Smoking behavior and circulating vitamin D levels in adults: A meta‐analysis

Abstract To determine the effect of smoking on circulating vitamin D in adults, we performed a meta‐analysis. Literature before 9 May 2021 was retrieved from electronic literature databases such as EMBASE, PubMed, and Cochrane. The quality of the included studies was assessed by two researchers against the Newcastle–Ottawa scale and JBI Evidence‐based Health Care Centre criteria. All eligible studies and statistical analyses were performed using STATA 14. Twenty‐four studies with 11,340 participants meeting the criteria were included in the meta‐analysis. The results of meta‐analysis showed that the level of circulating 25(OH)D in smokers was lower than that in nonsmokers. A subgroup analysis based on vitamin D supplement use showed that both smokers who used vitamin D supplements and smokers who did not use vitamin D supplements had lower blood 25(OH)D levels compared with the control group. In addition, subjects were divided into different subgroups according to age for meta‐analysis, and the results showed that the serum 25(OH)D level in each subgroup of smokers was lower than that in the control group. This meta‐analysis revealed differences in circulating vitamin D levels between smokers and nonsmokers, with smokers likely to have lower circulating vitamin D levels.


| INTRODUC TI ON
It is well known that smoking is an important cardiovascular and cerebrovascular risk factor. Now, some studies have proved that smoking is not only a major harm to cardiovascular and cerebrovascular health, but also a major cause of chronic obstructive pulmonary disease. Smoking is associated with accelerated decline in lung function, increased mortality, and worsening symptoms in asthma and COPD (Boulet et al., 2006). There are also meta-analyses that show smokers have an increased risk of fracture (Kanis et al., 2005;Shen et al., 2015).
Vitamin D is one of the most important fat-soluble vitamins in the body, the most important of which are D2 and D3. As the main circulating form of vitamin D, 25(OH)D plays an important role in regulating cell proliferation and immune response, in addition to regulating serum calcium and phosphorus levels (Umar et al., 2018;Vanherwegen et al., 2017). Studies have found that inhibiting the production of cAMP protein induced by 1,25(OH) 2D3 can enhance | 5821 YANG et Al. the growth of mycobacterium (Liu et al., 2007). A meta-analysis found that vitamin D deficiency increases the risk of ischemic stroke (Zhou et al., 2018). Vitamin D deficiency has also been related to type 2 diabetes (Berridge, 2017;Lucato et al., 2017) and may increase the risk of falls (Girgis et al., 2013). Vitamin D deficiency is now a major health problem in the real world.
Many studies have shown the effect of smoking on circulating 25-hydroxyvitamin D, although someone has considered that no statistically significant effect on circulating vitamin D levels (Stürmer et al., 2015). However, there is growing evidence of the negative effects of smoking on 25-hydroxyvitamin D and calcium metabolism (Cuomo et al., 2019;Cutillas-Marco et al., 2012;Hermann et al., 2000;Kim et al., 2018;Richard et al., 2017). A study by Jääskeläinen et al. (2013) of 5,714 subjects (47% males) aged 30-79 years found that smokers had lower serum 25(OH)D concentrations than nonsmokers. Hermann et al. (2000) found that current smokers had significantly lower serum 25-OHD levels than nonsmokers. Although the decrease was only 6.8%, serum levels were negatively correlated with the number of cigarettes currently smoked per day (β = −.16; p = .003). In addition, Thuesen et al. (2012) in a recent large population study showed that the odds ratios of severe vitamin D deficiency(25(OH)D < 10 ng/ml)/vitamin D deficiency(25(OH)D < 20 ng/ml) associated with daily smoking were 1.47 and 1.36, respectively. This cannot be explained by other confounding lifestyle factors. After adjustment for other confounders, the adverse effects of smoking on circulating vitamin D levels were still suggested (Brot et al., 1999).
There is no consensus on whether smoking causes damage to circulating vitamin D levels, which may have to do with differences in study design or participant characteristics (such as age, sex, weight, health status), and techniques (testing methods) between studies. In recent years, research has continued on whether low levels of vitamin D are common in smokers. To determine the impact of smoking behavior on circulating vitamin D levels, we conducted a meta-analysis of published studies to draw conclusions and provide public advice on clinical and public health issues.

| Selection criteria
All studies included in this meta-analysis met the following five criteria: (1) Diseases affecting vitamin D absorption or metabolism were excluded, (2) clear smoking status and circulating vitamin D data, (3) all subjects are 18 years or older, (4) reported the types of samples, and (5) compare the circulating vitamin D status of smokers with that of nonsmokers. Any of the following studies were excluded: participating in osteoporosis studies or not excluding diseases that affect vitamin D levels; data cannot be obtained, extracted, or documented (or if the author does not respond when attempting to contact); studies on minors; patients who use marijuana or drugs; and studies on vitamin D deficiency are ill-defined.

| Data extraction
The following information and clinical characteristics of the participants were extracted from the included study: first author, year of publication, research type, detection method, sample type, study location, average age, body mass index (BMI), number of participants, number of smokers and nonsmokers, circulating vitamin D levels in smokers and nonsmokers (dichotomous variables studies were used to investigate vitamin D deficiency in smokers and nonsmokers), and use of vitamin D supplements.

| Quality assessment
The quality assessment of the included studies used the Newcastle-Ottawa scale (NOS) based on cohort studies and case-control studies. We evaluated the article on the basis of selection, comparability, and outcome/exposure, with a possible total quality evaluation of 10. The Australian JBI Centre for Evidence-Based Health Care (2016) Quality Assessment Tool was used as the evaluation criteria for the cross-sectional study. All articles are evaluated by a reviewer.

| Outcome measures
Our primary outcome measure was adult circulating vitamin D levels, compared between smokers and nonsmokers, and performed a subgroup analysis based on meta-regression results.

| Statistical analyses
Continuous results were calculated using Cohen's method as mean difference (SMD) and 95% confidence interval (CI), and the combined effect values of dichotomous variables were RR and its 95% confidence interval, using an inverse-variance model. I 2 statistics were used to test heterogeneity among studies. A value of less than 50% of I 2 is considered low heterogeneity. In statistics, when there is significant heterogeneity, the size of the mixed effect is calculated by using the random effect model. Otherwise, a fixed effect model is used. Suppose p value < .05 or 95% CI without 0 (the 95% CI for the RR value does not include 1) was considered statistically significant. Based on the results of meta-regression, the main sources of heterogeneity were determined and a subgroup analysis was conducted. The Begg method was used to assess publication bias. Data analysis was performed using STATA 14.

| Characteristics of the chosen articles
Two thousand five hundred forty-one articles were retrieved from the database (Figure 1). After reviewing the summary, we selected 218 complete publications for further evaluation based on our inclusion criteria. Of these studies, 150 were excluded because patients with diseases that affect vitamin D absorption or metabolism were not excluded, and eight were excluded because the samples were not adult venous blood. There were 36 studies did not have the required data or did not contain sufficiently detailed data and that could not be obtained by contacting the authors. Finally, this meta-analysis included 24 articles (Alam et al., 2018;Bianchi et al., 2012;Chaudhuri et al., 2013;Cuomo et al., 2019;Cutillas-Marco et al., 2012;Díaz-Gómez et al., 2007;Hermann et al., 2000;Jorde et al., 2019;Kassi et al., 2015;Kim et al., 2018;Klingberg et al., 2015;Li et al., 2016;Lokki et al., 2020;Lu et al., 2020;Mulligan et al., 2014;Okan et al., 2020;Richard et al., 2017;Schierbeck et al., 2012;Schmitt et al., 2018;Stürmer et al., 2015;Supervía et al., 2006;Tonnesen et al., 2016;Xu et al., 2020) (Table 1). The 24 studies involved 11,340 participants: 1,399 smokers and 9,941 nonsmokers were included in this metaanalysis (when quitters are present, they are classified as nonsmokers

Author
Year Type of experiment

| Quality assessment
The quality of the studies was evaluated according to NOS and JBI standards. The literature in this meta-analysis included five cohort studies, two case-control studies, and 17 cross-sectional studies. We judged study quality in terms of selection, comparability, and results/ exposure of studies according to the Newcastle-Ottawa scale of observational studies ( Table 2). The total quality scores of the seven articles (cohort and case-control studies) ranged from 4 to 9, with a possible total of 10 points. The authenticity evaluation results of the cross-sectional study according to the JBI quality evaluation tool are shown in Table 2. Due to strict inclusion and exclusion criteria to exclude the influence of certain diseases on circulating vitamin D levels, these studies had low selection bias. Since the purpose of this study was to explore the effects of smoking on circulating vitamin D levels in adults, the exposure factors were identified as smoking, and subject information was obtained using questionnaires or written self-reports; no study using smoking as an exposure factor was subject blind.

| Vitamin D deficiency or insufficiency in smokers and nonsmokers in the dichotomous variable studies
People who are current smokers are more likely to have circulating vitamin D deficiency or deficiency than nonsmokers. We obtained a 95% confidence interval combined effect amount (RR 1.11, 95% CI: 1.03-1.19, p < .001). p < .05 and 95% CI did not contain 1. The model is fixed, inverse-variance, I 2 = 75% ( Figure 2). No source of heterogeneity was found in meta-regression. Sensitivity analysis suggested that study by Yawen Lu et al. was the main source of heterogeneity. When that study was removed, I 2 dropped to 38%.
To better rule out the effect of vitamin D supplement use on the results, we analyzed the use of vitamin D supplements in subgroups.
The use of vitamin D supplements was defined as 1, and nonuse of vitamin D supplements was defined as 0. The results are as follows ( Figure 3): The combined effect size was available in people who did not receive vitamin D supplementation (RR 1.29, 95% CI: 1.09-1.53, p < .05), I 2 = 0 in this subgroup. A combined effect size was obtained when vitamin D supplements were known to be used (RR1.17, 95% CI: 1.06-1.28, p < .05), I 2 = 72.5% in this subgroup.

| Circulating vitamin D levels in smokers and nonsmokers in continuous variable studies
People who are current smokers had lower circulating vitamin D levels than nonsmokers. Due to the different vitamin D units provided in the literature, the combined effect amount was expressed by SMD. We

| Publication bias
Begg's test was used to examine the publication bias of the litera-   (Ward & Klesges, 2001). The effect of smoking on bone metabolism is greater in the elderly than in the young, and the dose effect of smoking on bone has been shown in some literature (McCulloch et al., 1991;Ortego-Centeno et al., 1997;Välimäki et al., 1994). The same phenomenon has been seen in studies of smoking and circulating vitamin D levels (Rapuri et al., 2000). This is contrary to our observation that smoking has a greater negative effect on circulat-  Smith & Hansch, 2000).

| D ISCUSS I ON
There are several hypotheses about the specific mechanism by which smoking lowers vitamin D levels.
Skin synthesis is the main source of vitamin D in the human body, and vitamin D synthesis is affected by skin aging. Smoking (Ernster et al., 1995) and ultraviolet radiation from the sun (Pillai et al., 2005) are considered important factors that contribute to skin aging in humans. Photosynthesis of precholecalciferol in human skin, that is, where it binds to vitamin D-binding proteins (Holick et al., 1980;Kira et al., 2003;MacLaughlin et al., 1982). Increased skin aging significantly reduces the skin's ability to convert 7-DHC to precholecalciferol (Holick, 1995). López Hernández et al. (1995) found evidence of accelerated skin aging among smokers, finding that smoking, sun exposure, and age are important factors that increase skin wrinkling, and skin wrinkles are a significant sign of skin aging. Matrix metalloproteinases (MMPs) induce photoaging, and there is ample evidence to emphasize the effect of smoking on skin aging by activating MMPs (Holick, 1995;Lahmann et al., 2001). This is consistent with our findings that in people who do not take vitamin D supplements, the risk of vitamin D deficiency in smokers increases with age. Smoking can also affect the expression of cytokines and inflammatory mediators.
In the study of Tsutakawa et al. (2009), it was found that nicotine of never smokers after smoking cessation (Jorde et al., 2005). We conjectured that decreased vitamin D levels in smokers may be associated with decreased parathyroid hormone secretion, since PTH enhances the effect of renal 1-hydroxylase on vitamin D3 activation (Lips, 2001).
In addition, vitamin D3 synthesized in the skin and vitamin D2 extracted from food circulating in the body are first hydroxylated as they pass through the liver and then hydroxylated by the 1α hydroxylase CYP27B1 as they pass through the kidneys to form active vitamin D,1,25(OH)2D (Omdahl et al., 2002). Long-term smoking can lead to accumulation of heavy metals in the body. Tobacco plants contain lead and cadmium (Lugon-Moulin et al., 2006). The decrease in serum 1,25(OH)2D levels in smokers is related to the accumulation of cadmium in the kidneys (Brot et al., 1999;Kido et al., 1989).
The mechanism of this process may be renal damage caused by cadmium and lead poisoning. In fact, the increase in lead and cadmium in the body can damage renal tubular function and glomerular function (Cooper, 2006).
Smoking can also lead to disorders in the catabolism of vitamin  (Dong et al., 2012). When smokers' CYP24A1 genes were affected by smoke, lower vitamin D levels were linked to lung cancer risk.
Other studies have shown that tobacco alters the sense of smell and taste of food, leading to lower levels of vitamin D intake (Frye et al., 1990;Grunberg, 1982;Morabia et al., 2000;Redington, 1984).
It has been reported that other peripheral tissues, including respiratory epithelial cells, also contain 1a-hydroxylase and may serve as a source of 1,25VD3 (Hansdottir et al., 2008). The researchers found that acrolein and extract (CSE) from tobacco smoke significantly reduced the conversion of 25(OH)D3 to 1,25(OH)2D3 in airway epithelial cells by downregulating the expression of CYP27B1(CYP27B1 is the gene encoding the 25-hydroxyvitamin D3 1a hydroxylase responsible for converting 25VD3 to 1,25(OH)2D3) (Mulligan et al., 2014). One of the articles we included was a study of patients with chronic rhinosinusitis (Mulligan et al., 2014), showed that smokers, regardless of whether they had chronic rhinosinusitis, had lower circulating vitamin D levels than nonsmokers.
This study is the first meta-analysis of circulating vitamin D levels in smokers. There have been many independent studies, but this is the first time we have combined statistics on this subject.
Because diseases affecting vitamin D absorption or metabolism were excluded in advance, the results of this meta-analysis were not affected by adverse or beneficial effects of disease on vitamin D sources and metabolism. Confounding factors such as estrogen replacement therapy may partially mask the effect on women, and oral contraceptives or estrogen may prevent bone loss, increasing error F I G U R E 4 Vitamin D levels of smokers at different ages variability. So we also excluded menopausal women who were on hormone replacement therapy. However, for women of childbearing age using oral contraceptives, it is unclear whether these women of childbearing age using contraceptives have been excluded since no such information was provided in the article, which may be one of the sources of heterogeneity in the subgroup of this population.
In addition, several limitations of our study must be noted. Since vitamin D samples were taken from serum and plasma, this may lead to interstudy heterogeneity. However, meta-regression showed that the sample type was not the main source of interstudy heterogeneity, which had little impact on the results, and the few studies with the sample type of plasma were not suitable for further subgroup analysis. More research is needed to confirm whether vitamin D levels in smokers are reduced in serum and plasma, respectively.
Our study did not analyze the effect of exercise on VD levels. Some studies have suggested that physical activity has an effect on vitamin D levels, but studies until now have not established a direct relationship between physical activity and VD (Bell et al., 1988;Jacques et al., 1997;Klausen et al., 1993;Lucas et al., 2005;Ma Moun et al., 2005;Scragg et al., 1992Scragg et al., , 1995, or that sun exposure is a confounding factor, and there are conflicting results regarding the effect of physical activity on 25(OH)D metabolism. It was found that there was no significant relationship between physical activity and 25(OH)D after controlling for sunshine duration. In addition, since the included studies came from different countries and regions, and the included studies were mainly retrospective studies, the time, season, and latitude of blood sample extraction of subjects could not be unified, and these confounding factors might affect the results and produce heterogeneity among studies. However, in a single study, the subjects had the same season, latitude, and blood sampling time, and there were relevant measures to exclude the influence of confounding factors, so the study results were less affected by these factors. The association between smoking and circulating vitamin D levels is unlikely to be accidental, as estimates come from a large mixed sample (more than 11,340 participants) and these studies are heterogeneous in terms of several participant and methodological characteristics. Due to the exclusion of some diseases that affect vitamin D absorption or metabolism, there is inevitably a selection bias, and since all studies rely on self-reporting, there may be bias in determining smoking exposure.
Smoking-related vitamin D deficiency (VDD) may pose an even greater public health problem in our aging society. Given the cumulative and dose-dependent effects of smoking on VDD and its association with the vitamin D-PTH axis, an increase in smoking among these young populations is likely to lead to a significant increase in the future public health burden of osteoporosis or other cardiovascular, pulmonary, infectious, and immune-related diseases. Smoking has a significant adverse effect on circulating vitamin D levels, but vitamin D levels can be restored to nonsmoker levels after quitting smoking. In view of the adverse effects of smoking on vitamin D-PTH axis health and its associated public effects on bone metabolism and other systemic health, the prevention of VDD and smoking cessation must be emphasized.

ACK N OWLED G M ENTS
Studies included in this paper are all published human studies. None of the authors participated in any clinical study and therefore did not sign informed consent.
F I G U R E 5 Endocrine system of vitamin D

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest. Writing-review & editing (equal).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available in this article. No additional supplementary material is provided.