Estimation on local transmission of malaria by serological approach under low transmission setting in Myanmar

One-year longitudinal follow-up sample collections were done in Shwegyin Township (22° 20′ 0″ N, 95° 56′ 0″ E), one of the Tier I areas of Myanmar Artemisinin Resistance Containment (MARC) areas in 2015. Because of the nearby gold mines, many migrant workers were working in these areas and malaria was notified as one of the leading diseases in Shwegyin and evidence on drug resistant-malaria was documented [15]. It has been selected as one of the township for elimination programme initiated in Myanmar.

Required sample size has been calculated by PASS sample size software (version 13, NCSS, USA). As this study was carried out as a longitudinal observational cohort, group sequential Log-rank test has been applied assuming 90% power to detect the hazard rate 0.76 when the proportions sero-conversion in each group were 0.3 and 0.4 at a significant level (alpha) of 0.05 using a two-sided Log-rank test for the four sequential tests in each active sample collection. The required minimal sample size is 884. Loss to follow-up had to be assumed 20% in each time and the sample size had to be 1060 in minimum.

The participants in this study were recruited by randomized cluster sampling in Shwegyin Township. Inclusion criteria included minimum age of 5 year, both sex and local residents in the study area for more than 3 years. As the study aimed to conduct a follow-up longitudinal study up to 1 year, the migrant or mobile populations were excluded. The person who currently showed the severe sign and symptoms of malaria was also not included in this study.

At each visit, 1 mL of the whole blood was collected from the fore arm of the participants under aseptic condition. The sera was extracted from the blood and kept at − 80 °C until analysis. The remaining whole blood was used for malaria parasite detection as described below.

Randomized cluster sampling method was used in this study. According to the distribution of the local health centers, two out of the four centers were randomly selected. Then, all villages were listed, from which seven villages were randomly selected. Among them, sampling interval was calculated to get the required samples in the villages i.e., at least 151 per village. Household visit or public meeting places were used to collect the samples based on the convenience of the participants. Active case detection was done in every 3 months until 1 year and passive case detection was carried out by local health personnel.

For negative control, healthy sera were collected from the people of the Republic of Korea who had no exposure to malaria. A total of 96 healthy malaria-naïve individuals were collected and they were 6–13 years old with mean age of 10 years.

Rapid diagnostic test (RDT) (PfHRPII and Pv specific pLDH based assay, SDFK80, Standard Diagnostics, Republic of Korea), microscopy and molecular method using nested PCR amplifying the small subunit ribosomal RNA, were used for detection of asymptomatic malaria infection in all samples collected in every 3 months as described details previously [16, 17].

All malaria infections were treated according to the National Malaria Treatment Guideline of Myanmar. Briefly, artemisinin-based combinational therapy (ACT) with primaquine stat dose for falciparum infection, chloroquine followed by 14 days course of primaquine in vivax infection and chloroquine for malaria infection.

Among the collected samples, 685 (57.95%) attended all four visits. Among them, 270 samples were selected after excluding of the possible confounding factors, i.e., two cohorts in which composed of less than 15 years-old and more than 15 years-old samples in which equal distribution of male and female, and history of malaria (Fig. 1).

The procedures for protein array were described previously [18]. Briefly, amine coated slides were prepared. For microarray screening, 1 µL of optimized concentrations (25 ng/µL for PvMSP1-19, 12.5 ng/µL for PfMSP1-19, 100 ng/µL for PvDBPII and PvAMA1) was spotted into each wells of the arrays and incubated for 2 h at 37 °C. Then the slides were washed with PBS-Tween (0.1%) for 10 min followed by distilled water washing for 5 min. The array slides were then blocked with 5% BSA in PBS-T (PBS with 0.1% Tween) for 1 h at 37 °C. After washing again, the slides were probed with plasma collected from the patients and healthy control individual (1:25 dilution). Alexa Fluor 546-conjugated goat anti-human IgG (10 ng/µL, Invitrogen, Carlsbad, CA, USA) in PBS-T was used as the detection antibody and the signals were detected in a fluorescence scanner (InnoScan 300, Innopsys, Carbonne, France). The Mapix software was used for data acquisition and analysis. All samples were duplicated in the same slides. Mean values of two mean fluorescence intensity (MFI) were calculated.

To assess the cross reactivity of the recombinant antigens between falciparum and vivax, protein microarray using pooled clinical falciparum, pooled clinical vivax and negative control pooled samples, was used in the same array slide with duplicated spots followed by the same procedures described above [19].

Data was checked and analysed by using Microsoft excel and IBM SPSS Statistics (version 23, International Business Machines Corp., Armonk, NY, USA). Pearson’s Chi squared test was used to determine association with a p value of < 0.05 accepted as significant. Logistic regression was calculated using the selected independent variables to estimate the outcome variables. Mann–Whitney test was used to for nonparametric analysis on non-normal distribution. For microarray data, mean fluorescence intensity (MFI) data were normalized and transformed by vsn method with asinh (hyperbolic arc sine) using R-program [20]. The cut-off value was defined as two standard variation (SD)s above the transformed MFI of malaria naïve individual. Mapping for surveillance of the malaria cases by molecular method, PCR comparing with serological response against PvDBPII, PvAMA1, PvMSP1-19 and PfMSP1-19 was carried out. The map was generated as described previously using ARC GIS and SaTScan software [21]. Circular window was used to scan systematically adjusting with the percent positive among the participants in each villages. As the village level output is not provided by this software, final map was generated by the Photoshop CS3 software. In the map, the relative size of the circular windows for seropositive in each of the antigens in the study areas focusing on the geographical distribution of the malaria parasitaemia (by molecular) contracted with malaria serology positive (by protein microarray) were generated. The data include all four active sample collections in accordance with the sample size calculation by Sequential Rank Test.

The study was conducted only after receiving the ethical approval from Institutional Ethical Committee of the Department of Medical Research, Myanmar (Approval number 49/Ethics-2014). It was also registered with ClinicalTrial.gov (Identifier: NCT02708199). Written informed consents were taken in all of the participants. Participation in this study was entirely voluntary. All detected malaria cases were treated according to the national anti-malarial treatment guideline in Myanmar. The personal information collected in this study was kept confidential.

In this study, a total of 1182 local residents were recruited to study. Passive case detection was carried out by local health authorities and local malaria volunteers to detect the clinical malaria episodes. Active case detection and field blood sample collections were conducted one in every 3 months until 1 year. A total of 685 (58.0%) attended all four visits (Table 1).

All collected samples were checked for asymptomatic parasitaemia by RDT, microscopy and nested PCR. Although there was no RDT positive case, two cases in visit 1 were detected as P. vivax infections (parasite densities of 580 and 1200 parasites per µL, respectively). Among the participants, 39 showed asymptomatic infection by PCR and 12 were detected vivax infection in more than one visit (Table 1).

The purified recombinant target antigens were used to screen the four visits collected plasma samples of 270 participants by protein microarray. A total of 1080 plasma samples (1:25 dilution) and 98 healthy sera were included to analyze. Among the candidate serological markers, PvDBPII showed the highest seropositivity: 160/270 (59.3%) in at least one visit of sample collection, followed by PvMSP1-19 antigen 85/270 (31.5%), PfMSP1-19 antigen 73/270 (27.0%) and PvAMA1 antigen 65/270 (24.1%), respectively.

Similarly, overall seropositivity in four times visits (V1-V4) data showed the highest rate in PvDBPII (31.7%), followed by PvMSP1-19 (17.7%), PfMSP1-19 (14.5%) and PvAMA1 (9.9%), respectively. PvDBPII showed the highest seropositivity and stable level among the study population in all four visits. Similarly, PfMSP1-19 showed the stable antigenicity among the population. However, PvMSP1-19 seropositivity showed the highest antigenicity in V4. Interestingly, PvAMA1 seropositivity was lowest in V4 (Fig. 2).

When the age group (< 15 and > 15 year) was considered as a factor that may contribute to the seropositivity of the candidate antigens, only PfMSP1-19 and PvDBPII showed higher seropositivity in old age group (p = 0.000 and p = 0.004). However, there was no association between the seropositivity of the antigens and sex of the participants (Table 2).

Similarly, there was no association between the previous history of malaria and antibody seropositive rate except in PfMSP1-19 that showed the significant association (p = 0.0025). All antigens showed high seropositivity among the asymptomatic infection. PfMSP1-19 showed 3/4 (75.0%) seropositivity among asymptomatic falciparum infection while PvMSP1-19 showed 39/51 (76.5%) seropositivity among the asymptomatic vivax infection. Similarly, PvDBPII and PvAMA1 showed 37/51 (72.5%) and 38/51 (74.5%), respectively among the asymptomatic vivax infection (Table 2).

Among 270 samples, only 28/270 (10.4%) showed co-seropositivity against PvMSP1-19 and PfMSP1-19 antigens. On the other hand, PvMSP1-19 seropositivity alone accounted for 57/270 (21.1%) and PfMSP1-19 seropositivity alone for 45/270 (16.7%). When PfMSP1-19 and PvDBPII seropositivity rates were compared, 49/270 (18.2%) showed co-seropositivity. While PfMSP1-19 seropositivity alone accounted for 24/270 (8.9%), PvDBPII seropositivity alone was 111/270 (41.1%). Similarly, seropositivity rates of PfMSP1-19 and PvAMA1 were compared and 13/270 (4.8%) were seropositive against both antigens. Similar seropositivity rate against the PfMSP1-19 alone (60/270, 22.2%) and that against PvAMA1 alone (52/270, 19.3%) was observed.

Serological response among the four vivax antigens among vivax subclinical cases indicated that all antibody level were stable up to 1 year after infection except in PvMSP1-19 which showed significant reduction of mean fluorescence intensity in 9 months after treatment of infection (Fig. 3).

When mapping was done to detect the malaria geographical patterns of malaria in seropositivity compared with parasitaemia in the samples collected in active case detection by PCR, the PvMSP1-19 showed the similar pattern of the local transmitted areas with that by molecular method. PvAMA1 and PvDBPII showed the high seropositivity in the villages at which very few vivax asymptomatic cases were identified by PCR, suggesting the long-lasting antibody response against these antigens (Fig. 4).