Epidemiologic and Spatiotemporal Characterization of Influenza and Severe Acute Respiratory Infection in Uganda, 2010-2015

Abstract

Rationale: Little is known about the epidemiology of severe acute respiratory infection (SARI) or influenza in sub-Saharan Africa. Characterization of influenza transmission dynamics and risk factors for severe disease and mortality is critical to inform prevention and mitigation strategies.

Objectives: To characterize the epidemiology and transmission dynamics of influenza and risk factors for influenza-associated severe respiratory infection in Uganda.

Methods: Clinicians at 12 sentinel surveillance sites prospectively collected clinical data and upper respiratory tract samples from consecutive patients who met criteria for SARI and influenza-like illness (ILI). Samples were tested for influenza A and B viruses using real-time reverse transcription-polymerase chain reaction. Spatial and spatiotemporal cluster modeling was performed to identify loci of increased influenza transmission. Morbidity and mortality were assessed through chart review in a defined subset of patients. Univariable and multivariable analyses were used to identify risk factors for severe respiratory infection, prolonged hospitalization, and in-hospital mortality.

Measurements and main results: From October 2010 to June 2015, 9,978 patients met case definitions for SARI and ILI and had samples tested for influenza A and B. Of the 9,978 patient samples tested, 1,113 (11.2%) were positive for influenza. Among 6,057 patients with ILI, 778 samples (12.8%) were positive, and among 3,921 patients with SARI, 335 samples (8.5%) were positive. Significant clustering of influenza cases was observed in urban and periurban areas and during rainy seasons. Among 1,405 cases of SARI with available outcome data, in-hospital mortality was 1.6%. Infection with the 2009 pandemic A/H1N1 subtype and prolonged time to presentation were independently associated with SARI among influenza cases.

Conclusions: Influenza is associated with a substantial proportion of acute respiratory infection in Uganda. As influenza vaccination programs are developed in East Africa, timing campaigns to confer protection during rainy seasons should be considered, particularly among high-risk urban populations.

Keywords: Africa South of the Sahara; acute respiratory infection; epidemiology; influenza; spatiotemporal analysis.

Figure 1.

Influenza cases in Uganda by month, October 2010–June 2015. Data available for 1,111 cases. Vertical dotted lines indicate time-periods composing the biannual rainy seasons. Colored solid lines represent influenza types and subtypes identified: A/H3N2 (blue), A/H1N1pdm09 (red), A/unsubtypeable (green), B (purple). A/H1N1pdm09 = 2009 pandemic A/H1N1.

Figure 2.

Spatiotemporal influenza clusters in Uganda detected by SaTScan space-time scan statistic, 2010–2015. Both significant and nonsignificant clusters are included. Large circular lines indicate geographic area of each influenza cluster and do not signify the number of cases in the cluster; smaller solid circles indicate the location of subcounties composing each cluster. A/H1N1pdm09 = 2009 pandemic A/H1N1.

Figure 3.

Significant spatiotemporal influenza clusters in urban and periurban subcounties of Kampala and Entebbe detected by SaTScan space-time scan statistic, 2010–2015. Large circular lines indicate geographic area of each influenza cluster; smaller solid circles proportionally indicate the number of influenza cases in subcounties composing each cluster. A/H1N1pdm09 = 2009 pandemic A/H1N1.