We searched Pubmed/Medline, and Embase, for studies reporting the association between silicosis and pulmonary TB published up to 6 September 2022. There was no specific limitation for the starting date of search. The following search strategy was used: (“silicosis”[MeSH Terms] OR (“silicosis”[Title/Abstract] OR “silicoses”[Title/Abstract] OR “pneumoconiosis”[Title/Abstract] OR “pneumoconioses”[Title/Abstract]) OR “pneumoconiosis”[MeSH Terms]) AND (“tuberculosis”[MeSH Terms] OR “tuberculosis”[Title/Abstract]).

The records found through database searching were merged, and the duplicates were removed using EndNote X8 (Thomson Reuters, New York, NY, USA). Two reviewers (MA and BM) independently conducted title and abstract screening to exclude irrelevant studies. Two other reviewers (MA and BM) retrieved and assessed the full text of potentially eligible records. Inclusion criteria were: populations exposed to silica and silicosis patients; studies reporting comparative effect estimates, specifically case-control or cohort studies reporting relative risks or odds ratios across groups with and without silicosis (including different grades). Only English-medium papers were considered.

Studies without comparison between those with and without the disease, studies focused on latent TB infection, cross-sectional and mortality studies, and autopsy-based studies were excluded. Review articles, duplicate publications, conference abstracts, editorials, postmortem studies, and those in which the full text of papers was not found were also excluded.

The data extraction form was designed by PJ and the following data were extracted by BD, BM, MA, and FK: first author's name, type of study, year of publication, study period, the country/ies where the study was performed, study population, age, gender, occupation, mentioned confounding factors, diagnostic approach for silicosis and tuberculosis, duration and grade of silicosis, follow-up duration and technique, the estimated effect measures of silicosis and TB, and the total number of following items: TB population, silicosis population, TB population with and without silicosis, silicosis without TB, and the total number of non-silicosis individuals.

The Newcastle-Ottawa Scale (NOS) for observational studies was used to perform a study quality assessment.20 The NOS scale assesses the risk of bias in observational studies with three domains: 1) selection of participants, 2) comparability, and 3) outcomes. A study could be given a maximum of one point for each numbered item within the selection and outcome areas, and a maximum of two points allocated for comparability. The low, moderate, and high-quality studies were given scores of 0–3, 4–6, and 7–9, respectively.

Statistical analyses were conducted using Comprehensive Meta-Analysis software, version 2.0 (Biostat Inc., Englewood, NJ, USA). Pooled estimates and 95% confidence intervals (CI) for the relative risk of TB in silicosis individuals compared to those without silicosis ones were evaluated using the random effects model due to the estimated heterogeneity of the true effect sizes. The between-study heterogeneity was assessed by Cochran's Q and the I2 statistic. Publication bias was evaluated using Begg's tests (P < 0.05 was indicative of a publication bias) and forest plot.

Based on the Newcastle-Ottawa Scale (NOS), which was used to assess the quality of the cohort studies, the mean (standard deviation [SD]) NOS score was 7.5 (1.29), which is suggestive of a high methodological quality and a low risk of bias of the included studies (Table 1).

Five articles described cohort studies21,23-25,28 and two case-control studies.26,27 Publication dates ranged from 1984 to 2013. Most of the studies were conducted in South Africa (n= 4, 57.14%),21-23,26 one (14.28%) in Sweden,24 one (14.28%) in Hong Kong, 25 and one (14.28%) in Iran. 27 The total population was 5,884 persons. The male gender was dominant (n=5,333, 90.63%), and the mean age was 47.70 years (Table 2 and Table 3).

The diagnosis of silicosis was mainly based on chest X-ray findings; however, occupational history and physical examination were helpful. The most prevalent methods for tuberculosis diagnosis were sputum smear microscopy and culture (Table 2).

The follow-up duration reported in articles ranged from 3 to 20 years. The most common follow-up techniques were clinical evaluation, radiologic imaging, sputum examination, pulmonary function tests, and questionnaires (Table 2).

All studied populations were silicosis-confirmed patients, except one study,27 which stratified the population into silicosis and silica-exposed individuals without silicosis.

Disease severity classification was based on the level of exposure to silica27 and X-ray images were classified according to International Labor Organization (ILO) standards21,22,26 (Table 3).

All studies established the substantial effect of silicosis on tuberculosis development, with the estimated effect measures ranging from 1.86 to 8.623,25 (Table 3).

Several confounding factors were mentioned in our reviewed studies, including occupation, age, gender, height, HIV, smoking, duration of employment as an indicator for the time of silicosis or silica exposure, the severity of silicosis, co-existence of other pulmonary diseases, progressive massive fibrosis, baseline lung function, using anti-TB medications before silicosis diagnosis, and level of education (Table 3).

Since no data for confounding factors were available, we could not perform any stratification analysis.

One study evaluated tuberculosis risk factors in a silicosis cohort in Hong Kong and identified four TB risk factors as follows: no anti-tuberculosis treatment before the date of silicosis diagnosis (RR 4.51, 95%CI 2.46–8.24), progressive massive fibrosis (RR 3.78, 95%CI 2.25–6.36), small opacities exceeding 1.5 mm (RR 2.17, 95%CI 1.38–3.42), and caisson work (RR 1.56, 95%CI 1.01–2.41).25 Almost all samples had a history of mine work or dusty jobs. One study reported that 48.4% of their silicosis population had worked in the mines for 11 to 20 years.22

According to two studies22,27 older silicosis patients were mentioned as having a higher risk for TB. One study demonstrated that TB was most common in the age group of 36 to 45 years, and the age groups that were least susceptible to TB were those aged ≤35 years (2.6%) and 66+ years (7.4%).22

The frequency of TB increased along with the duration and intensity of silicosis/silica exposure. One case-control study showed that individuals with >15 years of moderate to severe silica exposure had the highest TB rate.27 Another study reported that the largest group of subjects with TB had a mining history of 11-15 years.22 Moreover, one study found that probable, early, and high-grade silicosis, as well as the period of employment, were significant risk factors for TB.26

No significant publication bias was found in the selected studies (Begg's test P>0.05 and forest plot qualitative analysis) (Fig. 2). Pulmonary silicosis increased the risk of pulmonary tuberculosis, with a pooled risk ratio of 1.35 (95%CI 1.18-1.53) with an I 2 of 94.30% (Fig. 3). Considering the large outlier effect size of Westerholm et al.,24 we repeated the analysis without this study; however, the result did not significantly differ (RR 1.85, 95%CI 0.45-7.60, I 2: 94.41%) compared with the previous estimate.

Fig. 2.

Forest plot indicating no publication bias of the selected studies.

(0.13MB).

Fig. 3.

Forest plot of the cohort studies indicating the impact of silicosis on the risk of tuberculosis.

(0.06MB).

The final result of our study is along with the Ehrlich et al. meta-analysis published in 2021 which demonstrated an increased risk of tuberculosis in silicosis patients.32 However, our study has some strengths and is different in some ways. We included studies focused purely on the relationship between silicosis and silica dust exposure and tuberculosis. Postmortem studies or studies in which the sample population had specific comorbidities (e.g., end-stage renal disease) were excluded to lessen the potential effect of the confounding factors. Also, we excluded the articles in which the full text was not found. Against this, we included two other studies that not previously included.22,23 Our calculated pooled risk ratio (1.35, 95%CI 1.18-1.53) was a little lower than Ehrlich et al. (RR 1.92, CI 95% 1.3-,2.73) but in line with its direction and statistical significance.

The major limitation of our study is its originality, as the data on the association between silicosis and tuberculosis has been known and ascertained for some time and also reported by international agencies NIOSH (2003), OSHA (2004) regarding the need for health surveillance for tuberculosis in silicosis patients and people at risk of silicosis. However, that association has resulted in very high rates of tuberculosis in countries with poor tuberculosis and silica exposure control and an update of data relating to scientific evidence can still have value.

Some other limitations of our study are associated with the studies we included and are as follows: First, there have not been any recent cohort and case-control studies on the topic and the most recent one was published in 2013. Further new investigations with a large number of populations are required to elucidate the issue. Second, the male gender was dominant in almost all of our included studies and only one study.27 had an equal male-to-female proportion. This could be because mine workers are mainly males but females can also be affected by working and living near workplaces.39 We suggest future studies to consider this issue. Third, several confounding factors were mentioned in our study which despite our inability to subgroup analysis, due to insufficient and inconsistent data, can inspire future research in this field.