In this study, the performance of the CareStart™ Malaria HRP-2/pLDH (Pf/pan) Combo Test was retrospectively evaluated on a large panel (n = 590) of samples obtained in returned travellers suspected of malaria. Overall sensitivity for the detection of P. falciparum was 88.8%, reaching 94.3% and 99.3% in samples with parasite densities above 100/μl and 1,000/μl respectively. Overall sensitivities for P. vivax, P. ovale and P. malariae were 77.6%, 18.4% and 30.4% respectively, reaching 90.2%, 31.7% and 50.0% for parasite densities above 500/μl. No positive results occurred among the Plasmodium negative samples; species mismatches occurred in 2.2% (11/495) of samples. Agreement between observers was good and test results were reproducible. The test was easy to perform with good clearing of the background. Test line intensities for the HRP-2 line were strong and medium in the majority of cases, but line intensities for the pan-pLDH lines were lower, especially in the case of P. ovale.
The evaluation of an RDT in a reference setting is a logical step preceding evaluations in field trials , but this retrospective approach has its limitations [7, 10]. For instance, it was not possible to explore discordant or unexpected test results by reviewing patient files or by testing for interferences such as the rheumatoid factor. In addition, test conditions in the reference setting are more favourable compared to the field setting. Further, an influence of sample storage was apparent for P. ovale, resulting in a possible underestimation of the actual sensitivity. For the other species, no such effect was noted, and a previous prospective study showed no effect of storage on the stability of the HRP-2 antigen . Finally, stringent criteria were used for defining test characteristics: the eight P. falciparum samples that showed only a pan-pLDH line were scored as "species mismatches", as this result represents a major diagnostic error, i.e. the misdiagnosis of a P. falciparum infection as a non-falciparum infection. However, these results could also be considered as a correct diagnosis of "malaria", thereby increasing the overall sensitivity up to 91.3%.
The CareStart™ Malaria HRP-2/pLDH (Pf/pan) Combo Test has recently been evaluated as part of the WHO/FIND Malaria RDT Evaluation Program . This evaluation showed for P. falciparum and P. vivax detection rates (percentage of naturally infected diluted human samples detected by the product) of respectively 97.5% and 90.0% at parasite densities of 200/μl, and detection rates of 100.0% and 95.0% at parasite densities of 2,000-5,000/μl .
The detection rates as demonstrated by WHO/FIND are slightly higher as compared to the sensitivities found in the present study, but the differences were not statistically significant. Comparisons are difficult as the present study included unprocessed clinical samples with a wild range of parasite densities and the WHO/FIND evaluation used series of diluted samples at fixed parasite densities (200/μl and 2,000-5,000/μl).
A similar product of the same company, the CareStart™ Pf/Pv Combo (detecting HRP-2 and P. vivax-specific pLDH) has been evaluated in Ethiopia . This study reported higher sensitivities for P. falciparum (99.4%). However, it included exclusively P. falciparum samples with parasite densities above 100/μl, which is above the detection threshold of most RDTs. In addition, small differences may also be explained by the fact that we included P. falciparum samples with pure gametocytaemia as positive samples: in the scope of travel medicine this is a reasonable option , but in the present study five out of 17 samples did not react with the HRP-2 line and hence decreased overall sensitivity. Another product of the same company, the CareStart™ Malaria pLDH (Pf/pan) Combo Test has been evaluated in a field study in Madagascar . This study reported sensitivities for P. falciparum that are comparable to the present one, including low values at parasite densities < 100/μl (60,0%) and increasing sensitivity at higher parasite densities (100% at > 500/μl). For P. vivax there were only nine samples included, making comparison difficult. The same product (CareStart™ Malaria pLDH (Pf/pan) Combo Test) and another product of the CareStart™ brand (Malaria pLDH (pan)) have recently been evaluated in Myanmar . Reported sensitivities for the detection of P. vivax were significantly higher than those found in the present study in case of the CareStart™ Malaria pLDH (pan) (91.0%), but for the CareStart™ Malaria pLDH (pan/Pf) they were in line with the present findings (78.5%). Studies evaluating other RDTs in non-endemic countries report similar sensitivities as those found for the CareStart™ Malaria HRP-2/pLDH (Pf/pan) Combo Test in the present study: for P. falciparum they ranged from 87.5-99.0%, with one exception of 76.2% [7, 10, 16–21]. For P. vivax, RDTs detecting pan-pLDH showed sensitivities of 33.5% and 62.0%-95.0%,[7, 10, 13, 18, 20, 22] compared to 46.0%-93.0%  for those RDTs targeting aldolase. The increase of sensitivities at higher parasite densities is a well-known phenomenon with breakpoints around 100/μl (P. falciparum) and 500/μl (P. vivax) [7, 10, 13, 22, 23].
Unlike the WHO/FIND evaluation, the present study included P. ovale and P. malariae samples, for which sensitivities were poor. Previous studies that included P. ovale or P. malariae species reported sensitivities for these species combined between 36 - 95% for pan-pLDH and 7 - 80% for aldolase based RDTs . In addition, a previous study with a study design similar to the present one reported sensitivities for P. ovale and P. malariae of 76.3% and 45.2% respectively . Even taken into account an underestimation of the sensitivity for P. ovale (due to an effect of sample storage), the presently found low sensitivities for P. ovale and P. malariae are of concern given the name and the claimed performance of the test. In a field study however, a related RDT of the same company, the CareStart™ two-band RDT targeting pan-pLDH, detected all six P. malariae infections, though parasite densities were not given .
Of note, in the present study there were eight P. falciparum samples (2.5%) that gave a reaction only with the pan-pLDH line but not with the HRP-2 line. Geographic origin of samples probably did not contribute to the species mismatch, as all eight samples were from patients on their return from sub-Saharan Africa where HRP-2 mutations have not yet been described . Low parasite densities close to the detection threshold may explain for the failure of HRP-2 detection of these samples. This species mismatch is of concern, because misdiagnosis of P. falciparum infection and therefore treatment with ineffective drugs can have fatal consequences .
In line with studies on other RDTs [7, 16, 27, 28], there was a correlation between line intensities and parasite densities, and there were, although to a lesser extent than observed for another RDT , diagnostic clues to parasite densities when scoring line intensities (such as a strong HRP-2 line intensity indicating a parasite density > 100/μl). The low intensities of the pan-pLDH lines, especially for the non-falciparum species, are of concern especially when extrapolating the present findings to field settings in both endemic and non-endemic settings, where disregarding faint test lines as negative results is a common mistake [29–33].
The agreement between observers was excellent for positive and negative readings and good for line intensity readings, and test reproducibility was also good. Of interest is the observation that reading beyond the reading time increased the number of positive pan-pLDH lines, with an increase in sensitivity (although not significantly) at the expense of a slightly lower specificity. Antigen-antibody reactions are time dependent, explaining for the increase in sensitivity, but on the other hand, the so-called backflow phenomenon (non-specific binding) may have interfered, explaining for the apparently false-positive reactions observed by one out of three readers in the Plasmodium negative samples .
The CareStart™ Malaria HRP-2/pLDH (Pf/pan) Combo Test was scored as easy to perform by the present team. The good clearance of the background facilitated reading of faint and weak test lines but the difference in color tone between control and test lines hindered reading of test line intensity. Slight improvements to the package insert and manufacturer's instructions are to be expected, especially with regards to interpretation of the test results. This is of particular concern in the case of the self-test kit, designed for travellers or expatriates, who are among the most inexperienced end-users [19, 35].
In conclusion, the present study demonstrated that the CareStart™ Malaria HRP-2/pLDH (Pf/pan) Combo Test performs well for the detection of P. falciparum and P. vivax infections, but poorly for P. ovale and P. malariae. The occasional species mismatches, in particular P. falciparum samples identified as non-falciparum species, are to be explored and improvements in test instructions should be made.