Article Text

Download PDFPDF

Reactivation of tuberculosis and vitamin D deficiency: the contribution of diet and exposure to sunlight
  1. A Sita-Lumsden1,
  2. G Lapthorn1,
  3. R Swaminathan2,
  4. H J Milburn1
  1. 1
    Department of Respiratory Medicine, Guy’s and St Thomas’ Foundation Trust, London, UK
  2. 2
    Department of Chemical Pathology, Guy’s and St Thomas’ Foundation Trust, London, UK
  1. Dr Heather Milburn, Chest Clinic, Guy’s Hospital, St Thomas’ Street, London SE1 9RT, UK; heather.milburn{at}


Background: As well as its role in the regulation of calcium metabolism, vitamin D is an immunoregulatory hormone. Epidemiological evidence also suggests a link between vitamin D deficiency and tuberculosis (TB). A study was undertaken to examine serum vitamin D concentrations before treatment in patients with active TB and their contacts from the same ethnic and social background and to investigate the relative contributions of diet and sunlight exposure.

Methods: Serum vitamin D concentrations were measured before treatment in 178 patients with active TB and 130 healthy contacts. The prevalence of vitamin D deficiency and its relation to skin colour, month of estimation and TB diagnosis were determined. 35 patients and 35 frequency-matched contacts completed dietary and sun exposure questionnaires to determine the relative contribution of these to serum vitamin D concentrations.

Results: There was a statistically significant difference in serum vitamin D concentrations between patients and contacts (20.1 vs 30.8 nmol/l, 95% CI 7.1 to 14.3; p<0.001) and significantly more patients had severely deficient concentrations (<21 nmol/l) than controls (114/178 (64%) vs 40/130 (31%), p<0.001). There was no association between serum concentrations of vitamin D and skin pigmentation. The healthy contacts showed a predictable seasonal pattern, rising to peak concentrations in the summer months, but this response was absent in patients with TB. Dietary intake was the same in both patients with TB and contacts matched for age, sex and skin colour, but patients with TB displayed a stronger correlation between serum vitamin D concentrations and dietary intake (r = 0.42, p = 0.016) than controls (r = 0.13, p>0.1). There was no difference in sunlight exposure between the groups.

Conclusions: Patients with active TB have lower serum vitamin D concentrations than contacts from similar ethnic and social backgrounds and with comparable dietary intake and sun exposure, and do not show the expected seasonal variation. These observations indicate that other factors are contributing to vitamin D deficiency in patients with TB and suggest abnormal handling of this vitamin.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Links have been made between tuberculosis (TB) and deficiency of vitamin D (25-hydroxycholecalciferol) following a number of observations. Serum concentrations of 25-hydroxycholecalciferol in patients presenting with TB are on average lower than in healthy matched controls,1 and the prevalence of TB is higher among those with low serum 25-hydroxycholecalciferol concentrations such as elderly subjects and Asian immigrants to the UK.2 In one study, patients with pulmonary TB who had higher vitamin D levels were shown to have less extensive radiological disease.3 Furthermore, Wilkinson et al4 found an association between vitamin D deficiency and vitamin D receptor polymorphisms and TB in Gujerati Asians. In addition, notifications of TB in England and Wales are reported to peak in the summer months, suggesting that a low trough of vitamin D concentration after winter contributes to reactivation of latent infection.5 Further work suggests, however, that the association may not be so clearcut because, although a summer peak of TB was confirmed in patients of Indian subcontinent ethnic origin, seasonality was not present in the white population.6

The above associations between vitamin D deficiency and TB may be explained by evidence for an immunoregulatory role for this vitamin. Impaired T cell function, including decreased production of the Th1 cytokines interleukin-2 and interferon γ, have resulted from deficiencies of protein, zinc and the active metabolite of vitamin D, 1,25-dihydroxyvitamin D3 (calcitriol).7 In vitro studies have shown that monocytes have receptors for calcitriol, and vitamin D metabolites can activate the antimycobacterial responses of human monocytes and macrophages, enhancing phagocytosis and granuloma formation.810 These metabolites can act synergistically with interferon γ to contain intracellular replication of Mycobacterium avium and M tuberculosis, and can induce nitric oxide synthase to suppress growth of M tuberculosis in human macrophages.1113

Many factors can influence serum vitamin D concentrations, but diet and sunlight exposure are particularly important.14 In this study we have measured serum vitamin D concentrations at presentation and throughout treatment in a mixed ethnic group attending an inner London chest clinic for possible TB to look for seasonality, ethnic differences and association with active disease. We have further investigated a group of patients with TB and their frequency-matched controls to assess the influence of dietary intake and sunlight exposure on vitamin D concentrations.



Serum 25-hydroxycholecalciferol concentrations were measured routinely in all those presenting to an inner London chest clinic with either a history of exposure to TB or a clinical history suggestive of active disease. Concentrations of parathyroid hormone were measured in 10 cases chosen at random to ensure that low vitamin D levels indicated true vitamin D deficiency. Further measurements were taken during and at the completion of treatment for TB. Data were collected at presentation throughout the year (note being taken of the month of presentation) from 178 patients with culture-positive pulmonary or extrapulmonary TB and 130 healthy contacts from similar ethnic and social backgrounds. Contacts were close household or social contacts with positive tuberculin skin tests or evidence of old primary disease on chest radiography but no evidence of active TB.

Skin colour of patients and contacts was defined as “dark”, ie, black (these were generally people of West Indian or African origin and included some South Indians and Sri Lankans); “mid”, ie, brown (generally those of Asian and South American origin); and “light”, ie, white (generally those of North European origin). Their demographic data are shown in table 1.

Table 1 Patient characteristics

As several factors are known to influence 25-hydroxycholecalciferol concentrations, we excluded infants under 2 years, women over 55 years, significant smokers (>10/day), the clinically obese (body mass index >30 kg/m2) and patients with concurrent disease, eg, carcinoma of the prostate, any disseminated carcinoma, uncorrected thyroid disease or diabetes mellitus and renal disease. These factors can themselves either depress or elevate serum levels of vitamin D and would therefore add unnecessary variables to the results. These exclusion criteria were applied to both cases and contacts.

Ethical approval was obtained for the study and subjects provided written consent to participate.

Measurement of serum 25-hydroxycholecalciferol and parathyroid hormone

Serum was extracted with acetonitrile (to precipitate protein) and 25-hydroxycholecalciferol measured by radioimmunoassay (Diosorin Ltd, Stillwater, Minnesota, USA). The interassay coefficient of variation of this method was 9.1% and the recovery of added 25-hydroxycholecalciferol to both normal sera and hypergammaglobulinaemic sera was >92%.

Intact parathyroid hormone was measured by a chemiluminescence assay on the Nicholas Advantage Specialty system. The intra-assay and interassay coefficients of variation were 5.4% and 9.2%, respectively.

Definition of vitamin D deficiency

Severe 25-hydroxycholecalciferol deficiency is generally defined by a serum concentration of 20 nmol/l or less.1 Others define clinical vitamin D deficiency as a serum concentration of <40 nmol/l as this level has been associated with severe skeletal effects.9 For the purposes of this study, we defined severe deficiency as a serum concentration of 20 nmol/l or less, deficiency as 21–39 nmol/l and adequate concentrations as 40–195 nmol/l. All subjects with vitamin D levels <40 nmol/l were given dietary advice and supplementation with Adcal D3 (calcium 600 mg as carbonate plus cholecalciferol 10 μg (400 units)), one or two tablets daily, depending on the degree of deficiency. Follow-up measurements were taken after a minimum of 3 months on replacement therapy.

Seasonal variation

Serum vitamin D levels were measured in patients and contacts at presentation. Patients and contacts presented at different times throughout the year and, because of the additional effects of ultraviolet radiation in the summer months, data are presented for each month of the year. Comparisons were made for patients and contacts for each month and for the six summer months compared with the six winter months.

Investigation of diet

To asses the involvement of diet, 35 patients with TB and 35 controls matched for age, sex and skin colour completed a validated food frequency questionnaire and a questionnaire to establish levels of sun exposure. The age of patients and contacts in these groups ranged from 9 to 55 years (median 30); 36 were male and 46 had dark skin. As there were no pre-existing food frequency questionnaires relating to vitamin D, one was developed using the USDA table of foods naturally containing vitamin D (mainly oily fish, lard, egg yolk and liver)15 and information from manufacturers’ packaging on supplemented foods. In the UK margarine has been fortified since 1925. More recently, some breakfast cereals have been fortified and soya milk also contains 0.75 μg of added vitamin D/100 ml. We are not aware of any other commonly available foods being fortified in the UK.

The International Unit (IU) is a measure of the activity of vitamin D and equates to 0.025 μg. Using this conversion, we calculated each individual’s average daily intake in μg. In 1991 nutrients were given new dietary reference values by the Department of Health (UK) and, from the Dietary Nutritional Survey of British Adults 1990, recommended nutrient intake (RNI) values were published for different age groups.16 17 There was no RNI for the 4–64 age group as it was concluded that most vitamin D was supplied by the sun in this group.

Measures of exposure to sunlight

Ultraviolet-B (UV-B) rays from the sun trigger the synthesis of vitamin D3 in the skin. A questionnaire was developed based on studies that have determined the variables affecting cutaneous vitamin D production from sun exposure (ie, season, latitude, time of day, cloud cover, smog and sunscreens). Sunscreens with a sun protection factor of 8 or more will block UV-B rays.18 At London’s latitude of 52°N there is no radiation of the appropriate wavelength (290–310 nm) from the end of October to the end of March, and exposure of the face, arms and legs to the sun for 1–2 hours daily during the summer months would be required to obtain sufficient vitamin D to maintain normal concentrations throughout the year.18 This estimate formed the basis for defining adequate levels of sun exposure.

Statistical analysis

Data are reported as mean (SE) unless otherwise stated. Unpaired Student t tests were used to compare serum 25-hydroxycholicalciferol concentration, adequacy of sun exposure and dietary vitamin D intake between patients and their contacts. Seasonal variations in vitamin D levels at presentation in patients and contacts were also compared using the unpaired t test. The χ2 test was used to compare numbers of patients and contacts with severe, deficient or adequate concentrations of serum vitamin D. Pearson rank correlation coefficients were calculated separately for patients and controls to assess the relationship between vitamin D intake and serum 25-hydroxycholecalciferol concentration. Statistical significance was defined as p<0.05 and 95% confidence intervals (CI) are given for comparisons between means.


Serum concentrations of 25-hydroxycholecalciferol

Mean 25-hydroxycholecalciferol concentrations were significantly lower in the patients with TB than in healthy contacts (20.1 (0.95) nmol/l vs 30.8 (1.7) nmol/l, 95% CI 7.1 to 14.3; p<0.001). Only 11 of the 178 patients with TB (6%) had adequate serum concentrations and 114 (64%) had concentrations <21 nmol/l. Of the 130 contacts tested, 35 (27%) had adequate concentrations while 40 (31%) had severe deficiency with concentrations below 21 nmol/l (table 2). The vitamin D status indicated by the serum level was confirmed in 10 cases chosen at random by elevated concentrations of parathyroid hormone (92 (2.9) μg/l; reference range 10–65). Calcium levels corrected for albumin were within the normal range and there was no significant difference in albumin levels between patients and contacts.

Table 2 Serum 25-hydroxyvitamin D concentrations (nmol/l) in patients with active TB and their contacts

All patients and contacts found to have deficient concentrations were given supplements. To date, concentrations have risen above 40 nmol/l on replacement therapy in only 10/48 patients (21%) but 38 continued to have deficient levels despite treatment for several months (fig 1), although 13 of these showed some improvement (ie, 25/48 (52%) failed to show any improvement in 25-hydroxycholecalciferol concentrations after a minimum of 3 months on replacement therapy). Of 10 contacts with deficient concentrations (where follow-up data are available), 5 returned to adequate levels after 3 months on replacement therapy, 3 increased significantly (at least doubled) and 2 did not improve.

Figure 1 Serum 25-hydroxycholecalciferol concentrations (nmol/l) in patients with tuberculosis (TB) before and after a minimum of 3 months on vitamin D replacement therapy. Concentrations in 10/48 patients recovered but in 38 patients (79%) the concentrations failed to return to normal, although 13 of these showed some improvement.

Effect of skin pigmentation

Patients and contacts were grouped for skin colour into dark, mid and light skin pigment groups. All three groups of patients with TB had deficient concentrations of serum 25-hydroxycholecalciferol and there was no difference between the different ethnic groups. A similar pattern was seen in the contacts, but mean concentrations were all consistently higher than in patients from the same racial groups with TB (table 2).

Seasonal variation

Patients and contacts presented at different times throughout the year and measurements at presentation are shown in fig 2. Serum 25-hydroxycholecalciferol concentrations in the healthy contacts displayed a predictable seasonal pattern, rising to a peak in July, August and September following the increase in UV-B in the summer months, with a statistically significant difference between patients and contacts in July (95% CI 1.9 to 31.0 nmol/l, p = 0.013) and September (95% CI 5.7 to 35.8 nmol/l, p = 0.027). This seasonal variation was, however, strikingly absent in the patients with TB (fig 2). Concentrations at presentation in this group were consistently low throughout the year, although there was a very small increase in mean concentrations of 3.4 (1.5) nmol/l for those who presented between the months of May and October compared with those who presented from November to April. The difference in mean concentrations for contacts for the same periods was 14.8 (3.0) nmol/l (95% CI 5.1 to 18.5; p<0.001).

Figure 2 Seasonal variation in serum 25-hydroxycholecalciferol concentrations in patients with tuberculosis (TB) and controls. Controls showed an expected increase to normal concentrations in the summer months but this response was absent in patients with TB. *p<0.05; **p<0.01.

Role of diet and exposure to sunlight

Of the 35 patients with TB who completed the dietary and sun exposure questionnaires, only 6 had adequate serum concentrations of vitamin D and 20 were severely deficient. Of the 35 healthy contacts, 10 had adequate concentrations, 10 had concentrations just below the reference range and 5 had severely deficient concentrations. Patients with TB had significantly lower mean serum 25-hydroxycholecalciferol concentrations than healthy matched contacts (21.1 (2.5) vs 33.8 (3.0) nmol/l, 95% CI 4.8 to 20.6; p = 0.002). There was, however, no difference in mean daily dietary intake of vitamin D or sunlight exposure (table 3). Interestingly, the correlation between daily dietary vitamin D intake and serum 25-hydroxycholecalciferol concentrations was stronger in patients with TB than in healthy contacts (r = 0.42, p = 0.02 and r = 0.13, p>0.1, respectively; fig 3).

Figure 3 Relationship between dietary vitamin D intake and serum vitamin D level (nmol/l). Patients with tuberculosis (TB): r = 0.42, p<0.02; controls: r = 0.13, p>0.1.
Table 3 Summer sunlight exposure, daily dietary vitamin D intake and serum 25-hydroxycholecalciferol concentrations in patients with tuberculosis (TB) and their contacts frequency-matched for age, sex and skin colour


This study confirms previous studies1 19 that patients presenting with active TB have significantly lower mean concentrations of serum 25-hydroxycholecalciferol than their contacts from the same ethnic and social backgrounds. There was no evidence that the vitamin D assay used was affected by hypergammaglobulinaemia which can sometimes be associated with active TB.

Our study also suggests that this association is independent of skin colour. A large study looking specifically at skin type, sun exposure and serum vitamin D concentrations concluded that vitamin D concentrations were not linked with phototype but with sun exposure behaviour.20 Many of our patients failed to increase their serum 25-hydroxycholecalciferol concentrations to adequate levels even after successful treatment for TB combined with at least 3 months and usually 6 months or more of vitamin D replacement therapy. Given that all these patients successfully completed treatment for TB with varying degrees of supervision, it is unlikely that this observation can be explained simply by non-compliance. Furthermore, patients generally received higher doses of vitamin D replacement than contacts. Antituberculous chemotherapy can itself lower 25-hydroxycholecalciferol concentrations,21 and this may have been a contributing factor in the failure to raise concentrations to normal on treatment, but 10 patients still failed to increase their vitamin D concentrations despite continuing with replacement therapy after antituberculous treatment ceased, thus raising the intriguing possibility of a group of patients with deficient handling of ingested and cutaneous vitamin D. Clinical disease only develops in around 10% of immunocompetent people infected with M tuberculosis, which suggests that host genetic factors influence the outcome of infection. Vitamin D acts via the vitamin D receptor that is present on activated T and B lymphocytes. Genetic variation in the vitamin D receptor has been associated with bone mineral density and circulating concentrations of 25-hydroxycholecalciferol.22 Among patients with TB in a large Gambian case-control study, those with the genotype associated with high circulating concentrations of 25-hydroxycholecalciferol were under-represented.23 Polymorphisms in the vitamin D receptor gene, especially when assessed in combination with 25-hydroxycholecalciferol deficiency, have been associated with an increased risk of TB in Gujerati Asians.4

We did not find any seasonal variation in new cases of TB in this study group, the majority of whom were from sub-Saharan Africa. There was a striking seasonal increase in 25-hydroxycholecalciferol concentrations in the contacts, but only a very small increase in the summer months in the disease group. This does not support the idea that reduced sunlight in winter is responsible for the seasonal peak found in some populations,5 but suggests possible abnormal uptake, metabolism and storage of vitamin D, thus increasing the risk of reactivation of disease. Furthermore, there have been recent reports of low 25-hydoxycholecalciferol concentrations in healthy subjects resident in South Asia and elsewhere in the tropics where there is abundant sunshine.24 25 Concentrations in women in these countries may well be low due to dress, but this is less likely to be the case in men. An alternative explanation could be that, rather than low vitamin D concentrations resulting in reactivation of TB, low concentrations of 25-hydroxycholecalciferol could be the result of TB-induced nutritional deficiencies. A large prospective study would be needed to find out whether vitamin D deficiency precedes the development of active TB.

There was a weak association between dietary intake and serum 25-hyroxycholecalciferol concentrations in the TB group but not in contacts, suggesting consumption of vitamin D may be less important for maintaining normal concentrations in healthy individuals. Only six patients with active TB had normal serum concentrations. All were Black Africans, and three came from the same family and had a diet rich in oily fish and also took supplements. This did not, however, protect them from developing TB, and it is likely that other nutrients may be involved in maintaining the integrity of the immune system. Protein malnutrition, for example, potentiates M tuberculosis H37Rv-infected monocyte macrophages to produce higher concentrations of transforming growth factor-β,7 a cytokine implicated as a likely mediator of immunosuppression in TB. Copper, zinc and selenium are also essential elements for a healthy immune system and zinc concentrations increase during TB treatment.26 Despite generally higher concentrations of serum 25-hydroxycholecalciferol in the control (contact) group, there was no difference in dietary intake. Patients with TB and contacts had a marginally higher dietary intake of vitamin D than the average daily intake reported in the last National Dietary Survey in the UK.17 In 1986–7, the average daily intake in 16–64 year old subjects was 3.78 μg for men and 3.09 μg for women.

There were comparable numbers in both contact and TB groups who had adequate levels of sun exposure. This was based on results of a questionnaire, and it is possible that this form of self-reporting did not reflect true exposure to sunlight. More accurate results could be found in groups who spend no time in the sun or cover up for religious reasons. This observation, however, is supported by a small study from India which concluded that sunlight exposure was adequate in patients with TB with vitamin D deficiency but reduced dietary intake.27

From our data it appears that patients with active TB from similar ethnic and social backgrounds and with comparable vitamin D intake and sun exposure have lower serum 25-hydroxycholecalciferol concentrations than their healthy contacts. This indicates that other factors contribute to vitamin D deficiency in the TB group. Furthermore, the lack of seasonal variation, together with failure of a significant number of patients to increase 25-hydroxycholecalciferol concentrations to normal despite replacement therapy, suggests abnormal handling of this vitamin. A further study is required to investigate vitamin D intake, metabolism and storage to unravel the relationship between vitamin D and TB.


The authors are grateful to Mrs Elizabeth Weekes, dietician, for advice in compiling the dietary questionnaire and to Drs Bruce McClaren and Dipak Kanabar for introducing some patients to the study.



  • Conflict of interest: None.

  • AS-L devised and validated the questionnaires and collected data. GL collected and analysed data. RS collected data and provided biochemical advice. HM devised and coordinated the study and wrote the paper. All authors contributed to the final manuscript.

  • Abbreviations: