The European Journal of Public Health Advance Access published online on November 28, 2007
The European Journal of Public Health, doi:10.1093/eurpub/ckm108
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Impact of high temperatures on hospital admissions: comparative analysis with previous studies about mortality (Madrid)
C. Linares1 and J. Díaz2
Department of Education for Sustainable Development of the Madrid City Council, Autónoma University General Foundation of Madrid, Madrid, Spain
1 PEAC. Centro Nacional de Epidemiología. Instituto de Salud Carlos III. Madrid
2 Escuela Nacional de Sanidad. Instituto de Salud Carlos III. Madrid
Correspondence: Dr Julio Díaz Jiménez. Escuela Nacional de Sanidad. Instituto de Salud Carlos III. C/ Sinesio Delgado 6. 28029 Madrid. Tel: +34 918222280, Fax: +34 913877809, e-mail: julio.diaz{at}uam.es
Received November 30, 2006, accepted October 1, 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Background: Heat wave prevention plans are traditionally implemented according to a temperature limit above which mortality begins to rise. Although these prevention plans are obviously designed to avoid deaths, it is also necessary to establish the impact of extreme temperatures on hospital admissions in order to put hospital alert plans into action for dealing with people affected by heat wave victims.
Methods: We used data on daily emergency admissions between May and September, from 1995 to 2000, in the Hospital General Universitario Gregorio Marañón in Madrid. The causes for admission were considered as ‘organic’ (International Classification of Diseases, ICD-9: 1–799), circulatory (ICD-9: 390–459) and respiratory (ICD-9: 460–519). We stratified them according to the following age groups: all ages, from 0 to 10, 18 to 44, 45 to 64, 65 to 74 and above 75 years. The methodology used was Autorregresive Integrated Moving Average (ARIMA) modelling, including variables related to atmospheric pollution, seasonality and trends.
Results: The results show that the temperature above which hospital admissions soar coincides with the temperature limit above which mortality sharply rises, which, in turn, coincides with percentile 95 of the maximum daily temperature series for summer months. The pattern of hospital admissions is completely different from that of mortality. The rise in hospital admissions due to all causes and age groups is clearly smaller than that detected for mortality.
Discussion: These results suggest that people die rapidly from circulatory diseases before they can be admitted to hospital. This datum is vital with regard to implementing prevention plans prior to the arrival of the heat wave, if they are to effectively reduce mortality.
Keywords: heat wave, hospital admissions, mortality, prevention plans, temperature
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
While studies addressing the impact of extreme heat on health have been conducted for many years, e.g. the review by Basu and Samet covers a total of 98 papers dating from as far back as 1957 to 20021 and from a public health point of view, the 2003 heat wave represented a real landmark in an essentially European context. The excess mortality registered in Europe in the summer of 2003 marks a turning point in the design and implementation of European prevention plans, which until then had been a distant cry from those in place in US cities.2 Indeed it can be said that, prior to the summer of 2003, only Lisbon and Rome were provided with genuine heat-wave alert systems.3,4 At present, practically all European cities possess extreme-temperature prevention and alert plans, which come into operation when weather forecasts indicate the likelihood of the safety thresholds set for the specific locality or town being exceeded.
Most of these prevention plans have determined the temperature above which alert plans are set in motion, in relation to the associated rise in mortality.5,6 But it has not been established whether this mortality increase temperature threshold is the same one above which hospital admissions start to rise.
Furthermore, although many studies have analysed the effects of temperature on mortality, as we previously stated, fewer ones consider the increase in hospital admissions and the medical reasons for these. In this sense we can highlight the studies conducted in the United States7–9 or in the United Kingdom,10 which clearly identify the pathologies involved as well as the special risk.11 But even fewer studies analyse effects of heat waves on mortality or compare them with hospital admissions occurring in this period. In some cases, there is a clear relationship between mortality and admissions. Thus, for example, in the analysis of the impact of the heat wave in England,12 an excess mortality of 17% was found along with a similar excess hospital admission rate (16%), or there was a sharp increase in emergency admissions,13,14 but in relation to the 1995 heat wave in England and Wales, other studies15 showed that an excess mortality of 11.8% was accompanied by a mere 1.9% increase in hospital admissions.
Two papers have recently been published16,17 that establish a clearly differentiated behaviour pattern between mortality due to heat waves and hospital admissions for the same reason, and one of these studies questions whether the excess mortality due to circulatory diseases observed during heat wave temperature extremes is real or mere artefact.17
The main objective of a heat wave prevention plan is obviously to prevent deaths, but the influence of these temperature extremes on hospital admissions should also be established, as well as which pathologies are affected by them and what increase in admissions is to be expected, in order to reinforce, where necessary, the hospital services involved.
A study for Madrid analysed the influence on mortality of temperature extremes on all age groups,18 in the over-65s and over-75s,19 in the 45–64 age group20 and in the under-10s,21 but did not deal with the effect of extremely high temperatures on hospital admissions. The objective of the current article is double: on one hand, to analyse whether the temperature above which mortality rises coincides with the temperature limit above which hospital admissions increase. And on the other, whether the mortality behaviour pattern analysed in these studies coincides with that of hospital admissions. To this end, we analysed the influence of extremely high temperatures on hospital admissions in Madrid and compared the results with those already obtained for mortality. We also analysed the effects of atmospheric pollutants during high-temperature events. Published studies shown the association between other pollutants such as NO222 or particles and mortality23 or hospital admissions,24 especially in summer. Moreover, during the 2003 heat wave in Europe, the effects of certain atmospheric pollutants such as ozone had a dramatic effect on mortality.25
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The dependent variables are made up of the daily number of emergency hospital admissions to the Hospital General Universitario Gregorio Marañón in Madrid. We analysed admissions between May and September, from 1995 to 2000, as extreme heat events are most likely to occur in Madrid during these months.26 We analysed 918 days and a total of 49 572 admissions. This hospital was taken as an indicator of the emergencies in the whole city, due the fact that this is the reference hospital for Madrid, taking most of the emergency admissions of the region.27 Causes of admission were defined according to the International Classification of Diseases (ICD), 9th Revision, and grouped into the following categories: total organic disease causes (ICD-9: 1–799), circulatory disease admissions (ICD-9: 390–459) and admissions due to respiratory system diseases (ICD-9: 460–519). This was stratified according to the following age groups: 0–10, 18–44, 45–64, 65–74 years and over-75s.
Meteorological variables included in this study (maximum daily temperature, minimum daily temperature and relative humidity at 7 AM) were provided by the Spanish National Institute of Meteorology, from the Madrid-Retiro observatory of reference. This observatory is located in the vicinity of the Gregorio Marañón Hospital.
The air-pollutant concentrations were computed as daily average values and were provided by the City Council. Daily mean concentrations of nitrogen oxides (NOx), sulphur dioxide (SO2), particles with a median aerodynamic diameter of <10 µm (PM10) and ozone (O3) were considered, as furnished by the 24 monitoring stations making up Madrid's Municipal Automatic Air Pollution Monitoring Grid.
We analysed the aforementioned data in three phases. In the first of these we created scatter plots among the environmental variables and the different causes of admission, which indicates the type of relationship (linear, curvilinear, etc.) existing among the variables. But this relationship observed in the scatter plots might be due to the existence of seasonalities or analogous autocorrelations between the series, and in order to avoid this, we conducted a pre-whitening of the series.28 Between the pre-whitened functions we established the cross-correlation functions (CCFs) that indicate the lags in which there is a non-spurious statistical association (the series were previously pre-whitened). We then used modelling to quantify the effect among the different environmental variables and the causes of admission according to the different age groups.
The scatter plots of the different variables were created as follows: first, different intervals were established between the minimum and the maximum values of the independent variable. After that, the mean value of each interval, previously defined, was represented in the x-axis. Next, in the y-axis the mean value of the dependent variable (causes of emergency hospital admissions) in the corresponding interval was represented. Finally, the points were adjusted using the loess method (Statistical Product and Service Solutions (SPSS) application for performing scatter plots). So, the points that appear in the final scatter plot represent a number of 118.260 emergency hospital admissions in total, corresponding to each corresponding interval of maximum daily temperature.
A prior studies of hospital admissions due to all causes24 showed that the relationship between hospital admissions and SO2 is of a logarithmic nature, and the logarithmic transformation of this variable was therefore required. In relation to NOx and PM10, the relationship established was linear, and so no transformation of these variables will be necessary.
Ozone levels fit a quadratic curve when related to hospital emergency admissions;24,27 minimum admission occurred when ozone was at a concentration of 50 µg m–3. This value observed as the basis for defining high- and low-ozone concentrations and corresponds with percentile 95 of the mean concentrations values of ozone in this study period:
|
|
|
|
The only thing that remains to be analysed through the scatter plots, and which constitutes the objective of this study, is the relationship between the different causes of admission according to the different age groups and temperature.
In order to ensure that the relationship between the temperature and the different admission variables is the result of a real cause–effect relationship from the statistical point of view, CCFs need to be established among the pre-whitened series, as was previously stated. We used was the Box–Jenkins pre-whitening method,28 which consists of establishing the correlations, with different lags in time between the residuals of the series.
The modelling method used to quantify the effects of the independent variables on hospital admissions was the ARIMA modelling system with external variables.29 There were several reasons for choosing this type of modelling. First, this methodology was followed in the analysis of mortality and extreme temperatures,18,19,21 and the results obtained in this study can therefore be compared with those obtained for mortality. Second, other studies conducted with both Poisson models30 or GAM models31 provide similar results to the ones obtained with ARIMA models.
The ARIMA modelling with exogenous variables consists of, including in a univariate ARIMA model,29 the environmental variables in the lags that proved to be significant in the CCFs. With regard to primary pollutants, previous studies27 show that the effect thereof occurs between lags 0 and 4, whereas in the case of ozone, this effect can be delayed up to lag 9. Temperature is included in the lags determined by the CCFs conducted. It is also monitored according to seasonalities on a yearly, half-yearly and quarterly basis by means of sine and cosine functions with periodicities of 365, 180 and 90 days, respectively. It is also monitored according to overdispersion and trend of the admissions series. In order to control weekly variation, dummy variables for weekdays were included in the models.
Goodness-of-fit was evaluated through simple autocorrelation functions (ACF) and partial autocorrelation functions (PACF) of the residuals, and we also used Akaike's information criteria.32 The analysis was carried out using SPSS statistics pack.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Table 1 shows the descriptive statistics corresponding to the environmental variables considered in this study. Table 2 shows the descriptive statistics of the different causes of admission analysed according to the different age groups.
|
|
The scatter plots designed for maximum daily temperature on x-axis and the different causes of admission on y-axis show a similar behaviour pattern for all the causes analysed and all the age groups. All these scatter plots are characterized by a V-shaped relationship with a minimum admission temperature corresponding to 34°C. The branch on the right, corresponding to heat, shows at the same time two slopes, one between 34 and 36°C and another starting at 36°C in which the slope increases along with the bigger one. Figure 1 shows this graph for organic causes and all age groups. This temperature (36°C) is the one used to determine a heat wave. That is to say, a ‘heat wave’, from the point of view of hospital admissions, is considered to occur when the maximum daily temperature surpasses the 36°C threshold. Mathematically, and in order to evaluate the impact of temperatures >36°C, we defined the following new variable:
|
|
|
|
|
As we mentioned in the Methods section and in order to establish a cause–effect relationship from the statistical point of view, we conducted the CCFs between the different causes of admission, according to age groups and temperatures. Of the total of 18 CCFs conducted, only 3 showed correlation coefficients with statistical significance. These correspond to: organic causes in all the age groups (the association being established in the zero lag); organic causes in the over-75s age group (lag 1) and respiratory causes (lag 0), also in the over-75s age group.
If the corresponding ARIMA models are used to calculate the quantitative impact of the different environmental variables considered upon hospital admissions, the estimator values of the ARIMA modelling, with their corresponding confidence intervals, will be the ones shown in table 3. Therein it is highlighted that the variable Thwave is only significant in the same case as the maximum temperature in the CCFs. In the remaining groups and causes, the variables related to atmospheric pollution are the ones that appear to influence hospital admissions. As a general behaviour pattern, it was found that primary pollutants have a short-term influence (lags 0–4) as opposed to the medium-term effects of ozone (lags 7–9). Besides the effect of ozone, for an increase of 10 µg m–3 over the value of 50 µg m–3, as a daily average, it is similar to a 1°C increase in the threshold temperature for hospital admissions, as can be seen on comparing the effects of organic diseases on mortality in all the age groups. The effect of ozone is greater than that of primary pollutants. It should be kept in mind that the values corresponding to SO2 refer to their neperian logarithm and therefore cannot be compared with the other pollutants.
|
Table 4 quantifies the effect of temperature on hospital admissions for each degree of maximum daily temperature above 36°C. As can be seen, this increase is very small in all age groups, reaching the highest values in the over-75s, resulting from respiratory system diseases. There is clearly no relationship between heat and admissions due to circulatory causes in all the age groups analysed.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The first result obtained from this article is that the temperature at which mortality soars (36.5°C) practically coincides with those obtained in this study for the sharp rise in hospital admissions (36°C). In both cases this temperature practically coincides with percentile 95 of the series of maximum temperatures in summer months of the climatology series (1985–2000). This ratifies what was already observed for mortality, i.e. the importance of acclimatization of people to the local conditions they live in,33 as they can fall ill or even die at temperatures to which they are unaccustomed. This coincidence of thresholds is therefore logical and is based upon the biological processes involved during high temperatures.34
The CCFs pattern, however, shows a clear difference between hospital admissions and mortality. The lack of any association, except in determined cases, and never due to circulatory diseases, contradicts what might initially be expected from the mortality behaviour pattern. High temperatures can increase platelet and red cell count, blood viscosity and plasma cholesterol level during heat stress, and mortality from coronary and cerebral thrombosis,35,36 and an association should therefore be detected with admissions due to circulatory diseases, which is not at all the case in our study. The only possible explanation would appear to be that the aforementioned pathologies involve rapidly fatal health outcomes with a short-time interval between exposure to high temperature and deaths.17 That is to say, the patients die before reaching hospital. Another explanation might be that there were no available hospital beds, this increase therefore not being detected, but extreme temperature events generally occur when there is little pressure due to admissions,16,24 with the resulting availability of beds. Moreover, the effect of this lack of beds would not only be noted in circulatory diseases, but also in ones associated with the respiratory system or in admissions for all causes, and here the results do present an association.
The fact that this association is seen in respiratory system diseases not only corroborates what we previously stated in relation to free hospital beds but also indicates that the same happens with admissions as with mortality, that is, that the effects of high temperatures on mortality resulting from respiratory diseases is more long term, because of the biological mechanisms involved, such as an exacerbation of previous pathologies, than those owing to circulatory diseases, and these people are therefore more likely to be admitted to hospital. In this sense, studies conducted in Madrid for over-75s and 65–74 age group clearly show that the mortality resulting from circulatory diseases occurs in lags 0 and 1, whereas for respiratory system diseases this reaches 4.19
Furthermore, pollutant behaviour does show a similar pattern in mortality and in hospital admissions, perhaps due to the fact that the effect thereof is more long term than that of temperature and that the biological mechanisms involved are of a different nature.27 Moreover, the associations detected between pollutants in summer and hospital admissions once again highlight the interaction between high temperatures and primary pollutants37–39 and the important role played by pollution in these situations.40 Particularly noteworthy is the case of ozone; first because atmospheric conditions associated with extremely high temperatures are usually accompanied by high concentrations of this secondary pollutant.41 The combined effect of ozone and high temperatures was especially obvious in the heat wave in Belgium in the summer of 199442 and this was corroborated by studies dealing with excess mortality in France in the summer of 2003, although other studies43 that compared the relative contributions of ozone and temperature to this joint excess risk found that the contribution of ozone varied according to the city, ranging from 2.5% in Bordeaux to 85.3% in Toulouse. These results confirmed that in urban areas ozone levels have a non-negligible impact over population health. The effects of ozone concentrations appear at medium-long term (lag 7 and lag 9); this fact is coherent with previous studies in Madrid24,27 and the subjacent biological mechanisms of ozone.
The idea that mortality resulting from high temperatures occurs so rapidly that the patients die before being admitted to hospital is confirmed by the quantitative results of the models. Thus, in the case of mortality according to all age groups and all causes, the rise in mortality detected for each degree that the temperature surpasses the 36.5°C threshold in Madrid was 21.5%,18 whereas for admissions, this did not reach 5%. This is much more evident in deaths resulting from circulatory diseases. In the 75-year age group in Madrid, there was a 34.1% increase for each degree by which maximum daily temperature surpassed the threshold, in the case of women over 75 years,19 whereas in the case of admissions, we did not even establish a statistically significant association. The only case in which the values of the estimators obtained for the admissions and mortality models are comparable is when respiratory diseases in over-75s are considered.
Studies by Kovats et al. and Mastrangelo et al.16,17 show results similar to those encountered herein, in relation to the differentiated behaviour between mortality and admissions. However, similar results have also been found by other authors such as Johnson in England and Wales.44 In the heat wave in London in 2003, these authors detected a 59% rise in mortality in over-75s, whereas admission in this same group only rose by 16%. More telltale with regard to the discrepancy between mortality and admissions is the case of Chicago. In the 1995 heat wave, a mortality increase of 147%45 was accompanied by a mere 11% rise in hospital admissions.7
To conclude, the results of this study show a differentiated behaviour pattern between mortality and hospital admissions in the case of extreme temperatures. The evidence found in the bibliography is conclusive in that the excess mortality caused by cardiovascular diseases is the main cause of deaths due to high temperatures and the fact that this effect is not reflected in admissions is because the patients die from the biological processes involved before being admitted.
The results obtained in this study are highly relevant from the prevention point of view. Prevention should involve alerting the social services when the first symptoms of any pathology make an appearance, above all in the over-65 age group. Prevention measures are seen to be of vital importance if we consider that the highest risk group comprises people who live alone. The risk population should be warned of the importance of hospital assistance at the least symptoms of indisposition.
Furthermore, the meteorological services can warn of the arrival of a heat wave up to 72 h previously and with a high degree of reliability. The results found herein appear to indicate that this is a crucial moment for prevention planning. We must not wait until the heat wave occurs, but rather should act when it is predicted by the meteorological services, between 48 and 72 h beforehand in the case of extreme heat events.
Conflict of interest: None declared.
Key points
|
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
1 Basu R, Samet JM. Relation between elevated ambient temperature and mortality: a review of the epidemiological evidence. Epidemiol Rev (2002) 24:190–202.
2 Kalkstein LS. Description of our Heat/Health Watch-Warning Systems: their nature and extent, and required resources. In: CCASh Workshop on Vulnerability to Thermal Stress (2002) May 5–7. Germany: Freiburg.
3 WHO. Heat waves: risks and responses. In: Health and Global Environmental Change (2004) Series No.2.
4 Pirard P. Heat wave: a climatic deadly phenomena that can be prevented. Enfermedades Emergentes (2003) 5:145–6.
5 Pascal M, Laaidi K, Ledrans M, et al. France's heat health watch warning system. Int J Biometeorol (2006) 50:144–53.[CrossRef][Web of Science][Medline]
6 Díaz J, García-Herrera R, Linares C, et al. The impact of summer 2003 heat wave in Iberia: how should we measure it? Int J Biometeorol (2006) 50:159–66.[CrossRef][Web of Science][Medline]
7 Semenza JC, McCullough JE, Flanders WD, et al. Excess hospital admissions during the July 1995 heat wave in Chicago. Am J Prev Med (1999) 16:359–60.[CrossRef][Web of Science][Medline]
8 Palecki MA, Changnon SA, Kunkel KE. The nature and impacts of the July 1999 heatwave in the midwestern United States: learning from the lessons of 1995. Bull Am Meteorol Soc (2001) 00:1353–67.
9 Schwartz J, Samet JM, Patz JA. Hospital admissions for heart disease: the effects of temperature and humidity. Epidemiology (2004) 15:755–61.[CrossRef][Web of Science][Medline]
10 Ellis FP, Prince H, Lovatt G, et al. Mortality and morbidity in Birminghan during the 1976 heatwave. QJM (1980) 49:1–8.
11 Koken PJ, Piver WT, Ye F, et al. Temperature air pollution, and hospitalisation for cardiovascular diseases among elderly people in Denver. Environ Health Perspect (2003) 11:1312–7.
12 Johnson H, Kovats RS, McGregor G, et al. The impact of the 2003 heat wave on mortality and hospital admissions in England. Health Stat Q (2005) 25:6–11.[Medline]
13 Trejo O, Miro O, de la Red G, et al. Emergency department activity during the 2003 summer heat wave. Med Clin (Barc) (2005) 125:205–9.[CrossRef][Medline]
14 Dhainaut JF, Claessens YE, Ginsburg C, et al. Unprecedent heat-related deaths during the 2003 heat wave in Paris: consequences on emergency departments. Crit Care (2004) 8:1–2.[CrossRef][Web of Science][Medline]
15 Rooney C, McMichael AJ, Kovats RS, et al. Excess mortality in England and Wales, in Greater London, during the 1995 heatwave. J Epidemiol Community Health (1998) 52:482–6.[Abstract]
16 Kovats RS, Hajat S, Wilkinson P. Contrasting patterns of mortality and hospital admissions during the hot weather and heat waves in Greater London, UK. Occup Environ Med (2004) 61:893–8.
17 Mastrangelo G, Hajat S, Fadda E, et al. Contrasting patterns of hospital admissions and mortality during heat waves: are deaths from circulatory disease a real excess or an artifact? Med Hypotheses (2006) 66:1025–8.[CrossRef][Web of Science][Medline]
18 García-Herrera R, Díaz J, Trigo RM, et al. Extreme summer temperatures in Iberia: health impacts and associated synoptic conditions. Ann Geophys (2005) 23:239–51.
19 Díaz J, López C, Alberdi JC, et al. Heat waves in Madrid 1986–1997: effects on the health of the elderly. Int Arch Occup Environ Health (2002) 75:163–640.[CrossRef][Web of Science][Medline]
20 Díaz J, Linares C, Tobías A. Impact of extreme temperatures on daily mortality in Madrid (Spain) over the 45–64 age group. Int J Biometeorol (2006) 50:342–8.[CrossRef][Web of Science][Medline]
21 Díaz J, Linares C, López C, et al. Impact of temperature and air pollution on child mortality in Madrid. J Occup Environ Med (2004) 46:768–74.[CrossRef][Web of Science][Medline]
22 Sunyer J, Castellsague J, Sáez M, et al. Air pollution and mortality in Barcelona. J Epidemiol Community Health (1996) 50(1 Suppl):76–80.
23 Katsouyanni K, Touloumi G, Spix C, et al. Short-term effects of ambient sulphur dioxide and particulate matter on mortality in 12 European cities: results from time series data from the APHEA Project. BMJ (1997) 314:1658–63.
24 Díaz J, Alberdi JC, Pajares MS, et al. A model for forecasting emergency hospital admissions: effect of environmental variables. J Environ Health (2001) 64:9–15.[Web of Science][Medline]
25 Institut de Veille Sanitaire (I.N.V.S). Vague de chaleur de l'étè 2003: relations entre la température, pollution atmosphérique et mortalité dans neuf villes francaises. Accessed: 26 April, 2005. http://www.invs.sante.fr/publications/2004/psas9_070704/index.html.
26 Prieto L, García-Herrera R, Díaz J, et al. Minimum extreme temperatures over Peninsular Spain. Glob Planet Change (2004) 44:59–71.[CrossRef]
27 Díaz J, García R, Ribera P, et al. Modelling of air pollution and its relationship with mortality and morbidity in Madrid, Spain. Int Arch Occup Environ Health (1999) 72:366–76.[CrossRef][Web of Science][Medline]
28 Makridakis S, Wheelwright SC, McGee VE. Forecasting Methods and Applications (1983) San Francisco: Wiley and Sons.
29 Box GEP, Jenkins GM, Reinsel C. Time Series Analysis, Forecasting and Control (1994) Englewood Cliffs: Prince Hall.
30 Tobías A, Díaz J, Sáez M, et al. Use of Poisson regression and Box-Jenkins models to evaluate the short-term effects of environmental noise levels on daily emergency admissions in Madrid, Spain. Eur J Epidemiol (2001) 17:765–71.[CrossRef][Web of Science][Medline]
31 Díaz J, García R, Prieto L, et al. Mortality impact of extreme winter temperatures. Int J Biometeorol (2005) 49:179–83.[CrossRef][Web of Science][Medline]
32 Akaike H. A new look at statistical model identification. IEEE T Automat Contr (1994) 9:716–22.
33 Curriero FC, Heiner KS, Samet JM, et al. Temperature mortality in 11 cities of the Eastern of the United States. Am J Epidemiol (2002) 155:80–7.
34 Havenit G. Interaction of clothing and thermoregulation (review). Exog Dermatol (2002) 1:221–68.[CrossRef]
35 Keatinge WR, Coleshaw SR, Easton JC, et al. Increased platelet and red cell counts, blood viscosity, and plasma cholesterol level during heat stress, and mortality from coronary and cerebral thrombosis. Am J Med (1986) 81:795–800.[CrossRef][Web of Science][Medline]
36 Pan WH, Li LA, Tsai MJ. Temperature extremes and mortality from coronary heart disease and cerebral infarction in elderly Chinese. Lancet (1995) 345:353–5.[CrossRef][Web of Science][Medline]
37 Roberts S. Interactions between particulate air pollution and temperature in air pollution mortality time series studies. Environ Res (2004) 96:328–37.[Medline]
38 Samet J, Zeger S, Kelsall J, et al. Does weather confound or modify the association of particulate air pollution with mortality? An analysis of the Philadelphia data, 1973–1980. Environ Res (1998) 77:9–19.[Medline]
39 Katsouyanni K, Pantazopulu A, Toulomi G. Evidence for interaction between air pollution and high temperature in the causation of excess mortality. Arch Environ Health (1993) 48:235–42.[Web of Science][Medline]
40 Rainham DGC, Smoyer-Tomic KE. The role of air pollution in the relationship between a heat stress index and human mortality in Toronto. Environ Res (2003) 93:9–19.[Medline]
41 Ordoñez C, Mathis H, Furger M, et al. Changes of daily surface ozone maxima in Switzerland in all seasons from 1992 to 2002 and discussion of summer 2003. Atmos Chem Phys Discuss (2004) 4:7048–88.
42 Sartor F, Snacken R, Demuth C, et al. Temperature ambient, ozone levels and mortality during summer 1994 in Belgium. Environ Res (1995) 70:105–13.[Medline]
43 Filleul L, Cassadou S, Médina S, et al. The relation between temperature, ozone, and mortality in nine French cities during the heat wave of 2003. Environ Health Perspect (2006) 114:1344–7.[Web of Science][Medline]
44 Johnson H, CooK L, Rooney C. Mortality during the heat wave of August 2003 in England and Wales and the use of rapid weekly estimates. In: WHO Family of International Classifications Networks (2004) Iceland: Reykjavik. 24–30. WHOFIC/04.081.
45 Whitman S, Good G, Donoghue ER, et al. Mortality in Chicago attributed to the July 1995 heat wave in Chicago. N Engl J Med (1996) 335:84–90.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Michelozzi, G. Accetta, M. De Sario, D. D'Ippoliti, C. Marino, M. Baccini, A. Biggeri, H. R. Anderson, K. Katsouyanni, F. Ballester, et al. High Temperature and Hospitalizations for Cardiovascular and Respiratory Causes in 12 European Cities Am. J. Respir. Crit. Care Med., March 1, 2009; 179(5): 383 - 389. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||