Influenza in outpatient ILI case-patients in national hospital-based surveillance, Bangladesh, 2007-2008

Abstract

Background: Recent population-based estimates in a Dhaka low-income community suggest that influenza was prevalent among children. To explore the epidemiology and seasonality of influenza throughout the country and among all age groups, we established nationally representative hospital-based surveillance necessary to guide influenza prevention and control efforts.

Methodology/principal findings: We conducted influenza-like illness and severe acute respiratory illness sentinel surveillance in 12 hospitals across Bangladesh during May 2007-December 2008. We collected specimens from 3,699 patients, 385 (10%) which were influenza positive by real time RT-PCR. Among the sample-positive patients, 192 (51%) were type A and 188 (49%) were type B. Hemagglutinin subtyping of type A viruses detected 137 (71%) A/H1 and 55 (29%) A/H3, but no A/H5 or other novel influenza strains. The frequency of influenza cases was highest among children aged under 5 years (44%), while the proportions of laboratory confirmed cases was highest among participants aged 11-15 (18%). We applied kriging, a geo-statistical technique, to explore the spatial and temporal spread of influenza and found that, during 2008, influenza was first identified in large port cities and then gradually spread to other parts of the country. We identified a distinct influenza peak during the rainy season (May-September).

Conclusions/significance: Our surveillance data confirms that influenza is prevalent throughout Bangladesh, affecting a wide range of ages and causing considerable morbidity and hospital care. A unimodal influenza seasonality may allow Bangladesh to time annual influenza prevention messages and vaccination campaigns to reduce the national influenza burden. To scale-up such national interventions, we need to quantify the national rates of influenza and the economic burden associated with this disease through further studies.

Figure 1. Global positioning system coordinated map of Bangladesh showing the surveillance sites and geographical distribution of influenza positive cases during May 2007–December 2008.

Figure 2. Proportion with rRTPCR confirmed influenza among ILI case-patients in different age groups.

Figure 3. Kriging prediction of month of onset of influenza season in different parts of Bangladesh.

We divided the total study area into 5 km by 5 km regions. Then we estimated the month of first appearance of influenza for 2008 in each region based on the distance weighted first positivity month of the surrounding locations using interpolation. We plotted the estimated values using a color scale. We used the observed months as integers. We also include contour lines in the plot to specify the areas where the interpolated values were the same. In this figure 5 = May, 6 = June and 7 = July and the intermediate values are the mid-months.

Figure 4. Seasonality of influenza in Bangladesh.

Figure 5. Correlation of percent positivity of influenza with the average monthly rainfall, average temperature, average relative humidity, and average sunlight hours.

Figure 6. Non-coinciding seasonality of seasonal influenza in human and avian influenza in poultry in Bangladesh.