Nigeria is a Federal Republic comprising 36 States and its Federal Capital Territory, Abuja [35] [36]. The states are grouped into six geopolitical zones, the North Central (NC), North East (NE), North West (NW), South West (SW), South East (SE) and South (SS). Nigeria covers an area of approximately 923,768 sq. km, and has a large low plateau intersected by two major rivers, the Niger and Benue, in the central region of the country (Figure 1). It shares borders with Benin in the west, Chad and Cameroon in the east, and Niger in the north. Its coast in the south lies on the Gulf of Guinea on the Atlantic Ocean and Lagos, the former capital, is an important port city. Nigeria is Africa's most populous country with the total population estimated to be 160 million in 2012, with approximately 50% living in urban areas.

To review and synthesize the current knowledge of the human distribution of LF in Nigeria, a systematic search for data in peer-reviewed published literature and national reports was carried out. The search was conducted using PubMed, JSTOR, Google, SCOPUS and other online scientific and historical databases. References were also obtained from the references listed within articles, and then from the references within those articles.

Studies and reports with data on the prevalence of i) LF infection as circulating filarial antigen (CFA) from using immunochromatographic tests (ICTs), antibodies by ELISAs, and microfilaria (Mf) from blood slides, and ii) disease cases (hydrocele, lymphodema) [2] [37] were identified and collated into a database. Information on the location/collection site (village, local government area (LGA) and State), and time period (month, year), was also collected for mapping and descriptive analyses. Specific information on whether MDA for LF had been administered prior to the LF prevalence measure was recorded and considered in the analysis. The range of methods used to detect LF in the different studies was recorded, as well as information on the mosquito species, which was cross-checked with the Nigerian Anopheles database [19].

The locations of the community or collection site were geo-referenced using the latitude and longitude coordinates obtained from references directly or by cross-checking the names with data from the GEOnet Names Server, Directory of Cities and Towns in the World databases [38] [39]. The coordinates of the midpoint of the LGA was used as a proxy for the locations that could not be allocated exact latitude and longitude coordinates. This is considered to be a limitation of the review and restricts any accurate detailed mapping. It is also acknowledged that LF prevalence distribution has a degree of bias as the data are based on the locations selected by the investigators in the original study, and does not take sampling methodologies between studies into account, which may affect the outcome.

In addition, selected data from the Federal Ministry of Health (FMoH) collected during LF mapping activities were collated and included in the database. The LF data available for this study were based on Mf prevalence rates collected in selected LGA sentinel sites during baseline surveys in 31 LGAs across 18 States of Nigeria. The WHO standard protocol was used to collect blood samples at night and examined for the presence of Mf. The coordinates of the midpoint of each LGA was used to map the LF prevalence. A national LF endemicity map by LGA was also available from the FMoH, which provided an overall CFA prevalence based on ICT survey in each State carried out between 2000 and 2010. Specific LGA data is not publicly available and not included in this database, however, the State-level information on the number of LGAs surveyed, prevalence range and year of survey is available in the recently published Master Plan for NTDs [35].

All the relevant information was entered into an Excel worksheet and data analysis was performed using Stata software (version 12, StataCorp, Texas, USA). All data were mapped using the geographical information systems (GIS) software ArcGIS 10.0 (ESRI, Redlands, CA) to produce maps of LF prevalence distributions, and to examine the geographical overlaps with loiasis-endemic areas, and the different intervention distributions.

To examine the potential extent of LF and L. loa co-endemicity, the recent map of the predicted loiasis prevalence produced from a Rapid Assessment Procedure of Loiasis (RAPLOA) based on eye worm history carried out between 2004 and 2010 across Africa, including Nigeria [7], was imported into ArcGIS. Three levels of predicted loiasis prevalence were digitised (i.e. outlined, shaded) based on the defined distribution boundaries, which included low <20%, medium 20–40% and high >40% prevalence areas; the latter is equivalent to mf prevalence of >20%. The different levels of loiasis prevalence and the overlap with LF prevalence distributions were highlighted to help identify potential low risk (i.e. loiasis <20%) and medium to high risk (i.e loiasis >20%) SAE areas.

In total, 41 studies [41]–[81] from 68 published and unpublished filariasis studies identified in the literature were found to have examined the prevalence, clinical manifestations and entomological aspects of LF in Nigeria (Table S1). The studies excluded from the review had reported data on L. loa and/onchocerciasis only. The majority of LF studies (n = 30) were conducted post 2000 [41]–[69], nine studies were conducted between 1980 and 2000 [70]–[78], and three studies were conducted pre 1980 [79]–[81]. The studies indicate that LF is present in 19 States across all six geopolitical zones of the country (Figure 2a). The majority of studies were from the NC geopolitical zone, with the most comprehensive studies carried out in Plateau and Nassarawa States [49] [53] [54] [63] [66].

The FMoH sentinel site Mf prevalence data were carried out more widely in 31 LGAs across 18 States. All information was added to the database (Table S1) and mapped with the other specific Mf data described above (Figure 2a). The FMoH national endemicity map indicated that out of the 774 LGAs in Nigeria, 541 were classified as endemic, 164 were classified as non-endemic and 69 remained to be mapped (Figure 2b). The related state-level data are found in Nigeria Master Plan for NTDs [35].

The range of methods used to detect the presence of LF in Nigeria included serological methods (using ICTs or ELISA), parasitological methods (blood films for Mf) and physical examination for clinical manifestations (lymphodema, hydrocele), and were used either alone or in combination. Studies carried out before the 1980s only used parasitological examination of blood films, whereas post 2000, a combination of serology and parasitological methods were most widely used (Table S1).

In total there were 258 individual data points from 152 unique locations where the prevalence of W. bancrofti was measured by CFA or Mf. The average CFA and Mf rates by State are summarized in Table 1 and 2. The overall mean CFA prevalence rate across the country was 14.0% (n = 134; range 0% to 66.0%), and the overall mean Mf prevalence rate was 8.2% (n = 162; 0 to 47.4%). The CFA and Mf prevalence rates by geopolitical zones indicate that the highest rates occur in the NC, NE, NW and SS zones. The highest CFA prevalence rate recorded was 66% recorded at Ogi-Utonkon, Ado LGA, Benue State, NC [64], while the highest Mf rate was 47.4% recorded at Zing LGA, Taraba State, NE [55] (Table S1).

Overall, there were marked differences in W. bancrofti prevalence at sites that had not received MDA i.e. pre-MDA, compared to those sites that had received MDA i.e. post-MDA. The average CFA prevalence in pre-MDA sites, was 20.3% (n = 68; range 0% to 66.0%), which was approximately 3 times higher than the average CFA prevalence in post-MDA sites, 7.6% (n = 66, range 0.2% to 31.5%) (Figure 3a,b). The average Mf prevalence in pre-MDA sites was 10.1% (n = 124; range 0% to 47.4%), which was approximately 5 times higher than the average Mf prevalence in post-MDA sites, 2.0% (n = 66, range 0% to 12.1%) (Figure 3c,d). The distribution of post-MDA sites occur in the two States of Plateau and Nassarawa, and are the result of an extensive MDA programme delivering the combination of ivermectin and albendazole as detailed in the study publications [49] [53] [54] [63]. Baseline LF mapping was conducted in 1999 and 2000 in 30 LGAs of the two States, and MDA launched in 2000, and monitored from 2000 to 2009. Details are contained in the specific reference [63] (Table S1).

The most extensive study on clinical manifestations was conducted by Nwoke et al. [82], who used hydrocele as a clinical marker to estimate LF prevalence. A rapid epidemiological mapping survey (REM-LF) was conducted across 25 States and 536 villages in Nigeria. Details of the specific study sites are not available, however, the survey found that hydrocele was absent in 339 (63.3%) villages, and present in 197 (36.8%) villages, which were found to have different levels of hydrocele severity. Hydrocele was absent in Jigawa and Kano (NW), and Ogun (SW) States. Very few hydrocele cases (1–3%) were found in northern Borno (NE), Kaduna and Zamfara (NW), Edo (SS), Imo (SE),and in Ekiti, Ondo, Osun, Oyo (SW) States. The highest hydrocele rates were found in the NE States of Adamawa, Bauchi, Gombe, Taraba and southern Borno, in the NC states of Kogi, Plateau, Nassarawa, and in the northern part of Akwa Ibom State in the SS (Figure 2c) [82].

The clinical signs that were reported included limb lymphodema, hydrocele. chyluria and elephantiasis, and were from a few specific areas of the country [45]–[49] [59]–[62][65] [68]–[71] [74]–[75] [77]. For hydrocele, there were 22 sites with prevalence data ranging from 0.1% to 50%, while for lymphodema, 12 sites recorded prevalence rates ranging from 1% to 49%. The prevalence of limb elephantiasis was also recorded in 5 sites, which ranged from 1.7% to 11.8%. The distribution of these study sites is shown in Figure 2c, together with the high hydrocele prevalence states described by Nwoke et al. [82], and highlight the geographical concordance with CFA and Mf distributions which occur in the central south and eastern regions of the country (Figure 2a).

LF prevalence was examined in relation to the L. loa distribution in Nigeria defined by the RAPLOA surveys reporting eye worm history, which were carried out in 381 villages between 2002 and 2010 [7]. Loiasis was found predominately in the southern region of the country, with the highest risk in east along the border with Cameroon, which had a localized area >40% in the States of Taraba and Benue (Figure 4). Overall, there was minimal geographical overlap with the number of LF prevalence sites determined by CFA. The majority of sites with medium to high LF prevalence rates >25%, were found in low loiasis prevalence areas (<20%) where the risk of SAEs are considered to be low (Figure 4a). Similarly, there was minimal geographical overlap with the number of LF prevalence sites determined by Mf, however, more sites with medium to high LF prevalence rates >25% were found in medium loiasis prevalence areas (20–40%) where the risk of SAEs is potentially high (Figure 4b). The overlap with the LF endemicity map available from the FMoH shows a combination of endemic and non-endemic LGAs in the low (<20%) to medium (20–40%) loiasis prevalence areas (Figure 4c). Only LF non-endemic LGAs were found in the high risk loiasis area (>40%), which is highlighted in the close up of the map in Figure 4d.

The onchocerciasis CDTi priority areas [31] are shown in Figure 5a, and illustrate that large areas across the central region of the country are being targeted with ivermectin treatment. Since 1992, ivermectin has been distributed annually to 80% of the total population at risk, estimated at ∼38,331,140 people in 430 endemic LGAs [35]. The LF endemic areas to potentially benefit from CDTi priority areas are extensive and include large areas of NC, NW and NE zones of the country. The LF programme could readily add albendazole to the ivermectin being distributed in these areas (Figure 5b). The potential risks associated with ivermectin treatment for O. volvulus are related to potential SAEs in areas where W. bancrofti and L. loa are co-endemic in the southern region of the county, especially in the States of Benue, Cross River, Ebonyi, Enugu, Osun, Ekiti and regions of Edo, Ondo and Ogun States (Figure 5c).

The distribution of LLIN coverage across the six different geopolitical zones is shown in Figure 5d. The highest LLIN coverage occurred in the NE (61.8%) and NW (58.2%) zones, followed by the SS (43.5%), SE (32.1%), NC (32.1%) and SW (20.3%) zones respectively. This shows that the highest LLIN coverage occurred in the northern region of the country where LF does not appear to be highly endemic (Figure 5e). The lowest LLIN coverage occurred across the southern region of the country, which coincides with areas where both LF and loiasis are considered endemic and co-endemic (Figure 5f).

This MOM work builds on the recent study carried out in Democratic Republic of Congo [24], which first used the new overlap mapping approach to collate and map all available country data on W. bancrofti, examine the extent of L. loa co-endemicity and determine the risk and benefits of different intervention strategies. Collectively the two studies address two important countries in Africa with respect to the elimination of LF, as they have the highest burdens of disease, collectively accounting for more than 170 million people at risk. Furthermore, their LF Programmes are yet to scale up to reach full national MDA coverage taking into account the co-endemic areas of L. loa, which may require alternative treatment strategies in selected areas, and coordination with other NTD elimination and malaria control programmes. We advocate that the MOM approach should be used more widely over time and space, and at different geographical scales to better monitor and understand the impact of single and multiple interventions, and to assess progress towards elimination of LF and other diseases. Such an approach is also necessary for national planning purposes as well as increasing the cost effectiveness and coordination of programmes where different strategies are deployed, and where there have been previous interventions which will impact on the goals of the LF programme.