Airport sentinel surveillance and entry quarantine for dengue infections following a fever screening program in Taiwan

Abstract

Background: Dengue has not reached an endemic status in Taiwan; nevertheless, we have implemented a fever screening program at airports for the early detection of febrile passengers with a dengue infection. This study is intended to assess the performance of the airport screening procedures for dengue infection.

Methods: We analyzed data from the national surveillance system of the Taiwan Centers for Disease Control. We included the imported dengue cases reported by sentinel airports and clinics as well as the domestic cases from 2007-2010.

Results: Approximately 44.9% (95%CI: 35.73-54.13%) of the confirmed imported dengue cases with an apparent symptom (febrile) in the viremic stage were detected via the airport fever screening program, with an estimated positive predictive value of 2.36% (95% CI: 0.96- 3.75%) and a negative predictive value > 99.99%. Fluctuations in the number of the symptomatic imported dengue cases identified in the airports (X) were associated with the total number of imported dengue cases (Y) based on a regression analysis of a biweekly surveillance (i.e., n = 104, R(2)(X:Y) = 0.61, P < 0.005). Additionally, the fluctuating patterns in the cumulative numbers of the imported dengue cases (X) with a 1-2 month lead time (t) was in parallel with that of the domestic dengue cases (Y) based on a consecutive 4-year surveillance (i.e., n = 48, R(2)(X(t-1):Y) = 0.22, R(2)(X(t-2):Y) = 0.31, P < 0.001) from 2007-2010.

Conclusions: A moderate sensitivity of detecting dengue at the airports examined in this study indicated some limitations of the fever screening program for the prevention of importation. The screening program could assist in the rapid triage for self-quarantine of some symptomatic dengue cases that were in the viremic stage at the borders and contribute to active sentinel surveillance; however, the blocking of viral transmission to susceptible populations (neighbors or family) from all of the viremic travelers, including those with or without symptoms, is critical to prevent dengue epidemics. Therefore, the reinforcement of mosquito bite prevention and household vector control in dengue-endemic or dengue-competent hotspots during an epidemic season is essential and highly recommended.

Figure 1

Epidemiological trends of the dengue importation into Taiwan. (A) The scatter plots show the relationship between the number of imported dengue cases, according to the viremic status as measured by a biweekly surveillance of the airport screening program, and the number of total dengue importations into Taiwan from 2007–2010. The straight line represents a line fitted by the means of a least squares regression. The coefficient of determination, R², was estimated to be 0.61 (n = 104, P < 0.0001). The 95% confidence interval of the slope is 1.42-1.94. (B) The scatter plots show the relationship, as measured by a biweekly surveillance, between the number of dengue importations detected in the airport screening program and the number of dengue importations detected in the community from 2007–2010. The straight line represents a fitted line by the means of a least squares regression. The coefficient of determination, R², was estimated to be 0.22 (n = 104, P < 0.0001). The 95% confidence interval of the slope is 0.44-0.96.

Figure 2

Trends of the dengue epidemics and the relationships among the cases in Taiwan from 2007–2010. (A) A comparison of the fluctuations of the monthly number of imported (Xt) vs. domestic cases (Y). (B) The diversity in the annual dominant dengue serotypes among the domestic and imported cases. (C) The scatter plots show the relationship of the number of imported dengue cases (Xt-0), (Xt-1), (Xt-2) or (Xt-3) vs. the number of domestic dengue cases, as detected by a monthly surveillance (Y). The respective straight lines represent the lines fitted by the means of the least squares regressions. The coefficients of determination, R², were estimated to be 0.059, 0.21, 0.31, or 0.21 (n = 48) with respective 95% confidence intervals for the slopes being −0.57-6.67, 2.68-9.19, 3.81-9.93, or 2.46-8.98. (C) The diversity in the annual dominant dengue serotypes among the domestic and imported cases.