Efficacy of long-lasting insecticidal nets before and after the selection of pyrethroid-resistant Anopheles gambiae s.l. and Anopheles funestus s.l. in northeastern Tanzania: a longitudinal meta-analysis of experimental hut studies at a single location

Background: The extent to which insecticide resistance is affecting malaria vector control in community and home is not fully understood. This study assessed the implications of insecticide resistance for entomological efficacy of long-lasting insecticidal nets (LLINs) against wild free-flying field Anopheles gambiae s.l. and Anopheles funestus s.l. in experimental hut trials (EHT) in northeastern Tanzania before and after the evolution of pyrethroid-resistance. Methods: Evaluations of LLIN efficacy were conducted according to World Health Organization (WHO) guidelines in ten EHT commissioned by WHO between 2006 and 2017, before and after resistance development in Muheza.The evaluation criteria were mortality, blood feeding inhibition, induced exophily, personal protection, mass killing effect and deterrence. WHO bioassays determined resistance levels while PCR molecular diagnostics detected resistance alleles and identified mosquitoes to species. Results: Anopheline mosquitoes were fully susceptible to pyrethroids until 2010 when they began to show resistance. The Voltage Gated Sodium Channel L1014S point mutation kdr was detected in An. gambiae s.l. at allelic frequency of 47%. Synergist tests with PBO restored efficacy only partially, indicating involvement of metabolic mechanisms. Meta-analysis of the 10 EHT showed that mortality of susceptible An. gambiae s.l. was 6.7 and 5.2 fold greater on zero-times (z = 6.6, p = 0.001) and 20-times washed LLINs (z = 2.3, p = 0.023) than against resistant An. gambiae s.l.. The mortality of unwashed and washed LLINs against susceptible An. funestus s.l. was 3.3 (z = 2.8, p = 0.004) and 2.6 (z = 2.9, p = 0.004) fold greater than against resistant An.funestus s.l. Resistant Anophelines were more likely to exit the huts as compared to susceptible Anophelines (z = 2.79, p = 0.005). The transition from susceptibility to resistance on changes to blood-feeding rates was non-significant for either species. Conclusions: Reduced mortality induced by LLINs after selection of pyrethroid resistance indicates that resistance may undermine household and community control of vector populations. Personal protection indicators such as proportions blood feeding seemed less affected by the transition to resistance. Meta-analysis, comparing the same net brands before and after selection of resistance, revealed the factors (decreases in mortality, static blood feeding rates, and increases in pyrethroid-induced exiting rates) most affected by resistance.