Mosquito fitness is a crucial factor of vectorial capacity and malaria transmission
. However, the question whether Plasmodium parasites affect the fitness of mosquito vectors they infect remains open. Ten years ago, a review pointed out conflicting results in previous studies according to the design of experiments, environmental conditions, and the parasite-mosquito species used
. This underlined the need to develop experiments using natural vector-parasite species combinations in controlled environmental conditions. In the last decade, additional studies aimed at deciphering the cost of infection in malaria vectors by using different species and methods and produced again contrasted results
[3–7]. Some studies showed a shorter lifespan in infected mosquitoes or no detected effect of infection depending on the parasite load and/or environmental conditions
[3, 5–7]. Recently, another study observed a longer lifespan in infected mosquitoes associated with a lower fecundity
. Concerning the most important vectorial system for human malaria: Anopheles gambiae/Plasmodium falciparum, some studies suggested an infection cost in this combination of species
[7, 8], others did not find it
[9, 10]. Further investigations are therefore needed to determine how infection affects the mosquito vector in this epidemiologically relevant system.
Measuring how parasite infection affects fitness traits requires the comparison of infected and non-infected mosquitoes, either collected from the field or produced in laboratory conditions. Comparison of field-collected infected and non-infected mosquitoes is limited by the difficulties to obtain a large number of infected mosquitoes
, and by the lack of a reliable method for ageing the field-collected mosquitoes
. In the laboratory, obtaining a large number of Plasmodium-infected mosquitoes is facilitated by exposing insectary-reared Anopheles to gametocytes through membrane feeding (gametocyte-infected blood from naturally infected patients in direct membrane feeding assays, DMFA, or gametocyte-containing parasite culture in standard membrane feeding assays, SMFA)
. Different methods were used so far for producing non-infected mosquitoes to be compared to the ones exposed to Plasmodium gametocytes. In DMFA, in parallel to the experimental infection of mosquitoes, other females were exposed to parasite-free blood from another volunteer
[10, 12]; in SMFA the control mosquitoes were fed on a parasite culture not producing gametocytes
. Also exposing the mosquitoes that ingested infectious gametocytes to high temperatures just after the blood feeding can limit the infection success and produce non-infected mosquitoes
. All these methods however, incur the problem of confounding factors: when using the blood of uninfected hosts, one cannot rule out that host effects are responsible for the potential differences in mosquito life-history traits
; also, parasite strain-specific effects, or temperature effects
[15, 16] may be confounded with the effect of infection. This underlines a need for a proper control for evaluating the cost of infection.
It was shown recently that heating P. falciparum gametocyte-infected blood before feeding by malaria vectors impedes the infection by killing or inhibiting the infectivity of gametocytes
[17, 18]. This potentially allows the comparison of infected to non-infected mosquitoes that fed on the same blood at the same time. However, before using this method for characterizing the cost of infection to mosquitoes, it is necessary to establish whether feeding on previously heated blood affects the survival and fecundity of mosquito females. Therefore, life history traits of female mosquitoes exposed to non-infected blood that was subjected or not to a heat treatment, were compared. This allowed validating the method of using heat inactivation of P. falciparum gametocytes for comparison of fitness traits between infected and non-infected An. gambiae.