Mass Drug Administration (MDA).

Malawi’s MDA programme included a combination of ivermectin (IVM: 200 μg/kg of body weight) and albendazole (ALB: 400 mg) to treat the entire at-risk population [11,12]. The drugs were donated by the Mectizan Donation Program (IVM) and GlaxoSmithKline (ALB) [19], and the logistics and transportation of these donated goods were managed by the LF Programme Manager and distributed from the Central Medical Stores in the capital city of Lilongwe, to the district-level headquarters, and onwards to all district health facilities.

One senior Health Surveillance Assistant (known as HSA) per health facility supervised a team who were responsible for estimating the number of drugs needed to reach the target population. This was achieved through community enumeration prior to MDA implementation activities and this information further helped to calculate MDA coverage rates and determine if the WHO threshold of 65% coverage of the total populations to interrupt transmission had been achieved [12]. Prior to the campaign, intensive social mobilisation activities took place to heighten community awareness and assist in reaching the required coverage. The senior Health Surveillance Assistant supervised the MDA campaign, which was implemented annually in December between 2008–2014 and typically lasted seven days with selected follow-up in areas of potentially low coverage. The IVM+ALB drug combination was distributed through a directly observed treatment (known as DOT) approach by volunteer Community Drug Distributors (known as CDDs), who were trained and supervised by their local Health Surveillance Assistants.

The first MDA activity was supported by the Ministry of Health in 2008 and implemented in nine southern districts that were co-endemic with onchocerciasis, which used the drug IVM as its primary preventive chemotherapy strategy. The LF MDA activities scaled up to reach all 26 endemic districts in 2009 and maintained until 2014. A summary of the national MDA implementation years, populations, and overall high reported coverage rates between 2008 and 2014 is shown in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of national mass drug administration (MDA) implementation years, implementation units (IUs), populations, and reported coverage rates between 2008 and 2014.

https://doi.org/10.1371/journal.pntd.0011957.t001

MDA coverage surveys to verify the reported rates were conducted between 2010 and 2014 in selected districts and all found to be above the required 65% threshold to interrupt transmission [12]. In Zomba, the WHO 30 cluster-survey protocol was conducted in 2010, and estimated coverage levels of 62.6% (95% Confidence intervals CI 56.5–68.7%) and 66.8% (95% CI 60.0–73.0%) in 2009 and 2010 respectively. An additional investigation into alternative coverage survey methods including convenience sampling in markets, and household surveys conducted by local headman and religious leaders found coverage at 74.3% (CI: 71.1%-77.4%), 76.3% (CI: 69.6%-83.0%), 77.8% (CI: 72.5%-83.1%), respectively [20]. In 2012, a WHO-cluster survey in Ntchisi district estimated coverage at 78.7%. In 2014, different survey methods were implemented in three districts; a WHO cluster-survey in Machinga district estimated coverage at 73% (95% CI 70.0–77.0%), a Lot Quality Assurance Sampling method in Balaka district estimated coverage at 86% (95% CI 79.0–93.0%), and a Probability Sampling with Segmentation method in Zomba district at 77% (95% CI 73.0–80%).

Supplementary interventions.

The disease control and elimination programmes for malaria, onchocerciasis and soil transmitted helminths provided additional supplementary interventions that are likely to have helped to reduce LF transmission [18,21–23].

Malaria vector control interventions: The National Malaria Control Programme’s interventions include the distribution of insecticide treated/long-lasting insecticidal nets (ITN/LLINs) and indoor residual spraying (IRS) supported by national and international technical expertise and funding [22,23]. The periodic nationwide mass distribution of ITN/LLINs every three years from 2010, supplemented by routine distribution of nets at all antenatal clinics across the country. Further, there was localized indoor residual spraying (IRS) initially for two years (2009 to 2010) in a single district of Nkhotakota before scaling it up to a total of eight highly endemic districts close to Lake Malawi and the Lower Shire Valley with overall good coverage rates (>85%) until the programme was stopped in 2012 due to the emergence of pyrethroid insecticide resistance and lack of viable alternative insecticides. The main mosquito vectors of malaria and LF are similar and include the Anopheles funestus and An. gambiae complexes [24]. Therefore the distribution and impact of the vector control interventions for malaria, which have shown an overall reduction in malaria infection [25], are likely to have impacted the main LF vectors and helped to reduce the risk of transmission, particularly in areas where multiple interventions had been implemented [21].

Additional preventive chemotherapy interventions: The National Onchocerciasis Elimination Programme has distributed IVM MDA annually across eight endemic districts of the Southern Region of Malawi (Thyolo, Mwanza, Neno, Blantyre, Mulanje, Phalombe, Chiradzulu and Chikwawa) since 1997. From 2008 to 2014 these activities were integrated with the LF Programme, targeting all persons aged five years and above. The continued IVM MDA in these eight co-endemic districts at high levels, which has nearly eliminated onchocerciasis, is likely to have had an impact and reduced the risk of LF transmission [18,21].

The National Soil Transmitted Helminth Control Programme focuses on hookworm, roundworm, and whipworm parasites and has distributed ALB across selected population groups (e.g., school-aged children) in all 28 endemic districts since 2012. This has resulted in a reduction in the targeted parasitic infections, and has likely to have reduced the risk of LF transmission in these selected populations [18].

Sentinel and spot check sites.

Sentinel site and spot check site surveys were conducted to assess the impact of MDA between 2008 (prior to starting MDA) and 2014 (prior to stopping MDA) using a combination of diagnostic methods to detect W. bancrofti microfilaria (MF) or circulating filarial antigen in the population. Table 2 summarises the data of each survey by each evaluation unit (EU), which comprised multiple adjacent districts with similar demographic and environmental characteristics as recommended by WHO guidelines [12]. In total 11 EUs were formed from the 26 endemic districts and used for pre-and post-MDA surveillance activities.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of impact assessments results by Evaluation Unit (EU) between 2008 and 2014.

https://doi.org/10.1371/journal.pntd.0011957.t002

The baseline sentinel site surveys were conducted in 2008 (six sites) and 2009 (18 sites) prior to starting MDA. Participants over five years of age had finger prick blood samples collected from 22:00 to 01:00 to detect MF using the WHO microscopy standard protocol [12]. Overall MF prevalence rates ranged from 0% to 9.1% (Table 2). The first mid-point impact assessment was conducted in 15 of the 24 sentinel sites in 2012 and found that the overall MF prevalence had reduced to less than 3% in all sites.

In 2014, a pre-TAS assessment to determine if MDA could be stopped was conducted in 24 sentinel sites and spot check sites using the ICT diagnostic tool [16]. Overall, the prevalence was found to be less than 5%, in all but one EU which included Chikwawa and Nsanje districts (15.7%-17.3%). All people found to have an ICT positive test were subsequently treated using the standard IVM+ALB drug combination (Table 2). On the recommendation of the WHO, the LF Programme initiated the first of three TAS activities in 2014.

Supplementary assessments.

The Malawi LF Elimination Programme worked collaboratively with a range of national and international partners to conduct additional assessments related to transmission and surveillance including i) vector control ii) entomological monitoring iii) integrated surveillance iv) targeted surveillance v) hotspot surveillance and vi) alternative surveillance. Summaries of the related assessments are described here, and summary results are available in Tables A-F in S4 Table.

Vector control: As part of the TAS 2 and TAS 3 programme activities described above, additional information on ITN/LLINs was collected. The children sampled were asked if their house had an ITN/LLIN (i.e., bed net asked in local language) and if yes, did they sleep under it the night before. The TAS 2 survey found 73.6% (range EU2 41.5%–EU7 72.1%) of children reported a bed net in their house and 61.7% (EU7 37.4%—EU4 91.5%) reported to have slept under it the night before the survey. The TAS 3 survey found 68.0% (range EU11 48.0%–EU1 90.9%) children reported a bed net in their house and 58.2% (EU31.6% - 83.0%) reported to have slept under it the night before the survey. These results indicate overall moderate to high ITN/LLIN coverage for this 6-7-year age-group of children, with a reduction in ownership and usage in TAS 3. The reason for this small change is not known as no additional questions were asked. It is possible it may be related to the timing of the national ITN/LLIN distributions. The information was shared with the National Malawi Control Programme so that low coverage areas may be followed-up according to protocols. A summary of the results is available in Table A in S4 Table.

Entomological monitoring. An entomological assessment was conducted between June and August 2017 in the high-risk districts of Chikwawa and Nsanje with support from national partners based at the Malaria Alert Centre of the University of Malawi. The study included five sites in Chikwawa District (Belo, Misili, Chakanira, Kumwembe and Mlooka) and two sites in Nsanje District (Gulumba and Mailosi), which were found to have high infection and/or clinical case rates in previous investigations. Mosquitoes were sampled from approximately 100 households across the seven sites using pyrethrum spray knockdown catches and Centers for Disease Control and Prevention (CDC) light traps [15]. Samples were counted, morphologically identified and members of the Anopheles species identified by polymerase chain reaction (PCR). Further, reverse transcription PCR (RT-PCR) was used to detect W. bancrofti parasites in a sub-sample of mosquitoes from each study site. Mosquito DNA was extracted and pooled into pools of five mosquitoes, a sample size that is in line with WHO guidelines [15]. The main LF vectors included An. funestus (n = 2832) and An. gambiae s.l (n = 215) and of these 38 An. funestus mosquitoes were found to be positive in the high-risk sites of Belo (n = 19/1848; 1.0%) and Chakanira (n = 14/953; 1.5%) in Chikwawa District. A summary of the laboratory methods and results are available in Table B in S4 Table.

Integrated surveillance: An arbovirus and LF integrated surveillance activity was conducted from November 2017 to January 2018 by the National NTD programme. Eight sentinel sites were established primarily to monitor arboviral infections due to a potential threat of dengue at the time. LF testing using FTS diagnostic tool was integrated into the activity. Testing was conducted at health facilities by trained staff and included the first 20 people who presented at the facility with acute febrile illness characterized by high fever, body pains and/or were malaria negative. A total of 1,196 people were tested (56.2% male) with three people found to be FTS antigen positive from Nkhotakota, Matawale in Zomba District and Ndirande in Blantyre city, and subsequently treated using the standard IVM+ALB drug combination. A summary of the results is shown in Table C in S4 Table.

Targeted surveillance: A household prevalence survey was conducted in six villages of Chikwawa district in 2013 based on reports of high numbers of lymphoedema or hydrocoele cases and varying sub-district levels of MDA coverage. In each village between 105 and 117 people were surveyed and infection measured using the ICT diagnostic tool and any positive people were followed up for MF testing. Additional information was collected on MDA intake in 2013, ITN/LLIN ownership as a supplementary intervention and if there were cases of clinical cases within the household.

A total of 795 individuals were surveyed and overall village MDA intake (range 69.2–90.2%), and ITN/LLIN (range 74.5–92.0%) coverage rates were high. In total, 15% of participants were ICT positive (12.4% females; 20.0% males). MF assessments were conducted in 66 ICT positive participants with 12% positive (9.0% females; 16.0% males). Median MF density was 22 MF/20 μL (range 0.3–58.5 MF/20 μL) in individuals who reported no MDA participation and 2.5 MF/20 μL (range: 0.7–8.0) in those who recalled participation in 2012. The percentage of clinical cases varied widely (lymphoedema 0.9–26.7%; hydrocoele 0 .0–3.4%). The village of Chakanira had the highest rates of infection (38.5% ICT; 19.4% MF) and clinical cases (26.7% lymphoedema; 3.4% hydrocoele) and was subsequently followed up during later TAS and MMDP activities. A summary of the results is available in Table D in S4 Table.

Hotspot surveillance: As part of a post-TAS surveillance activity a cross-sectional prevalence survey was conducted in the high-risk districts of Chikwawa and Nsanje in November 2020. Ten study sites considered to be potential transmission risk areas ‘hotspots’ were selected and grouped based on three different criteria; four sites with high baseline prevalence ‘Baseline hotspots’ (Belo, Mbnde, Gamba, Nchacha 18); four sites with high number of clinical cases ‘Morbidity hotspots’ (Chapananga, Lingawa, Kalemba, Mkango) and two sites with TAS positives ‘TAS hotspots’ (Muona, Nsenjere). In each site, between 300 and 350 people were targeted via household systematic random sampling and those who participated were tested using the FTS diagnostic tool. Additional information was collected on ITN/LLIN ownership (if no, state reason) and whether they slept under it the night before the survey, (if no, state reason) as a supplementary intervention. All FTS positive people were treated with the standard IVM+ALB drug combination.

A total of 3247 people (43.9% male; overall average age 25.2 years) were surveyed, with 20 people found to be FTS positive (0.6%) including 10 males (0.70% positive, average age 47.5 years) and 10 females (0.6% positivity, average age 35.6 years). Overall, there were similar prevalence ranges between the three hotspot groups; Baseline hotspots 0.6% (site range 0.0–1.7%); Morbidity hotspots 0.6% (site range 0.0–1.4%) and TAS hotspots 0.7% (site range 0.6–0.8%). The highest number of FTS positives were found in the Baseline hotspot of Belo, Chikwawa District (n = 5/301, 1.7%), and in the Morbidity Hotspot of Mkango, Nsanje District (5/365, 1.4%).

Overall, 70.7% people reported they owned an ITN/LLIN (n = 2297/3247; site range 53.0–83.8%) and 94.9% slept under it the night before (n = 2179/2297; site range 87.4–99.2%). Of the 950 people who did not own an ITN/LLIN, the main reason was they could not afford it (71.2%; n = 676/950) and of the 118 people who did not sleep under it, the main reason was that it was too hot (70.3% n = 83/118). The overall low W. bancrofti antigen prevalence indicated that transmission potential was low and likely to be further impacted by the overall high ITN/LLIN coverage. A summary of the results is available in Table E in S4 Table.

Easy Access Group surveillance: As part of a post-TAS surveillance, an operational research activity to determine the utility of screening an easy to access group (EAG) or population sub-group (i.e., pregnant women attending ante-natal clinic (ANC)) as a surveillance tool was conducted in Chikwawa District in September and October 2021 by national partners in collaboration with the LF Programme and international partners [27]. Two cross-sectional studies were conducted simultaneously. The first survey was conducted across 12 health facilities to screen pregnant women during their visit (i.e., ANC survey–survey 1). Each facility was grouped into transmission risk zones based on the distance from the Shire River, which is known to be a major breeding site associated with Anopheles vectors and shown to be correlated with disease [24,28]. The high-risk zone was 0-5km; moderate between 5-10km and low >10km from the riverbank. The second survey was conducted at the household level in the same catchment area of the sampled health facilities (i.e., Household survey–survey 2). In each site, approximately 70 households were selected at random from a central location and all household members aged above 5 years were invited to participate. For both surveys, participants were tested using the FTS diagnostic tool, and asked if the household had an ITN/LLIN. All positive people were followed up for MF testing and after were treated using the standard IVM+ALB drug combination.

For the ANC survey (survey 1), the high-risk zone included five clinics and 291 (26.8%) women; the medium-risk zone included four clinics and 470 (43.3%) women; and the low-risk zone included three clinics and 324 (29.9%) women. Of the 1,085 pregnant women tested (average age 24.1), only one from Chikwawa ANC was found to be FTS positive, but negative for MF (22 years of age, no ITN/LLIN). Overall, women reported moderate to low ownership of an ITN/LLIN (34.6%; site range 5.7%-58.8%).

For the Household survey (survey 2), a total of 2,900 individuals from 53 villages and 1,048 households were surveyed across the two districts. The high-risk zone included 1,107 (38.2%) individuals, the medium-risk zone 1,107 (34.0%) individuals; and the low-risk zone 808 (27.9%) individuals. Of the 2900 people tested, 4 were found to be FTS positive and two from high-risk zone (male aged 54; female aged 58) and two from the medium risk zone (males aged 25 and 45). Overall, people reported moderate to high ownership of an ITN/LLIN (53.0%; site range 27.6%-80.2%). A summary of the results from both surveys is available in Table F in S4 Table.

Clinical case estimates and characteristics.

Historical studies published in 2002 found a high number of clinical manifestations in four villages in Karonga district (Northern Region) and two villages in the Lower Shire Valley (Southern Region) [7]. However, these studies only provided morbidity prevalence information for a small area. Therefore, as a first step to quantify the clinical burden on a larger scale, the LF Elimination Programme mobilised the Community Drug Distributors (known as CDDs) to collected additional information on the number of lymphoedema and hydrocoele cases during LF MDA implementation, and in 2011 a total of 5208 cases covering 11.7 million people were found: 3890 hydrocoele and 1318 lymphoedema cases, with the highest number reported from Chikwawa and Nsanje Districts. A summary of the results is available in Table A in S5 Table. This collection of data on morbidity cases was an important first step in understanding the burden. However, a study conducted in 2013 to determine whether this information collected during MDA accurately reflected the magnitude and distribution of lymphoedema cases, noting that in one catchment area of Chikwawa District, the actual number of cases of lymphoedema was over twice the number recorded in the MDA report [29]. This highlighted the need for improved methods of morbidity data collection in Malawi.

Active patient searching and mHealth data collection: In 2014, a mHealth real-time reporting tool (MeasureSMS-Morbidity) was piloted in three health centre catchment areas of Chikwawa District in which, trained Health Surveillance Assistants identified and reported all LF morbidity cases through active house-to-house patient searching. The methods are described elsewhere [29,30]. The mHealth tool allowed for the rapid collection and collation of information on LF patients including their location (community/health catchment name), age (number), gender (male, female), clinical condition (hydrocoele, lymphoedema, or both), stage of the condition (classified as mild, moderate, or severe as per WHO staging) and episodes of adenolymphangitis (ADL) also known as acute attacks (number in last 6 months) [13]. Following the successful pilot, and with funding from international partners between 2015–2021, the patient searching activities using the mHealth tool were scaled up across 23 districts. Seven districts were mapped completely (all health centre catchments) and 16 districts partially (selected health centre catchments) to obtain a geographical range of information from areas across the country with different levels of endemicity. The aim of the latter was to optimise limited resources given the expense of detailed mapping, and to provide an opportunity to develop models for unmapped areas (Traditional Authority level). This exercise facilitated the development of the first national LF clinical case database and LF morbidity map in Malawi, which was the first of its kinds in any endemic country in sub-Saharan Africa [31].

The clinical case mapping covered an area of over 33,000km² populated by 5.6 million people, representing approximately one third of the country. Data were summarized by Traditional Authority (n = 90) and Fig 1B and 1C highlights the distribution of mapped areas, the number of clinical cases and prevalence per 100,000 population based on 2018 census data [32]. In total 8,856 clinical cases were identified, which comprised 6,333 hydrocoele cases in males (71.5%; average age 50.5); 2439 lymphoedema cases (27.5%; average age 50.9), and 84 hydrocoele and lymphoedema cases in men (0.95%; average age 58.8). The three known high endemic districts of Karonga (n = 1092), Chikwawa (n = 1386) and Nsanje (n = 1140) accounted for 40.9% (n = 3618/8856) of all cases reported. The number of cases by Traditional Authority ranged from 1 to 359, and the prevalence rates ranged from 3 to 1543 per 100,000 with the highest recorded in Chikwawa District. A summary of results is available in Table B in S5 Table.

The overall summary of the number of lymphoedema and hydrocoele cases by age group and gender is shown in Fig 3. For lymphoedema alone, more females (65.0%; n = 1585/2439) were reported than males (35.0%; n = 854/2439) and the number of cases increased by age for both genders with most cases aged >40 years (71.1%; n = 1734/2439) (Fig 3A). For hydrocoele, a similar increasing trend with age was found with most cases aged >40 years (72.8%; 4610/6333) Fig 3B). The hydrocoele cases reported in the <20-year age group were considered to be due to other causes but included here for reporting and specific follow-up purposes to confirm their diagnosis. A summary of the results is available in Table C in S5 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Summary of lymphoedema and hydrocoele cases by age group and gender.

A. Number of lymphoedema cases. B. Number of hydrocoele cases.

https://doi.org/10.1371/journal.pntd.0011957.g003

The overall summary of lymphoedema cases and ADLs /acute attacks reported by stage of condition (mild, moderate, severe) and gender is shown in Fig 4. Overall, similar trends were found between males and females. For lymphoedema stage (n = 2403), there was a decreasing trend with most cases reported as mild (56.0%; n = 1345/2403), or moderate (31.0%, n = 746/2403) and fewer as severe (13.0%; n = 312/2403) (Fig 4B). For ADLs /acute attacks (n = 1957; note early data collections excluded this variable), the average (mean) number reported was 1.8 in the last 6 months. There was an increasing trend with severe lymphoedema cases (2.5) reporting a higher average number of acute attacks than cases with moderate (2.1) or mild (1.4) lymphoedema (Fig 4B). A summary of the results is available in Table D in S5 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Summary of lymphoedema cases and ADLs/acute attacks by stage of condition and gender.

A. Number of lymphoedema cases. B. Average number of ADLs/acute attacks.

https://doi.org/10.1371/journal.pntd.0011957.g004