The potential contribution of mass treatment to the control of Plasmodium falciparum malaria

Abstract

Mass treatment as a means to reducing P. falciparum malaria transmission was used during the first global malaria eradication campaign and is increasingly being considered for current control programmes. We used a previously developed mathematical transmission model to explore both the short and long-term impact of possible mass treatment strategies in different scenarios of endemic transmission. Mass treatment is predicted to provide a longer-term benefit in areas with lower malaria transmission, with reduced transmission levels for at least 2 years after mass treatment is ended in a scenario where the baseline slide-prevalence is 5%, compared to less than one year in a scenario with baseline slide-prevalence at 50%. However, repeated annual mass treatment at 80% coverage could achieve around 25% reduction in infectious bites in moderate-to-high transmission settings if sustained. Using vector control could reduce transmission to levels at which mass treatment has a longer-term impact. In a limited number of settings (which have isolated transmission in small populations of 1000-10,000 with low-to-medium levels of baseline transmission) we find that five closely spaced rounds of mass treatment combined with vector control could make at least temporary elimination a feasible goal. We also estimate the effects of using gametocytocidal treatments such as primaquine and of restricting treatment to parasite-positive individuals. In conclusion, mass treatment needs to be repeated or combined with other interventions for long-term impact in many endemic settings. The benefits of mass treatment need to be carefully weighed against the risks of increasing drug selection pressure.

Figure 1. Rebound in transmission following mass treatment & comparison with trial data.

(A) & (B) Typical MDA model output over time in scenarios of (A) non-seasonal and (B) seasonal transmission with different baseline transmission intensities: high: baseline average annual slide-prevalence = 50% (dark blue); medium: baseline slide-prevalence = 15% (mid-blue); low: baseline slide-prevalence = 5% (light blue). A single round of MDA is carried out at year 0 at 80% coverage (actual coverage is lower due to exclusion of pregnant women). Gray dashed lines indicate predicted prevalence in the absence of MDA. (C) & (D) Comparison of model predictions with trial data : slide-prevalence in 2–10 year olds in control villages and intervention villages with MDA carried out (C) every 14 days or (D) every 28 days. Blue line = average model-predicted values in control villages; dark red line = model predicted value in MDA villages; shaded areas = range of 20 simulations of control and MDA prevalence; blue circles = prevalence data in control villages; orange triangles = prevalence data from chloroquine-treated individuals in MDA villages; orange squares = prevalence data from amodiaquine-treated individuals in MDA villages.

Figure 2. Strategy options for mass treatment.

(A) Three scenarios of seasonal mosquito densities simulated by the model: green line = highly seasonal: 96% of infectious bites occur in the peak 3 months; red line = moderately seasonal: 50% of infectious bites occur in the peak 3 months; blue line = not seasonal (for comparison). (B)&(C) Example simulations showing the impact of MDA on slide-prevalence in a scenario of highly seasonal transmission if carried out (B) blue line: prior to the rise in mosquito densities (month 3 in this simulation) as shown in (A), versus (C) orange line: just prior to the peak of the transmission season (month 6 in this simulation). Baseline slide-prevalence in the absence of MDA is shown in gray for comparison. (D)&(E) Model-estimated cumulative % infectious bites averted per person after 1 round of mass treatment over the following 2 years (D) comparing different antimalarial types and (E) comparing different screening criteria used to allocate treatment, in a scenario of moderate transmission intensity (baseline slide-prevalence = 15%) with moderate seasonal variation (as in Figure 2A). Mass treatment is carried out prior to the high transmission season. ‘Short-acting’ = 10 days prophylaxis, ‘long-acting’ = 30 days prophylaxis. MDA = no screening; MSAT PCR = PCR-positive individuals are treated; MSAT M = microscopy-positive individuals are treated; MSAT M1 = microscopy detects all those in the more infectious A and D states in the model and not those in the U state (microscopy sensitivity = 58% in the scenario shown); MSAT M2 = microscopy sensitivity is 75% for infected individuals regardless of infection state (D, A or U); MSAT M3 = microscopy sensitivity is 50% for infected individuals regardless of infection state (D, A or U).

Figure 3. Potential for elimination by mass treatment.

(A) Example simulation of a mass treatment programme in a non-seasonal scenario where multiple rounds of MDA are carried out, and the interval in between treatment rounds is less (every 4 months) or more (every 6 weeks) than the critical interval required to achieve <0.1% slide prevalence. (B) Model-estimated minimum frequency of mass treatment required in a large population to bring slide-prevalence to 0.1% or less for at least 50 consecutive days in non-seasonal settings: MDA versus MSAT with PCR as a screening tool. (C) Model-estimated annual mean slide-prevalence after 10 years of MDA repeated every year or every 6 months, according to baseline slide-prevalence prior to intervention. Moderate seasonality is assumed (see Figure 2A). D) example stochastic simulations of transmission over time assuming a small human population (n = 1000) with 1 round of MDA before the peak transmission season in year 1. A single simulation in which MDA succeeds in eliminating infection locally is shown in black. E) and F) Model-estimated probability of local elimination at different transmission intensities in a population of (E) 1000 or (F) 10,000, following MDA of different intensities. Results are based on 500 stochastic realizations per plotted point. Moderate seasonality is assumed (see Figure 2A). All simulations in this figure assume a low level of vector control at baseline which is maintained over time (the baseline slide-prevalence shown is in the presence of vector control).

Figure 4. Model-estimated combined impact of MDA and vector control and the potential for elimination.

(A) Vector control is scaled up in year 0, with or without 2 rounds of MDA during the first year. B), C) & D) Probabilities of local elimination in an isolated population of (B) 1000, (C) 10,000 or (D) 500,000 using vector control alone or in combination with MDA. Transmission is moderately seasonal and MDA is given prior to the transmission season.