Background
Malaria in pregnancy (MiP) remains a major public health challenge in areas of high malaria transmission. In pregnant women, malaria can cause mild to severe maternal anaemia, particularly in primigravid women, and placental infection can interfere with the maternal-fetal exchange of nutrients and oxygen, leading to preterm delivery and low birth weight, and consequently increasing neonatal mortality [
1,
2]. In order to prevent the adverse consequences of MiP, the World Health Organization (WHO) recommends the use of intermittent preventative treatment in pregnancy (IPTp) with sulfadoxine-pyrimethamine (SP)—a full treatment dose administered to pregnant women during routine ANC visits in the 2nd and 3rd trimesters. IPTp has a protective efficacy of nearly 25% against low birth weight and 21% against neonatal mortality [
3].
Plasmodium falciparum resistance to SP results from an ordered accumulation of mutations in two genes, namely
P. falciparum dihydrofolate reductase (
Pfdhfr) and
P. falciparum dihydropteroate synthase (
Pfdhps) that code for enzymes targeted by sulfadoxine and pyrimethamine, respectively. Resistance increases with the number of mutant alleles. The presence of five mutant alleles, the
dhfr/dhps “quintuple mutant”, including the
dhfr substitutions N51
I, C59
R, and S108
N and the
dhps substitutions A437
G and K540
E, are associated with a very high rate of failure when SP is used for the treatment of uncomplicated falciparum malaria [
4]. However, SP remains effective for IPTp, even where the prevalence of the quintuple mutant is high, and thus its use continues to be recommended by the WHO [
5,
6]. The efficacy of SP for IPTp appears to be compromised in the presence of a sixth mutation at
dhps581 (sextuple mutant; A581
G); infection with parasites harbouring the sextuple mutant has been associated with increased placental parasitaemia and inflammation and failure of IPTp-SP to improve birth weight [
7‐
9].
Previous studies have reported that the
dhps 581 mutation is common in Tanga Region of Tanzania, with up to 57% of parasites harbouring the sextuple mutant [
10]; however, there are limited data from other areas of Tanzania. In light of the possibility that IPTp-SP may provide no benefit to women in areas with a high prevalence of the sextuple mutant, it is critical to define the extent of this mutation. This study aimed to investigate the prevalence of the SP resistance mutations
dhps K540
E (a surrogate marker for quintuple mutant [
4]) and
dhps A581
G [
9] and
dhfr I164
L [
1] (surrogate markers for sextuple mutant parasites, with A581
G being the more widely reported) in the parasite population in the Lake Zone (Mwanza, Geita, Mara, and Kagera Regions) and Southern Zone (Lindi, Mtwara, and Ruvuma Regions) of Tanzania.
Results
A total of 3575 febrile patients were screened, and 1770 were found to be positive for malaria, yielding an overall positivity rate of 49.5% (Table
2). In this pilot work to determine the extraction efficiency of individual versus pooled extraction using the 120 participants from Kharumwa health centre, the proportion of mutant alleles varied only slightly as a result of batching DBS prior to extraction, with 9.4–10.7% alleles mutated at
dhps540, 2.8–3.9% at
dhps581, and none at
dhfr164 (Table
3). As these differences in allele frequencies were considered negligible, DBS from all other health centers were extracted in pools of 10.
Table 3
Distribution of mutants using pools from individually extracted (1×), 3-punch extracted (3×) and 10-punch extracted (10×)
1× | 88.5% (339/383) | 2.8% (13/464) | 0% (0/319) |
3× | 88.6% (1014/1145) | 3.8% (52/1380) | 0.3% (2/586) |
10× | 89.9% (563/626) | 3.3% (25/765) | 0% (0/401) |
A total of 1750 dried blood spot (DBS) samples were collected (117–160 samples per facility) and used in the final study analysis. Deep sequencing resulted in an average of 341,420 reads at
dhfr (range 208,769–644,927) and 168,260 reads at
dhps (range 49,419–313,645). The library from Kharumwa was re-sequenced during this project. Based upon this repeat sequencing (Table
4) and the 3 sequences in Table
3, the allele frequencies ±standard deviations for
dhps540 and
dhps581 mutants were 90.0 ± 2.1% and 3.2 ± 0.7%, respectively. The
dhps540 mutation was common across all 14 sites, with allele frequency ranging from 55 to 98.4%, with higher allele frequency at sites in Lake Zone compared to Southern Zone (Table
4). Frequency of the
dhps581 mutation ranged from 0 to 2.4%, with the exception of Kayanga health centre (Kagera Region, Lake Zone) where 24.9% of sequences were mutated (Fig.
1). The
dhfr164 mutation was detected only at Kanyanga health centre (0.6%).
Table 4
Summary of drug resistance allele frequencies
Lake Zone |
Kagera | Karagwe | Kayanga | 120 | 89% (154,320/173,457) | 24.9% (46,056/185,117) | 0.6% (3972/623,835) |
Kagera | Muleba | Kaigara | 120 | 91.2% (270,373/296,323) | 1% (31,77/313,669) | 0% (123/286,155) |
Mara | Musoma | Murangi | 117 | 86% (160,431/186,561) | 0% (24/196,781) | 0% (20/327,806) |
Mara | Tarime | Sirari | 120 | 98.4% (262,461/266,632) | 0% (71/285,418) | 0% (17/249,213) |
Geita | Geita | Katoro | 122 | 97.2% (254,986/262,202) | 0.2% (547/280,267) | 0% 19/310,706) |
Geita | Nyang’hwale | Kharumwa | 120 | 92.9% (176,055/189,500) | 2.4% (4860/201,156) | 0% (23/322,015) |
Mwanza | Misungwi | Misasi | 130 | 92.7% (168,879/182,105) | 2% (3773/193,348) | 0% (15/348,488) |
Mwanza | Sengerema | Mwangika | 121 | 93.2% (177,559/190,591) | 0.3% (692/202,087) | 0% (24/386,750) |
Mean allele frequency Lake Zone* | 970 | 82.6% | 3.9% | 0.1% |
Southern Zone |
Lindi | Lindi Rural | Kitomanga | 124 | 94.4% (162,087/171,760) | 0% (47/177,058) | 0% (11/225,710) |
Lindi | Nachingwea | Marambo | 129 | 77.3% (35,351/45,753) | 0.1% (57/50,926) | 0% (41/231,375) |
Mtwara | Masasi | Nagaga | 160 | 68.1% (107,370/157,670) | 0% (17/164,456) | 0% (46/247,292) |
Mtwara | Mtwara | Nanguruwe | 129 | 73.2% (104,952/143,305) | 0.8% (1180/154,626) | 0% (17/385,082) |
Ruvuma | Songea | Madaba | 120 | 92% (71,494/77,751) | 0% (11/83,706) | 0% (13/324,890) |
Ruvuma | Namtumbo | Namtumbo | 118 | 55.7% (26,134/46,901) | 0% (12/52,484) | 0% (22/318,552) |
Mean allele frequency Southern Zone* | 780 | 76.8% | 0.2% | 0% |
Discussion
In a survey of parasites from patients at 14 health facilities in the Lake and Southern Zones of Tanzania, the
dhpsK540
E mutation was very common [in 10 facilities (71%) more than 85% of alleles at
dhpsK540
E were mutated, and nowhere were fewer than 55% mutated], while the
dhpsA581
G mutation remained rare and focal, with frequency greater than 2.4% in only one facility, Kayanga Health Facility, where nearly 25% of alleles carried the
dhpsA581G mutation. The quintuple mutation (represented by the
dhpsK540
E) has been reported widely across Tanzania, with prevalence ranging from 64 to 98% in one recent report assessing seven regions [
10] and 77 to 95% in another [
19]. Previous reports from Tanga Region have found a high prevalence of parasites harbouring the
dhpsA581
G mutation (44% in Korogwe [
7], 51% in Muheza [
19], and 57% in Bondo [
10]). Although the Lake and Southern Zones of Tanzania are not immediately adjacent to Tanga Region, the absence of the
dhpsA581
G mutation in the majority of sites highlights that there may be considerable geographic micro-heterogeneity. This is supported by data from Kavishe et al., who similarly found the
dhpsA581
G to be present at high proportion in only Tanga and Kagera Regions, though the prevalence of mutants at their sites were higher than the allele frequencies reported here [
10]. This is a reassuring finding, suggesting that IPTp-SP retains efficacy in the majority of Tanzania, but highlighting the need for monitoring in multiple geographic sites.
The pooling and sequencing methodology presented here is a cost-effective alternative to individual allele-specific PCR. Using this 2-step pooling methodology, it was possible to perform 90% fewer DNA extractions and 99% fewer sequencing runs than if samples from each participant had been sequenced individually, saving a substantial amount of both time and money. The individual extraction and bi-directional Sanger sequencing of 1750 samples would cost on the order of $19,000. Here the same data has been compiled for under $4000.
Another advantage of the pooling method is that it allows direct calculation of the allele frequency, rather than prevalence. That is, it calculates the percentage of the parasite population bearing the mutation rather than the prevalence of individuals bearing parasites with mutated alleles. From an evolutionary point of view, the allele frequency of a mutation is more important than prevalence of the mutation because it approximates the likelihood that a mosquito will become infected with a mutant parasite after exposure to a given population.
Allele frequency and prevalence can be different when mixed infections are present and when individuals have infections with varying levels of parasitaemia. Traditionally, prevalence data have to be transformed by a complex equation to yield predicted allele frequencies [
20]. However, in general, the allele frequency approximates the prevalence of major strain in a human population, barring any biases [
16,
21].
This study has a number of limitations. Samples were collected only from the Lake and Southern Zones, neither of which are immediately adjacent to Tanga region, where the highest prevalence of the
dhpsA581
G have been reported, thus more studies in those areas are needed to better define the extent of the mutant. With regard to the pooling methodology employed here, while much more cost-effective and time-saving than traditional PCR, it is not possible to trace a parasite strain back to a participant. Also, some haplotypes tend to be amplified better than others, leading to PCR amplification bias. However, newer barcoding methods, such as primer ID [
22], compensate for amplification bias and could allow identification of individuals (although DNA would still have to be extracted from individual DBS).
While overall, these data are reassuring with respect to the efficacy of IPTp-SP, there are select areas with a high prevalence of the sextuple mutant, where IPTp-SP may no longer provide a useful benefit against malaria. Spread of this sextuple mutant will threaten the usefulness of SP for IPTp. Given the fact that the quintuple mutant is already found throughout Tanzania, and the sextuple mutant has been found in high prevalence in several sites, continued surveillance in multiple sites, particularly in and around Tanga and Kagera, is warranted to monitor for the spread of the sextuple mutant. The pooling technique presented here provides a highly efficient and cost effective means to screen many samples from multiple sites.
Authors’ contributions
JG, JMN, SM, OMM, DSI, RAW designed the study. DSI, SM, OMM, RAW, SL, NK, CK, and LAP oversaw the collection of the samples. SMD, KLT, NJH, JAB, JJJ, and SRM analysed the samples. JG, JMN, and SRM drafted the manuscript which all authors edited and approved. All authors read and approved the final manuscript.