Weather affects public health in many ways and heat is one of the primary weather-associated threats to human life [1
]. High ambient temperature is associated with increases in both mortality and morbidity [2
]. For example, the death toll of the 2003 heat waves in Europe was estimated to be around 50,000 additional deaths [6
]. Heat waves, defined as extended periods of extreme high temperatures, exert a stronger influence on human health than isolated hot days [7
]. Generally, the heat effect is present particularly among the elderly, while the younger population is less vulnerable to heat stress (e.g., [9
]). Furthermore, there are distinct differences between the various diseases: while the heat effect is highest for cardiovascular and respiratory diseases, other illnesses seem to be less affected [9
Studies frequently report that the temperature-mortality relationship is stronger than the temperature-morbidity relationship, which also implies that heat episodes cause stronger increases in deaths than in hospital admissions [11
]. Some authors also investigated the influence of heat wave characteristics on human health and showed that more intense and longer heat waves feature stronger effects [2
]. Furthermore, an earlier occurrence of heat episodes in summer also led to a stronger effect [13
The pathophysiological mechanisms of the heat impact on the human body involve dehydration, increased blood viscosity and a higher risk of thrombo-embolic diseases [16
]. Moreover, high temperatures may impair the endothelial function, which in turn enhances the risk for cardiovascular events [18
]. Even if dehydration is avoided, the work required to maintain thermoregulation is increased, which stresses the cardiovascular system [19
While the previous results are generally well documented, there are a number of aspects regarding heat waves and health that are less frequently studied. For example, subtypes of diseases may be affected by heat via different mechanisms, which can lead to weaker or stronger physiological reactions to heat [4
]. Some studies also report stronger heat-related impacts on women than on men [21
]. Further subjects of interest are regional differences, since the effect of heat shows large geographical heterogeneity among cities or regions [9
Understanding how heat affects human health in its various aspects is of major importance to prepare for the future and to minimize the temperature-related health impacts in a changing climate. Scientific knowledge of heat wave-associated mortality in Germany was claimed to be insufficient so far [25
], since most studies focus on single heat episodes or only a subset of cities (e.g., [26
In our study, we analyze the impact of heat waves on ischemic heart diseases in Germany in the period 2001–2010. All IHD deaths and hospital admissions during this time period are considered in a nationwide database, which offers the advantage to assess reliable area-covering heat effects and to enable regional intercomparisons. Differences between mortality and morbidity as well as IHD subtype and gender differences are studied. Furthermore, the effects of heat wave characteristics (such as intensity, duration, and timing in season) on IHD are considered. The paper is organized as follows: a description of the data and methods used is given in Section 2
, results are presented in Section 3
, followed by a discussion and concluding remarks in Section 4
4. Discussion and Conclusions
In this study we provide evidence that heat waves in Germany are associated with significant excess in IHD mortality. On average, mortality increases during heat wave days by 15%, which is in line with results from studies regarding heat wave effects on total and cardiovascular mortality in the range of 10%–13% for other European countries, e.g., Austria [22
], Czech Republic [23
], the Netherlands [40
], and Sweden [41
]. The slightly higher values in this study may be due to different heat wave definitions and the restriction on persons with IHD. The mean temporal course of mortality during heat waves documents that the used heat wave definition is suitable to capture heat waves adequately. The link between heat and mortality is immediate with short time lags between the onset of the heat event and the mortality increase. Similar results are obtained by [23
] for the Czech Republic and by [42
] for Italy.
In contrast to the mortality results, no heat wave effect on hospital admissions could be found. Previous studies dealing with morbidity show a wide range of heat effects. Some authors find equally strong effects compared to mortality [43
], but the majority of studies report weaker impacts [42
]. Smaller or even absent impacts are observed particularly in studies that investigate heat wave effects on mortality and morbidity simultaneously [11
]. The reasons for this behavior are still not well understood, one hypothesis states that deaths during heat events possibly occur very rapidly. Especially, older people living alone may not receive adequate care and medical attention during heat waves, and, therefore, may die at home before they have the chance to reach the next hospital [11
]. However, further research is necessary to verify this hypothesis. The analyzed IHD morbidity data set shows a strong weekly course, with most admissions on Mondays and Tuesdays. More hospital admissions on Mondays are found also in other studies and are at least partially due to events with uncertain dates being coded as taking place on Mondays [51
]. Furthermore, the observed weekly course is an indicator for the fact that many hospital admissions are being planned. Unfortunately, in our study it was not possible to separate planned from spontaneous (acute) hospital admissions, so that possible heat effects could be superimposed by other effects that are controlled by administrative factors. Therefore, we cannot exclude that using other morbidity indicators rather than hospital admissions, also IHD morbidity can be related to heat events.
Regional variations in the heat wave effect on IHD mortality were found to be relatively large across Germany. This is in line with the outcome of other studies, which also observe geographical differences of heat impacts, for example between coastal and inland areas in Italy [42
]. The obtained findings underline the importance of long-term multi-city or multi-region studies, as investigations of heat waves for single cities or districts can mislead in the representativeness of results. Parts of the observed pattern can possibly be explained by climatic factors. For example, the relatively low heat wave effects in some Northwestern regions may be related to the maritime influence of the North Sea, which prevents an occurrence of extremely high temperatures. The strong heat wave impacts in Western Germany are at least partially due to the exceptional heat waves of August 2003, July 2006, and July 2010, each time featuring above-average mortality increases in this region. Additionally, longer average duration of heat waves in the West may also contribute to the elevated impacts. On the other hand, not all observed structures can be explained by climatic factors. For example, heat waves of similar intensity, duration, and timing in season are noticed to cause varying mortality increases in different regions. Potential reasons for the observed differences include uncertainties regarding the meteorological and mortality data sets, as well as population-specific and socioeconomic issues. Firstly, the used weather stations are primarily representative for their local environment but to a lesser extent for a larger area (for which health data were available). The thermal environment is further modified by population density (urban heat island effect) and building properties. Therefore, the individual exposition to heat is difficult to determine, particularly in spatial heterogeneous regions. Additionally, meteorological conditions not captured by the applied heat wave definition, e.g., high humidity or exceptionally high day/night-time temperatures, may be able to influence the results. Different regional reliability of the mortality data may also contribute to the observed pattern [52
], as the compiling of the mono-causal mortality statistics was done by the federal states individually. Furthermore, differences in age structure, medical care quality, distance to the next hospital, socioeconomic status, and health awareness are of major importance and play a strong role for the spatial heterogeneity of heat wave effects. For example, the relatively low heat wave impacts in the city of Berlin compared to stronger effects in the surrounding rural region of Brandenburg, can at least partially be explained by differences in age structure (people aged above 60 years in 2010: 24.5%vs.
]) and the easier and faster access to medical care in the former area.
We were able to show significant gender differences regarding IHD mortality, with females being more vulnerable to heat than males. Most studies also find higher heat impacts for women [21
], while only some authors report the opposite effect (e.g., [54
]). Possible reasons for the gender differences involve population-specific factors (different age, social structures, and ethnic variability), as well as physiological mechanisms according to menopause [55
] or pre-existing diseases, such as diabetes [56
]. We observed smaller heat wave effects on myocardial infarctions (I20–I22) than on chronic ischemic heart diseases (I24–I25). Thus far, studies investigating heat influence on subgroups of IHD simultaneously are relatively rare. Some studies report smaller heat impacts on myocardial infarctions than on the overall number of ischemic diseases [4
], but others do not [20
]. In Germany, 55% of I20–I22 deaths and 43% of I24–I25 deaths are attributable to men. Moreover, people dying from chronic ischemic heart diseases are on average older than people dying from myocardial infarctions. Therefore, we assume that the differences between IHD subtypes are at least partly due to different gender and age structures. As for data privacy protection reasons our data set did not include age and gender for IHD subtypes, we were not able to follow up this hypothesis. Future research should incorporate age and gender specifics to verify this assumption.
We demonstrated that heat wave characteristics contribute substantially to IHD mortality. More intense heat waves lead to stronger mortality increases, a fact that is well-known for different regions of the world [7
] and can be attributed to the positive temperature-mortality relationship for high ambient temperatures (e.g., [9
] for 15 European cities). Additionally, the duration of heat waves is associated with the magnitude of excess mortality, which is also confirmed by previous investigations [8
]. In our study, we found no timing in season effect of heat waves on mortality, contrary to results by other authors [13
]. However, differences regarding intensity and duration between first and subsequent heat waves suggest at least a small masked timing effect. Seasonal acclimatization is a major reason for the timing effect and the elderly are known to have impaired acclimatization abilities. Since, in our study, the timing effect appears to be more present for men and persons with myocardial infarctions, we hypothesize that this is possibly due to the younger age of these groups, which leads to a better within-season acclimatization and reduced heat effects in later-in-season heat waves.
Previous studies have shown improved public health response due to the implementation of Heat Health Warning Systems (HHWS) [59
]. In Germany, heat warnings are issued by the national meteorological service since 2005. However, only in some administrative regions an intervention chain was triggered in case of a warning by the responsible health authorities. Rough estimates show that there are only minor differences in the heat wave impact on IHD in Germany between the periods 2001–2005 and 2006–2010 (before and after the introduction of a HHWS in Germany). From the short period that has been analyzed before and after the introduction of the warning system and the differences in the character of the heat waves, e.g., heat waves were on average longer and a little bit more intense in the second five years, we cannot draw conclusions about the effectiveness of the warning system based on the available data.
Uncertainties regarding the estimation of heat wave effects on mortality involve possible confounding factors such as air pollution and forward displacement of deaths as well as changing structures in population and socioeconomic factors. Air pollution (particularly ozone and particulate matter) is reported to contribute to the impact of heat on mortality (e.g., [62
]). We did not account for air pollution due to the lack of data. However, current research shows that heat wave effects in Europe are relatively robust to air pollution, with air pollutants contributing only minor to mortality during heat waves compared to temperature effects [65
The term “mortality displacement” describes an increase in mortality during heat waves followed by a subsequent reduction after the heat wave. Such a pattern is frequently observed for heat episodes and suggests a short-term forward shift of mortality in people with impaired health who presumably would have died a few days or weeks later [40
]. The magnitude of the mortality displacement effect varies strongly between individual heat waves. In a review, effects for major heat waves are reported to be mostly less than 15% [66
], but other authors refer higher values (e.g., [67
]). We did not investigate mortality displacement as we concentrated on daily mortality increases during heat waves and did not assess absolute numbers of deaths. However, the three weeks following the average heat wave show no indication of substantial mortality displacement (Figure 2
Changing structures in age and socioeconomic factors during the period 2001–2010 may also influence the results. The increasing number of people aged above 60 years (2001: 24.1%, 2010: 26.3% [53
]) is likely to enhance the subgroup of vulnerable persons. On the other hand, improved health care systems and health awareness accompanied by an increased use of air conditioning act to reduce the heat effects on mortality. However, since the study period of 10 years is comparatively short, the impacts of these potential changes are assumed to be limited.
In conclusion, the presented results show that (a) heat waves in Germany are associated with increased IHD mortality rates; (b) heat wave characteristics are strongly modulating the mortality results; and (c) heat impacts are different regarding individual regions, genders, and diseases.
The strengths of the current study consist in the estimation of heat effects specifically for ischemic heart diseases, which was analyzed relatively rarely to date, and its national-scale design which enables the estimation of area-covering heat effects and regional comparisons.
Climate change will most likely enhance the number and intensity of heat waves in Europe [68
]. Therefore, the number of deaths attributed to heat is expected to increase considerably. Demographic changes are projected to additionally elevate the heat effect (higher number of potentially vulnerable persons), while long-term physiological adaptation to climate change will act in the opposite direction. Studies investigating the heat-mortality relationship are of major importance to design preventive measures, which are able to minimize the future heat wave risks. Heat health warning systems, heat wave action plans, public adaptation strategies and an enhanced use of air conditioning could help to reduce the individual heat burden. Furthermore, the heterogeneity in the heat wave impacts suggests that the implementation of gender and region specific heat wave action plans could be an appropriate strategy.