Background
The World Health Organization (WHO) estimates that despite recent reductions in malaria incidence, malaria contributed to the morbidity and mortality of about 166 million and 564,000 people, respectively, in sub-Saharan Africa (SSA) in 2012 [
1]. However, other epidemiological estimates nearly double these figures [
2]. It is unclear how ‘presumptive diagnosis’ has influenced these statistics [
3], since patients presenting with febrile illnesses with undetected parasitaemia by microscopy or rapid diagnostic tests (RDTs) are inadvertently integrated into the statistics, confounding prevalence measures of malaria in the process [
4‐
6].
Appropriate diagnosis is not only important for ensuring prudent use of anti-malarial medications, and facilitating correct prognosis based on appropriate differential diagnosis [
1,
7,
8], but also for tracking malaria elimination efforts [
9‐
11]. Conventional microscopy [
12] and RDT-based malaria diagnostics [
13] used in malaria endemic regions of Africa have relatively low sensitivity and specificity in the detection of low-parasitaemia infections [
14,
15]. Although the limits of detection (LoD) of parasitaemia by microscopy (100 parasites/μL or higher depending on technical skills) [
16,
17] and RDT (50–400 parasites/μL) [
18] may be well below the estimated parasite density threshold (3,500 parasites/μL) for initiation of febrile episodes [
19], many clinicians in areas of high malaria endemicity are compelled to heuristically prescribe anti-malarials on the basis of clinical symptoms alone [
6,
7,
20‐
22]. Such diagnostic and anti-malarial prescription practices in the absence of malaria positive microscopy and/or RDT results [
6‐
8] may cause overlooking potentially severe non-malarial febrile illnesses that may coexist in malaria endemic tropical regions, such as bacterial, mycobacterial, fungal, and arthropod-borne viral infections [
20,
11,
23,
24]. Indeed, even reliance on RDTs, which may also detect sub-clinical
Plasmodium parasitaemia, has been implicated in undermining considerations of such differential diagnoses [
10].
While low-parasitaemia
Plasmodium diagnostics may be inappropriate for attributing febrile symptoms to malaria, effective malaria elimination efforts must be able to identify sub-clinical, low-parasitaemia
Plasmodium infections that contribute to the infectious reservoir [
12]. Molecular techniques are generally superior to antigen-detection tests and microscopy for detecting pathogens in biological samples [
12,
13,
25,
26]. Specifically, techniques such as nested Polymerase chain reaction (nPCR) can enhance the sensitivity of detecting low-parasitaemia
Plasmodium infections by increasing the total number of PCR cycles and by diluting the inhibitors present in the initial amplification reaction [
27]. Similarly, direct PCR with high resolution melting (dPCR-HRM) analysis can distinguish between different
Plasmodium species in a sample [
28] and can hence provide better information on the relative abundance and composition of
Plasmodium species in a study area.
This paper describes a novel molecular diagnostic technique that combines nested PCR with HRM analysis (nPCR-HRM) to detect and differentiate low-parasitaemia
Plasmodium infections. The sensitivity of the novel coupled nPCR-HRM approach was compared to the performance of separate and independent implementations of nPCR [
27] and dPCR-HRM [
28] assays. The novel technique was further applied to quantify misadministration of anti-malarial drugs among patients presenting febrile symptoms without
Plasmodium infections detectable by microscopy or RDTs in rural clinics on Rusinga and Mfangano Islands of Lake Victoria in Kenya.
Discussion
By combining the high sensitivity of nested PCR with real-time species differentiating HRM analysis, the reporting of
Plasmodium infection prevalence that includes low-parasitaemia infections can be synergistically enhanced to improve size estimates of its infectious human reservoir in regions where malaria elimination efforts are being scaled up. Indeed, nPCR-HRM was able to detect
P. falciparum at orders of magnitude lower concentrations than either nPCR or dPCR-HRM. The LoD of the coupled nPCR-HRM assay (236 parasites/mL) is however in the same range as TaqMan Probe assays validated using the same
P. falciparum DNA standard (313 parasites/mL) [
36]. Out of the 62 patient samples in which
Plasmodium parasites were detected by the novel nPCR-HRM technique, the well-established nPCR [
27] and dPCR-HRM [
28] molecular techniques detected
Plasmodium parasites in only 20 and 39 samples, respectively. The same is evident for
P. falciparum, where nPCR-HRM identified 26.9% prevalence among microscopically negative febrile illness patients, while nPCR and dPCR-HRM identified only 10.2% and 13.7% prevalence, respectively. These results are consistent with the lower sensitivity of nPCR previously reported in comparison to dPCR-HRM [
38], yet further demonstrate that nPCR-HRM is significantly more sensitive and specific than either nPCR or dPCR-HRM, and was also the only method able to detect double and triple infections with distinct
Plasmodium parasites.
Although other novel molecular methods for
Plasmodium detection have been reported recently, they either failed to undercut the detection thresholds of microscopy on clinical specimen [
39] or were limited in ability to differentiate
Plasmodium species [
25,
36]. Due to these limitations in detecting and differentiating low-parasitaemia
Plasmodium infections, no previous studies have assessed drug administration among subjects with malaria parasitaemia below the detection limits of standard microscopy and RDT.
The lack of diagnostic tools sufficiently sensitive to detect low
Plasmodium parasitaemia in rural clinics in malaria endemic regions of SSA, and that can differentially diagnose other non-malaria febrile illnesses, including bacterial, mycobacterial, fungal, and arboviral infections, is a big challenge [
12,
20‐
24]. While limited training of laboratory technicians has been suggested as a cause of poor malaria diagnoses [
40], clinicians are left with little option, as a result of diagnostic limitations in rural clinics, but to treat patients with anti-malarials and antibiotics based on clinical symptoms [
6,
7,
20‐
22].
Febrile illness misdiagnosis can be significant in malaria endemic settings, resulting in a myriad of downstream issues including improper treatment, chronic suffering of patients as the underlying cause of illness remains unknown, and drug wastage [
7]. Moreover, poor malaria diagnostics can undermine the diagnosis of non-malarial febrile illnesses [
10,
23,
24]. Due to the limitations of microscopy and RDTs in detecting low-parasitaemia malaria [
14,
15] and clinical heuristic treatment practices [
6,
7,
20‐
22], 197 febrile patients from both study sites were diagnosed as not having malaria, but were still given different medications on the basis of clinical signs. In endemic regions with high transmission intensity, such presumptive fever management may be justifiable, as test-based management may not be safe or cost effective [
10]. Because clinicians are limited to heuristic methods of clinical diagnoses in such cases, treatment patterns are likely to differ between clinics and clinicians. Indeed, while all low
Plasmodium parasitaemia cases at Sena Health Centre were treated with anti-malarial drugs, only 44.4% of low-parasitaemia patients at Tom Mboya Health Centre were treated with anti-malarial drugs. However, in both Health Centres, the majority of anti-malarial drugs were prescribed to patients in whom we did not detect
Plasmodium parasitaemia.
The discrepant prescribing patterns between the two study sites may be due to the over three times greater numbers of patients with microscopically positive malaria observed at Sena Health Centre than at Tom Mboya Health Centre during the respective sampling periods. Accessibility to other health facilities may also have influenced prescribing practices. Compared to Tom Mboya Health Centre on Rusinga Island, Sena Health Centre on Mfangano Island is more isolated from the mainland where patients access other nearby and specialized health centres that they may be referred to for assessment of differential diagnoses. Due to the relatively limited accessibility to such referral clinics for patients on Mfangano Island, clinicians may have had a stronger tendency to prescribe the more readily available anti-malarials at Sena Health Centre.
Individuals with extremely low-parasitaemia
Plasmodium infection are likely to have endemic stability to malaria infection and may be able to tolerate a low degree of parasitaemia with absence of clinical disease [
19]. Therefore, anti-malarial drugs do not treat febrile illness patients with sub-microscopic
Plasmodium infections for the underlying differential diagnosis of their febrile symptoms [
10]. Nonetheless, treatment of patients with sub-microscopic
Plasmodium infection, may contribute to integrated malaria elimination/control efforts by reducing the size of the infectious reservoir within communities [
9‐
11].
This study demonstrates how improper febrile illness diagnoses confound prevalence estimates of
Plasmodium infections and other differential diagnoses. Some non-malaria febrile patients are diagnosed and treated for malaria based on their clinical symptoms, and may be inadvertently integrated into epidemiological statistics [
3‐
6], while some low-parasitaemia
Plasmodium infections go unrecognized [
2]. As a result, malaria burden and infection estimates suffer from unknown large margins of error [
4]. The complementary use of sensitive and specific molecular based methods informs size estimates of the potential infectious reservoir of malaria in endemic regions, while also providing estimates of undifferentiated febrile illness burden.
Improved understanding of
Plasmodium infection rates will translate into the development of better rural malaria diagnostics and treatments, ultimately reducing gametocyte incidence rates in asymptomatic populations with low-parasitaemia malaria [
12]. While individuals with submicroscopic parasitaemia have significantly lower
Plasmodium transmission rates to mosquito vectors [
41,
42], they may still constitute a considerable proportion of the human infectious reservoir [
12,
42,
43], a phenomenon that probably prevails among individuals with endemic stability to malaria infection in the endemic study sites. Therefore, more accurate parasite-detection based diagnosis and treatment may reduce malaria transmission [
42]. Low-parasitaemia malaria can result in acute febrile illness [
5] and individuals that may have low-parasitaemia in the peripheral blood may still require anti-malarial treatment due to higher parasite loads that can still be sequestered in capillaries [
44]. In contrast, acute febrile illnesses have many aetiologies ranging from bacterial, fungal and mycobacterial infections to arboviral infections [
24]. These unknowns complicate appropriate febrile illness diagnosis and treatment. Clear guidelines and policies on the management of non-malaria acute febrile illnesses are required urgently, particularly in malaria endemic regions.
This study’s approach to low-parasitaemia
Plasmodium detection and species differentiation is a valuable molecular tool that can be used to complement microscopy and RDT and enhance understanding of specific
Plasmodium infection rates. While sensitive and specific, nPCR-HRM was not designed as a diagnostic tool for health clinics, as it requires real-time PCR infrastructure, specialized training to avoid the cross-contamination risks associated with nested PCR assays, and more man power in comparison to RDTs. Recent developments in both loop-mediated isothermal amplification (LAMP) [
25] and isothermal lateral flow-recombinant polymerase amplification (LF-RPA) [
26] assays show promise as true point-of-care tests with high specificity and sensitivity for detecting low-parasitaemia
Plasmodium infections. Nonetheless, nPCR-HRM can be employed for monitoring low-parasitaemia infection rates within malaria endemic regions to assess the progress of regional malaria elimination efforts.
Conclusions
This is the first study that employs low-parasitaemia
Plasmodium diagnostics to quantify both the over-prescription of anti-malarial drugs in non-malaria febrile patients and under-prescription of anti-malarial drugs in low-parasitaemia malaria patients. While patient febrile illness symptoms cannot be attributed to any of the sub-microscopic
Plasmodium infections identified in this study, consistent anti-malarial treatment limited to such cases should contribute to regional malaria elimination efforts [
9‐
11]. Other recent cohort studies have contrasted presumptive clinical diagnoses and anti-malarial drug treatments with malaria parasitaemia detectable by microscopy [
21,
23,
24], as microscopy [
8], alongside RDTs [
5], is the mainstay of present malaria diagnostic practices in most malaria endemic regions of Africa. In a low transmission area clinical cohort study in Tanzania, Reyburn and colleagues found that 38% of suspected malaria patients were treated for malaria without further testing and 48% of microscopically malaria negative febrile patients were presumptively treated with anti-malarial drugs [
21]. More recently, Crump and colleagues showed that only 2.7% of patients diagnosed with malaria in two hospitals in northern Tanzania were actually malaria positive as determined by microscopy [
24]. Considering
Plasmodium infection prevalence, we further demonstrate the inadequacy of rural diagnostics in detecting low-parasitaemia infections.
While patients with undiagnosed and untreated low-parasitaemia Plasmodium infections may not suffer from acute febrile illness as a result of malaria, they may contribute to a significant proportion of the infectious reservoir during their asymptomatic state. Conversely, the lack of diagnostics and treatment for unrecognized differential febrile illness diagnoses may have severe health implications that cannot be addressed properly. Epidemiological estimates need to take into account the considerable proportions of asymptomatic low-parasitaemia Plasmodium infections that contribute to the infectious reservoir, while also considering the health impacts of unrecognized differential diagnoses of febrile illness in malaria endemic regions.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
PNK carried out the molecular assay optimization, data collection and analysis, and drafted the manuscript. DOO facilitated sample and data collection. POM helped draft the manuscript. PS facilitated sample collections. DKM contributed to the initial idea of the study and helped draft the paper. JV contributed to the initial idea of the study, designed the methodology, analysed data, validated the assays and wrote the manuscript. All authors read and approved the final manuscript.