The study included children with HS, younger than 15 years, from a single center. Case-records over a period of 27-years (1985-2011) were retrieved. Diagnosis was based on the presence of spherocytosis in the peripheral smear, an increased osmotic fragility test and a negative direct-antiglobulin test, in the background of a suggestive clinical profile. Eosin-5'-maleimide binding assay was not available during the study-period [2]. Disease-severity was classified based on the British Committee for Standards in Haematology (BCHS) guidelines, including hemoglobin, bilirubin and reticulocyte count [1,5]. Records of height (Herpenden stadiometer, accuracy 1mm) and weight (digital scale, accuracy 100 g) were retrieved. Body mass index (BMI) was calculated. Z scores for height, weight and BMI were derived using WHO AnthroPlus [8]. Stunting was defined as a height-for-age Z score <-2SD. A weight-for-age (<10-years) or BMI-for-age (>10-years) Z score <-2SD defined ‘underweight’. Stunting or underweight was considered ‘severe’ when the Z score was <-3SD [8,9]. In splenectomized patients, repeated-measure-ANOVA was utilized to compare the Z scores at 3-points: initial presentation, 1-year following surgery, and at the last clinic visit. Statistical analysis was performed using SPSS v20.0 (IBM). The Institutional Ethics Committee approved the study.

Records of 82 children were retrieved for analysis. The mean age (SD) was 6.7 (2.8) years (range 0.08,15 y). Forty-two (51%) had severe disease; 18 (22%) moderately severe, 16 (20%) moderate and 6 (7%) had a mild disease. Records of weight were available for 78 children; 25 (32%) were underweight and 10 (12.8%) were severely underweight. Among the 25 underweight children, 22 (88%) had moderate/moderately severe, 2 (8%) had moderate and 1 (4%) had mild disease. Records of height were available in 69 children; 18 (26%) were stunted; 8 (11.6%) were severely stunted. Among the 18 stunted children, 16 (89%) had moderately severe/severe HS, while 2 (11%) had moderate HS. Underweight children had a lower hemoglobin (P=0.037) and a more severe disease (P<0.001). Children who presented at a later age were more likely to be stunted (P=0.006). The proportion of stunted children progressively increased across the age groups: <4 years (3/17; 17.6%), 4-8 years (5/25; 20%), 8-12 years (6/23; 26%) and >12 years (4/4; 100%) (P=0.006). Symptom-interval was greater (P=0.003) for children with stunting. In addition, they had a severe disease (P=0.01) (Table I). The median number of transfusions were not different in children who had normal weight and height in comparison to those who had growth failure (Table I).

TABLE I  Variables Influencing Growth Failure In Children With Hereditary Spherocytosis

Twenty six (31.7%) children had a total splenectomy at a mean (SD) age of 7.9 (3.7) years. The median (range) follow-up was 4.5 (0-3, 19) years. The mean (SD) hemoglobin [6.8 (1.6) g/dL] improved to 12.1 (2.5) g/dL following splenectomy (P<0.001). The pre- and post-splenectomy anthropometric data were available for 24 and 23 children, respectively. At diagnosis, 9/24 (37.5%) children were stunted and 10/24 (41.7%) were underweight. The prevalence of stunting was nearly similar at 1-year following surgery (9/23, 39.1%) (P=1.0), and did not reduce to a significant extent at the last follow-up visit (5/23, 21.7%; P=0.34). Reduction in the number of underweight children (10/24, 41.7%) was not statistically significant a year following surgery (5/23, 21.7%; P=0.21); prevalence remained the same at the last follow-up visit (5/23, 21.7%) (P=0.21).

The comparison of Z-scores for height and BMI at the stated time-points is illustrated in Fig. 1. Repeated-measures-ANOVA failed to demonstrate a significant increase in Z scores for height (P=0.84) or BMI (P=0.14) following splenectomy. Even in children who underwent splenectomy at an earlier age (<6 years), the height-for-age failed to improve significantly [mean (SD) Z scores at baseline: -1.6 (1.3); 1-year following splenectomy: -1.4 (1.8); P=0.6].

Fig. 1 Comparison of Z scores before splenectomy, one year following surgery, and at the last follow-up visit, for (a) height-for-age, and (b) BMI-for-age, in children with hereditary spherocytosis.

The current study reports 32% and 26% children with HS to be underweight and stunted, respectively. Underweight children were more likely to have either severe/moderately severe disease with a lower hemoglobin. Stunting was more frequent in children with severe disease, although the hemoglobin was similar to those with normal height. Hemoglobin improved significantly following splenectomy, yet growth failed to improve significantly.

The main limitation of the present study is the retrospective nature of the data. The anthropometry, though recorded in the clinic by trained health care workers, would have operator- or instrument-related bias over the long study period. In addition, the evaluation and the follow-up is likely to be non-uniform. Moreover, there were no controls to compare the growth pattern.

Severe HS, without regular transfusions or splenectomy, results in growth retardation and delayed sexual maturation [10]. Long standing anemia is the plausible cause of failure to thrive. A lack of acuity of symptoms contributes to delay in seeking healthcare, and exacerbates the growth-failure [4]. Splenectomy corrects anemia and the compensatory exuberant erythropoiesis that causes growth-failure. Bader-Meunier, et al. [7] reported an increase in height equivalent to 2SD in the growth curve following surgery in five pre-pubertal children with HS. Sub-optimal growth is often considered a relative indication for splenectomy in children with HS [5,11].

It was intriguing to observe a lack of improvement in growth following splenectomy, particularly as the hemoglobin levels had improved significantly. There could be several contributory factors. The size of the cohort could be a limiting factor. Mechanisms beyond anemia and chronic tissue hypoxia could be contributory. Though formal iron studies were not conducted, median number of transfusions did not differ in children with and without growth-failure, therefore, transfusional iron overload is unlikely. However, non-transfusional overload related to heterozygous state for the hemochromatosis gene has been previously reported in children with HS [12]. We could not test for this in this retrospective cohort. Interestingly, it was noted that the prevalence of growth-failure was similar to that previously reported from the community (underweight: 24-29%; stunting: 38%) [13,14]. Indeed, factors responsible for undernutrition in the community, including poverty, malnutrition and recurrent infections, would have likely contributed to the growth-failure in this cohort of patients with HS as well. We hypothesize that a continuing environment of pre-existing sub-optimal nutrition could have blunted a likely growth-spurt following splenectomy.

In conclusion, growth-failure was frequent in children with severe HS and in older children with delayed presentation. Notwithstanding the correction of anemia, splenectomy failed to alleviate growth-impairment at a median follow-up of 4.5-years. Given the risks of sepsis and thrombosis following splenectomy, studies performed on larger cohorts need to evaluate factors contributing to impaired growth, and to evaluate if growth-failure by itself should be a robust indication for splenectomy in similar settings [15]. 　　

Contributors: DB: planned the study. AD: collected the data and prepared the manuscript; AT: contributed to patient enrollment and RD: reported the hematology. All authors contributed to manuscript writing and its final approval.

Funding: None; Competing interest: None stated.