Modelling Anti-Ov16 IgG4 Antibody Prevalence as an Indicator for Evaluation and Decision Making in Onchocerciasis Elimination Programmes

Abstract

Background: Onchocerciasis is targeted for elimination in Africa through annual or biannual ivermectin mass drug administration (MDA). An immunodiagnostic test, based on the detection of human IgG4 antibodies in the blood to the Onchocerca volvulus-specific antigen Ov16, is one of the recommended tools for determining whether transmission is interrupted and mass treatment can stop. For different transmission settings, the relationship between post-MDA Ov16 antibody prevalence in children (measured 1 year after the last round of MDA) and the duration and coverage of MDA, the mf prevalence in the population, and the probability that onchocerciasis is eventually eliminated is explored through mathematical modelling.

Methodology: The ONCHOSIM model was extended with new output on the Ov16 antibody serostatus of individuals. Seroconversion was assumed to be triggered by the first worm establishing in the host, with seroconversion occurring either before maturation, after maturation or only after the start of mf production. We are mainly interested in seroconversion rates in children, and for now ignore the possibility of seroreversion to simplify the model.

Principal findings: Yearly repeated MDA leads to a strong reduction in the parasite acquisition rate in humans. This reduces the seroconversion rate in newborns and young children, while those who seroconverted before the start of control remain antibody positive. Both the microfiladermia prevalence in the population aged 5 years and above and the Ov16 antibody prevalence in children under 10 declined with increasing duration of MDA. The association between either of these indicators and the model-predicted probability of elimination was not influenced much by the assumed treatment coverage levels, but was found to depend on baseline endemicity levels, assumptions regarding the trigger of seroconversion, and diagnostic test characteristics (sensitivity and specificity).

Conclusions: Better understanding of the dynamics of Ov16 antibody responses is required for accurate interpretation of seroprevalence data and more precise estimation of endpoint for MDA. Our study demonstrates that this endpoint will be dependent on baseline endemicity levels, which should be taken into account in guidelines for defining when to stop MDA.

Fig 1. Model-predicted trends in mf prevalence for the population aged 5 years and above during 25 years of annual ivermectin mass treatment, with the first treatment provided at time = 0.

Average of 1,000 simulations (minus failed runs) per scenario. Results are shown for transmission settings with moderate and high transmission (ABR 10,150 and 18,078; average pre-control CMFL 10 and 55 mf/skin snip, respectively) and for different treatment coverage levels. The curves connect yearly model-predicted mf prevalence levels, always measured just before a treatment and exactly one year after the previous treatment. For clarity, we do not show trends between these yearly measurements (immediate drop in mf prevalence followed by a gradual increase until the next treatment). See S2 Appendix for a stochastic variant of the figure. The last measurement shown is 1 year after the last treatment.

Fig 2. Model-predicted trend in mf-prevalence in the population aged 5 years and above and Ov16 antibody prevalence in children aged 0–9 years in relation to the duration of annual mass drug administration, assuming a fixed coverage of 70%.

Average of 1,000 simulations (minus failed runs) per scenario. Results are shown for transmission settings with moderate and high transmission (ABR 10,150 and 18,078; average pre-control CMFL 10 and 55 mf/skin snip, respectively) and for different treatment coverage levels. The Ov16 antibody prevalence was estimated assuming that the Ov16 antibody test has a sensitivity of 80%, and specificity of 99%. Results are shown for each of three hypotheses regarding the seroconversion trigger, with the Ov16 antibody test becoming positive as soon as the first male or female worm establishes in the human body, before it matures (hypothesis 1); idem, but after maturation (hypothesis 2); or after the start of mf production (hypothesis 3).

Fig 3. Model-predicted age patterns of mf and Ov16 antibody prevalence in relation to the number of mass treatment rounds provided, assuming that treatment coverage is 70%.

Average of 1,000 simulations (minus failed runs) per scenario. Results are shown for transmission settings with moderate and high transmission (ABR 10,150 and 18,078; average pre-control CMFL 10 and 55 mf/skin snip, respectively) and for different treatment coverage levels. The Ov16 antibody prevalence was estimated according to hypothesis 2, assuming that the Ov16 antibody test has a sensitivity of 80% and specificity of 99%, and that seroreversion does not occur.

Fig 4. Probability of elimination in relation to the duration of mass treatment and the one-year post-treatment mf prevalence or Ov16 antibody seroprevalence.

A) The probability of elimination in relation to treatment duration, assuming that treatment would be discontinued after the indicated treatment duration. B) Probability of elimination in relation to the post-MDA mf prevalence in the population aged 5 years and above, measured one year after 1, 2, 3, … 25 treatment rounds, assuming that no further treatments take place. The lines connect outcomes for different durations of MDA, for the indicated coverage levels (60%, 70%, or 80%). C), similar to B, with probability of elimination shown in relation to the post-MDA Ov16 antibody prevalence in children aged 0–9, measured one year after the 1, 2, 3, … 25 treatment rounds. The Ov16 antibody prevalence was estimated according to hypothesis 2, assuming that the Ov16 antibody test has a sensitivity of 80%, and specificity of 99%. Note that the values on the X-axis in B and C is sorted from highest to lowest, for comparability with A. Results are shown for transmission settings with moderate and high transmission (ABR 10,150 and 18,078; average pre-control CMFL 10 and 55 mf/skin snip, respectively). See Fig 2 for the information on the mean mf and Ov16 antibody prevalence in relation to the duration of MDA.

Fig 5. Probability of elimination in relation to the average post-MDA mf prevalence or Ov16 antibody prevalence by pre-control CMFL.

The mf prevalence was assessed in the population aged 5 years and above and Ov16 antibody prevalence in children aged 0–9 (hypothesis 1–3), one year after 1, 2, 3, … 25 treatment rounds with a treatment coverage of 70%. The probability of elimination was assessed after each duration, assuming that treatment would be discontinued thereafter. The separate lines connect outcomes on different treatment durations for a given transmission setting and baseline CMFL. The Ov16 antibody prevalence was estimated assuming that the Ov16 antibody test has a sensitivity of 80%, and specificity of 99%. Note that the horizontal axis is ordered from highest to lowest.

Fig 6. Probability of elimination in relation to the average post-MDA mf prevalence or Ov16 antibody prevalence, by pre-control CMFL and age group tested for Ov16.

The mf prevalence was assessed in the population aged 5 years and above and Ov16 antibody prevalence in children aged 0–9 (hypothesis 2), one year after 1, 2, 3, … 25 treatment rounds with a treatment coverage of 70%. The probability of elimination was assessed after each duration, assuming that treatment would be discontinued thereafter. The separate lines connect outcomes on different treatment durations for a given transmission setting. The Ov16 antibody prevalence was estimated assuming that the Ov16 antibody test has a sensitivity of 80%, and specificity of 99%. Note that the horizontal axis is ordered from highest to lowest.

Fig 7. Probability of elimination in relation to the Ov16 antibody prevalence as measured one year after the last treatment, in relation to test characteristics, for scenarios with treatment coverage of 65%.

In the figures on the left, sensitivity is varied while specificity is fixed at 99%. In the figures on the right, the sensitivity is fixed at 80%, while specificity is varied. Note that the horizontal axis is ordered from highest to lowest. The lines connect outcomes for different treatment durations with otherwise the same assumptions. The Ov16 antibody prevalence was estimated according to hypothesis 2 in 0–9 year old children.