Spatial association of surface water quality and human cancer in China

Surface water quality deterioration poses significant challenges to human health, with carcinogenic effects being a key concern. While individual-level studies have linked drinking water quality to certain cancers, population-level analyses examining multiple pollutants and cancers are scarce, especially in China, a country with a high cancer burden. This study leverages China's extensive water quality monitoring network and cancer registries to investigate the spatial associations between various water pollutants and different cancer types at a national scale. The goal was to provide evidence on the systematic impact of surface water quality on cancer incidence, identify key pollutants in different river basins, and forecast the future water quality-related cancer burden. The study addresses the limitations of individual-level studies which struggle to capture complex exposure routes and the slow pace of evidence generation regarding carcinogens.

Previous research has primarily focused on individual-level studies linking drinking water sources and disinfection by-products to cancers such as esophageal, gastric, colorectal, and renal cancers. However, these studies often face challenges in determining individual exposure levels due to the complexity of exposure pathways (ingestion, dermal absorption, inhalation). Population-level ecological studies offer a more efficient approach. A notable example is Yang et al.'s work in the Huaihe River Basin, which established a link between water pollution frequency and digestive cancer mortality. The availability of comprehensive national-level data on water quality and cancer incidence in China provides a unique opportunity to conduct a large-scale ecological study to address this knowledge gap. However, the variation in geographic backgrounds, pollutant levels, and combinations across different water bodies, particularly the coexistence of multiple pollutants, necessitates a sophisticated analytical framework.

The study integrated nationwide data from China's National Surface Water Environmental Quality Monitoring Network and the National Cancer Registry System. Data from 2021 were used, encompassing 3632 water monitoring sections and 486 cancer registries, covering a population of 380 million. A spatial matching design was employed, using a 30km radius buffer around each water monitoring section to link water quality data to cancer incidence data. The study considered nine river basins in mainland China, with the Yangtze River Basin further subdivided into upstream, midstream, and downstream sections to account for environmental and health variations. Twenty-one basic pollution indicators were analyzed using the Environmental Quality Standard for Surface Water (EQSSW, 2002) and a stricter criterion based on the 75th percentile of each indicator's values. Spatial autocorrelation was assessed using global and local Moran's I. A novel design, "Cluster Analysis of Multi-pollutants in Space (CAMS)", was developed to grade water sections based on the number of pollutants exceeding the thresholds in high-high clusters or high-low outliers, accounting for the co-occurrence of pollutants and spatial interactions. Negative binomial regression was used to analyze the association between pollution indicators and cancer incidence, while an extreme gradient boosting algorithm with SHAP values was used to identify key pollution indicators in different river basins. Population attributable fractions (PAFs) were calculated to estimate the proportion of cancers attributable to poor water quality. Future projections were made based on linear regression of cancer incidence trends and assumptions about water quality.

The study revealed a dose-response relationship between the CAMS grade (reflecting the number of pollutants exceeding thresholds) and cancer incidence for several cancer types, including stomach, pancreatic, and kidney cancers. Higher CAMS grades were associated with increased incidence of these cancers. While some cancers showed no significant increase at lower CAMS grades, a joint effect became apparent at higher grades. The overall incidence of the 11 selected cancers increased by 12.7% in Grade 3 areas compared to Grade 0. Significant spatial autocorrelation was found for cancer-related pollution indicators. The Huaihe, Haihe, Songhua & Liaohe, and downstream Yangtze River basins, generally characterized by poorer water quality, showed high cancer incidence rates. Specific pollutants, such as permanganate index, petroleum, chemical oxygen demand, and fluoride, were identified as key drivers of cancer incidence in different basins. Notably, the Yangtze River Basin showed a gradient of increasing cancer incidence and decreasing water quality from upstream to downstream, reflecting China's east-west development disparity. The study estimated that 5% of all new cancer cases in the studied areas are attributable to poor surface water quality, with higher proportions observed in certain basins. Projections suggest a potential increase of 7.8% in new cancer cases by 2030, with a substantial portion attributable to poor water quality.

The findings demonstrate a strong nationwide spatial association between surface water quality and multiple cancer types in China, addressing a significant knowledge gap. The dose-response relationship strengthens the evidence for the widespread contribution of surface water quality to cancer. The CAMS design, by incorporating both the quality and spatial variation of multiple pollutants, provides a valuable tool for assessing complex environmental health risks, surpassing the limitations of analyzing individual pollutants. The variation in cancer incidence and key pollutants across river basins highlights the need for region-specific interventions. The study underscores the challenges in eliminating the impact of past pollution, even with substantial government investment in environmental restoration. The findings also highlight the interconnection of environmental and socioeconomic factors influencing cancer burden, such as agricultural practices (fertilizer use) and industrial discharge.

This study provides compelling evidence of the significant association between surface water quality and cancer incidence in China. The CAMS design offers a novel approach to assess complex environmental health risks, highlighting the importance of considering multiple pollutant interactions. The findings underscore the need for targeted interventions, stricter water quality standards, and sustainable development practices to mitigate the impact of water pollution on cancer burden. Future research could refine the CAMS methodology, explore the mechanisms underlying pollutant interactions, and investigate the long-term effects of water quality improvements on cancer rates.

The study's ecological design limits causal inference. Confounding factors other than water quality may influence cancer incidence. The reliance on annual average water quality data might mask short-term variations in pollution levels. The 30km buffer radius is an assumption, and the actual influence of water pollution might extend beyond this distance. The projections are based on certain assumptions regarding future water quality and population growth. The study focused on specific cancers with relatively high incidence rates.