There is a high prevalence of vitamin D deficiency reported in India: ranging from 75 to 90 % of the population, in all age groups including pregnant women and their newborns, infants, adolescents and adults( Reference Sachan, Gupta and Das 1 – Reference Sahu, Bhatia and Aggarwal 4 ); from north( Reference Sachan, Gupta and Das 1 , Reference Seth, Marwaha and Singla 2 , Reference Harinarayan 5 – Reference Zargar, Ahmad and Masoodi 7 ) and south India( Reference Balasubramanian, Shivbalan and Kumar 3 , Reference Harinarayan 5 ); in urban( Reference Sachan, Gupta and Das 1 – Reference Balasubramanian, Shivbalan and Kumar 3 , Reference Marwaha, Tandon and Reddy 6 ) and rural( Reference Sahu, Bhatia and Aggarwal 4 ) populations; and higher as well as lower socio-economic populations( Reference Marwaha, Tandon and Reddy 6 ). The reasons for this high prevalence, despite India being a tropical country, include low body surface area exposed to sun (due to the traditional, modest pattern of clothing and poor outdoor activity), skin pigmentation, dietary Ca deficiency and, possibly, particulate pollution. Goswami et al.( Reference Goswami, Gupta and Goswami 8 ) emphasized the role of duration of sunshine exposure and skin pigment in their study of 25-hydroxycholecalciferol (25(OH)D) levels among soldiers, physicians, nurses, pregnant women and people with depigmented skin. All groups had suboptimal 25(OH)D levels except the soldiers and people with depigmented skin.
There is no study on rural subjects in India, correlating their direct UVB irradiance exposure in all seasons of the year and serum 25(OH)D levels. Such a study would help public health policy planners to know whether sunshine can be realistically expected to be an adequate source of vitamin D for Indians and, if so, in which months of the year.
We documented some of the limiting factors for cutaneous vitamin D synthesis in a group of pregnant rural women by measuring the magnitude of UVB energy exposure and serum 25(OH)D in different seasons throughout the year. We also analysed the differences in radiant UVB energy in regions with differing amounts of particulate pollution.
Materials and methods
Study population and region
The study population was located in a rural area about 40 km from the city of Lucknow (latitude 26·8°N, longitude 80·9°E, altitude 128 m). Lucknow has three distinct seasons, namely winter (November–January), monsoon (July, August) and dry summer in the rest of the months. Pregnant women in the second trimester (n 139; mean gestational age 22 (sd 5) weeks) were recruited consecutively over one year. They belonged to poor socio-economic status and resided in huts or poorly ventilated houses of one or two rooms. The daily activity included household chores which were mostly performed in the courtyard and farm work during the seasons of farming. Outdoor activity and clothing were uniform for a season and dictated by the weather and necessities of farming. The body surface area exposed during the winter included the face, neck and the hands, and during the rest of the year it also included the forearms. We compared the skin colour tone of the face with a skin shade chart based on the Fitzpatrick classification( Reference Fitzpatrick 9 ). They all had skin type V. The mean daily Ca intake was low (214 (sd 150) mg) due to poor consumption of milk products (due to economic constraints), despite an effort made by families to provide more milk to pregnant women. A log of daily activities and clothing was recorded by the research staff once weekly throughout the year. The outdoor exposure was estimated between 10.00 and 14.00 hours, which included the peak hours of UV radiation exposure. Mean exposure time was 3·2 (sd 0·2) h/d during winter season (November–January) and 2·1 (sd 0·4) h/d during summer (February–October). The proportion of the body surface area exposed was 18·5 % in summer and 9·5 % in winter.
The study was approved by the Institutional Ethics Committee of Sanjay Gandhi Postgraduate Institute of Medical Sciences (2003). Verbal informed consent was obtained from all participants. Verbal consent was witnessed and formally recorded.
Measurement of UVB radiant energy
A portable UVB light meter (a silicon carbide photodiode with a 290–320 nm UVB filter) was used to measure the UVB radiation in millivolts. (Surface UVB energy peaks at 310 nm while the UVB energy directly related to vitamin D production peaks at 300–305 nm.) The instrument was designed, assembled and calibrated by one of the authors (P.U.). A silicon carbide detector made by Laser Components, UK (model JEC 0.3B) was used. The meter was calibrated by comparison with the integrated irradiance measured by a spectroradiometer which was calibrated by reference to a quartz–tungsten–halogen lamp irradiance standard against UK National Physical Laboratory standards. The calibration factor was 13 µW/cm2 per mV. The estimated uncertainty corresponded to approximately ±15 %. The angular field of view was 180°. A total of 474 UVB readings were taken at three locations, namely our hospital campus situated in the outskirts of the city, at a crowded city centre and in the villages of residence of our participants. Measurements were taken at different times of day between 09.00 and 16.00 hours, in all seasons of the year. During April and May (a time of dry heat with no cloud cover), readings taken at any one location on a particular day were then immediately accompanied by readings at the other two locations on successive days, so as to have the same environmental conditions during each set of three days. Similarly, to document the effects of clouds, during the monsoon season of July and August, consecutive cloudy and sunny day readings were taken.
Calculation of estimated cutaneous vitamin D formation
We employed the formula of Godar et al.( Reference Godar, Pope and Grant 10 ), which takes the following variables into account: standard erythemal dose (SED)( 11 ), a reflection of UVB energy received per hour; action spectrum conversion factor (ACSF)( Reference Pope, Holick and Mackin 12 ), a correction factor for solar zenith angle; geometric conversion factor (GCF)( Reference Model, Pope and Godar 13 ), a correction factor for geometrical body shape; skin type factor (STF), which accounts for relative efficiency of vitamin D synthesis for the given UVB energy depending upon the Fitzpatrick skin type( Reference Godar, Pope and Grant 14 ); age of the participants (AF); and fractional body surface area exposure (FBE)( Reference Lund and Browder 15 ). The formula employed is:
$$\eqalignno{ & {\rm Ch olecalciferol}\,{\rm (vitamin \ D}_{{\rm 3}} {\rm )}\,{\rm production}\,{\rm (IU{\scriptstyle /}d)} \cr & \,\,\, \;{\equals}{\,\rm SED{\scriptstyle /}d}{\times}{\rm ASCF}{\times}{\rm GCF}{\times}4900\,{\rm IU}\,{\rm for}\,{\rm skin}\,{\rm type}\,{\rm II} \cr & \,\,\,\,\,\,\,\,\, {\times}{\rm STF}{\times}{\rm FBE}{\times}{\rm AF}. $$
The production of 122·5 µg (4900 IU) vitamin D per SED was taken from experiments of Holick( Reference Holick 16 ). Furthermore, after certain duration of UVB energy exposure, there is cessation of cutaneous precholecalciferol formation and diversion to inactive metabolites such as lumisterol and tachysterol occurs. Various studies have shown a wide range of erythemal energy cut-off ranging from 0·5 to 1·5 minimal erythemal dose (MED), where cutaneous 7-dehydrocholesterol to precholecalciferol conversion plateaus off for various skin types( Reference Holick, MacLaughlin and Doppelt 17 – Reference Holick 19 ). Taking 1 MED (750 J/m2 for type V skin) threshold for our calculation, we capped the SED threshold for maximum vitamin D production in type V skin as 7·5 SED/d( Reference Godar, Pope and Grant 10 , Reference Tejasvi, Sharma and Kaur 20 ) because 1 SED is equivalent to 100 J/m2.
Serum 25-hydroxycholecalciferol
Participants were recruited during the different seasons throughout the year. Blood was sampled for analysis of serum 25(OH)D using a RIA kit (Diasorin, Stillwater, MN, USA; analytical sensitivity: 3·75 nmol/l, intra-assay CV: 8·6–12·5 %, inter-assay CV: 8·2–11·0 %). Vitamin D deficiency was defined as serum 25(OH)D<50 nmol/l.
Results
Surface irradiance: variation with season, time of day and location
Variation in UVB energy with the time of day, season and cloud cover is shown in Figs 1 and 2. Peak readings (between 12.00 and 13.00 hours) were at their lowest during the winter season from November to January and were at their highest during the summer season in May and June. The presence of clouds in the sky during the rainy season was associated with reduced radiant UVB energy.
Fig. 1 Peak UVB irradiance measured at 10.00 hours (
) and 13.00 hours (
) at different seasons of the year in Lucknow, India (latitude 26·8°N)
Fig. 2 Peak UVB irradiance measured at different times of day during different months of the year (
, November; 
, January; 
, April; 
, May; 
, August (cloudy); 
, August (sunny)) in Lucknow, India (latitude 26·8°N)
Peak UVB readings during April and May, between 12.00 and 13.00 hours, were taken at the hospital campus, the crowded inner-city location and the village areas. A total of eighteen UV energy readings (six readings at each location) at similar times at all three locations were taken on consecutive days. Significantly higher surface UVB irradiance was recorded in the hospital campus (mean 340 (sd 45) µW/cm2) and in village locations (mean 310 (sd 60) µW/cm2) than in the crowded inner-city location (mean 250 (sd 50) µW/cm2; P=0·024 for multiple group comparison, P=0·03 on post hoc comparison between the hospital campus v. inner-city location).
Average daily vitamin D synthesis in rural women
Using the method of Godar et al., the rural women’s average daily vitamin D synthesis during the winter season was calculated to be 19·25 µg/d (770 IU/d) and 37·25 µg/d (1490 IU/d) during summer.
Serum 25-hydroxycholecalciferol
Mean serum 25(OH)D was 28 (sd 15) nmol/l in women tested during winter (n 48) and 92 % of them had vitamin D deficiency (25(OH)D<50 nmol/l). During the rest of the year, mean serum 25(OH)D was 42 (sd 20) nmol/l (n 91) and 70 % of the women had vitamin D deficiency (25(OH)D<50 nmol/l). The highest 25(OH)D values were obtained in women sampled in the months of July and August (mean 56 (sd 20) nmol/l) and 60 % had values >50 nmol/l.
We did not find a significant difference in mean serum 25(OH)D levels between groups of women in early second trimester (≤20 weeks; 37·5 (sd 20) nmol/l) and late second trimester (>20 weeks; 38 (sd 18) nmol/l; P=0·9). The mean BMI of our participants was 21 (sd 2·2) kg/m2 and none of them had BMI≥27·0 kg/m2 (the obesity cut-off for Indians). Twenty-one participants had BMI≥23·0 kg/m2 (the overweight cut-off for Indians). There was no significant difference in mean serum 25(OH)D levels between groups of women with BMI<23·0 kg/m2 (39 (sd 18) nmol/l) and BMI≥23·0 kg/m2 (36 (sd 23) nmol/l; P=0·8).
Discussion
Our results demonstrate that in the summer months, the cutaneous production of vitamin D is equivalent to the daily ingestion of 37·25 µg (1490 IU) of cholecalciferol. Going by the findings of numerous studies in the literature as well as the recommendations of international expert groups, this amount should be sufficient to maintain serum 25(OH)D above 50 nmol/l( Reference Vogiatzi, Jacobson-Dickman and DeBoer 21 , Reference Holick, Binkley and Bischoff-Ferrari 22 ). Similar findings have been reported by other investigators studying in vivo cutaneous vitamin D production in South Asian individuals( Reference Farrar, Webb and Kift 23 , Reference Marwaha, Sreenivas and Talwar 24 ). After 6 weeks of exposure of 35 % body surface area, three times weekly, to 3·25 and 3·9 SED UVB energy, Farrar et al.( Reference Farrar, Webb and Kift 23 ) documented a rise in serum 25(OH)D to 23·8 and 31·5 nmol/l, respectively. Studying the effect of 30 min of natural sun exposure of 15–30 % of body surface area in Indian children every day for 4 weeks, Marwaha et al. ( Reference Marwaha, Sreenivas and Talwar 24 ) documented a rise in serum 25(OH)D from 23 (sd 14) to 33 (sd 17) nmol/l. Extrapolating to a longer daily exposure and a longer duration of study would give a post-exposure value comparable to our summer peak 25(OH)D. A web-based calculator of vitamin D formation per skin type, latitude, body surface area exposed and exposure time also gives similar results of vitamin D production as in our study( Reference Webb and Engelsen 25 ).
If type V skin can produce such a generous amount of vitamin D, why do 70 % of the women in our study still have 25(OH)D<50 nmol/l in serum? First, it must be borne in mind that daily doses of vitamin D, when ingested orally, take about 3 months to reach a plateau level in blood( Reference Heaney, Davies and Chen 26 ). A similar observation was documented with in vivo production in South Asians in response to UVB exposure for 6 weeks. Weekly measurements of 25(OH)D showed rising 25(OH)D even at 6 weeks, especially in the higher SED exposure group( Reference Farrar, Webb and Kift 23 ). These studies support our observation that the highest 25(OH)D values were obtained in the months of July and August (mean 56 (sd 20) nmol/l; 60 % of participants having values >50 nmol/l). It suggests that the severe depletion in our women during the winter months (mean 25(OH)D being 28 (sd 15) nmol/l) takes time to be compensated during the summer months of April, May and June. Just after it reaches a plateau, the cutaneous vitamin production is again hampered by the monsoon season and the following winter season, thus rendering our participants vitamin D-deficient for many months of the year. Marwaha et al.( Reference Marwaha, Yenamandra and Sreenivas 27 ) found a significant decrease in mean 25(OH)D levels during winter despite sun exposure. Second, a very important role is that of degradation of precholecalciferol into suprasterol I and II on prolonged exposure to sunlight( Reference Webb, de Costa and Holick 28 ). This is nature’s mechanism of preventing vitamin D toxicity. This mechanism of degradation of precholecalciferol on prolonged UVB exposure in our participants may partly explain their low observed 25(OH)D levels. All these studies reaffirm that it is the body surface area exposed which is the major limiting factor in our population. Third, our participants had an extremely low dietary Ca intake, which has been shown in both in human subjects and animal models to be associated with secondary hyperparathyroidism leading to depletion of vitamin D stores( Reference Clememts, Johnson and Fraser 29 – Reference Harinarayan, Ramalakshmi and Prasad 31 ). Fourth, the recent finding of genetic variation in the level of vitamin D-binding protein in those of African ancestry suggests that this should be explored in Indians( Reference Powe, Evans and Wenger 32 ).
Our studies on the comparison of the UVB irradiance in polluted v. non-polluted locations corroborate the findings of Agarwal et al., who also found lower irradiance in a crowded location full of vehicular traffic as compared with the countryside( Reference Agarwal, Mughal and Upadhyay 33 ). Particulate pollution is well known to block penetration by solar UV rays.
Our study has some limitations. The study was not longitudinal as each participant was not tracked through a whole year. Second, we did not have an objective measure of a personal UV meter to track the exposure time of each individual. The serum 25(OH)D values were not measured in all trimesters (although it must be mentioned that serum 25(OH)D levels have not been found to be statistically different between trimesters in a previous publication from north India( Reference Marwaha, Tandon and Chopra 34 )). Furthermore, our analysis may not be extrapolated to non-pregnant women. Despite these limitations, however, we think our study allows us to conclude that cutaneous vitamin D synthesis at latitude 26·8°N in our rural women is limited more by body surface area exposed to the sun than by duration of exposure, especially during winter when it is compounded by poorer UVB irradiance. The degree of vitamin D deficiency can be expected to be more pronounced in urban people as they have even less outdoor exposure and equally poor body surface area exposure. Particulate pollution plays an important role in limiting UVB irradiance in crowded city centres in India. It may be necessary for policy makers to recommend universal preventive supplementation at least during the winter months.
Acknowledgements
Acknowledgements: Technical help in the hormone laboratory from P.K. Awasthi and in the rural area from Arjun Singh is gratefully acknowledged. Financial support: This work was supported by the Department of Biotechnology, Government of India (V.B., grant number DBT/PR/3552/SPD/11/349/2002). The Department of Biotechnology, Government of India, had no role in the design, analysis or writing of this article. Conflict of interest: The authors declare no conflict of interest. Authorship: S.S. acquired the data, analysed the data and wrote the manuscript; P.U. conceptualized the study, standardized the methodology, analysed the data and approved the manuscript; M.S. planned the study, acquired the data and approved the manuscript; V.R. acquired the data and approved the manuscript; V.B. conceptualized, planned and executed the study, analysed the data and approved the manuscript. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the ethics committee of Sanjay Gandhi Postgraduate Institute of Medical Sciences. Verbal informed consent was obtained from all participants. Verbal consent was witnessed and formally recorded.
