State of health of populations residing in geothermal areas of Tuscany

OBJECTIVE The limited scientific knowledge on relationship between exposure and health effects in relation to geothermal activity motivated an epidemiologic investigation in Tuscan geothermal area. The study aims to describe the health status of populations living in Tuscany municipalities where concessions for exploitation of geothermal resources were granted. DESIGN This is an ecological study, so it is not useful to produce evidence to sustain a judgment on the cause-effect link. The major limits of this type of study are the use of the residence at municipal level as a proxy of exposure to both environmental and socioeconomic factors and the use of aggregated data of health outcomes that can lead to the well-known ecological fallacy. SETTING AND PARTICIPANTS Sixteen municipalities were included in the study area: eight are part of the so-called “traditional” geothermal area, defined as Northern Geothermal Area (NGA) and eight located in the Amiata Mountain defined as Southern Geothermal Area (SGA). In 2000-2006, the average resident population in the overall area was approximately 43,000 inhabitants. Thirty one geothermal power plants were active, with a production capacity of 811 MW, 5 of them with 88 MW located in the SGA. Statistical analyses on the entire geothermal area, NGA and SGA subareas, and the sixteen municipalities were performed. MAIN OUTCOME MEASURES MORTALITY DATA were obtained from Tuscany Regional Mortality Registry for the 1971-2006 period, analysing 60 causes of death, of interest for population health status or consistent with “Project SENTIERI” criteria. HOSPITAL DISCHARGE RECORDS of residents in Tuscany Region in 2004-2006, anywhere admitted to hospital, were analyzed considering only the main diagnosis, excluding repeated admissions for the same cause. The causes taken into account are the same analysed for mortality were considered. Age-standardized mortality rates (TSDM) and the temporal trends of TSDM for four periods (1971-1979, 1980- 1989, 1990-1999, 2000-2006) were computed. Age-standardized mortality/hospitalization ratios (SMR/SHR), with and without adjustment for the deprivation index based on 2001 census data, were calculated: mortality in the years 2000-2006 and hospitalization in 2004-2006. The expected number of events were computed using rates of residents in neighbouring municipalities (municipalities included in 50 km radius circle centred on the study area). Bayesian estimates of mortality/hospitalization ratios (BMR/BHR) at municipal level only and relating maps of the Bayesian risk estimators were elaborated. CONGENITAL MALFORMATIONS (MC) were analysed using data from Tuscan Registry of Birth Defect in 1992-2006 period, relative to outcomes of pregnancies in women resident in the municipalities of study area, wherever the birth or termination of pregnancy occurred. The ratio between observed and expected cases (O/A), with expected defined according to regional rate, were calculated and O/A Bayesian estimates (BMR) are showed only at municipal level . The low weight and the males/females ratio at birth were analysed using data from Tuscany Birth Certificates, covering period 2001-2007, excluding biths occurred in facilities outside Tuscany Region. For Low birth weight (<2,500 grams), very low birth weight (<1,500 grams), low birth weight in women with normal gestational age or greater than 36 weeks, gestational age less than 36 weeks, and the frequency of males, the observed/expected ratio was calculated, with the expected number defined according to regional rate. RESULTS ENVIRONMENTAL BACKGROUND High levels of arsenic in drinking water distribution emerges as a critical element, so that several municipalities resorted to granting exemptions for the parameters laid down by the Legislative Decree in force (DLgs 31/01). However, during the final phase of the study, new blast systems activated in the SGA decreased the arsenic levels in the water supply, reaching values not requiring derogations, which, instead, are still effective in some NGA municipalities. Air quality data, from Tuscany Regional Agency for Environmental Protection-ARPAT, show that geothermal activities are able to affect air quality, especially with hydrogen sulphide in NGA, and hydrogen sulphide and mercury in SGA. A significant contribution to the presence of mercury in air is due to previous metallurgical sites. Although mercury levels are below WHO guideline values, in SGA nearby Siena, values were significantly higher than in other geothermal areas, because of power plant PC2 (turned off in July 2011) in Piancastagnaio municipality. The hydrogen sulphide concentration levels were generally lower than WHO reference values, with occasional excesses over guideline value for health protection (150 μg/m3 as average of the 24 hours). Olfactory pollution was more critic with values exceeding 7-10 μg/m3 range even in areas without geothermal plants. POPULATION’S HEALTH STATUS This study evaluated health status of resident population in geothermal areas analysing geographic and temporal distribution of mortality, hospitalization and reproductive health outcomes (congenital malformations, low birth weight, sex ratio among newborns). In both geothermal areas mortality rates steadily declined from 1971 to 2006, in males and females, in line with the regional trends. In 2000-2006 period, in the overall geothermal area a significant mortality excess was observed for all causes among males (2,312 deaths, 2,146 expected), but not among females, using as referenc residents in neighbouring municipalities. The mortality excess among males was more evident for infectious diseases (25 deaths, 10 expected), especially tuberculosis (8 deaths, 2 expected), for respiratory diseases (218 deaths, 170 expected), in particular pneumoconiosis, including deaths from silicosis (51 deaths, 14 expected), and for nervous system diseases (72 deaths, 56 expected). Among females significant mortality excess for liver cirrhosis (35 deaths, 25 expected) emerged, while mortality from cardiovascular diseases and ischemic heart diseases were significantly lower than expected. In the NGA, mortality among men was lower than expected for all cancers (-15%), in particular for lung cancer (- 25%), while values significantly in excesses were observed for infectious diseases (11 observed, 4 expected) and respiratory diseases (90 observed, 73 expected), expecially pneumoconiosis (20 observed, 6 expected). Among females, significant mortality excesses for ovarian cancer (17 observed, 10 expected) and for circulatory disorders of brain (170 observed, 140 expected) resulted. In the SGA, mortality was more critical, accounting for majority of the excesses detected in overall Geothermal Area. In fact, only infectious diseases and pneumoconiosis were detected in excess in both the geothermal areas. In the SGA, excess of general mortality among males (1,431 deaths; 1,245 expected) but not among females emerged. Even for all cancers, an excess among males (505 deaths, 419 expected) was observed, in particular for cancer of stomach (53 deaths, 44 expected, not statistically significant after adjusting for DI), liver (39 deaths, 23 expected) and lung (124 deaths, 102 expected) cancer. Mortality in SGA was also in excess for respiratory diseases only among men (128 deaths, 97 expected), mostly due to silicosis (31 deaths, 8 expected), although steadily decreasing since 1971 as observed at regional level. Also tuberculosis resulted in excess in SGA (7 deaths, 1 expected). Among females acute respiratory disease mortality was significantly in excess (41 observed, 29 expected). Temporal trend showed a decline from the 70s to the 90s, with a rising trend in recent years in line with Tuscany region. It should be considered that pneumonia was the commonest cause of death of acute respiratory diseases, which allow for lower reliability of death certificate, especially among the elderly (> 64 years). Among females resident in SGA a mortality excess from digestive system diseases was observed (72 observed, 55 expected). The hospitalization in the overall Geothermal Area did not show any excess for all causes and all tumours in both genders. Statistically significant excesses for hospital admission from stomach cancer among males (49 observed, 38 expected) and females (42 observed, 28 expected), and from lymphohematopoietic tumours among females, particularly from lymphatic leukaemia (15 observed, 5 expected), were observed. As mortality analysis highlighted, also hospital admissions by geothermal areas and gender showed a worst picture in SGA than in NGA. In the latter, a significant excess of hospital admissions from all causes among females (1,357 observed, 1,284 expected) but not among males (1,193 observed, 1,141 expected) and an excess - close to statistical significance - from all tumours only among females (297 observed; 272 expected) were observed. Furthermore, statistically significant excesses of hospital admissions from digestive system diseases in both genders (M: 392 observed, 350 expected; F: 300 observed, 268 expected), from dementias (16 observed, 8 expected) and from lymphhematopoietic cancers among females, particularly from lymphatic leukaemia (9 observed, 2 expected), were observed. In the SGA, statistically significant excesses of hospital admissions for stomach cancer (M: 32 observed, 21 expected, not significant after adjusting by DI; F: 29 observed, 18 expected), for respiratory diseases (M: 408 observed, 351 expected; F: 339 observed, 277 expected) and for renal failure (M: 61 observed, 41 expected; F: 52 observed, 34 expected) were observed in both genders. Among males, statistically significant excesses of hospital admissions for liver cancer (23 observed, 14 expected), Parkinson’s disease (10 observed, 4 expected) and genitourinary diseases (312 observed, 282 expected) emerged; among females statistically significant excesses of hospital admissions for oesophagus cancer (4 observed, 1 expected), for circulatory disorders of the brain (235 observed, 209 expected), for chronic-obstructive pulmonary disease (70 observed, 42 expected), and an excess close to the statistical significance for cancer of the central nervous system (22 observed, 14 expected) was observed. Congenital malformations results, which take into account births and pregnancy interruption after prenatal anomaly diagnosis show a statistically significant excess of urogenital anomalies in the SGA. Analysis of low birth weight, preterm birth and ratio O/A in males, showed no deviations from the regional references. CONCLUSIONS Overall this descriptive epidemiological study showed that the health status picture of Geothermal Area is not dissimilar from the one observed in neighbouring communities, with the exception of a worst general mortality among males. However, some mortality and hospitalization excesses are worth of note, most evident in the geothermal area of Amiata (SGA) compared to that of the province of Pisa (NGA). The overall results, showing more critical traits for males and a substantial non alignment of mortality and hospitalization, are suggestive of an etiological role of occupational exposures or individual lifestyles, even if effects of environmental factors cannot be excluded. Among the causes of death and hospitalization which showed excesses, acute respiratory diseases, with mortality excess among females in the SGA, circulatory disorders of the brain, with mortality excess among females in the NGA, respiratory system diseases and renal failure, with hospitalization excess in the SGA in both genders, are proposed for further insights. In fact, these excesses observed in areas characterized by documented presence of pollutants of different sources and great concern of the local communities, reinforced the indications for improving environmental quality and environment and health monitoring. Recommendations to the regional and local authorities for ad hoc epidemiological investigations and public health interventions were released.