Areas with high soil percolation by herbicides have higher incidence of low-grade non-Hodgkin lymphomas

Abstract

Dear Editor, The incidence rate of non-Hodgkin lymphomas (NHL) doubled over the past two decades in both the US and most westernized countries [1]. Etiology remains largely unknown, but herbicide exposure (especially to phenoxyacids) has been considered a risk factor for indolent NHLs since the 1990s [2–4]. Although the fraction of agricultural workers in westernized countries is small and decreasing over time, herbicide exposure of the general population has been increasing [1]. Despite general population exposures may be lower than those in occupational settings, a relative risk as small as 1.2 could explain 15% of the current NHL risk, assuming that over 90% of the general population is exposed [5]. Based on preliminary unpublished observations on the geographical clustering of mortality from NHLs in the province of Pisa (western Tuscany, Italy), we ran an ecological study on the distribution of NHL cases in areas with high exposure to herbicides. We defined exposure to herbicides using an index calculated as follows: the 2,450,000 ha of the province of Pisa were subdivided into geofunctional areas marking autonomous systems characterized by very low intermigration, closed water system, and homogeneous soil composition [6]. Data on spraying of the 13 most used herbicides in the province of Pisa in 1988–1990 were extrapolated from the recordings of the Italian Health Ministry [7], as reported in Table 1. A modification of Mackay and Peterson’s model of fugacity was used to calculate herbicide percolation into the soil [8–11]. Land areas treated with herbicides were estimated on the basis of the 1980 Third General Census of Italian Agriculture [12], and herbicide use was assumed homogeneous throughout the whole province. The theoretical concentration of each active ingredient expected in percolation water in 1988– 1990 was estimated as micrograms per gram of soil. Assuming that all active ingredients had the same lymphomagenic impact, a cumulative concentration index (Cw) was used (millimeter per cubic meter). At first, we retrospectively investigated the correlation between the herbicide soil percolation index in 1988–1990 and mortality from NHLs in the province of Pisa in the period 1987–1992. Overall, 370 deaths from NHLs occurred in the study period. Mortality rates were calculated for each geofunctional area using data obtained from the Tuscan Mortality Registry. Soil percolation by herbicides and mortality rates of NHLs were then correlated using Spearman’s rank test (ρ), showing a moderate correlation (ρ=0.355; p<0.05). R. Fazzi (*) :D. Focosi :M. Petrini Division of Hematology, Department of Oncology, Transplants and Advances in Medicine, University of Pisa, Pisa, Italy e-mail: r.fazzi@hotmail.it