Indian infants and adolescents have a high prevalence of hypovitaminosis D and nutritional rickets [1-3]. Treatment includes therapeutic doses of vitamin D and calcium. In USA, Australia, and UK, the recommended dose and duration of vitamin D therapy is variable with either a high dose bolus therapy (Stoss therapy 200,000-600,000 IU of vitamin D as a single oral or parenteral dose or intermittent high doses) or continuous slow supplementation [4-6]. These recommendations are based on expert opinion and experience rather than sound evidence from randomized comparison of doses. Recent guidelines by Endocrine society recommend 50,000 IU weekly for 6 weeks [7]. Traditionally in India, a single dose of 600,000 IU of vitamin D is used for treating nutritional rickets in children [8]; no national recommendations exist. Recent reports have raised concerns of hypercalcemia following a therapeutic dose of 600,000 IU of vitamin D [9-11].

Scientific literature to provide evidence for the best therapy at minimum effective dose – that is feasible, economical and free of potential adverse effects – is scarce. The evidence for establishing the minimum effective therapeutic oral dose of vitamin D is limited to only two trials from Turkey [9,12]. These studies concluded equal efficacy of single oral dose of 300,000 IU of vitamin D in comparison to 600,000 IU for treating nutritional rickets in children. However, these studies had small sample size and did not measure vitamin D status. Vitamin D metabolism is influenced by several factors, including availability of sunlight and sun exposure, skin pigmentation, genetics, and socioeconomic status. Geographical location of Turkey (39.5ºN) carries a higher risk for hypovitaminosis D due to inadequate sunshine as compared to India (28ºN), sub-Saharan Africa, Latin America and Caribbean which have sufficient sunlight throughout the year [13]. Therefore, the results of these trials may not be applicable to tropical countries in Asia or Africa.

We hypothesized that an oral dose of 300,000 IU of vitamin D3 is not inferior to 600,000 IU for increasing serum 25(OH)D levels and achieving radiological recovery for treatment of vitamin D deficiency rickets of nutritional origin in children between 6 months and 5 years of age, in a tropical setting.

This non-inferiority randomized controlled trial was conducted at a tertiary care hospital attached to a medical school at Delhi, India from November 2010 to April 2012. A clearance from the institutional ethical committee and informed consent from parents were obtained.

Participant selection: Children between the ages of 6 months and 5 years presenting to pediatric outpatient or emergency with a combination of clinical evidence of rickets (wide wrists, bow legs, frontal bossing, rachitic rosary etc.) and radiological findings (fraying, splaying, and cupping at the epiphyseal ends of long bones in wrist/knee) consistent with the diagnosis of nutritional rickets [4,8,14] were eligible for inclusion. Critically ill children and those having coexisting fat malabsorption, liver or renal insufficiency and hypercalcemia were excluded. Children with history of having received vitamin D, calcium supplements, or other medications affecting vitamin D metabolism (e.g.; anticonvulsants, steroids, cancer chemotherapy) in previous 6 months were also excluded.

Data collection: Baseline assessment included a detailed socio-demographic and clinical history and physical examination at the time of enrolment. Anthropometry (weight, height/length, head circumference) was recorded as per standard techniques [15]. WHO Child Growth Standards were used as reference population [16]. The Z-scores for anthropometric parameters were calculated for each child using the "WHO Anthro software for PC" [17]. X–rays of the wrist and knee were obtained for all participants at enrolment, as per standard procedures [18]. At enrolment, a venous blood sample was obtained for the estimation of serum calcium, serum phosphorus, serum alkaline phosphatase, serum 25(OH)D, serum parathormone (PTH), serum albumin, serum glutamate amino transaminase (SGPT) and serum creatinine.

Follow-up: All children were asked to report for follow-up at 1 week (±3 d), 4 weeks (±1 week), and 12 weeks (±2 weeks) after enrolment. At each visit, an interval history was obtained for adverse effects such as headache, vomiting, abdominal pain, seizures or bulging fontanelle. A urine sample (5 mL) for estimation of urinary calcium to creatinine ratio was collected at 1 and 12 weeks after enrolment. At 4 weeks, a venous blood sample (2 mL) was taken for serum calcium estimation. Serum 25(OH)D and PTH levels were estimated at baseline and 12 weeks; simultaneous samples were also drawn for serum calcium, serum phosphorus and serum alkaline phosphatase. X-rays of the wrist and knee were repeated at 12 weeks. Follow-up was ensured by telephone or by personal visit.

Adverse effects – clinical (headache, vomiting, abdominal pain, seizures, symptoms of pseudotumor cerebri) and biochemical (hypercalcemia, urine calcium/creatinine ratio, hypervitaminosis D) – were also compared between the two groups. Hypervitaminosis D was defined as level greater than 150 ng/mL [20] and hypercalcemia was defined as serum calcium greater than 10.8 mg/dL [19]. A urine calcium-creatinine ratio of more than 2 was considered as sign of toxicity [21].

Sample size: Our past experience suggested that the serum level of 25(OH)D (primary outcome) follows log-normal distribution. Sample size for this study was thus calculated on the basis of geometric mean (GM) ratio and decided to set non-inferiority lower boundary for mean GM ratio to be 0.8 at 12 weeks. The geometric mean (GM) ratio indicates geometric mean of 25(OH)D in Group 1 divided by GM of 25(OH)D in Group 2. Using the coefficient of variation as 0.31 with one-sided 2.5% level of significance and 80% power, a sample size of 34 subjects in each group was considered adequate. Co-efficient of variation was calculated on results of a previous study by Soliman, et al. [22] assuming mean (SD) level of 28.2 (8.7) ng/mL of 25(OH)D after 3 months of receiving 10,000 IU/kg (maximum 150,000 IU) of vitamin D. Adding expected 10% lost to follow-up, 38 children per group were considered appropriate for this non-inferiority trial.

Statistical analysis: The baseline data were presented in mean and standard deviation for continuous variables and in proportions for categorical variables. Natural log transformation was applied for 25(OH)D, serum parathormone, and alkaline phosphatase because of skewed distribution of these parameters; results were presented in geometric mean and their 95% confidence intervals.

Two factor repeated measures analysis of variance (ANOVA) was applied to compare baseline and 12 weeks levels of serum 25(OH)D, PTH, calcium, phosphorus, and alkaline phosphatase. The interaction between the group and time was tested. Analysis of covariance (ANCOVA) was applied to compare the serum 25(OH)D levels at 12 weeks between the groups taking baseline value as covariate, and data expressed as geometric mean and 95% confidence intervals for log transformed variable; and mean ± SD for other continuous variables. Homogeneity of slope assumption of analysis of covariance was tested by including the interaction (group × baseline value). The value of serum 25(OH)D, serum parathormone, and alkaline phosphatase were expressed as relative change due to log transformation and calcium, phosphorus, and alkaline phosphatase were expressed as mean change. The least square means (estimated marginal means) resulting from the ANCOVA was used to calculate the one side 97.5% confidence interval for log transformed difference between the two treatment groups. The acceptable lower limit of one-sided 97.5% confidence interval of serum 25(OH)D level at dose 300,000 IU was set at 80% of that achieved in the group receiving 600,000 IU of vitamin D3. Intention to treat analysis was also used in full-analysis set considering baseline value as carried forward at 12 weeks.

Analysis was done with SPSS-20 (Chicago, IL).

Enrolment and baseline characteristics: A total of 312 children with rickets were diagnosed over 18 months, of which 106 met the inclusion criterion. Fig. 1 shows the enrolment and follow-up of children assessed in the study. Of the 76 children enrolled, 45% (n=34) were females. Baseline demographic, clinical, and biochemical features; and radiological score of children in the two Groups are shown in Table I.

Fig.1 Participant flow diagram.

TABLE I	Baseline Clinical and Biochemical Characteristics of Enrolled Children

The presenting complaints included not gaining height (n=43, 56.6 %), irritability (n=30, 39.5%), delayed gross motor development (n=29, 38.2%), bowing of legs (n=19, 25%) hypocalcemic convulsions (n=13, 17 %), bone pains (n=4, 5.5 %), delayed tooth eruption (n=1, 1.3%) and recurrent infections (n=61, 80.3%). Recurrent infections included respiratory tract infections (n=52) and diarrhea (n=9).

On examination, most frequent clinical findings were wrist widening (n=76, 100%), frontal bossing (n=72, 94.7%), protruded abdomen (n=64, 84%), rachitic rosary (n=62, 81.6%), Harrison sulcus (n=60, 79%), genu varum (n=46, 61.8%), genu valgum (n=1, 1.3%), wide open anterior fontanel (n=29, 38.7%), hypotonia of limbs (n=4, 5.5%), delayed tooth eruption (n=3, 3.9%) and enamel defects (n=2, 2.5%).

Biochemical abnormalities included raised alkaline phosphatase (n=76,100%), hypocalcemia (n=63, 82.9%) and hypophosphatemia (n=47, 61.8%). The median value (IQR) of serum alkaline phosphatase, serum phosphorus and serum calcium were 843.50 (582–333.5); 3.2 (2.5–4.3) mg/dL, and 7.8 (6.8–8.5) mg/dL, respectively. Serum parathormone levels were elevated (>65 pg/mL) in 43 (56.6%) children at enrolment. Serum 25(OH)D levels were categorized as normal, and deficient in 8 (10.5%) and 68 (89.5) children, respectively.

Radiological evidence of rickets included fraying and splaying in all children (n=76, 100%). Osteopenia, cupping and loss of demarcation between metaphysis and epiphysis were seen in 67% (n=51) of children. Mild, moderate and severe radiological changes were present in 29 (38.1%), 6 (7.9%), and 41 (54%) children, respectively. The distribution was comparable between the two groups (data not shown).

TABLE II  Comparison Between Baseline and 12 Weeks Value of Biochemical Variables 

Outcome measures: The changes in biochemical parameters in the two groups after 12 weeks of enrolment are compared in Table II. Children in both the groups doubled their baseline serum 25(OH)D level. Serum parathormone and alkaline phosphatase declined by approximately 60%. The adjusted ratio of geometric mean of serum 25(OH)D at 12 weeks between the Groups (taking baseline value as covariate) was 0.91 (95% CI: 0.65-1.29) (Table III).This is also diagrammatically depicted in Fig. 2. The results were almost similar when intention to treat analysis (ITT) approach was applied with baseline observations carried forward for 25(OH)D.

TABLE III Comparison of Biochemical Parameters Between the Groups, 12 Weeks After Intervention

Fig. 2 Ratio of geometric means from baseline to 12 weeks (main graph) and mean ratio between the doses (inset graph).

Almost half the children in both the groups showed normalization of alkaline phosphatase levels after receiving vitamin D3 treatment at 12 weeks [Group 1: 17/32 (53%); Group 2: 15/28 (54%)]. Secondary hyperparathyroidism (serum PTH >65 pg/mL) was present in 65.8% (25/38) children of Group 1 and 47.4% (18/38) children of Group 2 at enrolment. At 12 weeks, only 4 children in Group 1 and one child in Group 2 had raised serum PTH levels. Web Fig. 1 shows the distribution of children categorized according to serum 25(OH)D levels (both at baseline and 12 weeks after therapy). Web Fig. 2 shows the increase in serum 25(OH)D in each child in both the groups after 12 weeks of treatment.

Adverse effects: No child required a repeated dose of study medication or developed any signs of drug intolerance (nausea, vomiting, headache, persistent crying, etc.). No clinical adverse effects of Vitamin D3 therapy or infections were noticed in both the groups. During first week, 5 children had diarrhea (Group 1: 2, Group 2: 3). Four children developed respiratory tract infection in second week (Group 1: 2, Group 2: 2).

Laboratory parameters showed no signs of raised urinary calcium creatinine ratio at 1 and 12 weeks. The mean (SD) urinary calcium creatinine ratio in both the groups was comparable [Group 1: 1.2 (1.03), Group 2: 1.2 (0.86); P =0.92]. At 12 weeks also, all children had normal urinary calcium creatinine ratio, and mean (SD) levels were comparable in both the groups [Group 1: 1.0 (0.81), Group 2: 0.9 (0.7); P =0.8].

No child had evidence of hypervitaminosis D. Hypercalcemia was documented in 2 children (1 in each Group) at 4 weeks; and 3 children (1 child in Group 1 and 2 children in Group 2) at 12 weeks.

In this study on 76 children (6 mo–5 y) with nutritional rickets, a therapeutic oral dose of 300,000 IU of vitamin D3 was comparable to 600,000 IU for improving 25(OH)D status after 12 weeks of its administration. All children in both the groups demonstrated radiological healing at 12 weeks. Also, there was comparable improvement in radiological severity scores, and comparable decline of serum parathormone and alkaline phosphatase levels in both the treatment arms. We concluded that 300,000 IU of vitamin D3 is not inferior to 600,000 IU for treating nutritional rickets in under-five children. There is a potential risk of hypercalcemia with both regimes.

Our study had certain limitations. The diagnosis of rickets was on clinical and radiological parameters: biochemical indices were not included. Seven (9.2%) children were vitamin D sufficient (serum 25(OH)D >20 ng/mL) but had clinical and radiological evidence of rickets. A previous study by Voloc, et al. [7] has demonstrated poor correlation between clinical features and serum 25(OH)D levels. Radiological changes in rickets may also be due to hypocalcemia in face of sufficient serum 25(OH)D levels [26]. We did not differentiate between calcium deficient rickets and vitamin D-deficient disease. We also did not study the effect of vitamin D3 therapy on clinical improvement (nutritional status, reversal of deformities). A valid interpretation of our results on adverse effects is also not possible due to limitation of small sample size for adverse effects as an outcome measure. We did not estimate serum 25(OH)D levels at 1 week after therapy, which would have been a better indicator of immediate toxicity.

There was smaller increase in 25(OH)D levels in our study (10.8 ng/mL) as compared to that by Soliman, et al (21.95 ng/mL) [22]. Lower increase in mean levels of 25(OH)D may be related to complex mechanisms in pharmacokinetics and pharmacodynamics of vitamin D leading to difference in individual responses, altered vitamin D metabolism in Asian Indians [23], and vitamin D receptor (VDR) polymorphism or mutations [24]. It is also concluded that increase in serum 25OH vitamin D levels are dependent on baseline levels, dose of vitamin D, and weight of patients [25]. Due to a small post-supplementation increase in serum 25(OH)D levels, proportion of children having 25(OH)D level between 5–20 ng/mL remained almost same, before and after treatment.

Our results demonstrate that even a dose of 600,000 IU may not be enough to normalize serum 25(OH)D (beyond 20 ng/mL), in Indian children with rickets. We advised daily vitamin D supplementation for all children for 12 weeks, after completing the study duration of 3 months. We wonder, whether these children should have been started with routine daily vitamin D supplementation, immediately following the mega dose. This approach would have needed a strict monitoring for vitamin D intoxication.

We conclude that a therapeutic oral dose of 300,000 IU of vitamin D can be safely substituted for 600,000 IU for treating nutritional rickets in under-five children. None of the two regimes is effective in normalization of vitamin D status in majority of patients, 3 months after administering the therapeutic dose. Studies are needed to document the optimal strategy of vitamin D3 supplementation in children treated with mega-dose of vitamin D, for replenishing the body stores.

Contributors: The study was conceptualized by PG. Methodology was finalized with inputs from DS, SR, SVM, GM, and RKM. Data were collected by SR and HG. Laboratory support and interpretation was provided by SVM. GM was responsible for assessment and interpretation of radiological results. The manuscript was drafted by HG and SR with inputs from SVM, GM, and RKM. Statistical analysis was planned and conducted by PG and RKM. PG and DS critically reviewed the manuscript for intellectual content. All authors approved the final paper. PG shall stand as the Guarantor.

Funding: Indian Council of Medical Research, New Delhi; and University College of Medical College; Delhi.

Competing interests: None stated.