Background
Rapid diagnostic tests (RDTs) and microscopy are the cornerstone of confirmation of clinical malaria diagnosis in most endemic countries and are also widely used for prevalence surveys. They also have more limited use in other scenarios, including active and reactive case detection and screening pregnant women. RDTs have high sensitivity against clinical infections for
Plasmodium falciparum [
1,
2] as these are typically associated with higher parasite densities and thus higher levels of antigenemia. However, sensitivity when used for detecting asymptomatic infections is considerably lower.
Recently, a Plasmodium falciparum histidine-rich protein 2 (HRP2)–based RDT (Alere™/Abbott Malaria Ag P.f RDT [05FK140], now called NxTek™ Eliminate Malaria Ag Pf) with a tenfold improved analytical sensitivity as compared to average conventional RDTs (co-RDTs) was prequalified by the World Health Organization (WHO). From here on, we refer to this test as the highly sensitive RDT (HS-RDT).
An unprecedented wide range of studies has been conducted using the HS-RDT in a variety of transmission settings and use cases to investigate the practical benefits of its lower limit of detection (LOD). First and foremost, this test was intended to identify asymptomatic infections in mass screening and active case detection interventions, particularly in low-transmission settings. The diagnostic sensitivity of an RDT is driven primarily by two factors: (i) LOD of the test (i.e., its analytical sensitivity) and (ii) the malaria antigen distribution in the sampled infected population. It has been shown that conventional RDTs are more sensitive in high-transmission settings [
3], which is most likely because, on average, individuals have higher parasite densities and thus higher antigenemia [
4]. To assess the utility of the HS-RDT, we need to better understand the how it performs in comparison to co-RDTs or more sensitive nucleic acid amplification–based tests (NAATs) (e.g., polymerase chain reaction [PCR]) and in different transmission settings.
Questions remain around the utility of more highly sensitive RDTs for other use cases: for example, given that current RDTs perform well in clinical settings, will a more sensitive test increase the number of people with symptomatic malaria being correctly diagnosed? There may be individuals that test positive due to having chronic asymptomatic infections, but malaria is not the primary cause of their fever. Furthermore, given that HRP2 decays relatively slowly after parasite clearance, there is a risk that HS-RDTs will increase the numbers of false positives in individuals with recently cleared infections. Twenty-five percent of co-RDTs are estimated to remain positive for at least 20 days after the clearance of parasites, and this is expected to be greater for a more sensitive RDT [
5].
There has been interest in testing pregnant women for malaria during antenatal care (ANC) visits, where drugs more effective than those used in intermittent preventative therapy during pregnancy (IPTp) could be given upon testing positive [
6].
P. falciparum infections are typically harder to detect in pregnant women as the parasites commonly sequester in the placenta [
7] and treating asymptomatic infections has been shown to have positive impacts on both the mother and the infant [
8]. Therefore, more sensitive diagnostics to identify pregnant women to be treated and cleared of asymptomatic infections has a clear potential public health outcome. In a scenario where pregnant women are intermittently screened, we need to know how much more sensitive the HS-RDT is compared to a co-RDT in this population, and whether using a more sensitive test could reduce malaria burden in pregnant women and improve pregnancy outcomes. The testing of pregnant women at their first ANC visit has been identified as a potential sentinel surveillance strategy for monitoring population-level changes in prevalence, but questions remain as to whether this approach accurately captures these trends, and whether an HS-RDT would improve the accuracy of this strategy.
In this article we summarise published and available unpublished data to evaluate the performance of the HS-RDT across different transmission settings and use cases.
Discussion
The performance of any RDT depends on both the limit of detection of the test and on the distribution of target analyte in the populations where it is being used. In the specific case presented in this review, a test with a given LOD will have lower diagnostic sensitivity in infected populations with lower HRP2 concentrations and higher diagnostic sensitivity in populations with higher HRP2 concentrations. The performance of a test (HS-RDT) with a lower LOD than another test (co-RDT) will be more resilient to fluctuations in the target analyte, in this case HRP2 distribution in a given population. Furthermore, the improvement in sensitivity for a test with a lower LOD will depend on the proportion of the population whose analyte concentrations fall between the LOD of the new test and the LOD of the conventional test.
In asymptomatic cross-sectional surveys, the sensitivity of the HS-RDT in asymptomatic populations is estimated to be 56.1% compared to 44.3% with a co-RDT with PCR as the reference standard. We found a positive relationship between PCR prevalence and the sensitivity of the HS-RDT, indicating that it may perform relatively better in high-transmission settings. This is consistent with evidence that parasite densities are higher in high-transmission settings [
4]. The HS-RDT is estimated to detect on average 46% more infections than a co-RDT (Fig.
5). The results presented here show that the HS-RDT consistently outperforms co-RDTs when used for cross-sectional surveys.
For clinical diagnosis, the incremental benefit of using the HS-RDT compared to a co-RDT to test febrile individuals appears to be marginal, with all studies reporting a small number of additional PCR-positive cases detected with the more sensitive test. This is likely because the antigen concentrations of individuals with febrile malaria are higher than in asymptomatic populations and are in the range where the co-RDT already performs well (Fig.
2A, B). The HS-RDT did detect more false positives (PCR negative, HS-RDT positive) compared to the co-RDT, highlighting the importance of proper clinical management to investigate a range of causes of the fever for RDT-positive individuals. This is particularly relevant in settings where proper diagnostic skills to identify other causes of fever may be limited, as is often the case among community health workers and basically trained health staff. This is especially a concern in high-transmission areas with a high probability of having recent malaria episodes with persistent antigens. Conversely, the HS-RDT may also detect new infections sooner [
12] which is a clear benefit in highly susceptible populations such as infants. The risk–benefit of a more sensitive test for malaria case management needs to be better understood.
For screening pregnant women for the malaria, the results here indicate that the HS-RDT may be more sensitive than a co-RDT for detecting malaria in pregnant women in all but one study, and that the additional infections detected by the HS-RDT may have clinical significance for the mother and child. However, only the data aggregated across all stages of pregnancy from the Benin trial [
29] produced a statistically significant result. Currently, WHO recommends IPTp in areas with moderate to high transmission. However, there is widespread resistance to sulfadoxine-pyrimethamine, the antimalarial used in this intervention [
36]. This has led to calls for a screening-based approach where women are screened with an RDT regardless of symptoms and are treated with a more efficacious antimalarial if positive. Initial trials of this approach using a co-RDT have had mixed results and as such this intervention is not recommended by WHO [
37‐
41]. However, it has been considered that a more sensitive screening tool could improve the effectiveness of this approach [
6]. Additionally, trials have investigated the safety of giving antimalarials to women in the first trimester, which if recommended by WHO, would be more impactful if a highly sensitive and specific RDT was available in antenatal clinics.
The Senegal National Malaria Control Programme has recently started to evaluate the use the HS-RDT in reactive case detection strategy. However, there is currently limited direct evidence on the added value of an HS-RDT in active case detection. Results presented here indicate that the HS-RDT is more sensitive than the co-RDT in asymptomatic individuals, suggesting that its use in active case detection would likely improve the effectiveness of the intervention compared to using a co-RDT.
The HS-RDT has a more limited stability claim than most co-RDTs; it claims storage stability to 30 °C versus 40 °C for most of the WHO prequalified RDTs. It also has a more restricted shelf life of 12 months, compared to 24 months for other prequalified co-RDTs [
42]. A survey of many of the investigators of the studies described here suggests that many were aware of these limitations and managed them, but many studies also did not track temperature exposure. There is a possibility that HS-RDTs in some of the studies were compromised, perhaps leading to reduced incremental benefits of the HS-RDT over the co-RDT. Some studies compared the performance of the tests on frozen samples and in one case reconstituted specimens by combining retrospectively blood pellets with plasma [
30]. The impact of this on the performance of any given RDT is not properly characterized and may have influenced some of the study results. Conversely, different studies used different co-RDTs, which will have had different LOD for HRP2 and consequently the difference in comparative test performance may be affected by this. The LOD of HRP2-based RDTs has been shown to vary up to fourfold depending on the product and parasite culture strain investigated [
43]. This, together with lot-to-lot performance variations, could confound the gain in clinical sensitivity observed between the HS-RDT and co-RDTs. Furthermore, PCR is typically used as the reference assay for assessing the performance of a new RDT, however, there is known to be large variation in the sensitivity of different PCR assays. For example, a highly sensitive PCR assay may detect very low density infections that also have very low levels of HRP2. RDTs may miss these infections and as a result have lower estimated sensitivity values. An investigation of the impact of PCR sensitivity is presented in Additional file
3. Results indicate that the HS-RDT sensitivity estimates were relatively robust to the PCR assay sensitivity, however the co-RDT performed significantly worse in studies where a sensitive PCR assay was used.
The Alere™/Abbott Malaria Ag P.f RDT is the first malaria RDT that has been launched with the claim of being ‘highly sensitive’. However, incremental improvements to RDT performance have been continuous and ongoing since development of these tools began in the 1990s. The HS-RDT has been submitted to unprecedented evaluation, with dozens of peer-reviewed articles published since its launch in 2017. The HS-RDT has catalysed a better understanding of how to evaluate RDTs and even the development and adoption of quantitative antigen tests to support their evaluation [
17,
18,
43,
44]. The lessons learnt with the HS-RDT will help more quickly assess the implications of improvements in LOD for other tests in terms of sensitivity in different population groups and use cases.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.