The electronic search retrieved 12,430 studies. After removal of duplicates, screening of abstract & titles of all studies, 52 of them were considered for full-text review. Full-text scrutiny resulted in eleven studies for the final quantitative data synthesis. Total number of studies excluded and the reasons are described in the PRISMA flowchart (Figure 1).
Description of studies: Details of the individual studies such as sample size, mean age of participants, duration of occupational exposure, type of industry (responsible for occupational Pb exposure), simultaneous exposure to other heavy metals, and the outcome parameters reported are described in table 1. Majority of the included studies reported sample mean of the outcome variables of control groups within the recommended normal values (i.e., Serum calcium: 8.2 - 10.2 mg/dl, ionized Calcium: 4.60 – 5.08 mg/dl, serum PTH: 10–65 pg/ml and vitamin D: 25 - 80ng/ml) (24) suggesting the control group as apparently normal in terms of the outcome parameters (24). However, the mean Vitamin D of the control group from Dongre et al. 2013 was lower than the recommended values (7). None of the studies reported calcitonin levels, while a single study reported bone turnover/ resorption markers (5). Briefly, all studies recruited 100% male participants, except Batra et al. 2014 (4), where 11.3% were female participants. The sources of occupational Pb exposure among the exposed group vaired across Pb battery and its allied industries, Pb smelting process, e-waste recycling process, welding, and paint industries. The duration of occupational Pb exposure was not available for a majority of the studies, while few reported the exposure duration ranging from 5 to 34 years. The studies were restricted to Asian (India, China, and Bangladesh), European (Croatia, Austria, Turkey, and Israel), and African (Nigeria) countries, while no studies from Australia or American (both Latin and North) countries were available.
Risk of Bias assessment: The risk of bias during participant selection, ascertainment of exposure, and outcome parameters were evaluated using Newcastl Ottawa scale (23) (Table 2). Notable risk of bias in the participant selection included convenient sampling, inadequate representativeness of the exposed and control participants (in view of absence of precise records on exposure), non-description of precise source of control participants in some of the studies (7, 9, 25) and incomplete description of control participants (i.e., occupation, confirmation of their non-exposure) (4, 7, 9, 25). Additionally, as the interviewer(s) were non-blinded to the participant status (as either exposed or unexposed), the exposure assessment could have been potentially biased.. Therefore, none of the included studies was comparable to that of a well-conducted RCT.
Meta-analysis results of individual outcome parameters
Blood lead levels: Current systematic search identified eleven studies that reported BLL among occupationally Pb exposed & control workers (3-9, 12, 13, 25, 26). All identified studies reported significantly (p < 0.05) higher BLL among occupationally exposed workers as compared to control subjects. The results were congruent with the current pooled quantitative analysis, wherein individuals with occupational exposure to Pb had a mean BLL of 36.13 µg/dl (95% CI, 25.88 to 46.38, I2=99.75%, p<0.05) higher than controls (Figure 2). Subgroup analysis of studies involving Pb exposed participants with history of co- exposure to other heavy metal(s) did not alter the results (Supplement Figure 1). Explorative uni-variate meta-regression with participants’ mean age as the moderator did not significantly reduce the heterogeneity / alter results. Further, the assymteric funnel plot suggests publication bias (p=0.425), while the contour enhanced funnel plot suggests the presence of other biases as well. (Supplement Figure2).
Urinary lead levels: Two studies reported Urine Pb levels among occupationally Pb exposed & unexposed workers. The pooled mean difference between the groups was 10.89 µg/dl (95% CI, -1.01 to 22.79, I2=96.71). Subgroup, meta-regression and funnel plot analysis were not possible in view of fewer studies.
Serum calcium: Ten of the included studies reported serum calcium levels, while seven of them observed significantly lower calcium levels among occupationally Pb exposed group compared to control group (3, 4, 6-9, 12, 13, 25, 26). The Pb exposed group exhibited -0.72 mg/dl (95% CI -0.36 to -1.07, I2=95.12%) lower calcium levels as compared to the control group. (Figure 3A). Subgroup analysis of studies involving Pb exposed participants with a history of co-exposure to other heavy metal(s), did not significantly reduce the heterogeneity, however reduced the magnitude of the difference between the duo (-0.39 mg/dl, 95% CI -0.64 to -0.15, I2=72.04%) (3, 6, 8, 12, 13) (Supplement Figure 3). Meta-regression analysis with participant’s mean age as moderator, did not significantly change the results. Further, the assymetric funnel plot was suggestive of publication bias (p=0.779) and the contour enhanced funnel plot suggested the presence of other biases as well (Supplement Figure 4).
Serum ionized Calcium: Four of the included studies reported ionized serum calcium levels. All but one study reported statistically significant lower ionized calcium levels among occupationally Pb exposed group compared to control group (3, 7, 8, 26). The results were consistent with that of the serum calcium levels, i.e., the pooled mean difference revealed lower ionized calcium levels among occupationally Pb exposed group as compared to the control group (-1.20 mg/dl with 95% CI -2.38 to -0.02 and I2=99.76%) (Figure 3B). In view of limited studies, subgroup, meta-regression and funnel plot analysis were not executed.
Parathyroid hormone (PTH) (Paratharmone): Five included studies reported parathormone levels; all but one observed lower serum PTH levels among occupationally Pb exposed as compared to the control group (4, 5, 7, 8, 26). Occupationally Pb exposed group exhibited mean 37.97pg/ml (95% CI -14.36 to 90.29 and I2=99.80%) lower parathormone levels as compared the control group, with high levels of heterogeneity (Figure 4A). However, the results were statistically not significant. Subgroup and meta-regression analyses weren’t taken up, given fewer available studies reporting parathormone levels.
Vitamin D: Five studies reported Vitamin D levels, and all but one reported significantly lower Vitamin D levels among occupationally Pb exposed group compared to control group (4, 5, 7, 8, 25, 26). The pooled mean difference between the duo was -12.26 (95% CI -25.36 to 0.84, I2=99.10%) ng/dl with high heterogeneity (Figure 4B). Although the results indicate the point estimate of the effect measure for Vitamin D was lower among Pb exposed, the result was statistically not significant. Due to the limited number of studies, subgroup and meta-regression analysis were not executed.
Calcitriol: Two of the included studies reported differences in calcitriol levels (active form of vitamin D), with one study reporting higher (8) while the the other (5) reported contrastingly lower serum calcitriol levels among occupationally Pb exposed group compared to control group. The pooled mean difference between the duo was -5.98 (95% CI -8.40 to 20.35, I2=91.20%) pg/dl with high heterogeneity. The results are suggestive of lower calcitriol among Pb-exposed as compared to the control group; however, the result was statistically not significant. Due to a limited number of studies, subgroup and meta-regression analysis were not carried out.
Other outcome parameters: Akbar et al. (2014) alone compared serum osteocalcin (bone resorption marker), urinary deoxypyridinoline, and urinary hydroxyproline values between the two groups(5). Both bone turnover markers in urine (deoxypyridinoline and hydroxyproline) and bone resorption markers (osteocalcin) were higher among the Pb exposed as compared to the non-exposed control; however, the results of urinary bone turnover markers (deoxypyridinoline and hydroxyproline) were statistically significant (p< 0.05).