brother of Patient-1, delivered at term, birth weight 3,690 g, Hgb 15.5 g/dL (9.62 mmol/L; normal range for ages 8.44–12.35), Hct 47% (0.47, normal range for ages 0.39–0.59), MCV 117.7 fL (normal range for ages 93.1–115.4 fL), and erythrocyte morphology: macrocytosis and anisocytosis. Two weeks later, genetic analysis showed the same genotype as in Patient-1, which offered a genetic diagnosis of TRMA. Patient's family has been asked to come for thiamine therapy. After admission, at 3 weeks of age, laboratory tests indicated serious anemia with moderate leukopenia: Hgb 8.5 g/dL (5.28 mmol/L; normal range for ages 6.76–10.49), Hct 22% (0.22, normal range for ages 0.32–0.50), MCV 97.4 fL (normal range for ages 84.5–102.5), leukocytes 6.6 × 109/L (normal range for ages 6.9–19.6), thrombocytes 199 × 109/L (normal range 150–450), but normal blood glucose levels. Promptly, oral thiamine (50 mg once daily PO) was started. The response to thiamine therapy was not satisfactory regardless of the increase in dose (up to 150 mg once daily PO), until iron supplementation was added 2 weeks later (iron sulfate 13 mg twice daily PO). It seemed that iron supplementation was necessary for the complete recovery of erythropoiesis, and after 4 months, we were able to reduce the thiamine dose to 100 mg/d. At 16 months of age, Hgb levels were stable: Hgb 10.4 g/dL (6.45 mmol/L), Hct 29% (0.29), MCV 85.1 fL, with normal leukocyte count (8.9 × 109/L, normal range 6.0–16.0), normal blood glucose levels (3.8 mmol/L), C-peptide 0.17 nmol/L, insulin 10.2 pmol/L, HbA1c 29 mmol/L (4.8%), normal hearing, as well as normal physical and neuromotor development. No abnormal results on routine screening for other clinical features of TRMA were found.

Discussion

TRMA can manifest with any of three cardinal findings (megaloblastic anemia, diabetes mellitus, and/or sensorineural deafness) anytime between infancy and adolescence, usually during the first 4 years of life. ⁵ Variable age at onset was observed between subjects with the same mutation, including siblings, indicating intrafamily variability. ^{2 5}

In patients with TRMA syndrome, diabetes mellitus is nonautoimmune with median age at the onset of diabetes being 10 months. Insulin treatment is usually needed. Nonetheless, early treatment with thiamine has been shown to potentially improve glycemic regulation leading to a decrease in insulin requirements. Many patients may also be able to discontinue insulin therapy. ^{5 8 9} If started before the development of diabetes, thiamine therapy can delay the onset of diabetes.

Once treated with thiamine, patients with megaloblastic anemia become transfusion independent. ^{1 5 8} There are indications that thiamine therapy can become ineffective in some patients after puberty. ⁸ In a cohort of 32 patients, Habeb et al did not find thiamine therapeutic failure in pubertal and postpubertal patients regarding the control of erythropoiesis. ⁵ Nonetheless, one of the individuals in their cohort who were insulin independent required insulin again at the age of 11 years. ⁵ Further follow-up investigations are needed to assess changes in the efficacy of thiamine due to age and duration of disease in patients. Unfortunately, in case of thiamine insufficiency, there is no adequate alternative, so patients become insulin and/or transfusion dependent.

Hearing loss is a progressive feature, likely based on the particular mutation and time of initiation of thiamine therapy. The degree of internal hair cell loss has been shown to depend on the length and severity of thiamine deprivation. ⁴ Patients who began thiamine therapy early, at 1.5 months of age and had normal hearing after 30 months of follow-up, ¹⁰ as well as patients with preserved hearing at ages 15 and 30 years ³ have been reported, suggesting that early treatment (before 2 months of age) could be effective in preventing deafness. On the contrary, Potter et al reported on three affected siblings, of whom two were started on thiamine therapy at ages 3.6 and 0.6 months, and both developed sensorineural hearing loss at ages 2 and 3.9 years, respectively. ⁹ It was, thus, postulated that onset of hearing loss may be delayed but not prevented by thiamine treatment.

It seems that there is no effect of thiamine therapy on additional clinical features of TRMA syndrome, except hematopoietic abnormalities. ⁵ Data on long-term follow-up are limited, thus further studies and reports are needed to elucidate evolution of associated clinical features, response to therapy and long-term prognosis.

As already shown, when thiamine treatment is initiated early, the development of diabetes, anemia, and probably even hearing loss may be postponed. Both of our patients had severe anemia at diagnosis, effectively treated with thiamine. In Patient-1, thiamine was started before genetic confirmation of TRMA to cease the need for further transfusions and to treat hyperglycemia enabling insulin discontinuation. A positive response to treatment enabled specific genetic testing and made the diagnostic process simpler. Patient-2 developed severe anemia at the age of 3 weeks, proving that waiting for the genetic confirmation, even at neonatal age might cause hazardous delay of the medical intervention. After early thiamine introduction, Patient-2 never had elevated glucose level. At the age of 32 and 16 months, both of our patients have stable Hgb and glucose levels and normal hearing. In addition, Patient-1 has cone-rod dystrophy that appears to be unaffected by thiamine therapy, while Patient-2 had mild leukopenia that has been normalized with thiamine therapy. Regular six-monthly evaluations for other related conditions gave normal results. To our best knowledge, there are no clinical records of nonsense mutations present in our patients (c.373C > T; p. Gln125Ter), whereas other reported nonsense mutations have provided different phenotypes. In addition to anemia, our patients have a rather mild clinical presentation, with no significant additional findings and hearing preserved to date. Although there is no phenotype–genotype correlation found in patients with TRMA syndrome, such a phenotype as well as a good response to therapy might suggest milder nature of novel missense mutation found in our patients (c.1214C > G; p.Thr405Arg). Follow-up is needed to evaluate the effect of mutations and therapy on hearing in our patients, as on the development of possible additional features.

Conclusion

As an autosomal recessive condition, TRMA is most generally diagnosed in consanguineous families and isolated populations. However, TRMA should be considered in any patient with suggestive clinical features ^{11 12} as more than a dozen patients from European countries have been published. Considering that no side effects of oral thiamine have been documented and that no toxicity of oral thiamine has been demonstrated even if severely overdosed (Opinion of the Scientific Committee on Food for the Tolerable Upper Intake Level of Vitamin B1, European Commission, Health & Consumer Protection Directorate-General, 2001), ¹² treatment with thiamine should be started as soon as the diagnosis of TRMA is suspected. Early therapy can reverse the need for repeated erythrocyte transfusions in patients with anemia and decrease or even cease the need for insulin in diabetic children while awaiting genetic confirmation. We also recommend that thiamine should be started soon after birth in the siblings of TRMA patients to avoid potentially harmful disorders, as the first symptoms of the disease may be visible in neonatal age. Thiamine therapy should be stopped after receipt of negative results of the genetic analysis or continued in patients with a confirmed disease.