Table 1.
Characteristics of diagnostic tests used in identified studies.
Fig 1.
PRISMA flow diagram of the inclusion and exclusion of studies identified in the literature search.
Table 2.
Summary of identified age-stratified clinical VL incidence studies.
Table 3.
Summary of identified age-stratified infection prevalence studies.
Table 4.
Infection prevalence by different diagnostics and VL status.
Fig 2.
Age-specific visceral leishmaniasis incidence (cases per 1000 individuals per year) in different studies in the Indian subcontinent.
Fig 3.
Age-prevalence distributions of positivity on different diagnostic tests for Leishmania donovani infection in the Indian subcontinent.
(A) Seropositivity by Direct Agglutination Test (DAT) (numbers in parentheses denote cut-off for positivity), (B) seropositivity by rK39 Enzyme Linked Immunosorbent Assay (ELISA) or Rapid Diagnostic Test (RDT), (C) Polymerase Chain Reaction (PCR)/quantitative PCR (qPCR) positivity for parasite DNA, (D) LST positivity. All the prevalence studies include individuals with active or dormant asymptomatic infection, and some include a small number of active VL cases or exclude past VL cases. The year in which the survey was performed is shown in square brackets.* Prevalence includes a small number of active clinical VL cases.** Past VL cases excluded from prevalence.*** Data includes two individuals who were rK39 RDT+ but qPCR-.
Fig 4.
Comparison of age-prevalence distributions of positivity on different diagnostic tests for L. donovani infection in studies with multiple diagnostic tests.
(A) Bern et al [24,38]: prevalences of rK39 ELISA and LST positivity, (B) Hasker et al [23]: rK39 ELISA and DAT seroprevalences (C) Schenkel et al [68]: DAT seroprevalence and LST positivity, (D) Topno et al [70]: DAT, rK39 RDT and PCR test positivities. * Prevalence includes a small number of active clinical VL cases.** Past VL cases excluded from prevalence.
Fig 5.
Estimated conversion and reversion rates, λ and γ, from (A) age-independent (λ = b0 = constant) and (B) age-dependent (λ = b0 + b1 × age) reversible catalytic models for the age-prevalence distributions of infection for the different studies in Table 3. Crosses show the maximum likelihood estimates for (λ,γ) for each study, dots (λ,γ) estimates from longitudinal studies [23,24,50], and lines their 95% confidence intervals. λ estimates in (B) are at age 20yrs (i.e. λ = b0 + 20b1).