Malaria vaccine efficacy: the difficulty of detecting and diagnosing malaria

Abstract

New sources of funding have revitalized efforts to control malaria. An effective vaccine would be a tremendous asset in the fight against this devastating disease and increasing financial and scientific resources are being invested to develop one. A few candidates have been tested in Phase I and II clinical trials, and several others are poised to begin trials soon. Some studies have been promising, and others disappointing. It is difficult to compare the results of these clinical trials; even independent trials of the same vaccine give highly discrepant results. One major obstacle in evaluating malaria vaccines is the difficulty of diagnosing clinical malaria. This analysis evaluates the impact of diagnostic error, particularly that introduced by microscopy, on the outcome of efficacy trials of malaria vaccines and make recommendations for improving future trials.

Figure 1

Estimated vaccine efficacy as a function of sensitivity and specificity for a vaccine with a true efficacy of 80% and an attack rate of 5%. Adapted from [3].

Figure 2

The sensitivity of microscopy as a function of parasite density. The percentage of microscopists (n = 25) reporting a false negative slide versus the mean parasite density reported by the remaining microscopists (adapted from [38]).

Figure 3

Variable sensitivity can obscure differences between vaccinated and unvaccinated individuals in challenge studies. If blood smears are read by two different microscopists at time = t₁ and time = t₂, differences in sensitivity could result in no difference seen between the two study subjects. Microscopist A with an LOD of 50 p/μl would read negative smears at time t₁. Microscopist B with an LOD of 10 p/μl would read positive smears at time t₂. An initial number of merozoites are released from the liver at time t = 0 and assumed to grow exponentially with 16 merozoites produced from a single merozoite every 48 hours. The table inset gives the "window of detection" of blood-stage infection when the sensitivity varies between 10 and 100 parasites per microliter for different numbers of primary merozoites released from the liver.

Figure 4

Discrepancy in density between two microscopists reading a single slide as a function of parasite density. Adapted from [8]. Parasite density measurements less than 10,000/μl were assumed to have an error described by the equation: Error = 0.61 – 0.054*Ln(Density). Densities greater than or equal to 10,000/μl were assigned an error of 10%.

Figure 5

Percent error associated with vaccine efficacy increases as vaccine efficacy increases. Plot shows results for 100 and 1,000 individuals per arm. VE=1−D¯VD¯UV MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaacqWGwbGvcqWGfbqrcqGH9aqpcqaIXaqmcqGHsisldaWcaaqaamaanaaabaGaemiraqeaamaaBaaaleaacqWGwbGvaeqaaaGcbaWaa0aaaeaacqWGebaraaWaaSbaaSqaaiabdwfavjabdAfawbqabaaaaaaa@382A@ D¯V MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaqdaaqaaiabdseaebaadaWgaaWcbaGaemOvayfabeaaaaa@2F2F@ is the mean density in the vaccinated group and D¯UV MathType@MTEF@5@5@+=feaafiart1ev1aaatCvAUfKttLearuWrP9MDH5MBPbIqV92AaeXatLxBI9gBaebbnrfifHhDYfgasaacH8akY=wiFfYdH8Gipec8Eeeu0xXdbba9frFj0=OqFfea0dXdd9vqai=hGuQ8kuc9pgc9s8qqaq=dirpe0xb9q8qiLsFr0=vr0=vr0dc8meaabaqaciaacaGaaeqabaqabeGadaaakeaadaqdaaqaaiabdseaebaadaWgaaWcbaGaemyvauLaemOvayfabeaaaaa@3062@ is the mean density in the unvaccinated group. The frequency of densities in a cohort is described by a gamma distribution with a mean density of 3,000 parasites per microliter in the unvaccinated group and 1,500 parasites per microliter in the vaccinated group. Error associated with each density measurement is calculated using the error model in Figure 4 and propagated in the calculation of vaccine efficacy using standard error propagation methods.

Figure 6

Hypothetical density distribution of clinical episodes of malaria. The distribution of parasite densities among febrile cases is approximated by a gamma distribution with a mean of 3,000/μl in the unvaccinated group and vaccination reduces both the incidence of fever (25% reduction) and the mean parasite density (30% reduction) among the cohort. A cut-off value of 3,000 parasites per microliter is shown and the shaded area represents the cases which risk being misclassified due to the discrepancy between microscopists.

Figure 7

The observed vaccine efficacy (solid line) and the upper (red dashed line) and lower (blue dashed line) limit of possible outcomes as a function of threshold density based on the uncertainty associated with measurement of parasite density. Frequency of cases by density is assumed to follow the distribution in Figure 6. The upper and lower bounds are calculated using the error model in Figure 4.