It is well-established that the prevalence and incidence of influenza in most tropical countries like Madagascar are largely unknown. Clinically, influenza is not distinguishable from most other infectious diseases with fever in African countries. So, the load of influenza on morbidity and mortality is more often unknown in African region. Important reports on spatial and temporal data that describe the global circulation of influenza highlight the fact that there is virtually no data from Africa [Reference Finkelman1].
Regarding the recent pandemic period in Madagascar, the symptoms of the H1N1pdm strain resembled more those of a seasonal influenza strain than those of a pandemic or severe seasonal strain. But, in Madagascar, the difference between seasonal period and this pandemic period has been shown by the level of virus circulation. More people seemed to be affected by pandemic influenza virus than during the influenza seasonal period. Thus, the sentinel surveillance system implemented in 2007 has allowed identification of the peaks in most of the sentinel sites during this period [Reference Rajatonirina2]. Unfortunately, in Antananarivo the new tools had just been implemented when the pandemic was spreading, but in other sentinel sites, the percentage of fever syndromes reported was lower during the seasonal period than during the pandemic period. It is not wrong to believe that when the number of cases is high the burden of disease becomes increasingly important.
The recent pandemic was probably like a seasonal epidemic but its burden is still unclear in low-income countries. In Madagascar, the influenza burden has already been described during epidemic periods in 2002 [3], and the WHO-GOARN team showed that in Ikongo District, 54% of the reported deaths due to acute respiratory infections were in children aged <5 years, but the highest mortality rate was in persons aged ⩾60 years. We showed the same trend during the pandemic in Antananarivo. Is this an ‘interesting local exception’? – it is much more likely a seasonal-like pattern.
The limits of our study [Reference Rajatonirina4] were presented in the Discussion section. Thus, the mortality surveillance system in Antananarivo is known to be imperfect. However, in this low-income country it was a genuine opportunity to find a collection of mortality data for three consecutive years. Regarding analysis, mortality data have been compared month after month for the three years and a significant difference was found only for November. Therefore our Figure 2 focused on this month.
Is it surprising that the elderly die more often than younger people as shown by Rajatonirina et al. in Figure 2? The ‘J’ curve described by our data is usually associated with the mortality trend in the population. Data from Madagascar showed that the number of deaths and the mortality rate were higher in elderly people, especially in November 2009.
The comparison month by month did not reveal any difference for January as suggested by the letter of Alonso et al. [Reference Alonso and Schuck-Paim5] at about week 4. The variance of mortality rate was emphasized by Alonso et al. and would be in relation to different endemics, for example the plague during the rainy season from December to March. Unfortunately, influenza was not the unique threat which struck the Malagasy population. The comorbidity was unknown and this was also a weakness of this study. Data were not available because deaths had occurred more frequently outside the hospital setting. This is the reality for developing countries where healthcare services are not generally used by patients because of financial constraints.
The result shown by Schuck-Paim et al. [Reference Schuck-Paim6] are interesting and well-documented but it seems to be incorrect to compare Antananarivo with an equatorial region as suggested by Alonso et al. in their letter. The climatic diversity in Madagascar is great with some equatorial climate areas and also some temperate climate areas as in the capital located at 1300 m above sea level. Furthermore, the country is one of the poorest countries and this is relevant to the economic gradient suggested by Schuck-Paim et al. [Reference Schuck-Paim6] in explaining spatial heterogeneities.
In accordance with Alonso et al.'s statement ‘the 2009 experience is of poor value to predict the incidence patterns, transmissibility and burden of a severe pandemic in the future’, we particularly focus on low-income countries. Estimating the burden of influenza in low-income countries is difficult, laboratory confirmation of influenza infection is rarely conducted, so most influenza-related hospitalizations and deaths are not attributed to influenza. Now we must tackle the most crucial aspect of this topic, which is surveillance enhancement in order to improve the tools used and to provide the best data about seasonal influenza and its burden. Data from sub-Saharan Africa are insufficient to allow most countries to prioritize strategies for influenza prevention. Much needs to be done to increase awareness of the importance of influenza in low-income countries.
