Dengue and dengue hemorrhagic fever in the State of Pernambuco , 1995-2006 Dengue e febre hemorrágica do dengue no Estado de Pernambuco , 1995-2006

In Pernambuco, the first dengue cases occurred in 1987. After a seven-year interval without autochthonous cases, a new epidemic occurred in 1995. Important aspects of the dengue epidemics during the period 1995-2006 have been analyzed here, using epidemiological, clinical and laboratory data. A total of 378,374 cases were notified, with 612 confirmed cases of dengue hemorrhagic fever and 33 deaths. The mortality rate was 5.4%. The incidence rate increased from 134 to 1,438/100,000 inhabitants, corresponding to the epidemics due to serotypes 2 and 3, in 1995 and 2002, respectively. Dengue mainly affected adults (20-49 years); 40.7% were male and 59.3% were female. From 2003 onwards, the number of cases among individuals younger than 15 years old increased. Out of 225 dengue hemorrhagic fever cases, 42.7% primary and 57.3% secondary infections were identified (p = 0.0279). Neurological manifestations were also observed. From 2002 onwards, serotypes 1, 2 and 3 were circulating; serotype 3 was predominant. Key-words: Dengue. Dengue virus. Dengue hemorrhagic fever. Epidemiology.

Over the last two decades, dengue fever (DF) and dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS) have emerged as important arthropod-borne viral diseases associated with high morbidity among human populations in the Americas 32 .On these continents, the dengue virus (DENV) has shown hyperendemic characteristics, as the consequence of co-circulation of multiple serotypes 6 9 .The reemergence of dengue in Latin America occurred during the 1960s (Caribbean and Venezuela) and in Colombia in the 1970s 9 .
In Brazil, the first laboratory-confirmed dengue outbreak occurred in 1982 in the northern region (Boa Vista, State of Roraima), with isolation of serotypes 1 (DENV-1) and 4 (DENV-4) 31 .Following this, there were no further cases reported until 1986, when an epidemic of DENV-1 occurred in the State of Rio de Janeiro 36 , followed by the introduction of DENV-2 and DENV-3 in 1990 and 2000, respectively, resulting in several epidemics in this country 23 24 27 .After the introduction of DENV-3, explosive epidemics were observed almost all over the country 37 .From 1986 to 2006, more than four million dengue cases were reported in the country 19 .
In the State of Pernambuco, the first dengue outbreak occurred in 1987, with 2,118 cases reported.However, this outbreak was brought under control within the same year by the vector control measures undertaken.In 1995, after experiencing a seven-year interval without confirmed autochthonous dengue cases, a new epidemic occurred with the introduction of DENV-2.Since the introduction of DENV-3 in 2002, the three serotypes are now circulating in Pernambuco 3 .From 1995 to 2006, approximately 380,000 dengue fever cases were reported in this state.In the present study, we describe important epidemiological characteristics of the dengue/dengue hemorrhagic fever cases that were reported during the epidemics in the State of Pernambuco during the period from 1995 to 2006.

MATERIAL AND METHODS
Study site.The current study was conducted in the State of Pernambuco, in the northeastern region of Brazil, which covers an area of 98,938km 2 and had an estimated population of 8,457,267 inhabitants in 2006.The state has 185 municipalities, but the majority of the population lives in the Metropolitan Region, which includes Recife (the state capital) and 18 other municipalities 18 .
Study population.The 378,374 cases included in this study had acute febrile illness suggestive of dengue infection.Information on patients was recorded on standardized forms completed by health staff and reported to health surveillance authorities in the state.The cases were analyzed following clinical, epidemiological and laboratorial criteria that were in accordance with guidelines from the Brazilian Ministry of Health 20  Dengue case definition.Dengue fever (classical dengue) and dengue hemorrhagic fever were defined in accordance with WHO guidelines 39 .
Data sources and analysis.Data from reported dengue cases was available in the official information system of the Pernambuco State Health Department (SSPE).Entomological data was obtained from the SSPE vector control program.Laboratory diagnosis data was available from the Central Public Health Laboratory (LACEN-PE).Dengue incidence rates were calculated according to age group and gender for all municipalities, for each year, by using census population data 18 as the denominators.The mean incidence rates for the state, according to age group and gender, was calculated by using the median points of estimated populations for the years 1995 to 2006 18 .The statistical analysis was performed with the aid of the Stat Pac Inc. software for Windows (version 10.2; Bloomington, MN, USA), which is available at http://www.statpac.comand Microsoft Office Excel version 3.0 (Microsoft, USA).Odds ratios (OR) and 95% confidence intervals (CI) were calculated.The chi-squared test was used for comparisons of means and proportions and p < 0.05 was considered statistically significant.Sample collection.Specimens from 91,480 suspected cases were sent to LACEN-PE for dengue infection confirmation and/or virus isolation, together with individual standardized forms containing basic demographic data, symptom onset dates, sample collection data and clinical manifestations.Aliquots of serum samples were stored at -70°C and -20°C for virus isolation and serological tests, respectively.
Dengue diagnosis.The laboratory diagnosis of dengue was undertaken using virological and serological methods, depending on the dates of disease onset and specimen collection.A total of 120,000 serum samples received by the laboratory were subjected to dengue IgM-capture ELISA analysis using an in-house method (MAC-ELISA) 12 and/or a commercial kit (PanBio Pty Ltd, Brisbane, Australia) and testing was performed in accordance with the manufacturer's instructions.A total of 9,000 acute serum samples were processed for virus isolation using C6/36 cells 11 ; virus isolates were identified by the indirect fluorescent antibody assay 7 .The reverse transcription-polymerase chain reaction (RT-PCR) for detecting and typing DENV 13 was carried out on 840 acute-phase serum samples and specimens from fatal cases.The hemagglutination inhibition assay (HI) 2 was used on paired serum samples from 225 DHF cases, to classify the type of immune response, in accordance with World Health Organization (WHO) criteria 39 .

RESULTS
From 1995 to 2006, 378,374 dengue cases were reported in Pernambuco and 612 DHF cases were confirmed, with 33 deaths (Table 1).In 1995, the main vector, Aedes aegypti, was detected in 42 municipalities; in 1996 and 1999, it was confirmed in 125 and 173 municipalities, respectively; and in 2002 it reached all 185 municipalities, with an infestation index > 1%.In 1995, 9,982 cases were reported, corresponding to an incidence rate (IR) of 134.1 cases per 100,000 inhabitants.In 1996, the number of cases was 128% higher (IR = 307.1/100,000inhabitants) and in 1998, an incidence of 699.6/100,000 inhabitants was reached (Table 1).Over the next three years (1999-2001), a reduction of 32% in the number of cases was observed.The major epidemic caused by DENV-3 that occurred in 2002 affected all 185 municipalities, and the number of cases was 7.8 times greater than in 2001; 116,245 cases were reported (IR = 1,437.8/100,000inhabitants).In 2003, the number of cases was 4.5 times lower than in 2002 (IR = 329.6/100,000inhabitants).Thereafter, 2004-2006 was an endemic period.In all outbreaks, the majority of cases occurred in the first half of the year (Figure 1), usually peaking in the months of April and May.In the outbreaks of 1995 to 1997 and in the 2002 epidemic, the bulk of case occurrences were in the Metropolitan Region of Recife (MRR), with 42.5% of the total number of cases in these four years; other regions were responsible for the majority of cases in 1998-2001 and 2003-2006.Over this 12-year period, the incidence rate in the city of Recife ranged from 28 to 2,378/100,000 inhabitants.The 86,247 cases reported in Recife represented 22.8% of the total number of cases occurring in Pernambuco.Other municipalities in the metropolitan region, such as Jaboatão dos Guararapes, Olinda, Cabo de Santo Agostinho and Camaragibe, presented mean incidence rates of 259, 384, 722 and 1019 cases/100,000 inhabitants, respectively.The total number of cases reported from 1995 to 2006 in the five major regions of the State of Pernambuco were: 1) Metropolitan Region of Recife, 176,458 cases; 2) Agreste, 101,027 cases; 3) Mata, 55,245 cases; 4) Sertão, 31,812 cases; and 5) São Francisco, 14,101 cases.Table 2 shows the annual incidence rates in these geographic regions.From 1995 to 2001, there was co-circulation of DENV-1 and DENV-2.In 2002, after the introduction of DENV-3, circulation of all three serotypes was detected; nevertheless, there was predominance of DENV-3.The percentages of DENV isolation, by year, are shown in Table 1.The predominant DENV serotypes in each epidemic/endemic year are shown in Figure 1.

Figure 1 -Annual distribution of dengue fever cases reported in the State of Pernambuco, over the period 1995-2006, according to semester of occurrence and the dengue virus serotypes predominant in each year.
The overall age distribution showed that 55% of the dengue cases occurred in adults in the age range from 20 to 49 years old (Table 3).During the period 1995-2006, dengue incidence was consistently higher in adults, reaching up to 1,938/100,000 inhabitants in the 20-49 years age group, in 2002.When dengue cases were classified into three age groups, i.e. <5 years, 5 to 14 years and > 15 years old (Figure 2), it was observed that the > 15 years age group was the most affected.From 2003 onwards, the group less than 15 years old was more affected than in previous dengue outbreaks.In 2003, the incidence of cases among children younger than 5 years old was higher than in the 5 to 14 years age group.The dengue incidence rates in <1 year, 1-4 years and 5-14

Age
Reported

Figure 2 -Distribution of dengue cases reported in the State of Pernambuco, over the years 1995 to 2006, according to incidence rates for three different age groups (0-4 years; 5-14 years; > 15 years).
years age groups were 326, 210 and 199/100,000 inhabitants, respectively.Table 4 shows the annual distribution of the numbers and incidence rates of cases, by age groups.
Regarding gender, over this 12-year period, the male to female ratio was constant at 1:1.5.From 1995 to 2006, 153,120 cases (40.7%) were reported in male patients and 224,425 cases (59.3%) in female patients.A statistically significant difference in the frequencies of cases between female and male patients was found.Female individuals were more affected [OR = 2.15; 95% confidence interval (CI) = 2.129 -2.168; c 2 = 26,878.65(p <0.0001)].The gender and age distributions were very similar when the data for each year were stratified.
The first DHF cases were confirmed in 1996, after the introduction of DENV-2 into the State.From 1996 to 2006, 612 cases of DHF were confirmed, with 33 deaths.The case-fatality rate ranged from 2.0% to 16.7%, with a mean of 5.4% (Table 1).In the epidemic year of 1998, the ratio of DHF to DF cases was 1:1,120 and mortality was 2.1% (1/47); while in the epidemic year of 2002 (DENV-3), the ratio of DHF to DF cases was 1:342 and mortality was 5.9% (Table 1).The age distribution among DHF cases over this period is presented in Table 3.It was observed, among the DHF cases, that 81.7% (n = 500) occurred in individuals >15 years old, and mainly in adults between 20 and 49 years old (median age 47 years; modal age 23 years).However, the incidence rate in the 50 to 79 years age group was somewhat similar to that in the 20 to 49 years age group (Table 3).Only 18.3% of the patients (n = 112) were younger than 15 years old.Regarding gender, there were 236 (38.6%) male cases and 376 (61.4%) female cases, thus representing one male case to 1.6 female cases.There was no statistically significant difference between the genders [OR = 1.23 (CI 0.98 -1.53); c 2 = 3.29 (p = 0.070)].Regarding the type of immune serological response, the results from hemagglutination inhibition assays performed on paired serum samples from 225 DHF cases characterized 96 (42.7%)DHF cases as primary infection and 129 (57.3%)DHF cases as secondary (sequential) dengue infection, in accordance with the WHO criteria 39 .There was a statistically significant difference (p = 0.0279) in secondary infection as a risk factor for DHF.Among 13 fatal cases analyzed, one case of primary and five of secondary dengue infection were characterized.
The laboratory studies performed at LACEN-PE on 116,098 specimens (24,518 second samples) from 91,480 suspected dengue cases confirmed 48,300 cases (52.8%) by at least one of the following tests: virus isolation, RNA viral detection, positive anti-dengue IgM ELISA or fourfold rise in HI titer.Analyses conducted on the data from these positive cases showed that the main signs and symptoms presented were consistent with dengue fever (Table 5).The most common hemorrhagic manifestations in 468 laboratory-positive DHF cases were petechiae, metrorrhagia, bleeding gums, epistaxes, melena and hematemesis (Table 5).Neurological manifestations were reported in 32 patients with dengue fever; the most severe cases were: encephalitis (6 cases), meningoencephalitis (4 cases) and Guillain-Barré syndrome (3 cases).Anti-dengue IgM was detected in both serum and cerebrospinal fluid samples from these patients.

DISCUSSION
Over the last 20 years, the incidence of dengue infections in Brazil increased from a reported annual rate of 37 cases per 100,000 inhabitants in 1994 to 454/100,000 inhabitants in 2002 37 .In Pernambuco, 378,374 cases have been reported over the last 12 years, and the infection incidence rate increased from 134 cases/100,000 in 1995 to 1,437.8/100,000inhabitants in 2002, as consequence of DENV and disease dispersion in the State.Despite the efforts to keep the vector density at low levels, this objective was not achieved.According to official data 19 , the vector density (house index) was greater than 1% in 80-90% of the municipalities in the country, including in Pernambuco.In Fortaleza, Brazil, it has been observed that dengue outbreaks never occurred when the house index was < 1% 33 .
As demonstrated in Figure 1, the majority of cases in each outbreak were observed in the first half of the year, similar to what has been observed in other Brazilian states 37 .Nevertheless, it should be pointed out that dengue outbreaks may occur during both halves of the year, as seen in the State of Rio de Janeiro, during epidemics of DENV-1 (January to August 1990) and DENV-2 (October 1990 to May 1991) 25 .The major dengue epidemics in Pernambuco occurred in 1998 and 2002, although dengue activity was observed throughout the 12-year period (Table 1), with cocirculation of DENV-1 and DENV-2 until 2001 and inclusion of serotype 3 from 2002 onwards.The percentages of each serotype isolated from 1995 to 2006, which are shown in Table 1, indicate which DENV serotype predominated in each outbreak (Figure 1).Based on these results, it is highly suggestive that the 1997-1998 epidemics were caused by DENV-1, while DENV-2 was responsible for the 1995-1996 outbreaks and then DENV-3 caused the 2002 epidemic 3 .The replacement of one serotype by another, which is usually observed during epidemics, may be explained by the growing numbers of people who became immune to the former serotype 9 .Nevertheless, it has been shown that DENV-3 (genotype III, Sri Lanka/India) has been able to dislodge the other serotypes when introduced into a new region, thus showing its high infection capacity among the human population 4 9 26 .
Despite the large number of cases reported in Pernambuco during this period, the number of dengue infections that occurred was probably an underestimate, because the majority of the cases reported are those that were attended at public health services.On the other hand, it could be possible that there was overestimation of the number of cases, particularly during epidemics, considering the percentages of laboratory-confirmed cases (58.2%), among those subjected to specific tests.Taking into account the five major geographic regions of Pernambuco, the majority of the cases occurred in the Metropolitan Region (46.6%), which represents only 2.8% of the total area of the State and is the most densely populated (41.8% of inhabitants).Dengue outbreaks have started from this region and spread to the others.The metropolitan regions of large cities are usually responsible for the majority of cases occurring during epidemics 4 35 37 .However, when the case incidence rates were examined according to geographic regions (Table 2), it was observed that the Metropolitan Region of Recife and the Agreste region presented nearly the same incidence rate, i.e. 5,232.6 and 5,297.8/100,000inhabitants, respectively.The Agreste region reported 26.7% of the total number of cases (101,027/378,374).
During the study period, dengue infection affected all age groups, as seen in Table 4.However, the incidence rate was higher among adults, particularly in the 20 to 49 years age group (Table 3).The epidemiology of dengue in Pernambuco presents a feature common to other Brazilian states, in that it is characterized by dengue fever and DHF affecting mainly adults 37 40 .It is feasible that there may have been an underestimation of dengue cases among children in Pernambuco, because dengue infection in children is frequently milder and may be confused with other febrile and exanthematic diseases 9 28 29 .Thus, it is possible that the diagnosis was not made and/or children did not present significant symptoms during dengue infection and therefore did not seek medical assistance.Nevertheless, it is quite unlikely that DHF cases in children were misdiagnosed, although this issue deserves to be better investigated.The low apparent rate of dengue infection among individuals younger than 15 years of age found in the State of Pernambuco (Figure 2) has also been reported in D1: DENV-1, D2: DENV-2, D3: DENV-3.Source: Surveillance and Epidemiological Service of the Central Public Health Laboratory, Pernambuco State Health Department, Brazil.
Source: Surveillance and Epidemiological Service of the Central Public Health Laboratory, Pernambuco State Health Department, Brazil.
. The Research Ethics Committee of the Aggeu Magalhães Research Center, Fundação Oswaldo Cruz, Brazil, approved this study in accordance with CNS Resolution 196/1996.

Table 1 -Number of dengue/dengue hemorrhagic fever cases reported in the State of Pernambuco: incidence rates per 100,000 inhabitants, numbers of deaths, case fatality rates and dengue virus serotype predominance, according to year of occurrence, 1995-2006.
DHF: dengue hemorrhagic fever, DENV: dengue virus, NC: number of cases, IR: incidence rate, DHF/DF: dengue hemorrhagic fever/dengue fever, D: death, MR: mortality rate.Source: Surveillance and Epidemiological Service of the Central Public Health Laboratory, Pernambuco State Health Department, Brazil.