Background
Although global malaria mortality fell by 60% over 2000 to 2019, the progress has leveled off in recent years, with sub-Saharan Africa bearing the highest burden of the disease [
1]. High-transmission countries aim at getting back on track to reduce mortality and morbidity, while low-transmission countries aim at eliminating malaria by sustaining the control effort and preventing a rebound in transmission [
2,
3]. São Tomé and Príncipe (STP), an island nation located in Central West Africa, has made significant progress toward a low-transmission country through effective vector control interventions [
4,
5]. Malaria elimination in STP is promising with the benefits of the relatively isolated location, small population, and single vector and parasite species responsible for malaria transmission [
4,
6]. However, STP's progress towards malaria elimination is being threatened by a potential rebound in malaria cases, emergence of insecticide resistant vectors and increased human mobility [
4,
7]. In response to this situation, the Taiwan Anti-Malaria Advisory Mission has partnered with the government of STP to reinforce case follow-up by establishing a real-time electronic case management system, and preserving residual dried blood spots (DBSs) from malaria patients for implementation research. By integrating case surveillance data and parasite’s genetic information, the dynamic changes of parasites over the control period can be tracked, mainly focusing on the genetic diversity, anti-malarial drug resistance, and treatment effectiveness in
Plasmodium falciparum.
The genetic structure of
P. falciparum in STP was studied before 2004 by analysing the diversity levels of microsatellite loci [
8]. The authors detected differences in parasite populations across 1997, 2000, and 2004, showing that local malaria control strategies could cause dynamic changes in parasite populations [
8]. Over a decade later, with the expansion and transformation of the national malaria control program, changes in the parasite populations were expected but have not yet been proven. Therefore, this study tracked the genetic diversity of parasites using the merozoite surface protein 1 and 2 (
msp1 and
msp2). MSP1 and MSP2 are antigens targeted by host-immune responses during blood-stage invasion [
9,
10], and are polymorphic markers for identifying genetically distinct parasite subpopulations [
11,
12]. MSP1 can be divided into three allelic types, K1, MAD20, and RO33, based on the variable sequences in the block 2 region [
13]. MSP2 can be grouped into two dimorphic families, FC27 and 3D7/IC, with different repetitive patterns in the block 3 region [
14]. The genetic structure of parasite populations over time can be traced by genotyping these markers in this longitudinal study.
Malaria is diagnosed through passive case detection by microscopy in hospitals and district health centers and through mass screening by rapid diagnostic tests (RDTs, immunochromatographic malaria combo cassette test) [
7]. A 3-day course of artesunate-amodiaquine (ASAQ, first-line drug) was given to uncomplicated outpatients, and intravenous quinine was given to severe malaria cases and pregnant women during their first trimester, according to local regulations until 2018 [
7,
15]. To follow-up the treatment outcomes, local healthcare workers would visit patients on days 3, 7, 14, 21, and 28 after treatment to collect blood specimens, making blood smears for microscopic examination, and dried blood spots (DBSs) for other research analysis. The second-line drug, artemether-lumefantrine (AL, Coartem®, Novartis), was given to patients with remaining parasites at the follow-up day after the initial treatment [
7,
16]. Primaquine was given to patients showing gametocytes during follow-up. A small proportion of recurrent infections observed in STP raised the concerns of possible drug resistance in the parasites. Artemisinin resistance is known to be associated with mutations in the propeller domain of kelch 13 (
pfk13) [
17]. Resistance to ACT partner drugs, including amodiaquine (AQ) and lumefantrine (LF) used in STP, is associated with gene mutations in the multi-drug resistance I (
pfmdr1) and chloroquine resistance transporter (
pfcrt) [
18,
19]. Thus, this study genotyped these markers coupled with case follow-up data to screen the drug-resistance markers and investigate risk factors for treatment failures in STP.
Discussion
This study tracked the dynamic changes in genetic diversity, anti-malarial drug resistance of parasites, and treatment follow-up of malaria cases in STP. The prevalent parasite strains changed during the transition time from the low-transmission to the pre-elimination settings. Moreover, patients with younger age, higher parasite density at enrollment, and receiving quinine treatment were more likely to experience recurrence during follow-up, which was majorly due to recrudescent infections based on the genotyping results.
A previous study analysed 180
P. falciparum isolates from STP in 2000 and showed that the frequency of
msp1 K1, MAD20, and RO33 types was 50%, 44%, and 6%, respectively [
12]. This was similar to the frequency detected in the samples from 2010 to 2011 (K1 = 54%, MAD20 = 42%, RO33 = 4%) in this study. After the peak incidence in 2012, the frequencies of the MAD20 and RO33 allelic types increased and replaced K1 types as the prevalent alleles. Similar results were also found in
msp2. The proportion of
msp2 3D7/IC and FC27 alleles detected from 2010 to 2011 was similar to that in 2000 [
12], of which 3D7/IC was the prevalent allelic type (60–65%), and FC27 was the minor type (35–40%). However, after 2012, FC27 replaced 3D7/IC as the dominant
msp2 alleles. The changing of genetic makeups could be associated with malaria transmission intensity and was shown in studies from other island countries [
32,
33]. For example, a generally high parasite genetic variation in
msp1 and
msp2 was found after deploying multiple malaria control measures since 2004 on Bioko Island, the neighboring island of STP [
32]. A progressive decrease in parasite genetic diversity was observed due to the declined malaria transmission intensity after the introduction of ACT on the Grande Comore Island, located off the southeast coast of Africa [
33]. Generally, STP was under low transmission intensity, and the mean MOI was low across the study period. Still, changes of
msp1 and
msp2 compositions in parasites were detected and coincident with the transmission trend. Although with limited sample numbers, the reasons for this changing deserved further discussion and investigation.
The haplotypes that showed expansion after 2012 in STP were
msp1 MH3 and RH1, and
msp2 FH5 and FH6. The
msp1 MH3 and
msp2 FH6 were found before 2012 and increased in frequencies afterward. The
msp1 RH1 and
msp2 FH5 were not detected in the samples collected before 2012. Based on these findings, this study supposed the changes of the prevalent strains might be due to the population expansion of a few haplotypes that originally existed locally, or some new strains that may have evolved or emerged through the recombination or importation events [
34‐
36]. The expansion of a few parasite clones could be attributed to several factors, including selective antibodies, anti-malarial drug pressures, importation, or a founder effect at the beginning of the epidemics [
34,
37,
38]. Changes in the prevalent types before, during, and after the outbreak were also identified in Djibouti, Northern East Africa, and were suggested to be associated with the expansion of a few strains that were already prevalent during the epidemics in 1999 [
37]. Another study in Myanmar showed that the population structure of
msp1 and
msp2 had diversified drastically in 2013–2015, compared to the previous years, 2004–2006, and could be due to a higher level of intragenic recombination estimated in the recent population [
36]. For islands like STP, human mobility and malaria importation are of great challenges, especially importing from mainland Africa where malaria transmission is more intense [
39‐
41]. Examples from Bioko Island showed that much of the
P. falciparum parasites currently observed on the island could probably be attributed to imported cases from Equatorial Guinea [
40,
41]. Future studies could verify these plausible factors by utilizing more genomic, epidemiological, and mobility data to reveal the rebounded causes and threats in STP.
Anti-malarial drug resistance is another critical challenge to malaria control and elimination [
42]. According to the policy decision recommended by the WHO [
43], ACT should be changed if the proportion of treated patients remaining parasitaemic on day 3 exceeded 10%. In STP, the proportion of patients showing positive parasitaemia on day 3 after ACT treatment was 2.5% (63/2,576), showing that the current treatment policy was acceptable. However, by monitoring target-site mutations, this study found an increase in the
pfmdr1 86Y mutation compared to that in a previous study conducted in 2004 when ACT was introduced in STP [
44]. The prevalence of the
pfmdr1 86Y mutation in parasite isolates increased from 21% in 2004 to 87% in 2014–2016. This mutation is associated with the reduced sensitivity to AQ and chloroquine (CQ) in
P. falciparum [
45]. With the increased
pfmdr1 86Y mutation and nearly fixed mutation of
pfcrt 76 T, the local parasites may show increased tolerance to AQ, raising concerns regarding the use of AQ as the first-line ACT partner drug in STP. Other polymorphisms showed the same pattern of predominance (
pfmdr1 184F and
pfmdr1 D1246) as shown in the parasites isolated in 2004 [
44]. Overall, the prevalent haplotype of
pfmdr1 and
pfcrt was YFD (51.4%) and CVIET (92.8%), which was similar to the findings in the neighbouring Bioko island, where the
pfmdr1 YFD (45–59%) and
pfcrt CVIET (92%) were also the prevalent types from 2011 to 2014 [
46,
47].
The drug-resistance genotyping results found that most parasites detected after treatment were recrudescent infections, showing identical genotypes as the initial infections. However, a few recurrent infections (seven patients) showed substitutions of drug resistance types after treatment. The observed substitutions in these seven patients could be owing to two reasons. First, multi-clonal and new infections were detected in four post-treatment samples, showing that they were infected by additional or new parasite strains that carry the alternative alleles. Second, the substitutions of AL resistant types (pfmdr1 86Y → N86, and pfcrt 76 T → K76) were detected in patients treated by the second-line drug, AL, during follow-up. This suggested possible rapid selection against AL treatment. However, AL was not popularly used in STP compared to ASAQ. Therefore, the drug-resistance selection of AL will require further investigation with larger sample size.
This study found that young children had a higher risk of showing high parasitaemia and treatment failures, probably due to the lack of acquired immuno-protection and lower treatment compliance compared to adults [
48]. Notably, patients treated with quinine were more likely to show insufficient clearance of parasites during follow-up than those treated with ACT. Several possible reasons were suggested as follows. First, patients treated with quinine mostly had higher parasitaemia levels, prolonging the clearance time. The mean parasite clearance time of quinine treatment was approximately 4–5 days [
49,
50], which was slower than ASAQ treatment (2–3 days) in general [
51], and the difference may be more significant in cases with higher parasitaemia. Second, poor compliance and tolerability with the quinine regimen may occur, especially in children. The recommended quinine regimen in sub-Saharan Africa is 10 mg/kg administered three times daily for 7 days [
43,
49]. This prolonged treatment course may reduce patients’ compliance if they have to take medicines after being discharged from the hospital. Studies in Africa have found unacceptably high treatment failure rates for patients who did not complete the 7-day quinine regimen [
52,
53]. One study in Uganda investigated children administered a 7-day course of oral quinine, and found 69% (18/26) showed recrudescence during follow-up. The primary causes were poor adherence due to the caregivers forgetting to administer the drugs, the drugs being vomited up, or the children feeling better [
54]. This was similar to the observations in STP. The study results also proved that most of the recurrent infections in quinine group were due to the incomplete clearance of the initially infected parasites, with only a few due to new infections. The final reason was the possible development of resistance against quinine in the parasite population. The mechanism of quinine resistance has not been well elucidated [
53]. Conflicting results from the lack of resistance to varying degrees of resistance against quinine have been reported in Africa [
55,
56]. Notably, one in vitro study from Western Kenya showed polymorphisms in
pfmdr1 86Y, 184F, and
pfcrt 76 T were significantly associated with reduced quinine susceptibility in
P. falciparum [
57]. This suggests that the parasite strains in STP may have developed reduced sensitivity in quinine owing to the high prevalence of
pfmdr1 86Y, 184F, and
pfcrt 76 T detected in this study. However, the contribution of these genetic polymorphisms differed among strains; thus, the association with in vitro quinine susceptibility should be further assessed in the STP isolates [
57].
Quinine has been used since chloroquine was abandoned in STP due to high resistance in the 1990s [
49]. According to local regulations, it has been used for severe malaria treatment, at least until 2018 [
4,
7,
15,
16,
58]. One possible reason why quinine has long been used may be that the medicinal plant for quinine (Cinchona tree) is abundant on Sao Tome and Principe Islands [
59], and the communities are more adapted to quinine as one of the most-used anti-malarial drugs. However, following the WHO Guidelines for Treatment of Malaria [
43] and the current findings, it was suggested to replace IV quinine with IV artesunate. According to the latest WHO malaria report 2020 [
1], the local government has updated their anti-malarial drug policies in 2019, using IV artesunate to treat severe malaria now. Although no mutations were found in the artemisinin resistance marker in this study, the
pfk13 mutations have been reported in a few African parasites recently, for example, the emergence of
pfk13-mediated artemisinin resistance in Rwanda [
60] and Uganda [
61], which should be carefully monitored in all the African isolates in the future.
Overall, the increased outdoor mosquito density and pyrethroid resistance [
4], the changes of parasites’ antigenic alleles and drug resistance mutations, and a higher treatment failure in young children treated by quinine could be significant challenges for malaria elimination in STP. Moreover, surveillance data showed that malaria hotspots in STP are distributed nearby the Central Hospital (HAM), military camps, schools, markets, and airports located in the capital district, Água Grande. These are places that aggregate many people and may intensify malaria transmission. Although the higher case numbers observed in these hotspots may be due to the accessibility to medical services, poor environmental management in these densely populated locations pose a great threat to the rebound of malaria. Concerning malaria elimination, these specified risk factors and hotspots should be addressed in the policy decision-making for the national malaria control programmes in STP.