We extracted total yearly cases of malaria and population size in Amazonas from the database maintained by the Brazilian Ministry of Health (DATASUS, 2004), National Foundation of Health (6–8), and Secretary of Surveillance in Health (3,9). Data from FMT-AM were extracted from the malaria logbooks (for the years before the Foundation started publishing the reports) and from the Quarterly reports (for the years that the Foundation published the reports). All malaria cases diagnosed and referred for treatment are maintained (1989–1994) and quarterly reports are published by the FMT-AM (10). Quarterly reports published from 1995 to 2004 provided the total number of malaria diagnoses, case-patients admitted to the hospital, and number of deaths. Data from 2005 and 2006 were obtained by one of the authors (M.R.F. Costa) directly at FMT-AM (Subgerência de Arquivos Médicos e Contas Hospitalares).The hospital protocol is to exclude mixed infections by additional testing. We collected and tabulated data from these sources by year, parasite species, admissions, and percent admissions (Table). Percent admission was calculated as the total number of case-patients admitted to the hospital due to the specified parasite, divided by the total number of malaria cases caused by that same parasite in FMT-AM during that year, multiplied by 100.

Figure 1

Figure 1. Malaria diagnoses (in thousands) according to blood smears positive for Plasmodium falciparum or P. vivax in Amazonas, Brazil (open squares), and Fundação de Medicina Tropical do Amazonas (FMT-AM) (solid triangles), 1980–2006.

Malaria cases from all causes in Amazonas, 1980–2006, are presented in Figure 1. An irregular increase is noted since 1988, reaching a peak in 1999, followed by a decline in 2001, only to rise again in the following years. A decrease was observed in 2006, but the data are not final. Figure 1 also shows the total number of malaria cases diagnosed at FMT-AM; fluctuations observed are very similar in direction and relative magnitude to those found statewide.

Figure 2

Figure 2. Malaria diagnoses by parasite species, Plasmodium falciparum (solid circles) and P. vivax (open circles), at the Fundação de Medicina Tropical do Amazonas (FMT-AM), Amazonas, Brazil, 1989–2006. A) Total numbers of diagnoses...

The number of infections due to P. falciparum and P. vivax diagnosed annually at FMT-AM are shown in Figure 2, panel A. The percentage of diagnosed case-patients admitted to the hospital, by parasite and year, is presented in Figure 2, panel B. In 1989, 264 (20%) of the patients with a diagnosis of P. falciparum infection were admitted to the hospital, while only 26 (0.85%) of those infected with P. vivax required admission. While P. falciparum remains the main cause of malaria admissions, we observed a significant increase in P. vivax admissions: the mean percent admissions from 1989 to 1996 was 0.59% (standard deviation [SD] 0.18), increasing to 1.91% (SD 0.74) from 1997 to 2006. This relative increase in P. vivax malaria requiring admission to the hospital for treatment was disproportionate to the change in numbers of cases and to the relative frequency of P. vivax cases over P. falciparum malaria cases.

We presented the epidemiology of recent malaria cases in the State of Amazonas and city of Manaus, emphasizing the emergence of severe P. vivax malaria. Assuming that patients requiring hospital admission were sicker than those treated as outpatients, we observed that malaria transmission in this region was continuous and fluctuated in intensity. P. vivax was consistently the main cause of malaria, but the number of patients with P. vivax requiring hospital admission increased significantly in recent years. Changes in control operations were linked to the reestablishment of malaria in major urban areas of the Amazon basin, e.g., Belém (11). In Manaus, this could have had an impact on P. vivax disease manifestations and severity but did not seem to have affected the severity of disease caused by P. falciparum, perhaps because the current policy of early diagnosis and treatment has been reported to have a greater impact on disease caused by P. falciparum than P. vivax (12,13).

In this study, we assumed that case definition and criteria for admission at FMT-AM, for both P. vivax and P. falciparum malaria, were relatively constant (http://www.fmt.am.gov.br/). Our data showed that the likelihood of hospital admissions for case-patients diagnosed with P. vivax malaria increased substantially after 1996, while the percentage of P. falciparum admissions declined. The decreasing rate of admission for P. falciparum malaria during the later years of our study supports the interpretation that the criteria for admission to FMT-AM were not loosened. It is possible that referrals to FMT-AM from elsewhere in the region increased during this period, but that would likely affect P. falciparum admissions too. Based on these considerations, we interpret the data as suggestive of an increased illness associated with P. vivax infections in the region.

In this study we did not attempt to describe the specific disease manifestations that were the basis for admissions of individual patients. However, recent reports described a range of unusual manifestations of P. vivax infection elsewhere (14), consistent with the disease manifestations reported in Manaus (5).

Biologic aspects of the human host, vector, and parasite and changes in the environment contribute to the epidemiology of malaria. Our data demonstrate that malaria is a growing health burden in the Amazon Region of Brazil and that disease caused by P. vivax is a substantial and increasing threat to the health of the population in Manaus. More studies are needed to understand the complex mechanisms of this disease and its impact on susceptible populations.