The case fatality risk of 2009 Pandemic Influenza A (H1N1): a systematic review

Abstract

BACKGROUND: One of the immediate public health priorities during the 2009 pandemic was to establish the transmissibility and severity of the novel influenza A (H1N1) virus (pH1N1). Whereas transmissibility was estimated generally reliably from the early stages of the pandemic, there was greater difficulty in estimating severity. One measure of severity at the individual level is the risk of death among people who have the disease, and this conditional measure can be referred to as the case fatality risk (CFR). Our objective was to review published estimates of the CFR of pH1N1, identify challenges to real-time estimation of the CFR, and make recommendations for estimation of severity in the next pandemic. MATERIALS AND METHODS: Studies reporting estimates of the CFR of pH1N1 were retrieved from the Medline (PubMed) electronic database on April 20, 2012. Eligible articles reported a CFR in the population for the first wave or the first year of the pandemic. Studies that only reported the CFR in population subgroups, such as pregnant women, or those with chronic diseases, were excluded. We defined the CFR as the number of influenza-associated deaths divided by the number of pH1N1 cases or infections in a population. RESULTS: We identified 46 articles reporting estimates of the CFR. The CFRs were measured using cumulative incidence of infection derived from serologic data, estimated cases, medically attended cases, or laboratory-confirmed infections as the denominators. The earliest studies of the severity of pH1N1 provided high estimates of the CFR based on confirmed cases (CFRs of approximately 1000 deaths per 100,000 cases). After the peak of the pandemic, the CFRs based on symptomatic cases were similar in those studies using estimated infections and estimated cases as denominators (CFRs of approximately 10 deaths per 100,000 cases). In age-stratified analysis, point estimates of the CFRs increased with age from approximately 1 death per 100,000 cases in older children to approximately 100 deaths per 100,000 cases in elderly. CONCLUSIONS: The wide range in published estimates of the CFR led to challenges in identifying unbiased and comparable severity measure of the pandemic strain. While the choice of denominator could explain some of the variability, CFR estimates for the estimated case denominator still covered a very wide range. One clear limitation in estimating the CFR is the lack of consensus on the definition, computation, and estimation of the CFR, highlighted by the variable denominators and the difference between directly and indirectly estimated numerators. Real-time estimation of a well-defined measure of severity with sufficient comparability has to be ensured as preparedness against the next major global epidemic of respiratory infection.