In this double-blind randomized community trial, supplementation with zinc and vitamin A, either alone or combined, failed to show superiority over placebo. This was unexpected, as we could not confirm the results of the previous study in Jakarta, in which vitamin A and zinc supplementation was beneficial in terms of sputum conversion time. Results similar to ours with regard to zinc supplementation were obtained in a study in Tanzania, where such supplementation did not lead to a reduction of sputum conversion time compared to supplementation with a multi-micronutrient or placebo either. Also, zinc supplementation did not improve immune response among TB patients infected with HIV in a study in Singapore. The authors further concluded that in the absence of zinc deficiency, additional zinc supplementation was not beneficial. And to the best of our knowledge, the data on the effect of vitamin A supplementation in TB were inconclusive[17–19].
How can we explain the discrepancy between the results of the Jakarta study and the current results? The first explanation might be found in retinol concentrations and inflammatory response at baseline. In the Jakarta study, higher mean retinol concentrations (0.8 μmol/L and 0.9 μmol/L for the supplemented and placebo groups, respectively) were found than in the current study (0.7 μmol/L for all groups). The mean baseline CRP concentrations in the Jakarta study (53 mg/L for the micronutrient group and 44.1 for the placebo group) were almost twice as high as in the current study (mean CRP for all groups = 28.4 mg/L). Higher CRP concentrations reflect a stronger inflammatory response, which is known to lower the plasma retinol concentrations. Thus, the patients in the Jakarta study may have had a lesser degree of vitamin A deficiency, inflammation leading to the low plasma retinol levels. This might imply that higher dosages of vitamin A supplementation would have been needed for the TB patients of the current study to reach retinol concentrations necessary for a clinical effect. This was also proven by the concentration of retinol at 2 and 6 months of intervention, which could not elevate the retinol level in the plasma. This is in line with the study in Malawi which concluded that supplementations at the level of recommended daily allowance (RDA) did not meet their main outcomes. In contrast with this, the study in Tanzania which used multimicronutrient supplementations four to ten times higher than the RDA, reduced TB recurrences among the HIV-positive adults with TB, and increased T-cell counts of the HIV-negative patients. With regard to zinc on the other hand, the level at baseline were already slightly above the deficiency cut off (10.7 μmol/L). This may explain the lack of efficacy observed for zinc.
The second explanation may be connected to the differences in the chest radiographs and the grading of sputum AFB positivity, an important predictor of sputum conversion. The number of patients with cavities in our study was similar (38%) to that in the Jakarta study (37.5% of total patients in both group). More important however, is that in the present study, the proportion of patients with cavities among the four groups differed significantly at baseline (Table 1), with the highest proportion of cavities (51.5%) being present in the patients of the vitamin A supplementation group. This imbalance may have obscured the effect of vitamin A supplementation but would not explain the lack of effect in the zinc + vitamin A group. Findings of a Spanish study showed the association between the presence of cavity and the conversion time in the presence of HIV co-infection. It was concluded that cavity prolonged the conversion time, and the presence of HIV lowered the prevalence of cavity production in the lungs as captured in the chest x-ray. Thus, the absence (probably very small proportion) of HIV co-infection in the current study led to cavity formation in the lung and at the end prolonged the conversion time .
The severity of TB in our patients is witnessed by the fact that 40% had sputum AFB positivity grade 3 prior to treatment whereas in the Jakarta study 68% had grade 1 (for both supplemented and placebo groups) and this can be expected to lead to longer sputum conversion time. It may be that the supplementation effect does not become clear in such severe cases.
The third explanation for the difference between the current study and the previous study in Jakarta is related to the sample size. In Jakarta, the sample size was estimated "on the ability to determine a difference of retinol cosentration in plasma, with an alpha = 0.05 and 1-beta = 0.95 with use of a one-tailed test for concentrations of hemoglobin in blood, and of retinol and zinc in plasma." The findings in the Jakarta trial, of an effect of zinc+vitamin A supplementation on the sputum conversion time was unexpected as no sample size was estimated for this specific outcome. In the present study, designed for confirmation of the previous results, the effect on sputum conversion time was not found. Thus, the probability is real that the finding in the first study was not valid, due to underpowerment of the study to measure a difference in sputum conversion time.
Another explanation may be related to the observation that the patients in the current study were more malnourished than the patients in the Jakarta study. Mean BMI of our cohort was 16.5 Kg/m2 as compared to 18.5 Kg/m2 in the Jakarta study. There was no available population based study on adult malnutrition, however a recent study found 33% of under nutrition occurred among neighbors of the TB patients. Such a low BMI may reflect two processes. One would be protein energy malnutrition (which severely affects host defense) and the other wasting due to the catabolism induced by the acute phase response. As found in our subjects, the BMI increased along with reducing of inflammation. When the patients got healed, as shown in the total lesion area of lung, they gained more weight and also of other micronutrients, such as iron, zinc and vitamin A. (table 1 and 3) This might explain how the hemoglobin increased, although it was still in the borderline to be considered normal.
Neither of low intake or wasting is directly affected by supplementation of micronutrients and hence it may not be remarkable that we failed to show an effect in this study. Beside macronutrient deficiencies, there was also the possibility that additional micronutrients deficiency were present in the study population, which could not be corrected with zinc and vitamin A.
For example, vitamin D has been implicated in defense against TB, but vitamin D deficiency is a worldwide problem. Furthermore there might be a link between vitamin D deficiency with ethnicity. In the Jakarta study, the patients were mainly from Java, Sunda and Sumatra, with a minority of others, the current study consists mainly of patients indigenous to Timor and Rote islands. We found in the same study areas that ethnicity was associated with the development of TB and severity of TB[28, 29]. In line with this, a recent study showed that vitamin D receptor genetic polymorphisms were associated with the time to sputum culture conversion. Also it was shown that vitamin D receptor genotype independently predicted the sputum smear conversion time while on anti-TB therapy. One might speculate that ethnic background also plays a role in the response to supplementation with micronutrients.
We conclude that the patient groups studied here, were suffering from severe tuberculosis with cavities in at least one third of the patients, and high sputum positivity grade. The patients were vitamin A deficiency whereas zinc was at borderline but sufficient to support the immune system and perhaps more important, also severely malnourished. It was also important to notice that the supplementation could not elevate the level of retinol and zinc plasma higher than its borderline. Against this background we were not able to replicate the results of the Jakarta study that demonstrated a beneficial effect of vitamin A and zinc in tuberculosis. This means that the previous findings in Jakarta cannot be generalized. For further studies, a larger sample size will be required, and higher dosages of the zinc and vitamin A supplementations should be considered. Also, in patients with very low BMI, the effect of protein supplementations needs further study.