Pf-positivity by qPCR and us-qPCR
By standard Pf_18S rRNA qPCR, positivity was significantly higher in PNG compared to the other field sites (Chi2–test, p<0.001). The absolute number of additionally detected Pf infections by us-qPCR was highest in PNG, the site with intermediate transmission intensity. The relative proportion of additionally detected Pf infections differed substantially between Brazil (55%, 95% CI 25-82) and Thailand (33%, 95% CI 13-61) and PNG (30%, 95%CI 21-40) (Table 2, Fig. 1A, C), but owing to the small number of Pf samples at the sites of low endemicity, these differences in gains were not statistically significant (p=0.77).
Table 2: P. falciparum and P. vivax positivity in community samples from Thailand, Brazil and PNG, determined by species-specific 18S rRNA (qPCR) versus ultra-sensitive qPCR (us-qPCR).
|
P. falciparum positivity (%)
|
P. vivax positivity (%)
|
|
qPCR
(95% CI) (n/N)
|
us-qPCR
(95% CI)
(n/N)
|
p-value1
|
qPCR
(95% CI)
(n/N)
|
us-qPCR
(95% CI) (n/N)
|
p-value1
|
Thailand
|
1.3%
(0.7-2.5)
|
1.9%
(1.1-3.3)
|
0.074
|
3.90%
(2.7-5.6)
|
5.00%
(3.7-6.9)
|
0.008
|
|
(10/773)
|
(15/773)
|
|
(30/773)
|
(39/773)
|
|
Brazil
|
0.80%
(0.3-1.9)
|
1.70%
(0.9-3.1)
|
0.041
|
4.90%
(3.4-6.9)
|
6.50%
(4.7-8.7)
|
0.004
|
|
(5/651)
|
(11/651)
|
|
(32/651)
|
(42/651)
|
|
PNG
|
8.60%
(6.8-10.7)
|
12.20%
(10.1-14.7)
|
<0.001
|
8.00%
(6.3-10.1)
|
11.50%
(9.4-13.8)
|
<0.001
|
|
(71/828)
|
(101/828)
|
|
(66/828)
|
(95/828)
|
|
1McNemar´s Chi2-test used to determine if gain of infections by us-qPCR was significant.
- vivax prevalence by qPCR and us-qPCR
Pv was the prevailing species in Thailand and Brazil with 3.9% (30/773) and 4.9% (32/651) Pv_18S rRNA-positive individuals. Pv_18S rRNA positivity at these field sites was significantly lower than in PNG (8.0%, 66/828, Chi2 –test, p<0.001) (Table 2), where Pv and Pf prevalence rates were comparable (Table 2). The absolute number of Pv infections that were additionally detected by Pv_mtCOX1 us-qPCR was highest in PNG (Table 2), which in global comparison represents particularly high Pv prevalence. The proportion of Pv infection additionally detected by us-qPCR was similar in low endemic Thailand (23%, 95%CI 12-40) and Brazil (24%, 95%CI 13-40), but was higher in PNG (31%, 95%CI 22-41) (Fig. 1B).
By us-qPCR Pf/Pv co-infections were detected in 4.1% (34/828) individuals from PNG and in 0.6% (5/773) individuals from Thailand, whereas no co-infection was observed in Brazil.
- falciparum and P. vivax densities
Quantification by us-qPCR was reliable as shown by a good correlation with quantification by 18S rRNA qPCR (Additional file 2: Fig. S3). Characteristics of both us-qPCR assays that are relevant for quantification, agreed well between the different laboratories (Additional file 1: Table S1, Additional file 2: Fig. S1), hence, between-site comparisons of parasite densities by us-qPCR are appropriate. In contrast, densities of Pf and Pv by us-qPCR cannot be compared directly for several reasons: Firstly, both multi-copy assays target different numbers of gene copies per genome, thus, a comparison based on numbers of detected gene copies/µL would be disproportionate. Secondly, Pf densities were reported as parasites/µL blood by converting Ct-values determined by Pf_varATS us-qPCR into parasite numbers using a parasite trendline, whereas for Pv, owing to the lack of in vitro culture, a well-defined Pv parasite trendline was not available for absolute parasite quantification. Instead, Pv densities were derived from a dilution series of a plasmid standard and reported as Pv_mtCOX1 copy numbers/µL blood. While densities of Pf and Pv are not comparable between species, the median densities of either species are well comparable between field sites. For all countries, Pf and Pv densities by us-qPCR were stratified according to a sample’s positivity by either of the qPCR assays or both assays. (Table 3 and Fig. 1C, D).
Table 3: P. falciparum and P. vivax densities stratified according to positivity by standard 18S rRNA qPCR (qPCR), by any qPCR, or only by ultra- sensitive qPCR (us-qPCR).
|
P. falciparum Median Density1
parasites/µL (IQR)
|
P. vivax Median Density1
Pv_mtCOX1 copies/µL (IQR)
|
Country
|
qPCR(+)
|
qPCR(+) and
us-qPCR(+)
|
qPCR(-)
and
us-qPCR(+)
|
qPCR(+)
|
qPCR(+) and
us-qPCR(+)
|
qPCR(-) and
us-qPCR(+)
|
Thailand
|
238.9
(55.8-1519)
|
40.8
(0.5-536.2)
|
0.2
(0.2-0.2)
|
104.8
(24.9-246.7)
|
50.4
(4.9-190.8)
|
3.1
(0.9-7.6)
|
Brazil
|
4.02
(1.9-125.4)
|
0.42
(0.2-3.0)
|
0.2
(0.1-0.3)
|
152.6
(50.3-738.7)
|
79.8
(18.3- 464.4)
|
7.3
(3.7-22.0)
|
PNG
|
5.7
(1.0-70.0)
|
1.9
(0.2-22.5)
|
0.1
(0.0-0.4)
|
9.5
(3.1-145.9)
|
3.7
(0.9-16.4)
|
0.7
(0.3-1.2)
|
1 Log10 transformed median parasite density quantified by us-qPCR.
2 Owing to exhaustion of DNA, parasite densities of 5 of the Brazilian samples were quantified by 18S rRNA qPCR.
The median parasite density of Pf infections detected by Pf_18S rRNA qPCR was highest in low-endemic Thailand, but significantly lower in low-endemic Brazil and moderate-endemic PNG (Table 3). This difference in median density may be attributed to symptomatic cases in the Thai sample set. When all symptomatic Pf infections with densities >1000 parasites/µL (3/15) were excluded from the analysis of Thai community samples, the median Pf density decreased from 238.9 parasites/µL to 11.2 parasites/µL (IQR: 0.2-102.3 parasites/µL).These densities from asymptomatic Thai study participants are in the same range as densities observed in PNG and Brazil (Fig. 1C). In all three settings, parasite densities in infections detected only by us-qPCR were 10-20 times lower than densities in infections only positive by the less sensitive Pf_18S rRNA qPCR (p<0.001) (Table 3).
Just as for Pf, Pv parasite densities in infections only detected by us-qPCR were 8-10 times lower than densities in Pv infections detected by Pv_18S rRNA qPCR (p<0.001) (Table 3, Fig. 1D). The median Pv parasite density in Pv-positives detected by Pv_18S rRNA qPCR was 11-fold lower in community samples from PNG than in those from Thailand, and 16-fold lower than in Brazil (p<0.05) (Table 3). Similarly, the median Pv density in samples only positive by us-qPCR was significantly lower in PNG compared to Thailand and Brazil (p<0.05) (Table 3).
Age trends of P. falciparum and P. vivax infections
To identify population sub-groups that harbour particularly low density infections and where benefits from us-qPCR would thus be greatest, parasite prevalence and density was analysed with respect to age. In all study sites, Pv positivity varied by age. This trend was consistent by both diagnostic assays. The age-stratified prevalence rates generated by the two molecular assays did not differ significantly at any study site (Chi2; Thailand: p-value=0.96, Brazil: p-value=0.79, PNG: p-value=0.73). In Thailand and Brazil, Pv prevalence by both, standard and us-qPCR, peaked in adults aged 20-60 yrs (Fig. 2A-B), whereas in PNG prevalence was highest by both assays in adolescents aged 10-20 yrs (Fig. 2C). No Pv infections were found in the youngest age group in Brazil and Thailand (0-3 yrs), whereas Pv infections were detected by both assays in children from PNG aged 0-3 yrs (Fig. 2A-F). Pv infections were also absent in Thai children aged 3-5 yrs. At all three sites, median Pv densities measured by Pv_mtCox1 copies/µL blood were highest in the youngest age groups that included Pv-positive individuals, and tended to decrease by age (Fig. 2D-F). These trends were not statistically significant owing to the limited number of Pv-positive individuals in some age groups.
Age trends for Pf prevalence and parasite density in Thailand and Brazil could not be analysed because of the limited number of Pf-positive individuals. For PNG, the age distribution of Pf infection is shown in Additional file 2: Fig. S4.