Industrial air pollution and mortality in the Taranto area, Southern Italy: A difference-in-differences approach

BACKGROUND

A large steel plant close to the urban area of Taranto (Italy) has been operating since the sixties. Several studies conducted in the past reported an excess of mortality and morbidity from various diseases at the town level, possibly due to air pollution from the plant. However, the relationship between air pollutants emitted from the industry and adverse health outcomes has been controversial. We applied a variant of the "difference-in-differences" (DID) approach to examine the relationship between temporal changes in exposure to industrial PM₁₀ from the plant and changes in cause-specific mortality rates at area unit level.

METHODS

We examined a dynamic cohort of all subjects (321,356 individuals) resident in the Taranto area in 1998-2010 and followed them up for mortality till 2014. In this work, we included only deaths occurring on 2008-2014. We observed a total of 15,303 natural deaths in the cohort and age-specific annual death rates were computed for each area unit (11 areas in total). PM₁₀ and NO₂ concentrations measured at air quality monitoring stations and the results of a dispersion model were used to estimate annual average population weighted exposures to PM₁₀ of industrial origin for each year, area unit and age class. Changes in exposures and in mortality were analyzed using Poisson regression.

RESULTS

We estimated an increased risk in natural mortality (1.86%, 95% confidence interval [CI]: -0.06, 3.83%) per 1 μg/m³ annual change of industrial PM₁₀, mainly driven by respiratory causes (8.74%, 95% CI: 1.50, 16.51%). The associations were statistically significant only in the elderly (65+ years).

CONCLUSIONS

The DID approach is intuitively simple and reduces confounding by design. Under the multiple assumptions of this approach, the study indicates an effect of industrial PM₁₀ on natural mortality, especially in the elderly population.

[The Integrated Environmental Health Impact of emissions from a steel plant in Taranto and from a power plant in Brindisi, (Apulia Region, Southern Italy)]

OBJECTIVES

to estimate the environmental and health impact attributable to PM2.5 emissions from the ex-ILVA steel plant in Taranto and the ENEL power plant in Brindisi (Apulia Region, Southern Italy).

DESIGN

a SPRAY Lagrangian dispersion model was used to estimate PM2.5 concentrations and population weighted exposures following the requirements of the Integrated Environmental Authorization (IEA) of the two plants under study. Available concentration-response functions (OMS/HRAPIE and updates) were used to estimate the number of attributable premature deaths.

SETTING AND PARTICIPANTS

residents in the 40 municipalities of the domains of the VDS (assessment of health damage, according to the Regional Law n. 21/2012) of Brindisi (source: Italian National Institute of Statistics 2011 Census) and residents in Taranto, Statte, and Massafra (source: cohort study).

MAIN OUTCOME MEASURES

mortality from natural causes, cardiovascular and respiratory diseases, and lung cancer attributable to PM2.5. Incremental lifetime cumulative risks (ILCRs) for lung cancer associated to PM2.5 exposure.

RESULTS

there was a reduction of the estimated impacts from the pre to the post IEA-scenarios in both Taranto and Brindisi. In Taranto, ILCRs greater than 1x10^-4; were estimated in 2010 and 2012; the ILCR was greater than 1x10^-4; in the district of Tamburi (near the plant) also for the 2015 scenario. ILCRs estimated for Brindisi were between 1x10^-6; and 4x10^-5;.

CONCLUSIONS

the Integrated Environmental Health Impact Assessment confirmed the results of the VDS conducted according to the toxicological risk assessment approach. An unacceptable risk was estimated for Tamburi also for the 2015 scenario, characterized by a production of 4.7 million tons of steel, about half compared to one foreseen by the IEA (8 mt.).

Spatial distribution of kidney disease in the contaminated site of Taranto (Italy)

BACKGROUND

Exposure to heavy metals has been associated with kidney disease. We investigated the spatial distribution of kidney disease in the industrially contaminated site of Taranto.

METHODS

Cases were subjects with a first hospital discharge diagnosis of kidney disease. Cases affected by specific comorbidities were excluded. Standardized Hospitalization Ratios (SHRs) were computed for low/high exposure area and for modeled spatial distribution of cadmium and fine particulate matter.

RESULT

Using the high/low exposure approach, in subjects aged 20-59 years residing in the high exposure area a significant excess of hospitalization was observed in males and a non-significant excess in females. No excesses were observed in subjects aged 60 years and over. The analysis by the modeling approach did not show a significant association with the greatest pollution impact area.

CONCLUSION

Due to the excesses of hospitalization observed in the high/low exposure approach, a continuing epidemiological surveillance of residents and occupational groups is warranted.

Differences between weekend and weekday ozone levels over rural and urban sites in Southern Italy

Air quality data from a network of 11 monitoring stations in the Apulia region of southern Italy during the summer of 2005 reveal a high frequency of ozone law limit violations. Since ozone is a secondary pollutant, air quality control strategies aimed at reducing ozone concentration are not immediate. Herein, we analyse weekly changes in concentration levels of ozone (O(3)), nitrogen oxides (NO(x)), carbon monoxide (CO), and volatile organic compounds (VOCs), and evaluate how the differences in primary emissions cause changes in the production of ozone. The comparison between weekend and weekday levels of O(3) and its precursors are direct evidence for the existence of the "ozone weekend effect." This effect was observed at all stations with a considerable variation in the overall ozone magnitude, including both traffic stations and non-traffic stations. Data from VOC measurements at traffic stations primarily indicated elevated levels of benzene, toluene, and xylenes (BTX); all of these substances showed an overall decrease over the weekend. A single station indicated levels of non-methane hydrocarbon (NMHC) and PM10, both of which did not demonstrate any weekly cycle. Analysis of weekly and diurnal cycles of O(3), NO(x), CO, NMHC, and PM10 indicates that higher weekend ozone levels result from a reduction in the emission of nitrogen oxides on weekends in VOC-sensitive regimes. This indicates that a reduction in VOC and NO(x) levels would be more effective than NO(x) reduction alone. Our results underscore the need for improved and more efficient VOC measurements.