Datasets and drivers of catchment-scale river pollution: spatial correlations with district-level cancer incidence in Moldova

Background. Moldova has a legacy of intensive pesticide use associated with historical agricultural practices and obsolete pesticide storage, with documented contamination affecting river-associated water systems in major transboundary basins such as the Dniester and Prut (Sapozhnikova et al., 2005; Ivanova et al., 2021). Widespread pesticide application combined with inadequate waste management during the Soviet and post-Soviet periods resulted in long-term contamination structured along river networks and basin-scale hydrological organization. This study examines whether spatial and temporal patterns of cancer incidence and prevalence align with hydrologically structured downstream exposure rather than with conventional population-level epidemiological risk proxies, with emphasis on long-latency environmental signals.

Methods. A district-level ecological analysis was conducted using national cancer incidence and prevalence rates per 100,000 population aggregated across three periods, 2008 to 2012, 2013 to 2017, and 2018 to 2022. District-level mean rates were calculated for each period to avoid pseudoreplication. Hydrological vulnerability was defined by integrating official contamination site inventories with peer-reviewed river pollution data and downstream river basin organization. QGIS was used to link river catchments with administrative district boundaries. Seventeen vulnerable districts were compared with eighteen non-vulnerable districts. Group differences were assessed using Welch's t-tests and Mann-Whitney U tests, with effect sizes quantified using Cohen’s d. Correlations with alcoholism, liver hepatitis and cirrhosis, respiratory disease as a proxy for smoking, and population age structure were examined. Analyses were performed using R.

Results. From 2008 to 2017, vulnerable downstream districts showed lower cancer incidence and prevalence, with no statistically significant differences. In contrast, from 2018 to 2022, vulnerable districts exhibited higher cancer incidence by 14.2 percent and higher cancer prevalence by 27.6 percent. Both outcomes were statistically significant with p < 0.05 and showed the largest effect sizes, with incidence Cohen’s d equal to 0.93 and prevalence Cohen’s d equal to 1.59. Temporal trends were highly parallel between groups, with r equal to 0.969 for incidence and 0.999 for prevalence, indicating divergence driven by level rather than trend shape. Risk proxies showed weak and non-significant correlations.

Conclusions. The post-2018 emergence of excess cancer burden in downstream districts represents a delayed spatial pattern consistent with long-latency environmentally mediated exposure structured by river basin hydrology rather than by short-term population-level risk factors. This exploratory study demonstrates the value of hydrology-informed public health analysis for detecting environmentally structured disease patterns and identifying priority downstream corridors for targeted monitoring of river-associated water systems in transboundary basins.