Impact of urban heat islands on morbidity and mortality in heat waves: Observational time series analysis of Spain's five cities

Urban heat islands (UHIs) have become an especially relevant phenomenon as a consequence of global warming and the growing proportion of people living in cities. The health impacts that are sometimes attributed to the rise in temperature generated in an UHI are not always adequately justified. The objective is to analyse what effect UHIs have on maximum (Tmax) and minimum daily temperatures (Tmin) recorded in urban and non-urban observatories, and quantify the impact on morbidity and mortality during heat waves in Spain's five cities. Data were collected on natural-cause daily mortality and unscheduled emergency hospital admissions (ICD-10: A00-R99) registered in these 5 cities across the period 2014-2018. We analysed daily Tmax and Tmin values at urban and non-urban observatories in these cities, and quantified the impact of Tmax and Tmin values during heat waves in each of these cities, using GLM models that included Tmax only, Tmin only, and both. We controlled for air pollution and other meteorological variables, as well as for seasonalities, trend and the autoregressive nature of the series. The UHI effect was observed in Tmin but not in Tmax, and proved to be greater in coastal cities than in inland and more densely populated cities. The UHI value in relation to the mean Tmin in the summer months ranged from 1.2 °C in Murcia to 4.1 °C in Valencia (difference between urban/non-urban observatories). The modelling process showed that, while a statistically significant association (p < 0.05) was observed in inland cities with Tmax for mortality and hospital admissions in heat waves, in coastal cities the association was obtained with Tmin, and the only impact in this case was the UHI effect on morbidity and mortality. No generalisations can be made about the impact of UHI on morbidity and mortality among the exposed population in cities. Studies on a local scale are called for, since it is local factors that determine whether the UHI effect will have a greater or lesser impact on health during heat-wave events.

Does the meteorological origin of heat waves influence their impact on health? A 6-year morbidity and mortality study in Madrid (Spain)

BACKGROUND

In Spain, two synoptic-scale conditions influence heat wave formation. The first involves advection of warm and dry air masses carrying dust of Saharan origin (North African Dust (NAF) = 1). The second entails anticyclonic stagnation with high insolation and stability (NAF) = 0). Some studies show that the meteorological origin of these heat waves may affect their impact on morbidity and mortality.

OBJECTIVE

To determine whether the impact of heat waves on health outcomes in Madrid (Spain) during 2013-2018 varied by synoptic-scale condition.

METHODOLOGY

Outcome data consist of daily mortality and daily hospital emergency admissions (morbidity) for natural, circulatory, and respiratory causes. Predictors include daily maximum and minimum temperatures and daily mean concentrations of NO₂, PM₁₀, PM_2.5, NO₂, and O₃. Analyses adjust for insolation, relative humidity, and wind speed. Generalized linear models were performed with Poisson link between the variables controlling for trend, seasonality, and auto-regression in the series. Relative Risks (RR) and Attributable Risks (AR) were determined. The RRs for mortality attributable to high temperatures were similar regardless of NAF status. For hospital admissions, however, the RRs for hot days with NAF = 0 are higher than for days with NAF = 1. We also found that atmospheric pollutants worsen morbidity and mortality, especially PM₁₀ concentrations when NAF = 1 and O₃ concentrations when NAF = 0.

RESULTS

The effect of heat waves on morbidity and mortality depends on the synoptic situation. The impact is greater under anticyclonic stagnation conditions than under Saharan dust advection. Further, the health impact of pollutants such as PM₁₀ and O₃ varies according to the synoptic situation.

CONCLUSIONS

Based on these findings, we strongly recommend prevention plans to include data on the meteorological situation originating the heat wave, on a synoptic-scale, as well as comprehensive preventive measures against the compounding effect of high temperatures and pollution.

Population vulnerability to extreme cold days in rural and urban municipalities in ten provinces in Spain

BACKGROUND

The objective was to analyze whether there are differences in vulnerability to Extreme Cold Days (ECD) between rural and urban populations in Spain.

METHODOLOGY

Time series analysis carried out from January 1, 2000, through December 31, 2013. Municipalities with over 10,000 inhabitants were included from 10 Spanish provinces, classified into 42 groups by isoclimate and urban/rural character as defined by Eurostat criteria. The statistical strategy was carried out in two phases. First: It was analyzed the relationship between minimum daily temperature (Tmin) (source: AEMET) and the rate of daily winter mortality due to natural causes -CIE-10: A00 - R99- (source: National Statistics Institute). Then, It was determinated the threshold of Tmin that defines the ECD and its percentile in the series of winter Tmin (Pthreshold), which is a measure of vulnerability to ECD so that the higher the percentile, the higher the vulnerability. Second: possible explanatory variables of vulnerability were explored using Mixed Generalized Models, using 13 independent variables related to meteorology, environment, socioeconomics, demographics and housing quality.

RESULTS

The average Pthreshold was 18 %. The final model indicated that for each percentage point increase in unemployment, the vulnerability to ECD increased by 0.4 (0.2, 0.6) points. Also, with each point increase in rurality index, this vulnerability decreased by -6.1 (-2.1, -10.0) points. Although less determinant, other factors that could contribute to explaining vulnerability at the province level included minimum winter daily temperatures and the percentage of housing with poor insulation.

CONCLUSIONS

The vulnerability to ECD was greater in urban zones than in rural zones. Socioeconomic status is a key to understanding how this vulnerability is distributed. These results suggest the need to implement public health prevention plans to address ECD at the state level. These plans should be based on threshold temperatures determined at the smallest scale possible.

Gender differences in adaptation to heat in Spain (1983-2018)

In Spain the average temperature has increased by 1.7 °C since pre-industrial times. There has been an increase in heat waves both in terms of frequency and intensity, with a clear impact in terms of population health. The effect of heat waves on daily mortality presents important territorial differences. Gender also affects these impacts, as a determinant that conditions social inequalities in health. There is evidence that women may be more susceptible to extreme heat than men, although there are relatively few studies that analyze differences in the vulnerability and adaptation to heat by sex. This could be related to physiological causes. On the other hand, one of the indicators used to measure vulnerability to heat in a population and its adaptation is the minimum mortality temperature (MMT) and its temporal evolution. The aim of this study was to analyze the values of MMT in men and women and its temporal evolution during the 1983-2018 period in Spain's provinces. An ecological, longitudinal retrospective study was carried out of time series data, based on maximum daily temperature and daily mortality data corresponding to the study period. Using cubic and quadratic fits between daily mortality rates and the temperature, the minimum values of these functions were determined, which allowed for determining MMT values. Furthermore, we used an improved methodology that provided for the estimation of missing MMT values when polynomial fits were inexistent. This analysis was carried out for each year. Later, based on the annual values of MMT, a linear fit was carried out to determine the rate of evolution of MMT for men and for women at the province level. Average MMT for all of Spain's provinces was 29.4 °C in the case of men and 28.7 °C in the case of women. The MMT for men was greater than that of women in 86 percent of the total provinces analyzed, which indicates greater vulnerability among women. In terms of the rate of variation in MMT during the period analyzed, that of men was 0.39 °C/decade, compared to 0.53 °C/decade for women, indicating greater adaptation to heat among women, compared to men. The differences found between men and women were statistically significant. At the province level, the results show great heterogeneity. Studies carried out at the local level are needed to provide knowledge about those factors that can explain these differences at the province level, and to allow for incorporating a gender perspective in the implementation of measures for adaptation to high temperatures.

Temporal evolution of threshold temperatures for extremely cold days in Spain

In contrast to research on heat waves, there are no studies in recent years that analyze the temporal evolution of threshold temperatures (Tthreshold) for extremely cold days (ECD). It is unknown whether threshold temperatures have increased more quickly than the minimum daily temperature (Tmin) in recent years. The objective of this study was to analyze the temporal evolution of the minimum daily temperature (Tmin) in a group of Spanish provinces and compare it with the evolution of threshold temperatures. An ecological, retrospective time series study was carried out using daily observations between January 1, 1983 and December 31, 2018 (36 years) in 10 provinces that are representative of the different climate territories in Spain. For each representative observatory in each province, the values of Tmin were obtained for the winter months (November-March). The value of Tthreshold was determined for each province and each year, using dispersion diagrams for the pre-whitened series, with daily mortality due to natural causes displayed on the Y axis (CIEX: A00-R99) and Tmin grouped by 10 degree intervals on the X axis. To determine the temporal evolution of Tmin and Tthreshold for each province, linear models were fitted, with time as the independent variable. During the winter months, Tmin increased at an average rate of 0.2 °C/decade (IC95: 0.1-0.3), while Tthreshold remained practically constant during the period, at 0.1 °C/decade (IC95% -0.1 0.3). These values are much lower than those obtained in the case of heat, both in terms of the evolution of maximum daily temperature and that of Tthreshold. In conclusion, the fact that this trend has been maintained across time in a scenario of climate change, with a slow increase in minimum daily temperatures and constant values of Threshold, suggests a decrease in the number of ECD.

Analysis of vulnerability to heat in rural and urban areas in Spain: What factors explain Heat's geographic behavior?

INTRODUCTION

There is currently little knowledge and few published works on the subject of vulnerability to heat in rural environments at the country level. Therefore, the objective of this study was to determine whether rural areas are more vulnerable to extreme heat than urban areas in Spain. This study aimed to analyze whether a pattern of vulnerability depends on contextual, environmental, demographic, economic and housing variables.

METHODS

An ecological, longitudinal and retrospective study was carried out based on time series data between January 01, 2000 and December 31, 2013 in 42 geographic areas in 10 provinces in Spain. We first analyzed the functional relationship between the mortality rate per million inhabitants and maximum daily temperature (Tmax). We then determined the summer temperature threshold (Pthreshold) (June-September) at which increases in mortality are produced that are attributable to heat. In a second phase, based on Pthreshold, a vulnerability variable was calculated, and its distribution was analyzed using mixed linear models from the Poisson family (link = log). In these models, the dependent variable was vulnerability, and the independent variables were exposure to high temperatures, aridity of the climate, deprivation index, percentage of people over age 65, rurality index, percentage of housing built prior to 1980 and condition of dwellings.

RESULTS

Rurality was a protective factor, and vulnerability in urban areas was six times greater. In contrast, risk factors included aridity (RR = 5.89 (2.26 15.36)), living in cool summer zones (2.69 (1.23, 5.91)), poverty (4.05 (1.91 8.59)) and the percentage of dysfunctional housing (1.13 (1.04 1.24)).

CONCLUSIONS

Rural areas are less vulnerable to extreme heat than the urban areas analyzed. Also, population groups with worse working conditions and higher percentages of dwellings in poor conditions are more vulnerable.

Analysis of the impact of heat waves on daily mortality in urban and rural areas in Madrid

Backgraund

The objective of this study was to analyze and compare the effect of high temperatures on daily mortality in the urban and rural populations in Madrid.

Methods

Data were analyzed from municipalities in Madrid with a population of over 10,000 inhabitants during the period from January 1, 2000 to December 31, 2013. Four groups were generated: Urban Metropolitan Center, Rural Northern Mountains, Rural Center, and Southern Rural. The dependent variable used was the rate of daily mortality due to natural causes per million inhabitants (CIE-X: A00- R99) between the months of June and September for the period. The primary independent variable was maximum daily temperature. Social and demographic “context variables” were included: population &gt;64 years of age (%), deprivation index and housing indicators. The analysis was carried out in three phases: 1) determination of the threshold definition temperature of a heat wave (Tthresold) for each study group; 2) determination of relative risks (RR) attributable to heat for each group using Poisson linear regression (GLM), and 3) calculation of odds ratios (OR) using binomial family GLM for the frequency of the appearance of heat waves associated with context variables.

Results

The resulting percentiles (for the series of maximum daily temperatures for the summer months) corresponding to Tthreshold were: 74th percentile for Urban Metropolitan Center, 76th percentile for Southern Rural, 83rd for Rural Northern Mountains and 98th percentile for Center Rural (98). Greater vulnerability was found for the first two. In terms of context variables that explained the appearance of heat waves, deprivation index level, population &gt;64 years of age and living in the metropolitan area were found to be risk factors.

Conclusions

Rural and urban areas behaved differently, and socioeconomic inequality and the composition of the population over age 64 were found to best explain the vulnerability of the Rural Center and Southern Rural zones.

Key messages

Heat waves affect urban and rural areas differently. So results and conclusions from urban areas should not be extrapolated to rural population. Socioeconomic status and ageing are important variables which explain the different behavior of heat waves in different areas.

Analysis of the impact of heat waves on daily mortality in urban and rural areas in Madrid

The objective of this study was to analyze and compare the effect of high temperatures on daily mortality in the urban and rural populations in Madrid. Data were analyzed from municipalities in Madrid with a population of over 10,000 inhabitants during the period from January 1, 2000 to December 31, 2020. Four groups were generated: Urban Metropolitan Center, Rural Northern Mountains, Rural Center, and Southern Rural. The dependent variable used was the rate of daily mortality due to natural causes per million inhabitants (CIE-X: A00-R99) between the months of June and September for the period. The primary independent variable was maximum daily temperature. Social and demographic "context variables" were included: population >64 years of age (%), deprivation index and housing indicators. The analysis was carried out in three phases: 1) determination of the threshold definition temperature of a heat wave (Tumbral) for each study group; 2) determination of relative risks (RR) attributable to heat for each group using Poisson linear regression (GLM), and 3) calculation of odds ratios (OR) using binomial family GLM for the frequency of the appearance of heat waves associated with context variables. The resulting percentiles (for the series of maximum daily temperatures for the summer months) corresponding to Tthreshold were: 74th percentile for Urban Metropolitan Center, 76th percentile for Southern Rural, 83rd for Rural Northern Mountains and 98th percentile for Center Rural (98). Greater vulnerability was found for the first two. In terms of context variables that explained the appearance of heat waves, deprivation index level, population >64 years of age and living in the metropolitan area were found to be risk factors. Rural and urban areas behaved differently, and socioeconomic inequality and the composition of the population over age 64 were found to best explain the vulnerability of the Rural Center and Southern Rural zones.

The effect of cold waves on daily mortality in districts in Madrid considering sociodemographic variables

While there is much research that focuses on the association between cold waves and their impacts on daily mortality at the city level, few analyze the impact related to social context and demographic variables at levels lower than the municipal. The objective of this study was to determine the role of the percentage of people over age 65, income level and percentage of homes without heating in the analysis of the impact of cold waves on daily mortality between January 1, 2010 and December 31, 2013 in different districts of the municipality of Madrid. We calculated Relative Risks (RR) and Attributable Risks (RA) for each of 17 districts to determine correlations between the effect of cold waves and daily mortality due to natural causes (CIEX: A00-R99) using Generalized Linear Models (GLM) of the Poisson family (link log). The pattern of risks obtained by district was analyzed using binomial family models (link logit), considering socioeconomic and demographic variables. In terms of results, an impact of cold on mortality was detected in 9 of the 17 districts analyzed. The analysis of risk patterns revealed that the probability of detecting an impact in a district increases in a statistically significant way (p-value <0.05) with a higher percentage of homes without heating systems and a higher percentage of population over age 65. The results obtained identify the factors that should be considered in public health policies that target the district level to reduce the impact of cold waves.

Will there be cold-related mortality in Spain over the 2021-2050 and 2051-2100 time horizons despite the increase in temperatures as a consequence of climate change?

INTRODUCTION

Global warming is resulting in an increase in temperatures which is set to become more marked by the end of the century and depends on the accelerating pace of greenhouse gas emissions into the atmosphere. Yet even in this scenario, so-called "cold waves" will continue to be generated and have an impact on health.

OBJECTIVES

This study sought to analyse the impact of cold waves on daily mortality at a provincial level in Spain over the 2021-2050 and 2051-2100 time horizons under RCP4.5 and RCP 8.5 emission scenarios, on the basis of two hypotheses: (1) that the cold-wave definition temperature (T _threshold) would not vary over time; and, (2) that there would be a variation in T _threshold.

MATERIAL AND METHODS

The results of a retrospective study undertaken for Spain as a whole across the period 2000-2009 enabled us to ascertain the cold-wave definition temperature at a provincial level and its impact on health, measured by reference to population attributable risk (PAR). The minimum daily temperatures projected for each provincial capital considering the above time horizons and emission scenarios were provided by the State Meteorological Agency. On the basis of the T _threshold definition values and minimum daily temperatures projected for each province, we calculated the expected impact of low temperatures on mortality under the above two hypotheses. Keeping the PAR values constant, it was assumed that the mortality rate would vary in accordance with the available data.

RESULTS

If T _threshold remained constant over the above time horizons under both emission scenarios, there would be no cold-related mortality. If T _threshold were assumed to vary over time, however, then cold-related mortality would not disappear: it would instead remain practically constant over time and give rise to an estimated overall figure of around 250 deaths per year, equivalent to close on a quarter of Spain's current annual cold-related mortality and entailing a cost of approximately €1000 million per year.

CONCLUSION

Given that cold waves are not going to disappear and that their impact on mortality is far from negligible and is likely to remain so, public health prevention measures must be implemented to minimise these effects as far as possible.

Mortality attributable to high temperatures over the 2021-2050 and 2051-2100 time horizons in Spain: Adaptation and economic estimate

BACKGROUND

In recent years, a number of studies have been conducted with the aim of analysing the impact that high temperatures will have on mortality over different time horizons under different climate scenarios. Very few of these studies take into account the fact that the threshold temperature used to define a heat wave will vary over time, and there are practically none which calculate this threshold temperature for each geographical area on the assumption that there will be variations at a country level.

OBJECTIVE

To analyse the impact that high temperatures will have on mortality across the periods 2021-2050 and 2051-2100 under a high-emission climate scenario (RCP8.5), in a case: (a) where adaptation processes are not taken into account; and (b) where complete adaptation processes are taken into account.

MATERIAL AND METHODS

Based on heat-wave definition temperature (T_threshold) values previously calculated for the reference period, 2000-2009, for each Spanish provincial capital, and their impact on daily mortality as measured by population attributable risk (PAR), the impact of high temperatures on mortality will be calculated for the above-mentioned future periods. Two hypotheses will be considered, namely: (a) that T_threshold does not vary over time (scenario without adaptation to heat); and, (b) that T_threshold does vary over time, with the percentile to which said T_threshold corresponds being assumed to remain constant (complete adaptation to heat). The temperature data were sourced from projections generated by Coupled Model Intercomparison Project (CMIP5) climate models adapted to each region's local characteristics by the State Meteorological Agency (Agencia Estatal de Meteorología/AEMET). Population-growth projections were obtained from the National Statistics Institute (Instituto Nacional de Estadística/INE). In addition, an economic estimate of the resulting impact will be drawn up.

RESULTS

The mean value of maximum daily temperatures will rise, in relation to those of the reference period (2000-2009), by 1.6⁰C across the period 2021-2050 and by 3.3⁰C across the period 2051-2100. In a case where there is no heat-adaptation process, overall annual mortality attributable to high temperatures in Spain would amount to 1414 deaths/year (95% CI: 1089-1771) in the period 2021-2050, rising to 12,896 deaths/year (95% CI: 9852-15,976) in the period 2051-2100. In a case where there is a heat-adaptation process, annual mortality would be 651 deaths/year (95% CI: 500-807) in the period 2021-2050, and 931 deaths per year (95% CI: 770-1081) in the period 2051-2100. These results display a high degree of heterogeneity. The savings between a situation that does envisage and one that does not envisage an adaptive process is €49,100 million/year over the 2051-2100 time horizon.

CONCLUSION

A non-linear increase in maximum daily temperatures was observed, which varies widely from some regions to others, with an increase in mean values for Spain as a whole that is not linear over time. The high degree of heterogeneity found in heat-related mortality by region and the great differences observed on considering an adaptive versus a non-adaptive process render it necessary for adaptation plans to be implemented at a regional level.

Time trends in the impact attributable to cold days in Spain: Incidence of local factors

BACKGROUND

While numerous studies have shown that the impact of cold waves is decreasing as result of various processes of adaptation, far fewer have analysed the time trend shown by such impact, and still fewer have done so for the different provinces of a single country, moreover using a specific cold waves definition for each. This study thus aimed to analyse the time trend of the impact of cold days on daily mortality in Spain across the period 1983-2003.

METHODS

For study purposes, we used daily mortality data for all natural causes except accidents in ten Spanish provinces. The time series was divided into three subperiods. For each period and province, the value of T_threshold was obtained via the percentile corresponding to the cold day's definition for that province obtained in previous studies. Relative Risks (RRs) and Population Attributable Fraction (PARs) were calculated using Generalised Linear Models (GLMs) with the Poisson regression link. Seasonalities, trends and autoregressive components were controlled. Global RRs and ARs were calculated with the aid of a meta-analysis with random effects for each of the periods.

RESULTS

The results show that the RRs for Spain as a whole were 1.12 (95% CI: 1.08 1.16) for the first period, 1.15 (95% CI: 1.09 1.22) for the second and 1.18 (95% CI: 1.10 1.26) for the third. The impact of cold days has risen slightly over time, though the differences were not statistically significant. These findings show a clearly different behaviour pattern to that previously found for heat.

CONCLUSION

The results obtained in this study do not show a downward trend for colds days. The complexity of the biological mechanisms involved in cold-related mortality and the lack of robust results mean that more research must be done in this particular field of public health.

Time trend in the impact of heat waves on daily mortality in Spain for a period of over thirty years (1983-2013)

Many of the studies that analyze the future impact of climate change on mortality assume that the temperature that constitutes a heat wave will not change over time. This is unlikely, however, given the process of adapting to heat changes, prevention plans, and improvements in social and health infrastructure. The objective of this study is to analyze whether, during the 1983-2013 period, there has been a temporal change in the maximum daily temperatures that constitute a heat wave (T_threshold) in Spain, and to investigate whether there has been variation in the attributable risk (AR) associated with mortality due to high temperatures in this period. This study uses daily mortality data for natural causes except accidents CIEX: A00-R99 in municipalities of over 10,000 inhabitants in 10 Spanish provinces and maximum temperature data from observatories located in province capitals. The time series is divided into three periods: 1983-1992, 1993-2003 and 2004-2013. For each period and each province, the value of T_threshold was calculated using scatter-plot diagram of the daily mortality pre-whitened series. For each period and each province capitals, it has been calculated the number of heat waves and quantifying the impact on mortality through generalized linear model (GLM) methodology with the Poisson regression link. These models permits obtained the relative risks (RR) and attributable risks (AR). Via a meta-analysis, using the Global RR and AR were calculated the heat impact for the total of the 10 provinces. The results show that in the first two periods RR remained constant RR: 1.14 (CI95%: 1.09 1.19) and RR: 1.14 (CI95%: 1.10 1.18), while the third period shows a sharp decrease with respect to the prior two periods RR: 1.01 (CI95%: 1.00 1.01); the difference is statistically significant. In Spain there has been a sharp decrease in mortality attributable to heat over the past 10 years. The observed variation in RR puts into question the results of numerous studies that analyze the future impact of heat on mortality in different temporal scenarios and show it to be constant over time.

Spatial variability in threshold temperatures of heat wave mortality: impact assessment on prevention plans

Spain's current heat wave prevention plans are activated according to administrative areas. This study analyses the determination of threshold temperatures for triggering prevention-plan activation by reference to isoclimatic areas, and describes the public health benefits. We subdivided the study area - the Madrid Autonomous Region (MAR) - into three, distinct, isoclimatic areas: 'North', 'Central' and 'South', and grouped daily natural-cause mortality (ICD-10: A00-R99) in towns of over 10,000 inhabitants (2000-2009 period) accordingly. Using these three areas rather than the MAR as a whole would have resulted in a possible decrease in mortality of 73 persons (38-108) in the North area, and in aborting unnecessary activation of the plan 153 times in the Central area and 417 times in the South area. Our results indicate that extrapolating this methodology would bring benefits associated with a reduction in attributable mortality and improved effectiveness of public health interventions.

Mortality attributable to extreme temperatures in Spain: A comparative analysis by city

BACKGROUND

The Low Temperature Days (LTD) have attracted far less attention than that of High Temperature Days (HTD), though its impact on mortality is at least comparable. This lower degree of attention may perhaps be due to the fact that its influence on mortality is less pronounced and longer-term, and that there are other concomitant infectious winters factors. In a climate-change scenario, the studies undertaken to date report differing results. The aim of this study was to analyse mortality attributable to both thermal extremes in Spain's 52 provinces across the period 2000-2009, and estimate the related economic cost to show the benefit or "profitability" of implementing prevention plans against LTD.

METHODS

Previous studies enabled us: to obtain the maximum daily temperature above which HTD occurred and the minimum daily temperature below which LTD occurred in the 52 provincial capitals analysed across the same study period; and to calculate the relative and attributable risks (%) associated with daily mortality in each capital. These measures of association were then used to make different calculations to obtain the daily mean mortality attributable to both thermal extremes. To this end, we obtained a summary of the number of degrees whereby the temperature exceeded (excess °C) or fell short (deficit °C) of the threshold temperature for each capital, and calculated the respective number of extreme temperatures days. The economic estimates rated the prevention plans as being 68% effective.

RESULTS

Over the period considered, the number of HTD (4373) was higher than the number of LTD (3006) for Spain as a whole. Notwithstanding this, in every provincial capital the mean daily mortality attributable to heat was lower (3deaths/day) than that attributable to cold (3.48deaths/day). In terms of the economic impact of the activation of prevention plans against LTD, these could be assumed to avoid 2.37 deaths on each LTD, which translated as a saving of €0.29M. Similarly, in the case of heat, 2.04 deaths could be assumed to be avoided each day on which the prevention plan against HTD was activated, amounting to a saving of €0.25M. While the economic cost of cold-related mortality across the ten-year period 2000-2009 was €871.7M, that attributable to heat could be put at €1093.2M.

CONCLUSION

The effect of extreme temperatures on daily mortality was similar across the study period for Spain overall. The lower number of days with LTD meant, however, that daily cold-related mortality was higher than daily heat-related mortality, thereby making prevention plans against LTD more "profitable" prevention plans against HTD in terms of avoidable mortality.

Geographical variation in relative risks associated with cold waves in Spain: The need for a cold wave prevention plan

In general, there are few studies that analyse the impact of low temperatures on mortality and fewer still that use cold-wave-definition thresholds based on epidemiological and non-climatological criteria. Such a threshold definition, which took into account population features such as socio-economic and demographic characteristics, made it possible for a specific threshold temperature to be obtained for each of Spain's 52 provincial capitals in this study. Using generalised linear models with the Poisson regression link, and controlling for trend, autocorrelations and seasonalities of the series, and influenza epidemics, we obtained the impact of low temperatures on mortality in each provincial capital by calculating the relative risks (RRs) and attributable risks (ARs) for natural as well as circulatory and respiratory causes. The study showed higher minimum temperature thresholds in coastal areas, and an overall impact of cold on mortality in Spain due to natural causes RR=1.13 (95% CI: 1.11-1.16), circulatory causes RR=1.18 (95% CI: 1.15-1.22) and respiratory causes RR=1.24 (95% CI: 1.20-1.29) slightly greater than that obtained to date for heat. From a public health standpoint, there is a need for specific cold wave prevention plans at a regional level which would enable mortality attributable to low temperatures to be reduced. These plans have shown themselves to be effective in decreasing heat-related mortality, and we feel that they are essential for reducing cold-related effects on morbidity and mortality.

Geographical variation in relative risks associated with heat: Update of Spain's Heat Wave Prevention Plan

A decade after the implementation of prevention plans designed to minimise the impact of high temperatures on health, some countries have decided to update these plans in order to improve the weakness detected in these ten years of operation. In the case of Spain, this update has fundamentally consisted of changing the so-called "threshold" or "trigger" temperatures used to activate the plan, by switching from temperature values based on climatological criteria to others obtained by epidemiological studies conducted on a provincial scale. This study reports the results of these "trigger" temperatures for each of Spain's 52 provincial capitals, as well as the impact of heat on mortality by reference to the relative risks (RRs) and attributable risks (ARs) calculated for natural as well as circulatory and respiratory causes. The results obtained for threshold temperatures and RRs show a more uniform behaviour pattern than those obtained using temperature values based on climatological criteria; plus a clear decrease in RRs of heat-associated mortality due to the three causes considered, at both a provincial and regional level as well as for Spain as a whole. The updating of prevention plans is regarded as crucial for optimising the operation of these plans in terms of reducing the effect of high temperatures on population health.

Comparison of the effects of extreme temperatures on daily mortality in Madrid (Spain), by age group: The need for a cold wave prevention plan

A number of studies have shown that there is a time trend towards a reduction in the effects of heat on mortality. In the case of cold, however, there is practically no research of this type and so there is no clearly defined time trend of the impact of cold on mortality. Furthermore, no other specific studies have yet analysed the time trend of the impact of both thermal extremes by age group. We analysed data on daily mortality due to natural causes (ICD-10: A00-R99) in the city of Madrid across the period 2001-2009 and calculated the impact of extreme temperatures on mortality using Poisson regression models for specific age groups. The groups of age selected coinciding with the pre-existing age-groups analyzed in previous papers. For heat waves the groups of age used were: <10 years, 10-17 years, 18-44 years, 45-64 years, 65-74 years and over-75 years. For cold waves the groups of age used were: <1 year; 1-5 years, 6-17 years, 18-44 years, 45-64 years, 65-74 years and over-75 years. <1, 1-17, 18-44, 45-66, 65-74 and over-75 years. We controlled for confounding variables, such as air pollution, noise, influenza, pollen, pressure and relative humidity, trend of the series, as well as seasonalities and autoregressive components of the series. The results of these models were compared to those obtained for the same city during the period 1986-1997 and published in different studies. Our results show a lightly reduction in the effects of heat, especially in the over-45-year age group. In the case of cold, the behaviour pattern was the opposite, with an increase in its effect. Heat adaptation and socio-economic and public-health prevention and action measures may be behind this amelioration in the effects of heat, whereas the absence of such actions in respect of low temperatures may account for the increase in the effects of cold on mortality. From a public health point of view, the implementation of cold wave prevention plans covering all age groups is thus called for.

Influence of advections of particulate matter from biomass combustion on specific-cause mortality in Madrid in the period 2004-2009

Approximately, 20 % of particulate and aerosol emissions into the urban atmosphere are of natural origin (including wildfires and Saharan dust). During these natural episodes, PM10 and PM2.5 levels usually exceed World Health Organisation (WHO) health protection thresholds. This study sought to evaluate the possible effect of advections of particulate matter from biomass fuel combustion on daily specific-cause mortality among the general population and the segment aged ≥ 75 years in Madrid. Ecological time-series study in the city of Madrid from January 01, 2004 to December 31, 2009. The dependent variable analysed was daily mortality due to natural (ICD-10:A00-R99), circulatory (ICD-10:I00-I99), and respiratory (ICD-10:J00-J99) causes in the population, both general and aged ≥ 75 years. The following independent and control variables were considered: a) daily mean PM2.5 and PM10 concentrations; b) maximum daily temperature; c) daily mean O3 and NO2 concentrations; d) advection of particulate matter from biomass combustion ( http://www.calima.ws/ ), using a dichotomous variable and e) linear trend and seasonalities. We conducted a descriptive analysis, performed a test of means and, to ascertain relative risk, fitted a model using autoregressive Poisson regression and stratifying by days with and without biomass advection, in both populations. Of the 2192 days analysed, biomass advection occurred on 56, with mean PM2.5 and PM10 values registering a significant increase during these days. PM10 had a greater impact on organic mortality with advection (RRall ages = 1.035 [1.011-1.060]; RR  ≥  75 years = 1.066 [1.031-1.103]) than did PM2.5 without advection (RRall ages = 1.017 [1.009-1.025]; RR  ≥  75 years = 1.012 [1.003-1.022]). Among specific causes, respiratory-though not circulatory-causes were associated with PM10 on days with advection in ≥ 75 year age group. PM10, rather than PM2.5, were associated with an increase in natural cause mortality on days with advection of particulate matter from biomass combustion, particularly in the ≥ 75 year age group.

Seasonal variation of pharmaceutically active compounds in surface (Tagus River) and tap water (Central Spain)

Numerous studies have shown the presence of pharmaceutically active compounds (PhACs) in different environmental compartments, for example, in surface water or wastewater ranging from nanograms per litre to micrograms per litre. Likewise, some recent studies have pointed to seasonal variability, thus indicating that PhAcs concentrations in the aquatic environment may depend on the time of year. This work intended to find out (1) whether Tagus fluvial and drinking water were polluted with different groups of PhACs and (2) if their concentrations differed between winter and summer seasons. From the 58 substances analysed, 41 were found belonging to the main therapeutic groups. Statistical differences were seen for antibacterials, antidepressants, anxiolytics, antiepileptics, and cardiovascular drugs, with higher concentrations being detected in winter than in summer. These results might indicate that the PhACs analysed in this study undergo lower environmental degradation in winter than in summer. In order to confirm these initial results, a continuous monitoring should be performed especially on those PhACs that either because of an elevated consumption or an intrinsic chemical persistence are poorly degraded during winter months due to low temperatures and solar irradiation. It is especially important to identify which of these specific PhACs are in order to recommend their substitution by equally effective and safe substances but also environmentally friendly.

Influence of local factors in the relationship between mortality and heat waves: Castile-La Mancha (1975-2003)

INTRODUCTION

All the climate-change studies undertaken to date agree that one of the principal consequences of this phenomenon is the increase in heat waves, which, without exception, are linked to marked rises in mortality. The characteristics that modulate and determine the relationship between high temperatures and health must therefore be ascertained in the greatest possible detail, so that really effective prevention plans can be designed to address temperature extremes.

METHODS

We examined the effect of heat waves on daily non-accidental-cause mortality across all age groups in the Castile-La Mancha region (Spain) from 1975 to 2003. Quantitative analyses were performed using autoregressive integrated moving average (ARIMA) models, with other covariates, such as pressure trends, relative humidity, and duration and chronological number of heat waves.

RESULTS

Mortality increased significantly with respect to the mean, when temperatures exceeded the designated provincial thresholds in Castile-La Mancha. For each degree centigrade that temperatures exceeded these thresholds, the percentage increase in mortality amounted to increases of approximately 12% over the daily mean, albeit with clear provincial variations. The longest heat waves were associated with daily mortality, with those at the end of summer causing the lowest mortality. Meteorological situations most closely associated with increases in mortality were cyclonic conditions accompanied by low humidity.

CONCLUSIONS

Spatio-temporal variability in the temperature-mortality relationship must be studied in order to enable really effective heat-wave prevention plans to be drawn up. The influence of variables, such as heat-wave duration or time of appearance, is important in the total increase in mortality during temperature extremes. Since parameters, such as humidity or pressure trends, can play very different roles in different geographical settings, they should be analysed separately from temperature.

Mortality from cold waves in Castile--La Mancha, Spain

INTRODUCTION

As is known, the effects of extreme temperatures on mortality are characterised by an annual periodicity, with a rise centred in the winter months. The most recent epidemiological studies show that mortality caused by cold waves is, in many cases, comparable to that caused by the severest heat waves. This study sought to quantify the rise in mortality due to extreme cold and the factors that determine the relationship between these variables in Castile-La Mancha (Spain).

METHODS

We examined the effect of extreme winter temperature on daily non accidental cause mortality in Castile - La Mancha from 1975 to 2003, for all ages. Quantitative analyses were performed using ARIMA models, with other covariates, such as influenza, pressure trends, relative humidity, and cold wave duration and chronological number.

RESULTS

There were two mortality peaks: a short-term peak (with a lag of 3 to 7 days); and a longer term peak (of under two weeks). Excess mortality during cold waves was around 10% per degree centigrade below the threshold temperature for all the provinces except Guadalajara, where an increase of only 4.61% was detected. Mortality increased in response to rises in cold-wave duration and relative humidity. Cold waves occurring at the end of the "winter" season caused the greatest mortality.

CONCLUSIONS

This study confirms that daily mortality in Castile - La Mancha increases during cold waves. Efficient cold-wave prevention plans must therefore be implemented. Such plans should be based on in-depth knowledge of the causes that underlie and modulate the relationship between low temperatures and health effects.

Comparison between two methods of defining heat waves: a retrospective study in Castile-La Mancha (Spain)

INTRODUCTION

Following the 2003 heat wave, many European countries implemented heat-wave prevention plans. A number of aspects can prove fundamental in determining the effectiveness of such plans, and of these we sought to analyse the criteria used to define threshold temperatures and trigger a higher level of intervention.

METHOD

Retrospective study of the days on which heat-wave thresholds were exceeded during the period 1974-2003 was conducted. We compared when and at what level the heat-wave prevention plan would have been activated using a statistical-meteorological criterion (as applied by the Spanish Ministry of Health & Consumer Affairs) versus a temperature-mortality criterion.

RESULTS

The number of days on which the threshold was exceeded was far higher when the temperature-mortality criterion was applied. The temperature percentile at which a heat wave occurred was different for each province analysed and was inversely proportional to its respective ageing index. Using both criteria, there was an increase in heat-wave days per decade.

CONCLUSION

The establishment of a heat-wave threshold temperature must be based on knowledge of the cause-effect relationship between temperature and the health of a given population. Mortality is an appropriate indicator of population health. The future effects of climate change render it essential for this relationship to be studied on a local scale, so as to enable truly efficient prevention plans to be drawn up.