To collect preliminary information and verify the occurrence of the outbreak that was reported in the newspapers, the outbreak investigation team contacted the Civil Surgeon of Meherpur District and the Upazila Health and Family Planning Officers of Gangni and Mujibnagar subdistricts (upazilas). From 24 to 26 July 2010, the team visited Meherpur General Hospital and collected information on patient demographics (age, gender, and address), dates of admission and discharge, and the diagnoses from the hospital’s registry. In order to simplify case management of childhood illnesses including acute respiratory infections (ARI) and diarrheal diseases at the primary health care settings, the Government of Bangladesh introduced the Integrated Management of Childhood Illnesses (IMCI) strategy nationwide in 2002. According to this IMCI guideline, any child aged <5 years presenting with cough and fast breathing and/or difficult breathing for less than 30 days and with any general danger sign (unable to drink or breast feed, vomits everything, convulsions, lethargy, grunting, fever of 37.5 °C or above or feels hot, low body temperature of less than 35.5 °C or feels cold, movement only when stimulated, no movement at all, unconsciousness), or chest indrawing or stridor in a calm child, was categorized as a case of ARI.⁸ We assumed that the hospital physicians used this IMCI case definition during admissions to classify patients presenting with respiratory symptoms under the broad category of ARI. Following admission, the single pediatrician at the hospital re-assessed all admitted ARI patients and diagnosed children aged <2 years who initially developed coryza and/or low-grade fever followed by cough, fast breathing (tachypnea) or shortness of breath (dyspnea), and expiratory wheeze^8,9 as acute bronchiolitis. From the hospital’s registry, we listed all children aged <5 years who were admitted with ARI between 26 June and 26 July 2010. We also listed the patients diagnosed as cases of acute bronchiolitis between 26 June and 26 July 2010 and sought hospital records to note their clinical prognosis. We chose the 30-day time period to permit comparison of the monthly admission rates for acute bronchiolitis with previous months. We requested the Civil Surgeon to review the hospital records and then report to IEDCR the total number of patients discharged with a diagnosis of acute bronchiolitis who were classified using the same criteria by the pediatrician in the Meherpur General Hospital from January to December 2010.

To describe the outbreak in terms of persons, the team surveyed the case patients who were still in the hospital during the period of the investigation. We selected the mother as the primary survey respondent to provide information regarding her child’s health. If the mother was unavailable, we interviewed other family members. We explored clinical histories of these case patients, including exposure to possible risk factors, using a pretested, structured questionnaire that was jointly developed by IEDCR and ICDDR,B for investigating respiratory disease outbreaks. Physicians on the outbreak investigation team physically examined the hospitalized case patients looking for symptoms and signs of illness severity including central (bluish tongue) and peripheral cyanosis (bluish limbs).^10–12 The team recorded both the clinical features they observed and those reported in the patient information sheet by the attending physician during admission. We reviewed the patient treatment records to gather information regarding clinical management.

To determine if the acute bronchiolitis case patients surveyed in the hospital were similar to or different from those already discharged prior to the investigation and to identify possible clusters, we segregated the 101 cases in the line-list (identified from the hospital’s registry) based on the villages where they lived. The team defined a cluster of bronchiolitis if at least three admitted bronchiolitis cases came from the same village and lived within 15 min walking distance from each other according to the local community health worker. The team visited village communities with suspected clusters to survey and physically examine selected acute bronchiolitis cases who had been discharged from the hospital after June 26 until the beginning of the investigation. The team collected information on demographic and clinical profiles and exposures of these bronchiolitis cases from their mothers (or family members if the mother was not available) using the same questionnaire that was used for the hospital survey. Team members also visited the three government-run primary care hospitals of the district to seek additional cases of acute bronchiolitis in admitted children aged <2 years.

The investigation team collected nasal and throat swabs from all bronchiolitis case patients surveyed in the hospital and from those discharged case patients who were still symptomatic during the cluster investigation. Following collection, both samples obtained from the same patient were immediately placed in a single vial with viral transport medium containing DMEM (Dullbecco’s modified Eagle’s medium), 2.5% BSA (bovine serum albumin) fraction V, 1% glutamine, 2% HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), 1% penicillin–streptomycin, and Fungizone (250 mg/ml). The specimens were stored at 4–8 °C until transported within 72 h to IEDCR. The specimens were aliquoted in a biosafety level 2 safety cabinet and were stored in freezers at or below −70 °C until testing. At the Virology Laboratory of IEDCR, we performed RNA extraction and detected subtypes of influenza A and influenza B using a real-time reverse transcriptase polymerase chain reaction (rRT-PCR).¹³ Both RNA and DNA extraction were performed in a separate aliquot sent to the Virology Laboratory of ICDDR,B for molecular identification of other common respiratory viruses and bacteria. A real-time PCR respiratory pathogen panel assay developed by the Centers for Disease Control and Prevention (CDC), Atlanta, Georgia, USA, was used to detect respiratory pathogens including RSV, human parainfluenza virus 1, 2, and 3 (HPIV1, HPIV2, and HPIV3), human metapneumovirus (HMPV), adenovirus, Chlamydia pneumoniae, Mycoplasma pneumoniae, Streptococcus pneumoniae, Legionella pneumophila, and Legionella longbeachae.

We sought verbal informed consent from a parent of all participants. We suggested simple preventive measures including hand washing, and social distancing measures including cancellation of sports events to protect the wider community of the cluster villages. We also advised that parents avoid taking their ill child to public gatherings including mosques, weddings, and schools. This investigation was conducted in response to an outbreak and was approved by the Government of the People’s Republic of Bangladesh, but because this was an emergency response the activity was not reviewed by an independent human subjects committee.

Acute bronchiolitis cases were admitted to Meherpur General Hospital throughout the year. However, monthly acute bronchiolitis admissions increased between June and October, with the highest number of admissions occurring during July 2010 (Figure 1). We identified 101 children who were hospitalized and diagnosed with acute bronchiolitis between 26 June and 26 July 2010. Compared to the previous month, acute bronchiolitis admissions increased by more than 60% during the month of July 2010 (Figure 1), with more than half of these cases (59/101) admitted between 16 and 20 July 2010 (Figure 2). Of the 258 children hospitalized with acute respiratory infections (ARIs) during this month, 101 (39%) had a discharge diagnosis of acute bronchiolitis. All except nine cases lived in different communities within the three subdistricts of Meherpur (Figure 3). Fifty-six percent (57/101) of the cases lived in the urban subdistrict, while the rest lived in the two other rural subdistricts (Mujibnagar and Gangni). We surveyed 20 case patients admitted to Meherpur General Hospital and nine from the two villages where we had identified clusters. The mean age of the 29 case patients was 5.4 months (median 4 months); 65% (19/29) were males and the average monthly household income was US$ 53 (range US$ 22–124).

The main clinical symptoms included cough, fast breathing, shortness of breath, and feeding difficulty. About a quarter (7/29) reported a family history of atopy (defined as having asthma, eczema, or allergic rhinitis in the parents, or siblings or grandparents); 17% (5/29) of case patients had been affected at least once before with a similar illness that required hospitalization and 35% (10/29) were born prematurely. Thirty-one percent of respondents also reported that either another sibling or an extended family member had developed similar symptoms (cough/fast breathing/respiratory difficulty with a preceding mild fever/runny nose) within 1 week following the onset of illness in the child affected with bronchiolitis. All case patients were treated with supplemental oxygen, oral or inhalational bronchodilators, and nasal decongestants. Among the chest radiographs from nine case patients, the hospital’s pediatrician found patchy opacities or radiological evidence of bacterial pneumonia in only one case patient. Ninety percent (26/29) of the case patients received at least one broad-spectrum injectable antibiotic. There were no reported deaths from bronchiolitis or its complications among the 29 case patients (Table 1). As the hospital’s registry did not record patient outcome, we tried to assess the illness outcome of the remaining 72 cases by reviewing patient information sheets including copies of discharge and/or death certificates. We found that 75% (54/72) of the cases were discharged alive and 8% (6/72) were discharged against medical advice, a decision that families in low-income countries often make because of their inability to afford continued hospitalization, poor response to treatment, or dissatisfaction with the hospital environment.¹⁴ We could not retrieve the information sheets of 12 patients and the hospital records were insufficient to evaluate their outcomes.

Among the 29 acute bronchiolitis patients surveyed, we collected samples from 23 cases (including samples from all the 20 hospitalized cases and from three symptomatic discharged cases). Out of the 23 samples tested for different respiratory pathogens, 21 had detectable RNA for RSV, one had influenza A (H1N1), and one had influenza B. More than one virus was identified in 43% (9/21) of the RSV-positive cases. These co-infections included RSV with influenza A/H1N1 (n = 6), with influenza B (n = 1), and with adenovirus (n = 2) (Table 2). We did not detect any nucleic acids for bacterial pathogens in the specimens.