58 TRANSACTIONS OF THE ROYAL SOCIETY OF TROPICAL MEDICINE AND HYGIENE (1994) 88, 5%59 (Short Report1 The basic reproduction number for dengue fever in SBo Paulo state, Brazil: 1990-1991 epidemic C. A. MarqueslJ, 0. P. Forattini2 and E. Massad3r 5uperintendencia de Controle de Endemias do Estado de Slio Paula, Brazil; *Faculdade de Saude Ptiblica and 3Faculdade. de Medicina da Universidade de SGo Paula, LIMOIIHCFMUSP. Av. Dr Amaldo 455. 01246 Sao Paula, Brazil; 4Nhcleo de Pesquisa T&onBmica e Sistemritica em Entomologia Mkdica da Universidade de SBo Paula, Brazil Dengue fever reappearedas a major urban epidemic in Brazil in the State of Rio de Janeiro in 1986, the first out- ,b;be! of dengue virus reported in Brazil for several de- Of ;he 7117 cases of dengue fever reported in the State of S&o Paulo in the epidemic of 1990-1991, 6757 came from the region of Ribeirlo Preto, comprising 48 cities in the north-eastern part of the State (PONTES et al., 1992). The maximum averageBreteau index reported for the re- gion (in November 1991) was 9.02 (PONTES et al., 1992). This was the first record of dengue in this area; it was causedby serotype Den- 1. We analvsed the casenotification data from 12 cities of this region for the 1990-1991 epidemic. All the cases analysed had been confirmed by enzyme-linked immuno- sorbent assay for dengue-specific immunoglobulin M, and we calculated the basic reproduction number, Ra, for dengue virus in this region. Table 1 shows the cumu- Table 1. Cumulative number of initial cases of dengoe fever in 12 cities in the region of Ribeir5o Preto, Slo Patdo, Brazil (1990-1991)” lgarapava Morro Agudo Olimpia Miguelopolis f -n : 4 2 8 5 12 6 13 8 : n 1 5 2 6 3 10 4 13 5 6 5 7 6 I ?I 3 4 4 6 5 7 6 9 7 Bebedouro I ; : 4 7 7 10 10 12 Viradoudo I ?I 1 3 2 4 3 7 5 9 7 f n 1 7 3 11 4 15 5 17 8 Pitangueiras I n 8 : 9 6 11 10 6 6 12 12 7 7 Pontal Ituverava t n t n 1 1 1 1 2 2 4 3 4 3 5 4 7 5 7 5 8 9 8 9 E n 1 7 2 10 5 11 9 12 12 Jardinopolis I n 1 4 2 5 3 “n=Cumulative number of cases, t=time (days). lative number of initial cases over time for the 12 cities studied. Ro is defined as the number of secondary infec- tions caused by a single infected individual in an entirely susceptible population, and is a measure of the epidemic potential in a particular region and period of time. For a vector-borne disease, Ro may be understood as the num- ber of persons who will be infected from a single person initially infected by a mosquito. This number may be derived from entomological and parasitological par- ameters, from the observed proportion of the population infected at the end of the epidemic (KOOPMAN et al., 1991), or from the initial doubling time of cases at the be- ginnin have c % of the epidemic (ANDERSON 8z MAY, 1991). We osen the last method to estimate Ro for dengue virus in these 12 cities of Slo Paulo. According to this method, if we call the number of suscentible and infected individuals x(t) and y(t), respectively; the total popula- tion n(t) [n(t)=x(t)+y(t)], the natural mortality of the host population u, the contact rate p, and the inverse of the infectious period y, we have the following relationship: 4 Cd x(t)Y(t) -= B- -(P+YIY(c dt n(t) Noting that in the beginning of the epidemic n-+x, we can solve equation (1) for the first doubling time, td, when y(td)=&@), and obtain. RO= - [ 14.3 +1 1 (2) h+Yjtd The first doubling time of the number of cases, td, was estimated by least squaresfitting from the initial number of cases, and Ro was calculated with (l/u)=70 years and (li )=6 days. T able 2 shows the results of the above calculations for the 12 cities studied. The results shown in Table 2 are, on average, very close to that obtained for dengue fever in Mexico by KOOPMAN et al. (1991), Ro= 1.33. Table 2, Doubling time (tJ and basic reproduction number (IQ for dengoe fever in 12 cities in Brazil (1990-1991) City h RO Igarapava 6.93f0.34 Morro Agudo 6.93kO.53 Olimpia 5.61kO.75 Miguelopolis 5.05f0.70 Barretos 5.03kO.37 Bebedouro 5.OlkO.42 Pitangueiras 4.54kO.46 Jardinopolis 3.33kO.23 Guara 3.03kO.22 Viaradouro 2.97f0.23 Pontal 2.8OkO.22 Ituverava 2.79kO.20 Average 4.50f0.39 1~60+0~08 1.60f0.12 1.74kO.23 1.82f0.25 1,83+0.14 1.83f0.15 1.92kO.20 2.25f0.16 2.37kO.17 2.4OkO.18 2.48f0.19 2.49kO.18 2.03kO.17 The interpretation of the above result is that, on aver- age, each infective person infected 2.03 other persons via the mosquito vector at the start of the epidemic when the population was susceptible. The maximum basic repro- duction number found was 2.49 (which should be com- pared with the maximum value obtained in Mexico, 2.41). It should be mentioned that it is not possible to ex- clude sampling bias in the proportion of cases tested serologically since in the beginning more cases were likely to be tested serologically than in later stages of the epidemic, when clinicians were more experienced in dengue diagnosis. However, our aim was to estimate Ro, a value which is meaningful only in the early stages of the epidemic. Our results point to the already known fact that, in the regions studied, dengue virus is only moderately infec- tious, compared with other vector-borne infections like malaria, the basic reproduction number of which may reach considerably higher values. Estimation of Ro for the 1967-1968 Plasmodium vivax epidemic in Sri Lanka gave values > 100 in 4 of 6 areas (SIVAGNANASUNDRAM, 1973). The relatively low value of Ro for dengue in the regions studied indicates that a moderate decrease in the basic reproduction number will result in a large decrease in the proportion of the population that is infected. Also, due to this low average value of Rp, the dengue epidemic in the region studied wrl 4 robability of a be quite sensi- tive to reductions in the number of breeding sites or to the use of insecticides or screens.