Assessing the reliability of microscopy and rapid diagnostic tests in malaria diagnosis in areas with varying parasite density among older children and adult patients in Nigeria.

BACKGROUND: Current malaria control strategies are based on early diagnosis and appropriate treatment of malaria cases. The study aimed at comparing the performance of blood film microscopy and rapid diagnostic test (RDT) in Plasmodium falciparum detection in patients ≥6 years of age.
MATERIALS AND METHODS: A total of 154 consecutive pyretic patients aged 6-62 years were enrolled, sampled, and tested for malaria using RDT (first response) and microscopy by Giemsa staining. Genomic DNA was extracted after saponin hemolysis and nested polymerase chain reaction (PCR) was used to detect Plasmodium falciparum. The endpoints were sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).
RESULTS: Of the 154 patients, 80 (51.9%) had fever of ≥37.5°C. 106 (68.8%) were positive by First response® , 132 (85.7%) by microscopy, and 121 (78.6%) by PCR. The sensitivity, specificity, PPV, and NPV of first response compared to microscopic method were 82.2%, 100.0%, 100.0%, and 34.3%, respectively, while it was 75.4%, 75.0%, 95.3%, and 31.2%, respectively, when compared to PCR. The sensitivity, specificity, PPV, and NPV of the microscopic method compared to PCR were 92.3%, 50.0%, 90.91%, and 54.5%, respectively. There was a significant difference in the performance of RDT and film microscopy methods (P ≤ 0.05).
CONCLUSION: Microscopy performed better and is more reliable than first response (RDT) in areas with low parasite density among patients ≥6 years of age. Rapid diagnostic tests could be useful in aareas with high parasite density as an alternative to smear microscopy.

Background and Rationale

Malaria is presently one of the most common parasitic diseases and a major health problem worldwide. In Nigeria, malaria is endemic throughout the country and accounts for up to 60% outpatient visits to the health facility, 30% childhood mortalities, and 11% maternal deaths.[12] Efforts to achieve 50% reduction in malaria morbidity and mortality have been hampered by several ecological and anthropogenic factors ranging from parasite/vector, drug/insecticide resistance to a counterproductive attitude of the human populace.[3] The current malaria control strategies are mainly based on prompt and early diagnosis and correct treatment of malaria cases.[4] Clinical diagnosis is the least expensive, most commonly used method of malaria diagnosis and is the basis for self-treatment. However, the overlap of malaria symptoms with other tropical diseases other tropical diseases and can result in misdiagnosis. Microscopic examination of stained thick and thin blood films is currently the standard method for laboratory diagnosis of malaria. These techniques require considerable expertise and adequate quality control. On the other hand, its reliability is questionable, particularly at low levels of parasitemia and in the interpretation of mixed infection.[56] In comparison to expert microscopy, a wide range of poor specificity of local microscopy is reported.[78] Poor blood film preparation generates artifacts commonly mistaken for malaria parasites including bacteria, fungi, stain precipitation, dirt, and cell debris.[9] To improve the specificity of blood microscopy, fluorescent-based staining and fluorescent microscopy are now introduced in clinical laboratories but are expensive.

This facilitates malaria case detection, reporting, and confirmation of treatment failure.[10] In view of this, rapid diagnostic test (RDT) has been developed for situations in which reliable microscopy may not be available, accessible, and/or affordable. RDT detects malaria antigens in a small amount of blood, usually 5-15 μl using immune chromatographic assay formats in which monoclonal antibodies directed against specific parasite antigens are impregnated on a test strip or cassette. The result, usually a colored test line, is obtained in 5-20 min. The most common antigens currently targeted in malaria RDTs are histidine-rich protein 2 (HRP-2), Plasmodium lactate dehydrogenase (PLDH), and Plasmodium aldolase.[6] Several studies have reported excellent sensitivity and specificity for RDTs when compared to conventional microscopy.[1112131415] RDTs are cheap, simple to perform, and easy to interpret. In Nigeria, the National Policy on Malaria Diagnosis and Treatment states that RDTs should be deployed in situations where microscopy may not be possible due to lack of adequate laboratory facilities or as a complement to the use of microscopy in secondary health facilities.

The most recent method for diagnosis of malaria is by detection of parasite genetic material through polymerase chain reaction (PCR) technique. One important use of PCR method is in detecting mixed infections or differentiating between infecting species when microscopic examination is inconclusive.[16] In addition, the PCR technique could be useful for conducting molecular epidemiological investigation of malaria clusters or epidemics.[1718] A number of PCR assays have been developed for the detection of malaria DNA from whole blood as single, nested, allele-specific, or multiplex methods.[18] These assays have been used for the initial diagnosis, follow-up to treatment, and as sensitive standards against which other nonmolecular methods could be compared. PCR may not strictly be considered a rapid technique for the initial diagnosis of malaria as it is more expensive and time-consuming for a single sample than microscopy and RDT.[19]

A number of studies have been conducted assessing the performance of diagnostic tool in children, especially those ≤5 years of age;[1314] hence, we assessed RDT and blood film microscopic methods compared to PCR in detecting Plasmodium falciparum malaria in older children (≥6 years of age) to ascertain the analytical reliability of using RDT as an alternative to microscopy in malaria holoendemic areas of Enugu State, Southeastern Nigeria.

Materials and Methods

Health research ethics approval

Ethical approval was obtained from the University of Nigeria Teaching Hospital (UNTH) Ituku Ozalla, Enugu, Nigeria with reference number NHREC/05/01/2008B-FWA00002458-IRB00002323 and written, informed consent was obtained from parents and assent from the children.

Study setting

The study was conducted during the rainy season from March to November in three hospitals: District Hospital Nsukka (DHN), Bishop Shanahan Hospital Enugu-Ezike (BSHE), and Cottage Hospital Ugbuawka (CHU), located in three different local government areas in Enugu State, Nigeria. Enugu State is located in Southeastern Nigeria where malaria is present throughout the year with high malaria transmission during the rainy season. Two of the hospitals, Bishop Shanahan Hospital and Cottage Hospital Ugbuawka are situated in rural areas while District Hospital Nsukka is situated in an urban area. District Hospital Nsukka and Bishop Shanahan Hospital are secondary health facilities. Their laboratory units have two laboratory assistants headed by a laboratory scientist with 10 years working experience. Cottage Hospital Ugbuawka is a primary health facility with two laboratory technicians with over 4 years working experience.

Participants and eligibility

Outpatients (≥6 years of age) who had a temperature of ≥37.5°C and/or headache were recruited after informed consent was obtained. Patients who had taken antimalarials in the past 2 weeks prior to sampling were not enrolled. Descriptive characteristics of the subjects are shown in Table 1. All the enrolled patients were tested using RDT, microscopy, and PCR.

Table 1

Characteristics of the study patients and prevalence of malaria based on different diagnostic methods

Number of patients sampled	154
Mean age (years)	28.3±23
Sex male (%)/female (%)	53 (34.4)/101 (65.5)
Mean temperature (°C)±SD	38.5±1.20
No sampled at BSHE	49
No sampled at DHN	50
No sampled at CHU	55
No. Positive by microscopy (%)	132 (85.7)
No positive by RDT (%)	106 (68.8)
No positive by PCR (%)	122 (79.2)
MPD by microscopy±SD	5874.32±8120.55

No = Number, SD = Standard deviation, BSHE = Bishop Shannahan Hospital Enugu-Ezike, DHN = District Hospital Nsukka, CHU = Cottage Hospital Ugbuawka, MPD = Mean parasite density

Rapid diagnostic testing

RDT kit First Response®, a histidine-rich protein 2 (HRP2) combo cassette test from Premier Medical Corporation (23 Ziks Avenue Enugu. First Response® kits were supplied by the study group to the three study hospitals for uniformity and reliability of the results. All the laboratory staff was trained on how to use the kit in a 1-day workshop. All the patients were tested with First Response®. Each kit was labeled with the patient's number. The patient's index finger was swabbed and pierced with a sterile lancet provided with the kit. 5 μl of blood was scooped with a loop by the laboratory assistant and added to the sample well in a cassette followed by three drops of sample diluent. Results were read within 20 min. Faint test lines were considered to be positive. Each patient was sampled 3 mL of venous blood into well-labeled ethylenediamine tetraacetic acid (EDTA) tubes and transferred in a cold chain to Safety Molecular Pathology Laboratory (SMPL) within 5 h of sampling where blood film microscopy and PCR testing were performed.

Blood film microscopy

About 20 μl of blood was used in preparing thick smears of blood in a clean slide, which were dried and stained with 5% Giemsa stain for 30 min. Dried slides were viewed at ×1000 with oil immersion by the certified Medical Laboratory Scientists. Parasites were counted against 200 white blood cells (WBCs) from the thick film. Parasite density was estimated assuming total WBC count of 10,000/mL.[20] The pictures of the slides were captured and results were recorded electronically. Slides were stored in a secure slide box and were reconfirmed by another scientist. The laboratory personnel were blinded of the first response results.

Estimation of parasite load by microscopy

We adopted a method that assumed WBC count of 10,000 cells/μl for each subject.[20]

The formula for parasite density per μl of whole blood = no. of parasites counted/WBCs counted* 10,000 cells/μl.

Polymerase chain reaction assay

DNA extraction

Malaria DNA was extracted by saponin hemolysis method[21] from each blood sample on arrival to the SMPL. 1 mL of whole blood was transferred into a 15 mL sterile tube and centrifuged at 4,000 rpm for 5 min. The plasma was discarded and the cell pellets washed thrice in 5 mL of physiological saline. The cells were then resuspended in 2 mL of cold 0.5% saponin solution in phosphate buffer saline (PBS), mixed thoroughly by vortexing until all the cells were lysed, and incubated on ice for 20 min. The sample was centrifuged at 4,000 rpm for 5 min and the supernatant discarded. The pellet was resuspended in 1 mL of ice-cold 0.5% saponin and incubated on ice for 15 min. The tube was further centrifuged at 4,000 rpm for 5 min with the supernatant discarded but cells resuspended in 0.5 mL of PBS for another round of washing. The cell pellet was finally resuspended in 50 μl of Tris-borate-EDTA buffer (AE buffer), vortexed gently for 15 s, and incubated at 99°C for 10 min. The lysate was finally centrifuged at 8,000 rpm for 2 min and the DNA-rich supernatant was transferred to a 1.5 mL tube and stored frozen until required for PCR testing.

Plasmodium falciparum amplification by nested polymerase chain reaction

A set of high-performance liquid chromatography (HPLC) purified primers were used for nested PCR in Applied Biosystem 2730 thermal cyclers (Paisley, United Kingdom). A commercially optimized 2x PCR master mix (Southampton, United Kingdom) was used in setting up a 25 μl reaction volume. The first round PCR used rPLU5 forward primer 5′-CCTGTTGTTGCCTTAAACTTC and rPLU6 backward primer - 5′-TTAAAATTGTTGCAGTTAAAAC primers at final concentrations of 300 nM to amplify 1,200 bp genus (Plasmodium)-specific target of 18S ribosomal RNA gene. The thermal profile for first-round PCR was initial denaturation at 95°C for 10 min, 35 cycles at 94°C for 45 s, 58°C for 30 s, 72°C for 30 s, and final extension at 72°C for 5 min. Second-round PCR reactions were set up using the first-round PCR product and primers R2, rFAL 1 forward primer 5′ TTAAACTGGTTTGGGAAAACC, and rFAL 2 backward primer 5′ACACAATGAACTCAATCATGA specific to P. falciparum for amplifying 205-bp amplicon of the same gene at final concentration of 400 nM in 25 μl reaction volumes. The thermal profile for second round PCR was initial denaturation at 95°C for 10 min, 35 cycles at 94°C for 45 s, 58°C for 30 s, 72°C for 30 s, 58°C for 30 s, and final extension of 72°C for 5 min. For each round, the reaction mixture of 25 μL contained 2.5 μL of 10 x reaction buffers, 5 μL of magnesium chloride 0.75 μL of each primer, 0.2 μL of deoxynucleotide triphosphates (dNTPs), 9.05 μL of water, 0.25 μL of Taq polymerase, and 5 μL of DNA extract. The second-round PCR products were visualized in 2% agarose gel using SYBR Safe or Web Green (Promega, UK and WebScientific, UK) after staining with ethidium bromide. The whole reaction was done in duplicate as a quality control measure. All samples that were negative with RDT but positive with microscopy or negative for microscopy but positive with RDT were reconfirmed with PCR.

Patients’ management

The First response® result was immediately forwarded to the clinician to guide on the treatment decisions. Patients who were symptomatic but with negative first response result and all who tested positive for both first response and microscopy were all treated with artemether-lumefantrine (the current first line treatment for uncomplicated malaria) each patient received a complete regimen of AL treatment in the following order. First dose at 0 h, second dose after 8 h, then 3rd, 4th, 5th and 6th doses were taken 12 hourly for the next 48 h. Patients with body weight above 35, 25-35 and 15-25 kg received 4, 3 and 2 tabs per dose. Each AL tablet contains 20 mg artemether and 120 mg lumefantrine orally on the same day.

Statistical analysis

The overall performances of diagnostic methods were compared using Fisher's exact test (when three methods were compared) or a chi-square test (when two methods were compared). The results for the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were obtained using PCR as a gold standard; a subanalysis was also done comparing RDT to microscopy. P ≤ 0.05 was considered to be statistically significant for all differences within all comparisons. The statistical analyses were made using the Statistical Package for the Social Sciences (SPSS) Version 16.0.

Results

Demographics

A total of 160 patients were recruited; 6 withdrew consent before sampling and 154 patients were sampled. 53 (34.4%) were males while 101 (65.5%) were females. Out of the 154 patients sampled, 80 (51.9%) had fever of ≥37°C with a mean of 38.5°C. The age range was between 6 years and 52 years, the mean age was 28.3 ± 23 years. 31 (20.1%), 45 (29%), 30 (19.4%), and 48 (31.1%) persons were in the age of 6-12 years, 13-25 years, 26-40 years, and above 40 years, respectively. The mean parasite density by microscopy was 5,874.32 ± 8120.55 parasites/μL. The patients’ characteristics and prevalence of malaria based on the diagnostic methods are contained in Table 1.

Overall result of diagnostic techniques

Comparison of First Response® [rapid diagnostic test] with microscopy as the “gold standard”

Out of the 154 patients enrolled, 106 (68.8%) tested positive for both First Response® and microscopy while 25 patients (16.2%) tested negative for both First Response® and microscopy. Samples that were First Response®-positive but microscopy-negative were 0 (0%) while those that were First Response®-negative but microscopy-positive were 23 (14.9%). The analytical performance characteristics: Sensitivity, specificity, PPV, and NPV were 82.17%, 100.00%, 100.00%, and 34.29%, respectively. Test of significance (chi-square test) showed statistically significant difference when First Response® was compared to the microscopy method (P ≤ 0.05). The data are summarized in Table 2.

Table 2

Overall sensitivity, specificity, and predictive value of microscopy and RDT using PCR as the standard and a subanalysis of RDT only using microscopy as the standard

Diagnostic methods	TP (N)	FP (N)	TN (N)	FN (N)	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
Parameters for assessment
TFM	120	12	12	13	92.3	50.00	90.91	54.55
RDT	101	5	15	33	75.37	75.37	92.25	31.32
Standard (PCR)	122	0	32	0	100	100	100	100
Subanalysis of RDT only using microscopy as the standard
RDT	106	0	25	23	82.17	100	100	34.29
Standard (TFM)	132	0	22	0	100	100	100	100

TP = True positive, FP = False positive, TN = True negative, FN = False negative, PPV = Positive predictive value, NPV = Negative predictive value, TFM = Thick film microscopy, N = Number of patients

Comparison of First Response® with polymerase chain reaction as the “gold standard”

101 (65.6%) patients tested positive for both First Response® and PCR while 15(9.7%) patients tested negative for both First Response® and PCR. Samples that were First Response® -positive but PCR-negative were 5 (3.2%) while those that were First Response® -negative but PCR-positive were 33 (21.4%). Sensitivity, specificity, PPV, and NPV were 75.37%, 75.00%, 95.28%, and 31.25%, respectively. Test of significance (chi-square test) showed no statistical significant difference whenFirst Response® was compared to the PCR method (P ≤ 0.05). The data are summarized in Table 2.

Comparison of microscopy with polymerase chain reaction as the “gold standard”

Of the 154 patients, 120 (77.9%) tested positive for P. falciparum for both the microscopy and PCR methods while 12 (7.8%) tested negative in both the methods. Samples that tested positive by microscopy but negative by PCR were 12 (7.8%) while microscopy-negative, PCR-positive samples were 10 (6.5%). The sensitivity, specificity, PPV, and NPV were 92.31%, 50.00%, 90.91%, and 54.55%, respectively. Test of significance (chi-square test) showed that there was no significant difference between microscopy and PCR methods (P ≤ 0.05). The data are summarized in Table 2.

Stratification of parasite density in thick blood smear and correlation with rapid diagnostic test, polymerase chain reaction, and study area

Out of the 20 patients that were negative for microscopy 2 and 6 were positive for RDT and PCR, respectively. The parasite count range of 101-1,000 had the highest positivity (75) and out of the 75 patients, 72 and 73 were positive for RDT and PCR, respectively. With regard to the study area, patients from District Hospital Nsukka contributed to 73.3%, 25.3%, and 0% of the patients with parasite count of 1-100, 101-1000, and >1000 respectively. The contributions of other patients from BSHE and CHU are shown in Table 3.

Table 3

Stratification by parasite density in thick blood smear and correlation with diagnostic methods and study area

Frequency	0	1-100	101-1,000	>1,000
Parasite count
No. of patients observed	22	30	79	23
Mean parasite count/μL (range)	0	70 (25-100)	560 (130-910)	15,520 (1,000-47,500)
No positive for RDT	2 (9.1%)	8 (26.7%)	73 (92.4%)	23 (100.0%)
No negative for RDT	20 (90.9%)	22 (73.3%)	6 (7.6%)	0 (0%)
No positive for PCR	6 (27.3%)	25 (83.3%)	70 (88.6%)	15 (65.2%)
No negative for PCR	16 (72.7%)	5 (16.7%)	9 (11.4%)	8 (34.8%)
No of patients from BSHE	6 (27.3%)	3 (10.0%)	28 (35.4%)	12 (52.2%)
No of patients from DHN	8 (36.7%)	22 (73.3%)	20 (25.3%)	0 (0%)
No of patients from CHU	8 (36.7%)	5 (16.6%)	31 (39.2%)	11 (47.8%)

With regard to the three different hospitals, there was a significant difference in the performance of first response (RDT) and microscopy when compared among the three different hospitals (P < 0.001). In BSHE, DHN, and CHU, the sensitivities of first response (RDT) were 95.3%, 40.0%, and 90.7%, respectively, while the sensitivities of microscopy were 97.0%, 90.9%, and 89.1%, respectively. The specificity, PPV, and NPV of RDT and microscopy are shown in Table 4.

Table 4

Stratification of sensitivity, specificity, and predictive values, and correlation with study areas

Diagnostic method	Sensitivity	Specificity	PPV	NPV
Microscopy
BSHE	97.0	35.0	55.4	75.6
DHN	90.9	44.0	70.6	43.7
CHU	89.1	71.0	47.9	85.5
RDT
BSHE	95.3	79.9	92.7	78.5
DHN	40.0	73.9	85.0	28.6
CHU	90.7	93.7	89.1	80.4

BSHE = Bishop Shannahan Hospita Enugu-Ezikel, DHN = District Hospital Nsukka, CHU = Cottage Hospital Ugbuawka

The cost of testing with RDT, microscopy, and PCR were N = 300, N = 500, and N = 6,000. The costs of each test and turnaround time in SMPL are summarized in Table 5.

Table 5

Cost and turnaround time for RDT, film microscopy, and PCR

Test	Cost (N)	Turnaround time (min)
RDT	300	20
Film microscopy	500	45
PCR	6,000	1,440

Discussion

The new treatment guidelines for malaria in Nigeria relied on the 3Ts (test, treat, and track). The policy is geared toward slowing and/or preventing resistance to the recommended drug, artemisinin-based combination therapy — ACT (artemether-lumefantrine) for treatment of uncomplicated malaria. Hence, each patient must be tested before treatment. The key factors that influence testing include analytical reliability of the test method, cost of test/affordability, accessibility, and turnaround time. Since RDTs may not be very sensitive in detecting malaria, especially in areas with varying transmission intensities as suggested by some studies[2223] and both RDT and microscopy have their limitations in detecting malaria[222425] infections, there is a need to use a more accurate and sensitive diagnostic method such as PCR in assessing the accuracies of these diagnostic methods even though most evaluations of RDTs have used microscopy as the gold standard.[262728] Hence, in this study the reliability of first response and microscopy were assessed using PCR as the gold standard and for a balanced comparison, a subanalysis using microscopy as the gold standard was performed on first response. In this study also, Plasmodium falciparum was preferred because this species causes the most malaria morbidity in Nigeria.

When microscopy was used as the gold standard, the specificity and PPV of First Response® were very high. The specificity (100%) and PPV (100%) was higher than reported elsewhere.[1622] Harani et al. in 2006 reported a similar specificity of 98.3% for P. falciparum using an RDT kit but a lower predictive PPV of 78.0%.[27] The sensitivity and NPV of 82.17% and 34.29% were lower in our study than reported elsewhere.[1628]

When PCR was used as the gold standard, First Response® showed a lower sensitivity (75.37%) and specificity (75.00%) than when compared to microscopy but had a high PPV of 95.28% and an unacceptably low NPV of 31.25%, similar to that obtained when compared to microscopy. On the other hand, microscopy showed a higher sensitivity (92.3%), PPV (90.91%), and NPV (54.55%) than first response but had a lower specificity of 50.00%. The study also revealed that First Response® had 101 (65.6%), 5 (3.2%), 15 (9.7%), and 33 (21.4%) true positive, false positive, true negative, and false negative samples, respectively, against 120 (77.9%), 12 (7.8%), 12 (7.8%), and 10 (6.5%), samples, respectively, by microscopy. First Response® had more false negative samples, i.e., 33 (21.4%) compared to microscopy, i.e., 12 (7.8%) but a lower false positive, i.e., 5 (3.2%) than microscopy, i.e., 12 (7.8%). This suggests that First Response® has a high probability of not accurately detecting falciparum infection in sick patients while microscopy will likely detect falciparum infection in patients who are free from Plasmodium infection. An important implication of this scenario is that it will lead to an erroneous diagnosis of malaria resulting in either denying patients of the appropriate treatment or exposing them to treatment, which could lead to the development of drug resistance. With a true positive of 120 (77.9%), microscopy had a sensitivity of 92.3% and PPV of 90.91%. On the other hand, RDT with a high false negative of 33 (21.4%) and low false positive of 5 (3.2%) had a low NPV of 31.25% and high PPV of 95.28%.

It was observed that at high parasite count of ranges 101-1,000 parasites/μL and >1,000 parasites/μL, RDT detected 73.3% and 100.0% positive cases, respectively, while in low parasite count of ranges 1-100 parasites/μL, RDT detected 26.7% positive cases against 83.3% detected by PCR. This result suggests that there could be a relationship between parasite density and RDT performance. The study also revealed varying parasite densities in the study area, as patients from DHN showed low parasite density. They contributed to 73.3%, 25.3%, and 0% of the patients with parasite count range of 1-100 parasites/μL, 101-1,000 parasites/μL, and >1,000 parasites/μL. RDT performed poorly in this particular area (DHN) with low parasite density and sensitivity and specificity of 40.0% and 73.9%, respectively, while its sensitivity and specificity were higher (95.3% and 79.9%) for BSHE, and higher (90.7% and 93.7%) for CHU. There was poor performance of RDT in patients residing within Nsukka, which is an urban area compared to the patients residing in Ugbuawka and Enugu-Ezike where the major occupation is mainly farming and people live on farms. This study has shown a variation in the performance of RDT with respect to the study area and parasite density. The possible reason for this poor performance by RDT in areas with low parasite density may be its inability to detect Plasmodium at very low densities. It has been shown that the performance of RDT is affected by suboptimal sensitivity at low parasite densities[29] and another study revealed that it has low sensitivity for parasites below 100 parasites/μL and has insufficient accuracy.[6] When the performance of microscopy was assessed with regard to the study area, the difference in its performance was not statistically different.

Both methods showed low specificity and NPV while First Response® also showed low sensitivity. Several factors in the manufacturing process as well as environmental conditions may be responsible for the low performances observed in the use of RDT and microscopy in the diagnosis of malaria. First, the choice of PCR method as the “gold standard” for comparison might influence the outcome of our finding as previous studies used blood film microscopy as their comparator. Second, undue exposure of RDTs to >70% humidity and/or >30°C temperature, especially in the supply chain, may affect the quality of testing while occasional false negative results may be caused by deletion or mutation of the HRP-2 gene in the infecting parasite.[30] Third, it has also been suggested that natural anti-HRP-2 antibodies in some humans may also cause false negative results despite significant parasitemia with RDT.[31] Lastly, the low sensitivity observed in this study may also result from the use of patients >5 years of age. This is supported by a study by Batwala et al. in 2010, which revealed that the sensitivity of RDT was significantly higher than that of other techniques and was excellent in children <5 years of age, i.e., 97.7% [95% confidence interval (CI): 88-99.9] compared to those ≥5 years, i.e., 83.7% (95% CI: 69.3-93.2).[15]

The lower specificity value observed for microscopy could have been caused by subjectivity in the reporting of blood films in addition to interobserver differences in blood film reporting, which are known to influence blood film outcomes. Most importantly, the existence of low density infections and inappropriate use of antimalarials resulting into low parasitemia could affect the performance of microscopy.

The use of Nested PCR as the gold standard instead of microscopy agrees with studies by Andrade et al. in 2010[32] and Batwala et al. in 2010.[15] Andrade et al. in 2010 showed that Nested PCR was the gold standard for diagnosis of both symptomatic and asymptomatic malaria in the Brazilian Amazon because it detected a major number of cases and presented the maximum specificity while microscopy showed a low performance of 65.1% for correct diagnosis.[32] In another study by Coleman et al. 2006, it was observed that although PCR performance appeared poor when compared to microscopy, data indicated that the discrepancy between the two methods resulted from poor performance of microscopy at low parasite densities rather than poor performance of PCR. Hence, the study concluded that PCR appears to be a useful method for detecting Plasmodium parasites during active malaria surveillance in Thailand.[7]

Data from this study highlights the problem of using a less-than-perfect diagnostic test as a reference standard. Microscopic results were initially considered as the reference standards for true positive and true negative results, with all subsequent statistical analyses based on this assumption. Although the PCR method is ultrasensitive, reliable, and amenable to high throughput, it requires a bigger capital outlay to establish as well as a balanced skill mix to operate daily.

The study also revealed a turnaround time of 20 min, 45 min, and 1,440 min for First Response®, microscopy, and PCR, respectively. Cost of carrying out each test per patient was Naira 300, 500, and 6,000 for First Response®, microscopy, and PCR, respectively. For routine malaria diagnosis, PCR method with a turnaround time of 1,440 min and cost of Naira 6,000 may not be a better option. In addition, the urgency and importance of obtaining results quickly for patients with suspected malaria limits the usefulness of PCR in routine clinical practice.

Author's Contribution

EE is the principal investigator. EE sponsored this research financially as part of her Ph. D program. In collaboration with the hospital staff, she recruited and sampled the patients. She provided all data for the research work. EE also participated in molecular testing of the blood samples, data analysis, and wrote the manuscript. EO is the Chief Executive Officer (CEO) of Safety Molecular Pathology Laboratory where all the testing was done. He made a substantial contribution to the conception and design of the work and is the key data analyst. CV is the supervisor of the research work. She contributed in the general coordination of the research work.

Financial support and sponsorship

Nil.

Conflicts of interest

The authors would like to state they there are no conflicts of interest.