Malaria prevalence in Commune 5 in Tumaco (Nariño, Colombia)

Introduction

Malaria elimination programs face several challenges, including urban malaria and disease diagnosis. Both in Colombia and worldwide, malaria transmission is mainly rural, but cases of malaria in urban and peri-urban areas have been continuously reported over the last decade, most of them in the Pacific region.^1,2 These problems are compounded by the diagnosis of asymptomatic and submicroscopic malaria.

Despite the fight against malaria, the World Health Organization (WHO) observed an increase in cases worldwide, increasing from 227 million in 2019 to 241 million in 2020 in the 85 countries where the disease is endemic. Similarly, in terms of mortality, 558,000 deaths were registered in 2019 and 627,000 in 2020. However, in the Americas region, the WHO reported a reduction in cases, from an estimated 893,000 in 2019 to 652,000 in 2020, 77% of them in Venezuela, Brazil and Colombia. Meanwhile, there was a reduction in deaths, from 509 in 2019 to 409 in 2020.³

It is worth mentioning that 91% of the Colombian territory has eco-epidemiological conditions that promote malaria transmission and it is estimated that 22% of the population lives in these areas. In 2020, 81,363 cases were registered (annual parasite index (API) 8.4 cases per 1,000 inhabitants) and 5 deaths were confirmed in the country. The predominant parasites were Plasmodium vivax (49.7%) and Plasmodium falciparum (49.5%).⁴ On the Pacific coast, in the department of Nariño, where Tumaco is located, 17,421 cases were reported (API 31.9 cases per 1,000 inhabitants).⁴ According to the Vector-Borne Disease Prevention and Control Program, cases of urban malaria have been identified mainly in the neighborhoods that are part of Commune 5 in Tumaco, where 1,024 cases were reported in 2015 (API 42.2 per 1,000 inhabitants) and 770 in 2016 (API 31.7 per 1,000 inhabitants).

In Colombia, the strategic action route for comprehensive care, health promotion, prevention, surveillance, control and elimination of malaria has been based on the “2019-2022 National Malaria Strategic Plan” which, among other aspects, includes disease diagnosis through the use of optical microscopy and rapid diagnostic tests (RDTs), and the elimination of urban and peri-urban malaria in 18 municipalities on the Pacific coast.⁵

Optical microscopy can detect from 50 to 100 parasites/μL but it can lead to diagnostic errors when parasitemia is low.⁶ The purpose of RDTs is to detect one of three antigens in parasites: lactate dehydrogenase (LDH), aldolase, and histidine-rich protein 2 (hrp2), with significant differences from each other.⁷ The detection limit for RDTs is between 100-200 parasites/μL.⁸ It should be noted that hrp2 is specific for P. falciparum and that pfhrp2-based RDTs often detect the pfhrp3 antigen, rendering test interpretation difficult.⁹ Test performance may be affected by the antigenic variability of the target protein, the persistence of the antigen in blood after parasite elimination, and parasite density below the detection threshold. The presence of parasites with deletions or mutations of the pfhrp2 gene can lead to false-negative results, which has implications for RDT implementation, case diagnosis and treatment, and malaria control and elimination efforts.¹⁰

To overcome the limitations of optical microscopy and RDTs, molecular detection of malaria parasites through the polymerase chain reaction (PCR) was used, as it can detect parasitemia below 0.05 μL.⁷ On the other hand, PCR provides better information on the prevalence and distribution of parasitic species in endemic areas compared to that provided by optical microscopy and RDT.¹¹ PCR is a more sensitive diagnostic method to detect asymptomatic infections with very low parasite density. It may be useful in studies of submicroscopic infections, although its utility depends on the epidemiological importance of low-density infections, which have not yet been characterized.¹²

Regarding urban malaria, there are challenges in terms of what is meant by “urban”, “peri-urban”, and “rural”, because of the lack of established definitions to describe the socioeconomic and ecological contexts where the disease is transmitted.^1,13 This situation is compounded by the failure to establish the origin of malaria cases, to identify mosquito breeding sites, and to confirm the vector's transmission capacity in these environments.¹ Despite these difficulties, it has been established that urban and peri-urban malaria may be transmitted by population traveling from rural to urban and peri-urban areas.^14,15 Local transmission has also been suggested in urban areas where the presence of Anopheles, as a competent vector, and infected people converge.¹⁴ In Latin America, peri-urban transmission has been reported in Brazil (Porto Velho)¹⁶ and Peru (Iquitos, Sullana, Piura y Lima).^17–19 In Colombia, urban malaria has been identified in municipalities on the Pacific coast, where Tumaco is located. Peri-urban malaria is caused by the displacement of rural populations to peri-urban areas as a result of the armed conflict, presence of paramilitary groups, illicit crops, and illegal mining.¹

Most studies conducted in Colombia have addressed the prevalence of malaria in rural areas. Research focused on urban malaria has been scarce, reporting on total prevalence and the prevalence of asymptomatic and submicroscopic infections.^1,13,20–26 Also, research aimed at detecting the pfhrp 2/3 gene deletion has been limited (Guapi 6%).²⁷ Given the limited research on malaria prevalence carried out in the Colombian territory and the detection of pfhrp 2/3 genes in urban areas, this study was aimed at estimating malaria prevalence, detecting pfhrp 2/3 genes, and exploring associated factors in the urban area (Commune 5) in Tumaco during 2019.

Methods

Study design

Between July and December 2019, a prevalence study was conducted in Commune 5 of Tumaco, a malaria-endemic urban area. This Colombian municipality is divided into communes and, in turn, each commune is made up of neighborhoods. A stratified (by neighborhoods) multistage random sample with proportional allocation was established. In the first stage, the primary sampling units were the Commune 5 neighborhoods. In the second stage, the secondary sampling units were groups of adjoining blocks. A grid was placed on the map of each selected neighborhood. Each cell included the residential units that make up a block. Sampling was carried out proportional to size, taking into account the number of inhabitants per block. Using the Teaching Sampling package of the statistical package R, the blocks in each of the selected neighborhoods were chosen. In the third stage, in the selected block, the house in the lower-left corner (southwest point) was located with the help of the Global Positioning System (GPS). In this dwelling began the collection of information. Then the selection of the next dwelling to the right or to the left depended on the outcome of a coin toss. Once the address was established, the collection of information continued until completing 12 contiguous houses. The units of analysis and observation were all individuals in the selected households (Figure 1).

Figure 1. Diagram of the study sampling technique.

PSU: primary sampling units, SSU: secondary sampling units.

To calculate sample size, a prevalence of 4.22% was considered, as reported by the “Vector-Borne Disease Prevention and Control Program in the Nariño Department” study in 2015. Accuracy was estimated to have a 15.0% expected relative standard error,²⁸ 95% confidence, and a 1.5 cluster design effect. Adjusting for a non-response rate of 10%, the final sample size was 1,424 individuals from households in Commune 5.

Officials from the departmental and municipal Vector-Borne Disease Prevention and Control Program, community leaders, and researchers informed the population of Commune 5 about the study. This ensured community participation in the study.

The inclusion criteria for the study were: (1) inhabitants of the selected dwelling who voluntarily wish to participate and who at the time of the visit were in the dwelling, (2) minors who have the authorization of the responsible adult or guardian, (3) people who have adequately filled out the informed consent. The exclusion criteria were: (1) people with mental problems unable to sign informed consent and (2) those for whom a blood sample was not possible to obtain.

Study area

Tumaco is located in southwestern Colombia (1° 48' 24” N; 78° 45' 53” W), on the Pacific coast, Nariño Department. It spreads over 3,778 km² and the urban area is about 38 km². This area is made up of three islands (El Morro, La Viciosa and Tumaco), which are divided into five communes and 82 neighborhoods²⁹ (Figure 2). The municipality is located 2 meters above sea level and has a warm humid weather. It has an average annual temperature of 26.1°C, a relative humidity of 84.8%, and annual precipitation ranging from 109 to 373 mm³. In 2021, its population was estimated at 257,042 inhabitants, out of which 33.7% lived in the urban area.³⁰ Its economy is based on the production of African palm, cocoa and coconut and, to a lesser extent, tourism. In the urban area of Tumaco, all five communes are endemic to malaria. In this city, 70% of malaria cases are reported from Commune 5 in the last ten years. Larval habitats correspond to wells, ponds, puddles, and lakes for shrimp farming. Commune 5, with 26,000 inhabitants, is located on the continental area of Tumaco on the low-tide land and nearby estuaries. The presence of Anophelines spp. breeding sites (ponds, puddles, and lakes for shrimp farming) and high density of the population make this commune of high risk of malaria transmission.

Figure 2. Study area in Tumaco (Nariño, Colombia).

Sources: Google Maps: Maps Data: Google. © 2022 INEGI (retrieved on March 10. 2022). Google Earth: Maps Data: Google. © 2020 Maxar Technologies (retrieved on December 13. 2020).

Results

Entomological study

The entomological study was conducted in 25 households in the Obrero, Los Ángeles, Candamo, Unión Victoria, Ciudadela, Nuevo Milenio and Once de Noviembre neighborhoods. The presence of immature and adult Anopheles mosquitoes was registered.

37 Anopheles mosquitoes belonging to the albimanus species were collected. The highest number of mosquitoes was identified in the Los Ángeles and Candamo neighborhoods. No malaria vector mosquitoes were identified in the Nuevo Milenio and Once de Noviembre neighborhoods. Between 18:00 and 19:00 hours, the highest activity occurred in the peri-domestic area, with a 0.018 human-biting rate (HBR) per hour and indoor between 21:00 and 22:00 hours, with a 0.007 HBR (Figure 3). The average human biting rate (HBR) at night was 0.74 and the HBR was 0.06 for all studied areas.

Figure 3. Indoor and Peri-Domestic Human-Biting Rate (HBR).

Commune 5 neighborhoods, Tumaco (Nariño, Colombia). October 2019.

83 breeding sites were found. In the Commune 5 neighborhoods of Tumaco, An. albimanus was found in artificial ponds and wells. Out of those, two had Anopheles mosquito larvae, one in the Unión Victoria neighborhood and the other one in the Obrero neighborhood. One breeding site was a pond and the other one was a well. 32 larvae were collected in both breeding sites. 27 of them were in late stages (3rd and 4th) and were identified as An. albimanus.

Discussion

This study conducted in Commune 5, urban area of Tumaco, estimated malaria prevalence, pfhrp2 gene amplification, and explored factors associated with infection. The prevalence of malaria was 2.97%. It was higher in the 10 to 19 age group (4.42%) and 0 to 9 (4.06%), decreased with age, and was higher in cases of asymptomatic and submicroscopic infection. The pfhrp2 gene was found and the disease was positively associated with households inhabited by three or more people.

The blood samples were taken during the dry season, between July and September 2019. Malaria in the urban area of Tumaco has not shown a clear seasonal behavior.

The study found malaria in Commune 5, an area believed to have high transmission according to reports from the Vector-Borne Disease Prevention and Control Program in the Nariño Department (API 2014 84.0 cases per 1,000 inhabitants; API 2015 220.5 and API 2016 165.8) (Pilar Perez, Vector-Borne Disease Prevention and Control Program in the Nariño Department, personal communication). The decrease in the prevalence of malaria over time is possibly due to the intensification of actions to search for and eliminate breeding sites, massive use of mosquito nets, and social participation as measures implemented for Commune 5. Diagnosis and treatment actions also have contributed, despite the difficulties that exist in accessing services. In this way, it has been reduced human contact (infected) —a vector that has contributed to the decline in the burden of this disease.

The higher prevalence of malaria in people under 20 years of age suggests that the infection may have been acquired locally. Unlike Africa, which has a high transmission and where the population, most affected in terms of morbidity and mortality is under 5 years of age, in areas of low transmission such as Colombia, malaria occurs in older age groups. Among the aspects that are considered to think that the transmission occurs in the urban area are confirming that the infection in the residents of these areas; evidence of the presence of adult Anopheles that are biting the inhabitants of urban dwellings and identifying breeding sites near them.⁴⁰

In South America, the prevalence of symptomatic and asymptomatic infections have ranged from 0.1% to 33% in urban areas of Porto Velho,¹⁶ Manaos,⁴¹ and Mâncio Lima⁴² in Brazil. In Colombia, prevalence has ranged from 0% to 5.8%. In Quibdó, the estimated asymptomatic prevalence in a group of schoolchildren was 0% (95% CI: 0-1.4).⁴³ Two cases of submicroscopic infection were reported in Yesquita, Silencio and Roma neighborhoods of Quibdó.⁴⁴ In the Santa Mónica de Guapi neighborhood (Cauca Department, on the Colombian Pacific), four successive measurements reported asymptomatic prevalence of 2.7%, 1.2%, 0.6% and 0.3%, respectively.²⁷ In Buenaventura, the prevalence was 4.4%²¹ and in the California neighborhood in Tumaco it was 5.8%.²⁶ These figures are not comparable with those reported in this study due to differences in terms of study design, type of sampling used, selection of individuals, diagnostic techniques, time of year in which the study was conducted, and the area where it was carried out. Concerning this last aspect, although there is no consensus on the definition of urban and peri-urban, the presence of symptomatic and asymptomatic malaria in Commune 5 of Tumaco was demonstrated, turning it into a potential human reservoir and a challenge for disease elimination strategies.

The persistence of urban malaria may be caused by people moving back and forth between urban and rural areas. People travel from urban to rural areas, where the endemic is higher, either for work or leisure. People also travel from endemic rural areas to urban areas to access health services or to engage in commercial activities, causing them to stay for a long period of time there.¹⁵ Rural areas are the main source of malaria infection in urban areas.⁴⁵

In urban areas, malaria infection is not expected to be prevalent due to better-designed households with better basic utilities, better access to health services, and the absence of mosquito breeding sites.^34,46 However, in the study area, most households were found to belong to a low socioeconomic stratum, lack sewage, have few mosquitoes and breeding sites, and have low biting rates, which could in part contribute to disease transmission.

In this study, households with higher occupancy showed higher malaria risk. Studies in Rwanda and Tanzania found that mosquitoes are more attracted to houses with many people^47,48 because there is a higher production of cues that attract them and increase the risk of transmission, compared to households with lower occupancy.⁴⁹

Something that stands out is the high percentage of households in which people slept with nets (98%). Although no distinction was made between untreated nets, those treated at some point in time, and those treated with insecticide, the bivariate analysis allows us to see their protective effect when households with and without malaria cases were compared. If nets are a major part of malaria elimination strategies, implementation, monitoring and evaluation of their use and efficacy should be routinely carried out in Commune 5 of Tumaco.⁵⁰

In terms of case detection, as in other studies, PCR had the highest diagnostic performance,⁴⁰ followed by microscopy. When comparing microscopy/RDT results with PCR results, lack of concordance was observed. Those with a positive diagnosis through microscopy/RDT were not detected through PCR and those with a positive diagnosis through PCR were not detected through microscopy/RDT. It is possible that people who were positive according to microscopy/RDT did not have the parasite's DNA so it could not be detected through PCR, as described by Mudare et al.,⁵¹ or the sample may have had a low amount of parasitic DNA or it may have degraded. Another possibility is that there were changes in the primer ringing site, which would prevent amplification of the gene fragment under study. On the other hand, a positive microscopy with a negative RDT is probably because the limit of detection of RDT. The same rational is true for negative RDT with positive PCR. Cases of negative thick smear with positive RDT could be explained because Pfhrp2 antigen could be detected in the blood until 30 days after a good malaria treatment.

The species detected through the various diagnostic tests was P. falciparum, a finding that also corresponds to the most frequently reported species for malaria on the Colombian Pacific coast.⁴

In this study, differences were found in terms of the prevalence of malaria among men and women. As reported in other studies, malaria was more prevalent in men. This could be due to the fact that they spend more time doing outdoor activities, arrive to their households late in the evening, and have an indifferent attitude towards malaria prevention.⁵²

In terms of age groups, the highest prevalence of malaria was observed in the 10 to 19 age group and as age increased, prevalence decreased. This result is similar to that reported in other studies. Older people are likely to have a lower risk of becoming ill because they may develop some immunity if they have been previously exposed to the infection.^20,21 Less than 1% of participants had malaria in the last year and none in the last month. It is possible that in areas of low endemicity, such as Colombia, the lowest prevalence in older age groups is due to the fact that the population that may have less exposure was interviewed. For example, women who remain at home dedicated to household chores and individuals of school age who go to school. Recent studies in areas with high and low malaria transmission rates have revealed that asymptomatic infections contribute to malaria transmission.⁵³ Furthermore, asymptomatic malaria may precede symptomatic disease, and may confer partial immunity against infection.⁵⁴ In this study, participants were sampled at a single point in time. Those who were classified as asymptomatic had no history of illness in the last month and had not used antimalarial drugs in the last 15 days. For them, it is difficult to establish if they were in the incubation period of the disease or if they were carriers of a low number of parasites. Resolving a situation like this requires follow-up. In this regard, a study of 268 people followed for 29 months in western Kenya, in a high-transmission area, discovered that asymptomatic infections were highly likely to be followed by symptomatic disease. ⁵⁵ However, this does not seem to be the case in Commune 5 of Tumaco, which has a low transmission and where complicated malaria and malaria mortality do not occur.

This study found the presence of the pfhrp2 gene in samples from the urban area of Commune 5 in Tumaco. The findings suggest that a low proportion of malaria infections are not detected through the RDTs used in Colombia. In South America, gene deletions have been reported in Peru (prevalence between 20.7% and 41.0%) and Brazil (one case).⁵⁶ In Colombia, these deletions have been reported in the Amazonas Department, where they reached a prevalence of 67%⁵⁷; in Guapi (Cauca Department), where 6% of the parasite simples analyzed showed pfhpr2/3 gene deletion.²⁷ The factors promoting the appearance of these parasites in some regions still remain unknown.⁵⁸ The estimates obtained in this study do not exceed the minimum WHO criteria (>5%); however, it is important to monitor the hrp2-based RDTs used in the country. On the other hand, continuous monitoring related to submicroscopic infections and negative pfhrp2 results should be performed when using RDT. It is suggested that submicroscopic infections with parasites with pfhrp2/3 gene deletion could have infectious potential for mosquitoes and thus favor malaria transmission.⁵⁷

In the context of our study due to insufficient resources, we did not evaluate the deletion of these genes, we only amplified a fragment of these genes by means of PCR. However, in order to improve the management of clinical cases and design strategies for diagnosis, the detection of deletions in the hrp2/3 genes is relevant, since, as has been described, these can affect the accuracy of PDR in malaria endemic regions and of course in malaria control programs.

On the other hand, in the Commune 5 neighborhoods of Tumaco, An. albimanus was found in artificial ponds and wells, as described in the Buenaventura urban area.²⁰ Breeding sites were found in all the neighborhoods that were part of the study, except in Nuevo Milenio and Once de Noviembre. The differences between neighborhoods are caused by the urbanization level. Stilt houses predominate in some neighborhoods, while there is no sewage system in others, which leads to the construction of ditches where water stagnates and mosquitos breed. Although there is ongoing entomological surveillance, mosquito control interventions need to be strengthened and the community should be involved. An. albimanus had greater hematophagic activity in peri-domestic areas between 18:00 and 19:00 hours, a time when people engage in social activities outdoors and when they can possibly contract the disease,²¹ and indoors, between 21:00 and 22:00. However, given the low densities of An. albimanus and the limited observation, it was not possible to determine whether the mosquito was highly active at other times at night.

An. albimanus is considered a primary malaria vector in Colombia and particularly in Tumaco.⁵⁹ The presence of An. albimanus in different neighborhoods of Commune 5 provides partial evidence of the possible local transmission of malaria. This mosquito has been considered an opportunistic species due to its ability to adapt to different breeding sites.⁶⁰ Despite the small number of adults collected, there is a risk of transmission, taking into account that the maximum biting activity occurs in the peridomicile between 18:00 and 19:00 hours.

This study results should be studied with caution, given the limitations. First, the samples to be analyzed through PCR and pfhrp2/3 gene detection were taken on filter paper and were extracted after prolonged storage (due to the prioritization of PCR for Covid-19), which may have limited the amount of DNA to be used or may have promoted its degradation. Second, some samples collected on filter paper had little blood so the amount of gDNA was also variable. Third, the vast majority of individuals detected who were found to have malaria had no symptoms, which may not have allowed the researchers to detect the prevalence of pfhrp2/3 genes in symptomatic individuals. Fourth, the presence of illegal armed groups did not allow for an adequate entomological study because only partial observation was allowed in some neighborhoods, while in others the research group was not allowed. Fifth, due to study design, the participants were only questioned at sampling and there was no follow-up, so it is not possible to know if those who were asymptomatic later became symptomatic. Sixth, during survey application a memory bias may have been present, which could have altered the respondents' way of answering the questions.

Particular attention requires the limitation related to entomological aspects. The results presented are only an approximation to the subject. The difficulties presented did not allow them an adequate study of the ecology of the vector and to characterize the aspects related to the vector's biting density, feeding patterns, biting activity, parity rates, and resting behavior, among other issues that represent important factors for malaria transmission within the Commune 5.

A strength of this research was the participation of the “Vector-Borne Disease Prevention and Control Program in the Nariño Department” officials and the municipality of Tumaco. Their intervention will promote the use of these results for decision making in activities related to the pre-elimination of urban malaria in Tumaco.

In summary, the prevalence of malaria was estimated at 2.97% and the prevalence of asymptomatic infection was 2.16% in the urban area of Commune 5 of Tumaco. Parasites positive for the pfhrp2 gene were found. An. albimanus larvae were found in breeding sites and adult forms were found indoors and in peri-domestic areas. All of the cases were caused by P. falciparum. The disease was positively associated with households inhabited by three and more people. Based on these results, pre-elimination of urban malaria should include early diagnosis strategies and continuous active surveillance of asymptomatic infection, which requires the inclusion of molecular diagnostic tests. In addition, the RDTs used to diagnose should be monitored for potential pfhrp2 gene deletion, which will be focused on considering alternatives. Additionally, work should be coordinated with other sectors to establish activities aimed at improving basic sanitation and conduct continuous educational activities aimed at preventing the disease.

Data availability