A onda de calor de Agosto de 2003 e os seus efeitos sobre a mortalidade da população portuguesa

RESUMO - Portugal continental, como outros países europeus, foi afectado por uma onda de calor de grande intensidade no Verão de 2003, com efeitos na mortalidade da população. O excesso de óbitos associados à onda de calor foi estimado pela comparação do número de óbitos observados entre 30 de Julho e 15 de Agosto de 2003 e o número de óbitos esperados se a população tivesse estado exposta às taxas de mortalidade médias do biénio 2000-2001 no respectivo período homólogo. Os óbitos esperados foram calculados com ajustamento para a idade. O número de óbitos observados (O) foi superior ao número esperado (E) em todos os dias do período estudado e o seu excesso global foi estimado em 1953 óbitos (excesso relativo de 43%), dos quais 1317 (61%) ocorreram no sexo feminino e 1742 no grupo de 75 e + anos (89%). A nível distrital, Portalegre teve o maior aumento relativo do número de óbitos (+89%) e Aveiro o menor (+18%). Numa área geográfica contínua do interior do território (Guarda, Castelo Branco, Portalegre e Évora) houve aumentos relativos superiores a 80%. Em termos absolutos, o maior excesso de óbitos ocorreu no distrito de Lisboa (mais cerca de 396) e no do Porto (mais cerca de 183). As causas de morte «golpe de calor» e «desidratação e outros distúrbios metabólicos» tiveram os aumentos relativos mais elevados (razões O/E de, respectivamente, 70 e 8,65). Os maiores aumentos absolutos do número de óbitos ocorreram no grupo das «doenças do aparelho circulatório» (mais 758), nas «doenças do aparelho respiratório» (mais 255) e no conjunto de «todas as neoplasias malignas» (mais 131). No período da onda de calor e no período de comparação, a percentagem dos óbitos que ocorreu nos hospitais (52% e 56%), no domicílio (32 e 33%) e em «outros locais» foi semelhante. A discussão sobre os factores que condicionaram a obtenção dos valores apresentados, relativos ao excesso de óbitos por sexo, grupo etário, distrito, causa e local da morte, permite concluir que os mesmos se afiguram adequados para medir a ordem de grandeza e caracterizar o efeito da onda de calor na mortalidade. O erro aleatório, medido pelos intervalos de confiança, e alguns possíveis erros sistemáticos associados ao período de comparação escolhido não deverão afectar de modo relevante as estimativas.

ABSTRACT - Like other European countries, Portugal was affected by a heat wave of high intensity during the summer of 2003. In spite of the alert that was launched and of the interventions, which were carried out, it was recognised that still during the month of August, the heat wave had caused severe effects regarding the mortality of the population. This report presents the statistics of those effects, which were based on the definite mortality data that are now available. The excess of deaths associated with the heat wave were estimated by means of comparing the number of deaths observed between July 30 and August 15 and the number of expected deaths. This, if the population was exposed to the average mortality rates of the 2000-2001 biennium in what concerns the respective homologue period (30.07 to 15.08). Adjustment for age group was used to derive the number of expected deaths, because an alteration of the population structure was present, namely in the 75 year-old and + age groups within that biennium and 2003. Although, estimates have been carried out using other periods of comparison, the authors consider that the biennium of 2000- 2001 produced the most adequate estimation. The number of deaths observed (O) was higher than the expected number (E). Between the days of July 30 and August 15, 3 significant statistical peaks of mortality occurred, which took place on day 2 (418 total deaths), on day 8 (464 total deaths) and on day 13 (439 total deaths). The global excess of deaths associated with the heat wave period was estimated in 1953 deaths (IC95: 1866-2039), out of which 636 (IC95: 530-744) occurred in males and 1317 (IC95: 1202-1435) in females. These figures corresponded to a relative excess of 43% of the expected deaths in both sexes, 27% in males and of 61% in females. The death excesses took place, mainly in the older age groups. In the 75 age group and more, there were about 1742 deaths whilst in the 65-74 age group there were about 161 deaths. One should point out that the significant statistically excess of mortality was not observed in the younger age groups, contrary to what was verified in the heat waves of 1981 and of 1991. The effects of the heat affected all the districts of Mainland, although with different intensities. The district of Portalegre had the highest increase in the number of deaths (+ 89%), whilst the minimum number occurred in Aveiro (+ 18%). Four districts, which had relative increases above 80% (Guarda, Castelo Branco, Portalegre and Évora), constituted a continuous geographical area in the inland of the territory. The districts ranked differently when the absolute increase of the number of deaths was considered. The highest death excesses occurred in the district of Lisbon (about 396 deaths) and in Oporto (about 183 deaths), which are higher populated areas. The districts of Bragança (about 24 deaths) and Vila Real (about 31 deaths) had the lowest death excesses. Some of the death causes were associated with the observed excess of mortality. Namely, the «heat stroke» had the relative highest increase (ratio O/E = 70.0), «dehydration and other metabolic disorders» also had a high relative increase (ratio O/E = 8.65), both causes with ratios much higher than the rest. On the contrary, the highest absolute number of deaths increase was observed in the group of the «diseases of the circulatory system» (758 deaths) and, among these, the «cerebrovascular diseases» (370.2 deaths), the «ischaemic heart diseases» (144.5 deaths) and the «heart failure» (118.0 deaths). The group of the «diseases of the respiratory system» (255 deaths) and the group of «malignant neoplasms» (131.2 deaths) held the following positions. The percentage of the deaths that occurred in the various locations was similar in the heat wave period and in the comparison period (hospitals — 2003: 52.6%; biennium 2000- 2001: 56.0%; at home — 2003: 32.2%; 2000-2001: 32.6%; other locations — 2003: 15.2%; 2000-2001: 11.3%). The number of deaths associated with the heat wave cannot be directly determined, it can only be estimated using the number of expected deaths within the same period of time, if the heat wave had not occurred. In fact, the deaths associated with specific causes, which are directly related with the «exposure to excessive natural heat», represent only a small proportion of the total number of deaths. The figures presented for the total number of deaths, per sex, age group, district and death cause seem to be adequate in order to indicate the greatness of the effect of the heat wave in mortality. It should be pointed out that these estimations have a degree of uncertainties related with randomness, which were meant to be measured by the respective confidence interval. There may be still present some systematic errors associated with the period of comparison that was chosen (2000-2001) and with other circumstances, which, however, should not affect the estimates in a relevant manner.