Currently available anti-filarial drugs (diethylcarbamazine – DEC, albendazole and ivermectin) differ in their mode of action. Available evidence suggests that ivermectin and albendazole have direct killing effect on microfilariae (mf) whereas DEC acts through immune mechanisms 8. The heterogeneity in response to DEC 9 also indicates that DEC does not kill the parasite directly. In addition, the immune-mediated effect is expected to last longer, which may influence model predictions. Estimates from a model based on analysis of data from a hospital trial showed that a single dose DEC (6 mg/kg) can reduce the mf-production of adult worms by 67% at 12 months post-treatment and 87% at 24 months post-treatment, and reduce the mf-density in the peripheral blood by 57% at 12 months and 52% at 24 months post-treatment 910. The increasing effect of DEC on mf-production up to 24 months post-treatment is suggestive of slow immune mediated damaging effect of DEC on adult worms, which occurs at an extended time scale. However, reliable efficacy estimates, at least for DEC, need to be obtained through meta-analysis by generating data from clinical trials with different periods of follow-up. Also, models should account for the heterogeneity in drug response by individuals as well as for the average duration of the drug effect ('effective duration'). In LYMFASIM the heterogeneity in drug-response by individuals and the 'effective duration', measured as the period during which the female parasite recovers from treatment, are considered as stochastic variables. However, only the former is being used for model prediction.

Since efficacy estimates (measured as percent reduction of mf load with respect to baseline) vary between clinical trials, the most reliable estimates are based on rigorous meta-analysis of data from randomized controlled trials. Such estimates are available for ivermectin 11 and also for combination therapies of albendazole with DEC or ivermectin, albeit based on a limited number of trials 12. Similar analyses for DEC and albendazole monotherapy may be useful to more confidently predict the synergistic effect of combination therapy, and determine its superiority over a single dose regimen.

DEC is known to have also a macrofilaricidal effect, i.e., to be effective against adult worms 131415. There are similar indications for albendazole alone 16 or in combination with DEC 171819. However, ivermectin seems to have no macrofilaricidal effect 20, although model based estimates of its effect on mf-production by adult worms suggests that it may have a damaging effect on the reproductive system of female worms 911. In the absence of diagnostic tools to measure adult worm burden accurately, only indirect estimates have been used in both models 4567. This uncertainty will have greater impact on the predicted durations of MDA for LF elimination given the fact that the average and distribution of adult worms' life expectancy has a large influential effect on the population dynamics 21.

When co-administration of albendazole with either DEC or ivermectin was initially proposed, albendazole was not intended or expected to enhance the efficacy of either DEC or ivermectin. However, limited trials with the recommended combination of albendazole with ivermectin suggest that co-administration of ivermectin with albendazole is superior to either of the two drugs given alone. In the case of DEC co-administered with albendazole, the results have not been so equivocal, although the combination therapy appears to have an edge over either drug given alone 22. While additional studies may be useful to reliably obtain efficacy estimates as well as to determine the superiority of the combination, if any, current policies are clearly in favour of the combination.

Repopulation of the blood by mf depends on the effects of anti-filarial drugs on host immunity and worm fecundity. Drugs can improve host immunological responsiveness without affecting the viability and fecundity of adult worms (DEC 8; ivermectin 23). This in turn can clear mf during repeated treatments and thus reduce microfilaraemia. Drugs may also express their action in different ways such as sterilizing, killing or affecting the fecundity of the adult worms. These effects can affect the repopulation of mf in the blood. Anti-fecundity effects may be manifested in two ways: (i) by an overall reduction in reproductive capacity across all worms, or (ii) by sterilization (temporary or permanent) of a proportion of the female worms. Temporarily sterilized worms may resume their reproduction and release mf after a period of recovery. Both these effects may vary between individual worms and therefore may be stochastic in nature. That anti-fecundity effects may vary between worms has been incorporated into LYMFASIM, but its implications have not yet been fully explored or understood.

Clinical trials of single (DEC or ivermectin) or combination (albendazole plus DEC or ivermectin) drug therapies are limited to provide estimates for the change in efficacy (measured as percent reduction of mf load with respect to baseline) after repeated treatments. However, available trial results suggest that efficacy of single drug therapy (DEC or ivermectin) increases with multiple single dose treatments at a constant rate and reaches a plateau after a few treatments 2425. This effect does not necessarily imply that the drug has become less efficacious but that the percent reduction in microfilaraemia effected by treatment may also depend on mf load at the point of treatment, with the highest reductions recorded for heavier microfilaraemia, i.e., microfilaricidal efficacy declines with decreasing mf load (density-dependent efficacy). (For an individual, whose initial mf load is 100, a decrease to 5 following treatment would mean a reduction by 95%; whereas for a person with 5 mf ending up with 1 mf, the reduction will be of the order of 80%). Such density-dependent efficacy is also apparent from a study in which four repeated full (each with 6 mg/kg for 12 days) courses of treatments were given over a period of one year: efficacy did not vary among those who received one (90.4%), two (95.7%), three (97.4%) and four (100%) courses of treatment. However, those who required at least 2 courses of treatment, had a higher pre-treatment mf-density than those who needed only one-course (6 mg/kg × 12 days) of treatment 26. Similar observations are also evident from community based trials with single dose DEC (6 mg/kg) administered monthly 2728, or semi-annually 29 for one year and annually for four years 30 and DEC (6 mg/kg) or ivermectin (400 μg/kg) for 10 years (Fig. 1A31).

The efficacy of co-administration of albendazole (400 mg) with DEC (6 mg/kg), or ivermectin (150 μg/kg) is constant and independent of the number of repeated treatments (albendazole plus DEC: 92%, 95%, 98% and 94% following, respectively, the first, second, third, and fourth treatments 19; albendazole plus ivermectin: 87% and 93% following, respectively, the first and second treatments 32) i.e. the efficacy is maximum upon first treatment, decreasing afterwards with the increasing number of treatments and decreasing mf-load (density-dependent efficacy). The results of field trials with repeated administration of DEC plus ivermectin or DEC plus albendazole are also suggestive of an initial constant rate of increase and decline after few repeated treatments (>90% reduction in each of four annual rounds of MDA with DEC plus ivermectin 30; 70–96% reduction over four rounds of MDA with DEC plus albendazole 33; Fig. 1B31).

The currently available predictions of the prospects of elimination by both EPIFIL and LYMFASIM are based on the assumption that single drug efficacy (DEC, ivermectin, and albendazole) is constant and independent of the number of repeated treatments 4534. But with the evidence of change in efficacy after repeated treatments with single or combination therapies, as well as the consideration that drug resistance may emerge (see section on population genetics of drug resistance below), the chemotherapy components of the models need to be updated as density-dependent processes, relaxed by the effects of interventions, can considerably alter the prospects of elimination 3536.

Drug resistance is one of the causes for treatment failure in the control of many parasitic diseases. Large scale MDA with ivermectin in areas of the Onchocerciasis Control Programme in West Africa has raised concerns about the possibility that individuals who have been repeatedly treated may exhibit suboptimal responses to ivermectin (their rate of skin repopulation by mf of Onchocerca volvulus is faster than expected 3738). In populations of Wuchereria bancrofti in Ghana, the frequency of a phenylalanine to tyrosine change at position 200 of the beta-tubulin gene (TYR200, a marker for resistance to benzimidazoles in veterinary nematodes), was found to be significantly higher in microfilariae obtained from patients treated once or twice with albendazole plus ivermectin than in mf from those untreated by the GPELF 39, whereas, analysis of W. bancrofti populations in India revealed absence of this genetic change, suggesting albendazole susceptibility 40. These results indicate that selection for albendazole and ivermectin resistance may already be occurring in some localities. Though strong evidence for resistance to DEC is lacking 4142, more studies are required to understand whether or not selection under chemotherapeutic pressure is indeed occurring, to elucidate the possible mechanisms of resistance, as well as to measure the frequency of putative resistance markers in a broad spectrum of epidemiological settings for sound incorporation of worm genetic heterogeneity with respect to drug susceptibility in the models. Recently, EPIFIL has been modified to account for parasite population genetics, and has been used to examine the consequences of a hypothetical spread of benzimidazole resistance on the impact of MDA with albendazole in combination with ivermectin or DEC 43, or the impact of multidrug resistance (to ivermectin and albendazole) under various assumptions of the genetics of drug resistance 44. The LYMFASIM model can also be extended in a similar way; its microsimulation approach may prove to be more effective for accounting of any genetic variability between individual parasites within hosts that may be identified by molecular genetics to improve its predictive value.

Programmes report only coverage of drug distribution; no data on actual consumption of the drug are currently available. So far both models consider coverage as a proxy for the actual consumption, but there are considerable gaps between coverage (measured as the proportion of the population that received the drug) and consumption (measured as the proportion of the population that actually consumed the drug) 45. Quantification of their relationship 46 and its incorporation into models would enhance the robustness of model predictions.

A better understanding of the processes regulating infection in the human host is essential for predicting the long-term impact of MDA and assessing the prospects of elimination. Age-infection (prevalence and intensity) profiles derived from cross-sectional surveys have been used to interpret the processes that shape the age-infection profiles. Four types of age-intensity profiles can emerge from cross-sectional data 63. The shape of age-infection profiles is determined by the following processes: (i) age-dependent exposure, (ii) parasite-induced host mortality, (iii) heterogeneity within the host population, (iv) clumped infection, (v) density-dependent parasite mortality, and (vi) density-dependent parasite establishment. (If age profiles are investigated by using indirect measures of infection, which they usually are, i.e., microfilariae rather than adult worms profiles, then density-dependent female fecundity will also be important). Increase in infection (mf) prevalence/intensity with age either with (linear-saturating) or without (linear) reaching a plateau at adult age suggests that acquired immunity may not be very important, whereas an increase with age to reach a peak followed by a decline (convex pattern) suggests the operation of either acquired immunity or a reduction in parasite acquisition as a consequence of decreasing exposure with increasing age 6263. The former also being accompanied by a decrease in the age at which peak infection is reached (or peak shift 64). Application of both EPIFIL and LYMFASIM models to Pondicherry data and EPIFIL to East African data showed that the models had to include acquired immunity to explain the observed convexity in the age-infection profiles of W. bancrofti mf, although there is indication of transmission dependency 65. Despite strong indication for the role of acquired immunity from Pondicherry (Fig. 2366) and East Africa 65, a recent meta-analysis of age-prevalence data from India and African countries casts doubts over the role of immune mechanisms in regulating LF infection in humans 67. The impact of this uncertainty on the predicted duration of control would have serious implications on the decision to stop/continue MDA programmes. Premature cessation of MDA programmes may reduce the ability of developing an immune response and thus increase the likelihood of resurgence. This will be further worsened if drugs do not act on adult worms or their productivity. Therefore both EPIFIL and LYMFASIM models need to be validated in different epidemiological settings to explore the plausible mechanisms/processes regulating infections in humans.

At present both EPIFIL and LYMFASIM models have considered very simple assumptions of disease dynamics 12. Fits of the EPIFIL model to the Pondicherry situation indicated that disease progression (lymphoedema/hydrocoele) is a consequence of worm burden, but not necessarily associated with immunopathological mechanisms. However, available evidence suggests that, although progression to hydrocoele may be associated with worm burden, progression to lymphoedema is due to immunopathological reactions and secondary bacterial infections 68. Further progression of lymphoedema depends on the frequency of acute episodes of adenolymphangitis (ADL), and vice versa 69, and therefore models should also consider other modalities of disease progression. LYMFASIM should be extended to include the relationship between infection and disease, and the estimation of these parameters for quantification and validation. Prevention of disease incidence can be a long term outcome of GPELF with the goal of elimination of LF as public health problem. It would also be important to assess cost and cost-effectiveness of MDA for disease prevention, and of morbidity management (hydrocoelectomy/limb care) for disability prevention.