Introduction Lymphatic filariasis (LF) is a disease of the poor that is prevalent in 73 tropical and sub-tropical countries. LF is caused by three species of filarial worms – Wuchereria bancrofti, Brugia malayi and B. timori – and is transmitted by multiple species of mosquitoes. The disease is expressed in a variety of clinical manifestations, the most common being hydrocele and chronic lymphedema/elephantiasis of the legs or arms. People affected by the disease suffer from disability, stigma and associated social and economic consequences. Marginalized people, particularly those living in areas with poor sanitation and housing conditions are more vulnerable and more affected by the disease. Estimates made in 1996 indicated that 119 million people were infected with LF at that time, 43 million of them having the clinical manifestations (principally lymphedema and hydrocele) of chronic LF disease [1]. Earlier severe resource constraints and lack of operationally feasible strategies in the endemic countries left a significant proportion of the LF endemic population living unprotected and exposed to the risk of LF infection. Despite a long-standing and gloomy outlook for these individuals, the situation turned around dramatically in the 1990s for 2 principal reasons: 1) advances made in point-of-care diagnostics and 2) the finding of the long-term effectiveness of anti-filarial drugs given in single doses that permitted development of the strategy of annual two-drug, single-dose mass drug administration (MDA) to control/eliminate LF [2], [3]. As LF had already been postulated to be an eradicable disease [4] and with the success experienced in LF elimination activities in China [5] and elsewhere, the World Health Assembly (WHA) in May 1997 formulated resolution WHA 50.29 urging all endemic countries to increase their efforts and determination to control and eliminate LF. In response, the WHO was able to launch the Global Programme to Eliminate LF (GPELF) in the year 2000, largely because the manufacturers of albendazole (ALB) and ivermectin, two of the principal drugs used in the GPELF MDAs, donated these drugs for as long as needed to eliminate LF [3]. The principal strategy of the programme has been two-fold: 1) to implement MDA programmes in all endemic areas to achieve total interruption of transmission and (2) to provide effective morbidity management in order to alleviate the suffering in people already affected by filarial disease. The GPELF targets elimination of LF, at least as a public health problem, by the year 2020 [6]. The programme to implement MDAs targeting LF (GPELF) completed 13 years of operations in 2012 [7]. With its ambitious goal to eliminate LF by the year 2020, it is essential that progress toward this goal be assessed repeatedly in order to set benchmarks to guide future programmatic planning. How to define and assess this progress remains a challenge, but two strategies have been suggested. The first is to measure reduction in the burden of LF disease (i.e., hydrocele, lymphedema, microfilaraemia and associated subclinical disease) over the past 13 years – i.e., a clinical perspective; the second is to measure reduction in the risk of acquiring infection for populations living in (formerly) endemic areas – i.e., an epidemiologic perspective. In the present report we have pursued the first alternative – to model the decreased burden of LF (defined for the purposes of our calculations as hydrocele, lymphedema, and microfilaraemia) in order to assess the progress towards LF elimination from inception of the MDA programme through 2012 (i.e., during GPELF's first 13 years). In a parallel study, others have recently modeled the programme's progress from the alternative, risk-of-infection viewpoint (Hooper et al., submitted). Methods A simple ‘force-of-treatment’ model was formulated to estimate the impact of MDA on LF infection and disease. Model parameters: Individual countries and regions as the geographic units of assessment The GPELF aims to provide MDA (using ALB+either ivermectin or diethylcarbamazine [DEC]) to entire endemic populations at yearly intervals for 4–6 years. Such a programme, if implemented effectively (i.e. treating at least 65% of the total population during each MDA), is expected to interrupt transmission and eliminate LF [8]. Because the status of MDA activities in all of the 73 endemic countries at the time of this analysis (through 2012) ranged from no MDA at all in some countries to full completion of the MDAs in others, for the present study each country was evaluated separately. First, programme impact was determined for each endemic country; then, the burden of LF remaining in each of the five endemic WHO regions – Southeast Asia (SEAR), Africa (AFR), Western Pacific (WPR), Eastern Mediterranean (EMR) and America (AMR) - was calculated by summing the remaining LF burden for all the endemic countries within each region. Model parameters: Key elements in assessing programme progress Calculating progress of the MDA programme under GPELF – whether by burden or risk estimates – is affected by a number of important specific factors, namely; (1) the number of countries that have successfully completed implementing the MDA programme, (2) the number of countries currently implementing the programme and the geographical coverage or proportion of the endemic population targeted so far in each country, (3) the treatment coverage of the population targeted for MDA in each country, and (4) the duration of the programme (i.e., the number of rounds of MDA implemented) in each country. For the present analysis, all of these data have been sourced from the WHO PC data bank [9]. Model parameters: Calculation of the decrease in LF burden to assess programme progress There are 3 essential steps to assessing the decrease of LF burden since 2000: first, the establishment of the LF base-line burden (in 2000); then, estimation of the MDA impact for countries or IUs where MDAs have taken place during 2000–2012; and, finally, estimation of current burden for countries or IUs where no MDA has taken place. (i) Establishment of base-line data The MDA programme under GPELF was started in the year 2000. To quantify the impact of the MDA programme, first, a base-line disease burden was estimated, considering the year 2000 as the base-line year. After extensive review of the literature in the mid-1990s, Michael et al. (1996) [1] and Michael and Bundy (1997) [10] estimated the LF prevalence and burden for different endemic regions. LF epidemiology is such that, without specific intervention or environment-altering measures, prevalence is unlikely to change over a short period (few years) of time. Hence, for this work the LF prevalence during 1996 to 2000 period is considered to remain unchanged. However, the absolute number of people affected by the disease will have increased because of population growth in the endemic areas. Taking the above factors into account, the base-line LF burden was estimated by extrapolating the prevalence data defined earlier [1] to the population of the endemic countries in the year 2000 (Table 1). As the LF burden estimation for individual countries was not always possible due to paucity and availability of data on prevalence, base-line LF burden estimates were made following the earlier approach of Michael et al. (1996) [1], and utilizing the convention that all the endemic countries for which no specific information was available, within each endemic region, have an approximately similar average prevalence of microfilaraemia and chronic disease. 10.1371/journal.pntd.0003319.t001 Table 1 Burden of LF in 1996 and 2000 considered as base-line to understand the impact of MDA (2000–2012) under GPELF. LF burden 1996 LF burden 2000 WHO Region Total Population endemic countries Mf carriers Lymphoedema cases Hydrocele cases Total infected Total Population endemic countries Mf carriers Lymphoedema cases Hydrocele cases Total infected SEAR 1335 41.91 9.49 14.53 61.86 1506 47.40 10.74 16.47 70.00 AFR 474 25.78 4.31 9.43 37.06 568 30.91 5.17 11.31 44.44 WPR 1113 11.14 1.52 1.87 13.32 1261 12.62 1.72 2.12 15.10 EMR 100 0.0598 0.0100 0.0199 0.0897 116 0.0700 0.0117 0.0233 0.1050 AMR 179 0.1252 0.0179 0.0179 0.1610 199 0.1397 0.0200 0.0200 0.1796 Total 3200 79.01 15.35 25.87 112.50 3650 91.14 17.66 29.94 129.82 All figures in millions. The 1996 estimates were based on the work done by Michael et al. (1996). The 1996 data were extrapolated to the populations of each endemic country in 2000 to derive the baseline estimated for GPELF. (ii) Estimation of MDA impact on LF burden for all countries or IUs with MDA in place Since the decrease in LF burden is a direct result of the treatment provided to populations during the MDA, the model to estimate this burden decrease can be described as a ‘force-of-treatment’ model (see below). To quantify this force-of-treatment, a ‘treatment index’ (TI) was constructed. The TI is defined as the average number of treatments taken by persons in areas included in MDA. It takes into account three key parameters – the size of the population targeted, the treatment coverage and the number of rounds of MDA implemented. These data can be sourced from the WHO PC data bank [9]. The TI is calculated as the total number of treatments consumed divided by the size of the population of IUs included in MDA. How to interpret what the TI implies about the effect of the programme's MDAs on LF burden can be determined from considering the empiric observations reported in earlier studies of endemic populations treated with the same treatment regimens as those used in the current MDAs; these were reviewed and are summarized below and in Figures 1 and 2. 10.1371/journal.pntd.0003319.g001 Figure 1 Empiric observations defining the relationship between number of treatments per person and % reduction in Mf prevalence 1 year later. 10.1371/journal.pntd.0003319.g002 Figure 2 Empiric observations defining the relationship between number of treatments and % reduction in hydrocele prevalence 1 year later. For microfilaraemia, two of the principal anti-filaria drugs used in MDA campaigns – DEC and ivermectin – have been recognized to exhibit remarkable, rapid effects on decreasing microfilaraemia. The anti-microfilarial effect of both drugs is further fortified when they are administered in combination with ALB, a broad spectrum anti- helminth drug that affects both adult worm viability and production of microfilariae [11]. The impact of treatment on microfilaraemia is evident from the first round of MDA and increases with each round of treatment year after year. While one round of mass treatment has been reported to reduce the Mf prevalence (assessed ∼1 yr post treatment) by 26% to 41%, 5–6 rounds led to 88%–90% reduction [12]–[21]. A review by de Kraker et al. (2006) [22] highlighted that both the drug combinations used in GPELF – ALB+DEC and ALB+ivermectin – strongly reduce the LF infection levels, but even 4–6 rounds of single-dose DEC alone can cause reduction of mf prevalence by as much as 86% [13], [23]. Hence, in the present effort to establish the relationship between the number of treatments and the % reduction in microfilaraemia prevalence, results were included from all the community level studies that administered annual single dose treatment (Figure 1), regardless of the specific MDA regimen employed. This empirically derived relationship between the number of treatments given and the decrease in microfilaraemia prevalence (Figure 1), in fact, defines the relationship between the TI and mf prevalence, since the TI is the population-level equivalent of the number of treatments administered at the individual-level. For microfilaraemia, there is a steady increase in reduction of prevalence as the treatment index increases, such that the reduction was close to 95% at a treatment index of about 6.0. For hydrocele, a similar review was undertaken of available information on the effect that treatment with anti-filarial drugs has on hydrocele prevalence [13], [24]–[29]. Treatment with DEC single dose was common to all of the studies providing results that were used in the analyses. Only one study each evaluated single dose of DEC+ivermectin [13] and ivermectin alone [29] and in both the studies the impact of these drugs was similar to that of DEC. The number of treatments given in these studies ranged from 2 to 12 and in most of the studies treatments were given at yearly or half-yearly interval. A model fitting the non-linear relationship (Fig. 2) was used to define the relationship between the number of treatments and % reduction in prevalence of hydrocele - again, defining the TI for the effect of MDA on hydrocele prevalence (Figure 2). This reduction increased progressively up to 4 treatments, but beyond that the treatment appears to have little additional impact; also, the maximum reduction seen with repeated treatments was approximately 60% (Figure 2). For lymphedema, different from microfilaraemia and hydrocele, information is scanty on the impact of annual MDA on lymphedema. Studies in Indonesia [30], [31], China [32], and Polynesia [24], all showed reduction in lymphedema prevalence, but all used more prolonged courses or different treatment regimens from those used in the GPELF MDAs. Post-GPELF, three studies evaluated the impact of MDA on lymphedema. In Ghana, one round of MDA with ivermectin and ALB showed no impact on lymphedema [33]. Administration of annual, single-dose DEC for 4 years in Papua New Guinea reduced the lymphedema prevalence by 20% [13]. Seven years of treatment in India showed 14% reduction in lymphedema prevalence in communities treated with annual DEC and 15% reduction in communities treated with ivermectin [29]. In light of these outcomes, a cautious and conservative approach was adopted for estimating the impact of MDA; it is postulated that for a TI of ≥3 (equivalent to nearly 4 rounds of MDA) lymphedema prevalence will be reduced by not more than 14%, the least reduction observed with annual MDA [29]. A TI 1.9 billion treatments were delivered, prevented 7.4 million cases of hydrocele and 4.3 million cases of lymphedema. While these estimates on the number of hydrocele cases prevented are similar to the estimates in the present study, there is less agreement on the number of lymphedema cases prevented. The estimated 5.49 million lymphedema cases prevented in this study, after 13 years of MDA and delivery of 6.37 billion treatments, was lower, likely because of both the different strategies for calculating the effects and the conservative approach adopted in assessing the impact of MDA on lymphedema. The estimated 5.49 million lymphedema cases prevented in this study was a minimum number, and the actual reduction may be much higher. Of the various factors influencing the outcome of MDA programmes, treatment coverage is particularly important [8]. In this study, the impact of MDA was assessed using the reported treatment coverage – i.e. the treatment coverage reported by the country level programme managers and compiled in WHO's PC data bank [9]. There are, however, a number of reports suggesting that the programme-reported treatment coverage in the South-east Asia region, particularly in India, may be higher than the actual treatment coverage in the communities. For example, while programme-reported treatment coverage in India was generally in the range of 58% to 90%, various independent studies showed treatment coverage that varied widely and ranged from 90% in different parts of the country [58]–[74]. The data from these published studies give rise to an average ‘evaluated’ treatment coverage rate of 51.0%, less than the 71.33% average reported national coverage [9]. Since the TI used to calculate programme impact in our model incorporates programme coverage, it is necessary to understand the effect of this difference between reported and evaluated coverage. For India, the TI based on reported coverage was 5.27, but only 4.21 when based on ‘evaluated’ coverage – a difference of 20%. Interestingly, however, when those different TI's were applied to the model (Figs. 1 & 2), the effect was minimal, because for TI's >4, little or no additional benefit was achieved on the 3 parameters measured (microfilaraemia, hydrocele, lymphedema/elephantiasis). In other words, the initial rounds of MDA will exert greater impact on these manifestations compared to later rounds, a finding already reported empirically and shown in various studies [12], [13], [15], [17]–[20]. However, if the treatment coverage rate is high, a higher TI can be achieved in the early rounds of the programme, and fewer rounds of MDA may be required to maximize both impact and cost-effectiveness. It is possible that preventive chemotherapy as well as other interventions implemented against other vector-borne diseases have added to the impact of LF MDA and caused further reduction in LF burden in some countries. Principal among these other interventions are the ivemectin distribution under the African Programme for Onchocrciasis Control (APOC) and the malaria control measures of insecticide treated nets (ITN) and indoor residual spraying (IRS). Currently, ivermectin is distributedfor onchocerciasis control in as many as 26 countries in Africa, covering nearly 130 million population [75]. Most of the 26 countries are co-endemic for LF also and while less than half of this LF-endemic population is under specific treatment as part of the GPELF, many are likely receiving benefit from the ivermectin being used for onchocerciasis control, as has been demonstrated specifically in a number of countries in West Africa [76]–[80]. Similarly, the malaria control measures have been shown to reduce LF transmission considerably and remain promising adjuncts to the MDA of the GPELF activities [81]–[83]. While these coincident intervention measures have, and will continue to have, positive impact on the LF elimination efforts, quantification of their impact remains a daunting challenge. The reduction in LF burden achieved during the GPELF's first 13 years is almost certainly higher than shown through our analyses both because of the additional, on-going intervention measures and because of our conservative approach to estimating the impact on chronic disease. Though, there can be little question that impressive gains in decreasing LF burden have been achieved as a result of 13 years of MDA in the GPELF, still, however, a considerable burden of LF remains – estimated at 36.45 million Mf cases, 16.68 million cases of lymphedema and 19.43 million cases of hydrocele (Table 4). Extension of MDA to all at-risk countries and to all regions within those countries where MDA has not yet started is absolutely necessary to reduce the number of microfilaraemia cases and transmission. Such an extension of MDA will also reduce a proportion of hydrocele and lymphedema cases, but the burden of LF disease needs also to be approached directly. Techniques for effective morbidity management – both medical and surgical – are available but not nearly so widely implemented as they could or should be. The present model's calculations take into consideration only those burden-reducing benefits coming pari passu with MDA implementation. When appropriate morbidity management strategies are finally introduced and accelerated, the burden of LF disease will fall even more dramatically (and the model can be adapted accordingly).
