IJCS | Volume 31, Nº2, March / April 2018

130 Soares et al. Cardiovascular Mortality X Gross Domestic Product Int J Cardiovasc Sci. 2018;31(2)123-132 Original Article Being negative indicates an inverse relationship, that is, the higher the GDPpc, the lower the mortality from DCS. This evidences that the improvement in the socioeconomic indicators preceded the reduction in cardiovascular deaths. The behavior of the CBVD subgroup was similar to that of the DCS, regarding both the correlation index of GDPpc and the time lag. Regarding the IHD subgroup, the correlation indices, although significant, were not that close to the negative maximum value, and the optimal time lag was also a little shorter, around 18 years. These differences in IHD as compared to DCS and CBVD might be due to the lowest mortality rates from IHD in almost all municipalities throughout the study period. 21 This might have caused greater fluctuations in the IHD rates than in the others, which is even more evident when we observe that the municipalities with smaller populations have the lowest correlation indices and the greatest variations in optimal time lag. The increase in GDPpc might have influenced on the reduction of the deaths from DCS. This impact varied in the different Rio de Janeiro Statemunicipalities, in the Rio de Janeiro State regions, and even in the municipalities within the same region. In some municipalities, such as Carmo and Cordeiro in the Mountain region, and Nilópolis in the Metropolitan Belt, the 100-dollar increment in GDPpc was related to a reduction of more than 50 deaths per year fromDCS. In other municipalities, however, such as Angra dos Reis in the Ilha Grande Bay region, Macaé in the Northern region, and Cantagalo in the Mountain region, that same increment in GDPpc related to a reduction of less than 10 deaths per year from DCS. In two of those municipalities, that phenomenon can be explained by the great elevation in the GDPpc of Macaé and Angra dos Reis in the study period, because, despite having a reduction in death from DCS similar to that of other municipalities, their great elevation in GDPpc made the variation in deaths as compared to the GDPpc increase smaller. The CBVD as compared to the IHD stand out as the group with the highest reduction in the number of deaths per year, although the higher mortality rates fromCBVD in the initial years of the study should be considered. In addition, one can infer that the costs to prevent and reduce mortality from CBVD are lower than those estimated for IHD, because the reduction in the incidence of stroke, the major cause of death from CBVD, is closely related to the improvement in primary health care and arterial hypertension control, conditions affected by the global economic improvement reflected in GDP increase. 29-31 By providing important details when analyzing the Rio de Janeiro State municipalities, this study corroborates the clear inverse relationship between cardiovascular mortality rates and GDPpc. The inverse relationship of those variables has been suggested in the study 32 relating the Brazilian GDPpc between 1947 and 2004 to the mortality from IHD in the Rio de Janeiro State between 1980 and 2002, also showing the time lag between those variables. In addition, the use of the Human Development Index (HDI) showed an inverse relationship with the mortality rates from CBVD in the administrative regions of the Rio de Janeiromunicipality, and to every 0.05 reduction in the HDI, there was a 65% increase in the number of deaths from CBVD. 33 One limitation of this study is the quality variation in the completion of death certificate over time and in the municipalities studied. However, death certificates are the best mortality data source available. Amore serious limitation was the difficulty to obtain the economic data of the years before 1980, because they sometimes had a decennial periodicity and only those of the years of the IBGE census could be found, which determined the use of interpolation for the unavailable years. The compensation of the number of deaths from DCS, CBVD and IHD considering the deaths from IDC might have caused inaccuracy in the estimated mortality rates. Another limitation is the analysis with possible maximal time lag of 29 years, because, in some municipalities, the optimal time lag coincided with that value, and, thus, the actual value might have been greater; however, that happened in only 5 of 56 municipalities. Conclusion From 1979 to 2010, there was an important reduction in mortality from DCS in the Rio de Janeiro State municipalities, especially in the CBVD subgroup. The decrease in mortality from DCS was preceded by periods of GDPpc elevation, and the evolutionary variation of that indicator showed an important correlation with the reduction in mortality. A regional pattern for that correlation that indicated the importance of improving the population life conditions to reduce cardiovascular mortality could not be identified. Author contributions Conception and design of the research: Soares GP, KleinCR, Souza e SilvaNA, OliveiraGMM. Acquisition of