The relationship of polluted air and drinking water sources with the prevalence of systemic lupus erythematosus: a provincial population-based study

Abstract

Environmental exposures interact with genetic factors has been thought to influence susceptibility of systemic lupus erythematosus (SLE) development. To evaluate the effects of environmental exposures on SLE, we conducted a population-based cohort study across Jiangsu Province, China, to examine the associations between the living environment including air and water pollution, population density, economic income level, etc. and the prevalence and mortality of hospitalized SLE (h-SLE) patients. A total of 2231 h-SLE patients were retrieved from a longitudinal SLE database collected by the Jiangsu Lupus Collaborative Group from 1999 to 2009. The results showed that: It existed regional differences on the prevalence of h-SLE patients in 96 administrative districts; The distribution of NO₂ air concentration monitored by atmospheric remote sensors showed that three of the ultra-high-prevalence districts were located in the concentrated chemical industry emission area; h-SLE patient prevalence was positively correlated with the excessive levels of nitrogen in drinking water; The positive ratio of pericarditis and proteinuria was positively correlated with the prevalence of h-SLE patients and pollution not only induced a high h-SLE patient prevalence but also a higher mortality rate, which might be attributed to NOx pollution in the air and drinking water. In summary, our data suggested that NOx in air and drinking water may be one of the important predispositions of SLE, especially for patients with renal involvement.

Figure 1

Environmental exposures might be triggering the development of h-SLE. (A) Population density and distribution of h-SLE patients in Jiangsu Province (prepared by WQ in ArcMap 10.2, https://www.esri.com/zh-cn/arcgis/products/arcgis-pro/resources). (B) Significant difference analysis of the h-SLE patient prevalence of population density (person/km²) group. Level 0:230–560; Level 1: 560–700; Level 2: 700–1000; Level 3: 1000–2000; Level 4:2000–20,100. (C) Normalized relationship between the prevalence of h-SLE patients and the number of patient cases in Jiangsu Province, including 10 in the south of the Yangtze River and 5 in the north of the Yangtze River. The prevalence of h-SLE patients conformed to the normal distribution after BOX-COX conversion (λ = 0.053) and Shapiro–Wilk Test (P value: 0.186 > 0.05). Statistical analysis was performed with one-way ANOVA. *P < 0.05, ns not significant.

Figure 2

Association between the concentrated discharge of exhaust gas and the development of h-SLE. (A) NO₂ concentration distribution in the air of Jiangsu Province in August 2005 (prepared by WQ in ArcMap 10.2, https://www.esri.com/zh-cn/arcgis/products/arcgis-pro/resources). (B) Normalized graph of h-SLE patient prevalence and air NO₂ concentration in August 2005 in Jiangsu Province from 1999 to 2009. (C) Significant difference analysis of the prevalence of h-SLE patients in NO₂ concentration group. Level 0:250–420; Level 1: 420–585; Level 2: 585–750; Level 3: 750–910; Level 4:910–1075. The prevalence of h-SLE patients conformed to the normal distribution after BOX-COX conversion (λ = 0.053) and Shapiro–Wilk Test (P value: 0.186 > 0.05). Statistical analysis was performed with one-way ANOVA. **P < 0.01, ***P < 0.001.

Figure 3

Relationship between the distribution of h-SLE patients and drinking water sources. (A) Distribution of groundwater chemistry types (2000s) and h-SLE patients in northern Jiangsu (prepared by WQ in Microsoft Excel 2019, referring to the cation distribution map of hydrochemical facies for the phreatic aquifer in the 2000s in the paper, https://doi.org/10.1007/s12665-015-4575-4). (B) Normalized relationship between the prevalence of h-SLE patients and the cases of h-SLE in different drinking water sources in northern Jiangsu. (C) Prevalence distribution of h-SLE patients with different chemical drinking water sources in Northern Jiangsu. (D) Significant difference analysis of the prevalence of h-SLE patients in water chemistry type. The prevalence of h-SLE patients conformed to the normal distribution after BOX-COX conversion (λ = 0.053) and Shapiro–Wilk Test (P value: 0.186 > 0.05). Statistical analysis was performed with one-way ANOVA. **P < 0.01, ***P < 0.001.

Figure 4

Proportion of clinical manifestation indicators of h-SLE patients with 5 prevalence levels in Jiangnan and Jiangbei. (A) Proteinuria. (B) Pericarditis. (C) Serositis. (D) Anti-dsDNA antibodies. The patient prevalence (cases per 100,000 people) was divided into 5 levels: level 0: prevalence < 2 (432 cases); level 1: 2 < prevalence < 3(443 cases); level 2: 3 < prevalence < 5(496 cases); level 3: 5 < prevalence < 10(503 cases); level 4: prevalence > 10(357 cases).