A total of 1996 subjects were screened from February 1, 2017 and followed up till February 27, 2021 at three tertiary care hospitals in India. Of these 1150 subjects were excluded and 846 were randomly assigned to the vitamin D group (n = 424) and placebo group (n = 422). (Figure.1) Baseline and demographic characteristics did not demonstrate any significant difference between the two groups. Baseline BMI is shown in Table 1, with an overall mean of 19.3 ± 3.7 kg/m2. The baseline BMI amongst the two groups did not display any statistically significant difference (p-0.95). A majority of the participants in both groups were deficient in Vitamin D with average value of 14.2 ± 8.7ng/ml (Normal:25–80 ng/ml). Vitamin D levels amongst the two groups also did not reveal any statistically significant difference (p = 0.156). Similar trends were observed for calcium levels (p = 0.951), phosphate (p = 0.335), urea (p = 0.175), creatinine (p = 0.160), iPTH (p = 0.377) and hemoglobin (p = 0.957). Analysis of the chest radiographs revealed 46% of individuals having moderately affected lungs with no significant difference amongst the two groups at baseline (p = 0.151). Sputum smear profile at baseline showed 49% of subjects had 1 + bacilli (AFB), 21% had 2 + and 30% had 3+, with no significant difference seen among both groups (p = 0.46).
Table 1
Baseline and demographic characteristics
Demographic Profile ( n = 846) (MEAN ± SD) | |
Observations | Vitamin D arm (424) | Placebo arm (422) | p-value | Overall (n-846) |
Gender | Male – 276(65%) | Female – 148 (35%) | Male – 279 (66%) | Female – 143 (34%) | Male – 555 (65%) | Female – 291 (35%) |
Age (in years) | 29.2 ± 11.6 | 29.4 ± 11.4 | 0.821 | 29.3 ± 11.5 |
Weight (Kg) | 48.7 ± 9.5 | 49.1 ± 9.8 | 0.553 | 48.9 ± 9.6 |
Height (cm) | 158.3 ± 11.1 | 158.9 ± 10.8 | 0.457 | 158.6 ± 10.9 |
BMI (Kg/m2) | 19.3 ± 3.9 | 19.3 ± 3.6 | 0.955 | 19.3 ± 3.7 |
Calcium (mg/dl) | 8.8 ± 0.6 | 8.7 ± 0.6 | 0.951 | 8.8 ± 0.7 |
Phosphate(mg/dl) | 3.5 ± 1.0 | 3.4 ± 0.9 | 0.335 | 3.4 ± 1.0 |
25OH Vitamin D3 (ng/ml) | (3–63.5 ) 12.4 | (2.6–61.1) 11.6 | 0.156 | 14.2 ± 8.7 |
iPTH | (1–229.3 ) 17.8 | (4–115.8 ) 15.4 | 0.377 | 21.3 ± 16.3 |
Hemoglobin (gm/dl) | 11.9 ± 1.9 | 11.8 ± 1.8 | 0.957 | 11.9 ± 1.9 |
Urea (mg/dl) | 24.3 ± 6.7 | 24.9 ± 6.5 | 0.175 | 24.6 ± 6.6 |
Creatinine (mg/dl) | 0.7 ± 0.3 | 0.8 ± 0.3 | 0.160 | 0.7 ± 0.3 |
Chest X-ray severity | Normal − 8 Mild – 117 Moderate- 202 Severe − 97 | Normal − 6 Mild – 144 Moderate- 194 Severe − 78 | 0.151 | Normal − 14 Mild – 261 Moderate- 396 Severe − 175 |
Sputum Smear Profile | 1 + − 212 2 + − 92 3 + − 120 | 1 + − 200 2 + − 86 3 + − 136 | 0.461 | 1 + − 412 2 + − 178 3 + − 256 |
Primary Outcome:
For the final analysis, we took patients who had successfully completed their six months of treatment and 18 months of follow up (697 out of 846). 149 subjects were censored of which 87 were lost to follow-up, 57 individuals provided insufficient sample, 3 had adverse event and 2 patients expired.
There were 14 (4%) relapses in vitamin D arm and 19 (5.5%) relapses in placebo arm (n = 697 patients, n = 149 censored), which was however not statistically significant (p = 0.29). Bivariate Cox regression revealed a hazard ratio of 0.68 (95% CI: 0.34 to 1.37) having a log rank p-value of 0.29 (Table 2), thus implying a 32% reduction in relapses amongst individuals receiving vitamin D, which is not statistically significant. Upon testing for drug resistance, none of the relapsed participants were found to be harboring resistant organisms. All the relapsed patients were again initiated on ATT. Most of the individuals relapsed were male (n = 25) compared to female (n = 8).
Table 2
Primary outcome of the study
No. of subjects cured at 6 month | No. of Relapse in Placebo arm (n-343) | No. of Relapse in Vitamin D arm (n-354) | Hazard Ratio (Bivariate Cox Regression) | 95% CI | P- Value |
n-697 | 19 (5.5%) | 14 (4%) | 0.68 | 0.34–1.37 | 0.290 |
Secondary Outcomes:
Median time to sputum smear conversion in both groups was two weeks (95% CI 1.7 to 2.2 in vitamin D arm; 95% CI 1.7 to 2.6 in placebo arm) with hazard risk ratio of 1.06 (95% CI 0.92 to 1.23) and log rank p value 0.35 (Table 3). There was no significant difference observed between vitamin D and placebo groups with respect to median time to sputum smear conversion (n = 779 patients, n = 67 censored). Median time to culture conversion for both groups was four weeks (95% CI 3.6 to 4.3 vitamin D arm; 95% CI 3.7 to 4.3 placebo arm) with hazard risk ratio of 1.06 (95% CI 0.91 to 1.24) and log rank p value of 0.42. Again, there was no significant difference seen in the median time to culture conversion (n = 703 patients, n = 143 censored) between the two groups (Table 3).
Table 3
Secondary outcomes of the study
Group (s) | Subject(s) | Median Time to Conversion (In Weeks) | 95% CI | Hazard Ratio (Bivariate Cox Regression) | 95% CI | p-value |
Time to Sputum smear conversion (n-779) |
Placebo | 385 | 2 | 1.7–2.6 | 1.0 | 0.92–1.23 | 0.35 |
Vitamin D | 394 | 2 | 1.7–2.2 | 1.06 |
Time to Culture Conversion (n- 703) |
Placebo | 347 | 4 | 3.7–4.3 | 1.0 | 0.91–1.24 | 0.418 |
Vitamin D | 356 | 4 | 3.6–4.3 | 1.06 |
The intervention arm saw a significant increase in vitamin D levels compared to the placebo arm following supplementation, which was maintained throughout the follow-up at 6th, 12th, 18th and 24th month (p-value < 0.001 at each interval).
None of the subject in either group developed hypercalcemia. There was no significant difference observed in calcium level during follow up at 6th month (p = 0.62), 12th month (p = 0.52), 18th month (p = 0.17) and 24th month (p = 0.13).
Further sub group analysis was done to see the relationship between BMI (Asian Indian Criteria and relapse rate.(6) It was observed that no relapse occurred in the BMI range of ≥ 25 while 63% (n = 21) of relapse cases were seen in BMI range < 18.5, 33% (n = 11) in BMI range 18.5–22.9 and 3% (n = 1) relapse case was seen in BMI range 23–24.9. On Cox regression analysis, hazard ratio was 0.50 (95% CI: 0.2–1.3, p-0.19) which shows more than 50% of relapse cases were observed in < 18.5 BMI but results are not statistically significant (p = 0.19).
Safety profile (non-serious and serious adverse events) and mortality profile (all-cause deaths) analysis between the two groups (Table 4) showed that 10 (1.1%) patients died during the study (five in the vitamin D group and five in the placebo group). No death was directly attributable to the study intervention. Excluding death, seven serious adverse events were recorded (4 in placebo and 3 in vitamin D arms). Two patients in vitamin D arm had drug induced liver injury, with 1 patient developing liver abscess, while 4 participants in the placebo arm had serious side effects (one developing ethambutol ototoxicity, one developed allergic reaction to the placebo, one had drug induced liver injury and one developed DRESS syndrome). A total of 14 adverse events were recorded, eight in the vitamin D group and six in the placebo group, all having renal calculi, with none requiring a change in medical therapy (Table 4).
Table 4
| Vitamin D (n-424) | Placebo (n-422) | COX Regression | p-value |
Variable | No. of Events | Event / 100 person-month | No. of Events | Event / 100 person-month | Hazard Ratio (Vitamin D vs Placebo) 95% CI |
Death | 5 | 0.10 | 5 | 0.11 | 0.96 (0.2–3.3) | 0.95 |
Serious Adverse Event | 3 | 0.06 | 4 | 0.06 | 0.96 (0.1–4.7) | 0.96 |
Adverse Event (Renal Calculus) | 8 | 0.17 | 6 | 0.13 | 1.3 (0.4–3.7) | 0.61 |
Hypercalcemia | 0 | 0 | 0 | 0 | NC | - |
Hypervitaminosis D | 0 | 0 | 0 | 0 | NC | - |
*NC indicates that the risk ratio could not be calculated due to no events in both the groups |