Background: Malaria is the most widespread infectious disease in the world, especially in developing countries. The World Health Organization (WHO) recommends that malaria should be diagnosed biologically before antimalarial treatment starting. Several tests are available for its diagnosis, albeit with varying degrees of sensitivity and specificity. The limitations of the thick drop technique and the advent of molecular diagnostic techniques, which have revolutionized therapeutic approaches in several biomedical fields, led us to evaluate the performance of conventional and molecular malaria diagnostic tests at three experimental sites in Côte d'Ivoire. Methodology: We collected blood, saliva and urine samples in Anonkoua-kouté, Port-Bouët, and Ayamé from 93 patients with microscopically confirmed uncomplicated P. falciparum malaria. These patients, aged over 2 years, gave their informed consent before blood, saliva and urine samples were taken. P. falciparum genomic DNA extracted from these samples was amplified by nested PCR using primers specific to certain P. falciparum genes (Pfk13 propeller, pfdhfr and pfcrt genes). Computational statistical analyses requiring various functions and/or scripts of the R software (v 4.1.0) were performed on the data relating to the parasite density of malaria patients subjected to conventional (thick drop and blood smear) and molecular diagnostic procedures. Results: Significant variability in parasite density was observed at all three sites (p<0.05), with a higher parasite density at the Port-Bouët site. The conventional malaria diagnostic system performed moderately well at all three sites and for all patients (AUC=61.4%-62.9%). This performance tended to improve markedly when malaria patients were discriminated by age in young and adult patients (AUC=77.3%-78%), suggesting the susceptibility of the classical malaria diagnostic method to this anthropomorphic parameter. ROC analysis supported a very high performance of the molecular diagnosis of malaria for the biomarkers pfcrt and pfdhfr in the three biological fluids (AUC=100%) in contrast to the molecular biomarker pk13 propeller (AUC=50%). Considering biological fluids, ROC analysis suggested very high performance of malaria molecular diagnostic process for blood sample (AUC=100%) in contrast to saliva (AUC=79.8%) and urine (76.4%) which exhibited moderate performance. Conclusion: Considering as whole, malaria molecular diagnosis, although depending on molecular biomarkers typology as well as biological fluid, performs better than conventional diagnosis, which performances result strongly linked to patient’s age.
Malaria is a tropical disease caused by a protozoan of the genus Plasmodium. The clinical manifestations of malaria are very diverse and resemble flu-like symptoms. It is potentially fatal if not treated quickly and appropriately. Its diagnosis is therefore a medical emergency 1. This diagnosis is confirmed by the detection of Plasmodium in the blood by microscopic examination using the classic techniques of blood smear (FS) and thick drop (GE), which remain the reference methods in terms of sensitivity and specificity 2. These techniques make it possible to confirm the disease, identify the plasmodial species involved and determine and assess parasitaemia, which affects both prognosis and treatment 3. It also enables patients with febrile illnesses to be better managed, thereby avoiding over-treatment 4. However, although microscopic examination remains the most widely used parasitological diagnosis of malaria in most large care facilities and hospitals, the quality of microscopic examination is often insufficient to ensure good sensitivity and specificity of malaria diagnosis, disrupting the therapeutic outcome 3. This is because this examination requires a relatively long reading time and a qualified biologist 2 and its performance is dependent on the parasitaemia of the infected subject. In addition, it is an invasive method that inevitably uses needles, making this classic diagnostic technique stressful for certain groups of people. Studies have shown that despite the use of thick drops as a source of plasmodial DNA, real-time PCR remains more sensitive than microscopy and more accurate for detecting mixed infections 2, 5 The advent of new sequencing techniques has greatly contributed to the development of genomics and revolutionized molecular medicine. In recent years, these events have contributed to the development of new molecular techniques for the detection of Plasmodium in biological fluids. Basing on this, we embarked herein to assess the performances of conventional and molecular malaria diagnostics. Of note, this work aims to evaluate the degree of similarity and/or dissimilarity between malaria classical and molecular diagnosis in different experimental sites and biological fluids i.e. blood, urine and saliva. Performances comparative analysis between malaria classical and molecular diagnosis methods were performed by a multivariate computational statistical approach in R programing environment.
Study sites
This prospective study took place at Anonkoua-Kouté health center as well as at Port-Bouët and Ayamé general hospitals from February to August 2015. These sites are in the south of Côte d'Ivoire, where the climate is dominated by annual rainfall in excess of 1,700 mm, with temperatures varying between 27 and 33°C. Malaria is seasonal, predominating in the rainy season from June to September, with peaks in prevalence and incidence in October-November. Plasmodium falciparum is the dominant species, accounting for over 90% of the parasite formula. The main malaria vectors in this study area, i.e. Southern of Côte d'Ivoire are members of the An. gambiae sl and An. funestus sl complexes 6. Of note, Anonkoua-kouté health center and Ayamé general hospital were selected because of their high annual incidences of malaria. In addition, these health facilities have been considered for several years as the main sites for multicenter clinical efficacy testing by the Malaria Unit of the Institute Pasteur de Côte d'Ivoire. Port Bouët General Hospital was chosen for this study not only because of its consistently high annual incidence of malaria, but also and above all because of its marshy environment used for market gardening.
Study population and sample collection
For the computational multivariate statistical analysis clinical and anthropomorphic data regarding 253 patients clinically suspected of having malaria at Anonkoua-kouté health center (53), Port-Bouët (103) and Ayamé (98) general hospitals were sampled during our study period. After informed consent, blood samples were collected from patients over 2 years of age with an axillary or rectal temperature greater than 37.5°C and microscopically confirmed uncomplicated P. falciparum malaria.
Drawing blood and making confetti
In the laboratory at each of the study sites, 2-5 mL of venous blood from each participant was drawn and collected in a tube containing an anticoagulant (EDTA). Two drops of blood were used to make a thick drop (GE) and a blood smear (FS).
In each patient with microscopically confirmed malaria, 50 μL of whole blood was deposited on Whatman 3 MM filter paper using a micropipette with filter cones. The paper containing the blood spots (blood confetti) was dried for approximately 60 to 120 minutes at room temperature in a dust-free environment.
Collecting saliva and making confetti
Ten to fifteen minutes after rinsing the mouth with tap water, 2 to 5 mL of saliva from each participant was collected in a sterile bottle.
Using a micropipette and filter cones, 50 μL of saliva was deposited on Whatman 3 MM filter paper. The resulting confetti was dried for approximately 60 to 120 minutes (min.) at room temperature and protected from dust.
Collecting urine and making confetti
After blood and saliva collection, 5 to 10 mL of urine from each patient was collected in a sterile bottle. Using a micropipette and filter cones, 50 μL of urine was deposited on Whatman 3 MM filter paper. The resulting confetti was dried for approximately 60 to 120 min at room temperature and protected from dust.
Microscopic examination of parasites
Blood samples were taken and thickened (GE) and smeared (FS). The slides were then stained with 10% Giemsa for 10 minutes, dried and examined under the 100X objective of an oil immersion microscope. Parasite densities (PD), expressed as parasites number per microliter (parasites/µL) in blood sample, were calculated from the thick drop readings. The number of parasites per microliter of blood was established in relation to the number of leukocytes, set at 8,000 per microliter of blood. The parasite density (PD) was therefore obtained using the formula:
N = number of parasites counted and X = number of leukocytes counted. If the number of parasites N is≥we count at least X = 200 leukocytes and the parasite density will be : DP = N x 8000 / 200 = N x 40. If the number of parasites N is < 10, at least X = 500 leukocytes are counted and the parasite density will be: DP = N x 8000 / 500 = N x 16.
The FS was mainly used to identify Plasmodium species.
Quality control of parasite densities was carried out by a second microscopist (a technician from the site laboratory) on 10% of the slides examined and selected at random to detect any discrepancies between the results. In the event of a discrepancy, a third microscopist recounted the trophozoites.
Molecular biology techniques
The filter paper spots were cut into thin layers under sterile conditions to extract genomic DNA from the confetti.
Extraction of plasmodial DNA from blood confetti
Plasmodium DNA was extracted with methanol from blood confetti 7. Briefly, thin cuts of blood confetti were immersed in 1 mL of wash buffer (950 µL of 1X PBS plus 50 µL of10% saponin) and then incubated at 4°C overnight. The wash buffer was removed and then washed before adding 150 µL of methanol. After a 20 min incubation, the methanol was gently removed and the samples were dried at room temperature for 2 hours before adding 300 µL of sterile water. The samples were then heated to 99°C in a thermo-mixer for 30 min to elute the DNA. After removing the confetti debris, we aliquoted extracted DNA into a 1.5 mL Eppendorf tube and stored at -20°C.
Extraction of plasmodial DNA from saliva and urine confetti
Extraction of plasmodial DNA from urine and saliva confetti was performed using the Chelex®100 method 8; 9. Briefly, 180 µL of 5% (w/v) Chelex-100 solution (Bio-Rad, catalogue no. 1422832) was placed in a 1.5 mL centrifuge tube and heated at 100°C for 5 min. We added the fine cuttings from each confetti to the hot Chelex solution. The closed tube was gently vortexed for 30 sec and then reheated for 10 min. The tubes containing the samples were then centrifuged at 12,000 x g for 1.5 min, and the supernatant collected in a new micro-centrifuge tube (1.5 mL). We centrifuged this supernatant again at 12,000 g for 1.5 min. The supernatant from this second centrifugation was gently collected in a new micro-centrifuge tube and used immediately in the PCR amplification reaction or stored at -20°C. P. falciparum genomic DNA extracted from these samples was amplified by nested PCR using primers specific to certain Plasmodium falciparum genes (Pfk13 propeller, pfdhfr and pfcrt genes). Computational statistical analyses requiring various functions and/or scripts in R software (v 4.1.0) were carried out on data relating to the parasite density of malaria subjects subjected to conventional (thick drop and blood smear) and molecular diagnostic procedures. Similarly, data relating to the age, gender, temperature and weight of the malaria patients in the study sample were reported and taken into account for the statistical study.
Molecular analysis was carried out on three biological fluids (blood, saliva and urine) in which mutations in three P. falciparum molecular markers were assessed, observed and quantified, namely the pfdhfr, Pfcrt and pfk13 propeller genes in the malaria populations analyzed.
Statistical analysis
The computational statistical analyses required various functions and/or scripts in the R software (v 4.1.0).The results of the descriptive statistical analysis concerned the calculation of certain positional parameters (mean, median and absolute frequency) and dispersion parameters (standard deviations and coefficients of variability).
The analytical statistical analyses involved several packages and scripts from the R software (version 4.1.0). The R packages used to visualize the results of the statistical analysis were: (i) readxl to connect Excel with the R software, (ii) Rcmdr, a graphical interface that facilitates interactivity with the R software, (iii) ggpubr for elegant visualization of the data in R, (iv) multcompView for visualization of paired comparisons, (v) ggplot2 for producing more sophisticated graphs with a more modern design, (vi) pROC for displaying and producing ROC curves.
Several analytical statistical tests were carried out; (i) Shapiro's normality test, which was used to verify the normality of the distribution of parasite density in the malaria population at the three sites examined, (ii) the ANOVA test, which was used for the analysis of variance, (iii) the Bartlett or homogeneity of variance test and the Fligner-Killeen test, which were used to assess the statistical significance of the homogeneity of variance with respect to the parasite distribution in the malaria populations, (iv) the Tukey statistical test, which was used to make a multiple comparison of the means of the parasite density of the malaria populations at the different experimental sites, (v) the Pearson correlation test, which was used to assess the association between parasite density, temperature, age and weight of the malaria populations.
A statistical difference and/or association was considered significant for p<0.05.An ROC analysis was performed to show the performance of a classification model at all classification thresholds. For this analysis, we set the pathogenic parasitaemia threshold for the three sites studied in children (0-12 years) at 3,000 trophozoites per mm3 of blood and 1,000 trophozoites per mm3 of blood in adults (13 years and over).
We performed transformation and/or normalization of parasite density and as well, patients’ anthropomorphic data according to 11 with the purpose to develop regression a linear model linking parasite density to malaria population anthropomorphic parameters.
Descriptive analysis of Anonkoua-kouté site malaria population
In Anonkoua-kouté, 52 patients were included in the study, 21 of them male (40.39%) and 30 female (57.69%). Patient’s age ranged from two (2) to 53 years, with a mean age of 15 ± 13 years. Of note, male and female malaria populations mean age was estimated to 11 ± 8 and 17 ± 15 years respectively (Table 1). Patients’ weight ranged from7 to 83 kg (Table 2). Male and female malaria patients mean weight was estimated at 32.95 ± 18.60 kg and 39.89 ± 25.27 kg respectively (Table 1).
Malaria patients in Anonkoua-kouté site exhibited a parasite density ranging from 2200 to 110,000 parasites/µL, with an estimated mean of 24,751.923 parasites/µL, a median of 16,000 parasites/µL and a coefficient of variability (CV) of 1.06.
Descriptive statistical approach in characterizing Ayamé experimental site malaria population
The 98 malaria patients identified at this site comprised 60 women (61.22%) and 38 men (38.78%). Age of malaria patients in Ayamé experimental site ranged from two (2) to 76 years, with a mean age of 16 ± 15 years. Mean ages for the male and female malaria populations was estimated to 12 years (CV= 1.08) and 18 years (CV= 0.89) respectively (Table 2). For this experimental site, malaria patients weight range from nine (9) to 100 kg (Table 2). Average weight regarding male and female malaria patients was estimated to 30.45 ± 23.48 and 42.1 ± 24.37 kg respectively (Table 2).
Parasite density ranged from 600 to 20,000 parasites/µL, with an average of 30,798.367 parasites/µL, a median of 11,500 parasites/µL and a coefficient of variability of 1.62.
Descriptive survey of Port-Bouët experimental site malaria population
A total of 103 malaria patients were recorded at Port-Bouët site, 47 (45.63%) of whom were male and 54 (54.37%) females. The age of the patients ranged from tow (2) to 62 years, with a mean age value of 18 ± 14 years. Of note, male and female malaria populations mean age value was estimated to 19 ± 16 and 18 ± 13 years respectively (Table 3). Average weight of male malaria patients was estimated to 42.13 kg ± 24.11 kg, while those of female malaria population was estimated to 42.03 ± 23.46 kg (Table 3). Parasite density in malaria population ranging from 853-100,000 parasites/µL with an estimated average of 8175.466 parasites/µL, a median of 3200 parasites/µL and a coefficient of variability (CV) of 1.56.
Assessment of parasite density distribution in malaria patients at Anonkoua-Kouté
Variance analysis (ANOVA) by discriminating malaria population by sex revealed non-significant difference in variance in terms of parasite density distribution among patients in Anonkoua-kouté (p=0.14) (Figures 2A and B). The same test (ANOVA test) by comparing young (age raking from 0 to 12 years) and adult (age over 13 years) malaria patients population showed a non-significant variance difference in terms of parasite density distribution (p= 0.62) (Figures 2D and E). Furthermore, variance homogeneity test suggested homogeneous distribution of parasite density in malaria male and female populations as well as in young and adult malaria patients in Anonkoua-kouté (p>0.05) (Figures 2C and F respectively). In addition, Shapiro's normality test showed an abnormal distribution of parasite density in Anonkoua-Kouté experimental site (p<0.05) (Figure 5A).
Evaluation of parasite density distribution in malaria patients’ population at Ayamé
Herein, we analyzed the distribution homogeneity of parasite density in Ayamé experimental site basing on the following anthropomorphic parameters i.e. patients gender and patients age. Considering gender parameter, ANOVA test revealed a non-significant variance difference in terms of parasite density distribution between male and female malaria populations (p=0.16) (Figures 3A and B). However, by considering age anthropomorphic parameter (age between 0 and 12 years, and age over 13 years), ANOVA test revealed a significant variance difference in assessing parasite density distribution between young and adult malaria populations (p=0.03) (Figures 3D and E). In addition, the comparative statistical analysis suggested young malaria patients as exhibiting higher parasite density rate than adults one. Analysis of variance homogeneity by accessing parasite density distribution in male and female populations as well as in young and adult malaria patients, suggested a significant difference of variance (variance heterogeneity) in terms of parasite density distribution in Ayamé experimental site (p<0.05) (Figures 3C and F). Shapiro's normality test revealed an abnormal distribution of that parasite density in Ayamé site (p<0.05) (Figure 5B).
Assessment of parasite density distribution in malaria population in Port-Bouët experimental site
We assessed parasite density distribution in Port-Bouët experimental site basing on malaria patients anthropomorphic features i.e. gender and age. (Figure 4). The analysis of variance and/or ANOVA test, for both (i) age and (ii) gender parameters revealed a non-significant variance difference in terms of parasite density distribution with p= 0.11 and p=0.08 respectively (Figures4A, B, DandE). Bartlett’s test (variance homogeneity test) suggested variance heterogeneity in the distribution of parasite density (p<0.05) for gender and age anthropomorphic parameters in Port-Bouët experimental site (Figures 4C and F). In addition, for the same experimental site, Shapiro's normality test supported an abnormal distribution parasite density in the above-mentioned Port-Bouët experimental site (p<0.05) (Figure 5C).
Comparative analysis of parasite density distribution in malaria patients in Anonkoua-kouté, Ayamé and Port-Bouët sites
Analysis of variance (ANOVA) revealed a significant variance difference between malaria patients in the three study sites (p<0.001). Tukey multi-comparative test by comparing parasite density averages showed a significant difference between Port-Bouët and Anonkoua-kouté experimental sites (p<0.001) as well as between Port-Bouët and Ayamé sites (p<0.001) (Figure 6). The same analysis suggested a non-significant variance difference in parasite density by comparing Ayamé and Anonkoua-kouté experimental sites (p=0.25) (Figure 6). In other words, the parasite density parameter was more marked for patients in Port-Bouët (p<0.05) than in Anonkoua-Kouté and Ayamé (p>0.05) (Figure 6). Analysis of the homogeneity of the variance in parasite density among patients at the three experimental sites showed heterogeneity in the variance of parasite density (p<0.001). The Fligner-Killeen test also showed a significant difference in variance (p<0.001) in parasite density among patients in Anonkoua-kouté, Ayamé and Port-Bouët.
ROC analysis assessing the performance of malaria classical diagnosis procedure in Anonkoua-Kouté, Ayamé and Port-Bouët experimental sites
For the present ROC analysis, we fixed different parasite density threshold for young (age raking from 0 to 12 years) and adult (patients with age over 12 years) malaria patients. For malaria young population in Anonkoua-Kouté experimental site, ROC curve analysis recorded 27 malaria patients as true positives against two (2) false positives. For malaria adult population in the same site, 22 malaria patients were recognized as true positive against zero (0) false positives (Table 4). ROC curve analysis performed in Ayamé experimental site for young malaria population, acknowledged 43 malaria patients as true positive against 11 false positives. The same ROC curve analysis in the same site for malaria adult population exhibited 41 malaria patients as true positives against three (3) false positives (Table 4).
Port-Bouët experimental site exhibited 29 true positives malaria cases against 18 false positives for young malaria population, while55 malaria patients were recognized as true positives against one (1) false positive in the adult malaria population (Table 4). Interestingly ROC curve analysis in the three (3) experimental sites disregarding malaria patients’ age anthropomorphic parameter revealed weak area under curve (AUC) coefficients (61.4%-62.9%) (Figure 7A), suggesting weak performance of malaria classical diagnosis procedure. Of note, malaria classical diagnosis method exhibited a moderate performance when, malaria patients were categorized in young and adult population (Figure 7B). Considering as a whole, age parameter is a limiting for malaria classical diagnostic performance in terms of fixing parasite density threshold in declaring patient malaria status.
Correlation between parasite density, temperature, age and weight in malaria patients’ population
Pearson correlation analysis between patients' anthropomorphic parameters and parasite density revealed a relative significant negative correlation (r2 ≈-0, 2) (p: 0.04-0.1) between patients ‘age parameter and parasite density across the three experimental sites (Supplementary Tables 1, 2 and 3). Pearson's correlation test revealed a positive and significant correlation between patient age and weight across the three experimental sites (r2 > 85, p≤0.05) (Supplementary Tables 1, 2 and 3).
The same analysis also showed a significantly negative correlation between age and/or weight and temperature at Ayamé and Port-Bouët sites by contrast to Anonkoua-Kouté site (Supplementary Tables 1, 2 and 3). Similarly, the Pearson correlation test revealed a positive and significant correlation between temperature and parasite density at the Anonkoua-Kouté site (p≤0.05), unlike to Ayamé and Port-Bouët sites (p>0.1). Considering as a whole, this analysis supported the consistency and relative statistical significance (0.04< p <0.1) of the correlation coefficient between age and parasite density on the one hand and weight and age on the other (p<0.05), across the three experimental sites (Supplementary Tables 1, 2 and 3).
Assessment of mathematical relationship between parasite density and malaria patients’ age parameter
Because we previously showed a relatively significant and constant negative correlation between patient age and parasite density, we embarked here in developing a linear regression model between these two parameters. For that bio-mathematical model, we assumed patients parasite density parameter (PD) as the response variable, while patients’ age anthropomorphic parameter referred to the explanatory variable. Mathematical function liking parasite density and patient age anthropomorphic parameter is as follows: Parasite Density (DP) = -0.2Age + 0.2 (PD = -0.2Age +0.2). Mathematical model estimators’ coefficients exhibited strong statistical significance performance in predicting parasite density by patients age (p<<0.05). Of note, statistical significance of above-mentioned estimators, -0.2 and 0.2 of linear regression model are estimated respectively to p= 0.001 and p=2e-16. In other words, these estimators suggested that linear regression model as an excellent predictive function linking parasite density to patients’ age. Fisher test associated to this linear regression model showed significant variance difference between parasite density (model response variable) and patients age (explanatory variable) (F=10.63 DF=248, p=0.001) (Figure 8). The standard error of the model is estimated to 0.2 for a coefficient of determination R2 =0.2. In summary, the linear regression statistical model suggested significant statistical link between parasite density and patients’ age in all analyzed experimental sites.
ROC curve analysis in evaluating statistical performances of malaria molecular diagnostic process
Malaria molecular analysis was performed on three biological fluids i.e. blood, saliva and urine in which three molecular markers of Plasmodium falciparum resistance to antimalarial drugs were evaluated (Table 5). We measured effectiveness this technique in terms of success and failure of sequencing of the amplification products. In the case of blood, 822 DNA, fragments (86.98%) were successfully sequenced, compared with 123 (13.02%) cases of failure.
The same analysis in saliva reported 481 (74.23%) DNA fragments successfully sequenced compared with 167 (25.77%) cases of failure. In urine, 407 (74.54%) DNA fragments were successfully sequenced against 139 (25.46%) failures. ROC curve analysis based on malaria biomarker availability in analyzed biological fluids exhibited AUC coefficient values as following 100%, 79.8% and 76.4% for blood, saliva and urine respectively (Figure 9A). In addition, ROC curve analysis assessing classical malaria diagnosis process performance in blood by discriminating malaria patients in young and adult, in the three experimental sites exhibited an AUC value =92.7%, suggesting high performance of malaria molecular diagnosis process (AUC=100%) in comparison to conventional diagnosis technique when assumed blood as reference biological fluid (Figure 9A and 10).
Furthermore, the same analysis (ROC) revealed saliva as a potential alternative to blood in malaria molecular diagnosis process, exhibiting an AUC ≈ 80% higher than those of urine (AUC=76.4%) (Figure 9A). Pfcrt and pfdhfr molecular markers were discriminated with high sensitivity and excellent accuracy (high performance, AUC=100%) in the three biological fluid for all analyzed patients from the three experimental sites by contrast to Pfk13 propeller biomarker that exhibited bad performance (AUC =50%) in malaria molecular diagnosis procedure (Figure 9B). Taking together, ROC analysis suggested very high performance of malaria molecular diagnostic technique based on the discrimination of mutations in pfcrt and pfdhfr molecular markers in blood, by contrast to the conventional technique based on blood smears and parasite density.
Patients’ age average was estimated to 15 ± 13 years, 16 ± 15 years and 18 ± 14 years respectively in Anonkoua-Kouté, Ayamé and Port-Bouët, suggesting a relatively young study population. Patients’ statistical descriptive analysis revealed several cases of malaria in children. This result can be explained by the fact that children are more likely to be expose to uncontrolled waste dumps and sewage. These results are in line with those of the 12, which suggested the susceptibility and/or vulnerability of children to malaria parasite infection.
Of note, more than 50% of analyzed malaria population are female in the three-process experimental This result is in agreement with those of 13. In fact, the trend towards feminization of malaria in the study sites could be because of women are more exposed to mosquito bites because of their household and secular activities 14. It could also because women mainly in Cote d’Ivoire visit health centers more than men, especially for pregnancy monitoring 15 as well as for health check reasons 13, 16.
Statistical analysis showed significant variability in parasite density across the three experimental sites. The same analyses showed parasite density distribution heterogeneity, suggesting classical malaria diagnostic technique versatility, since this diagnostic method based on quantifying parasite density in blood (GE and FS). Indeed, parasite density variability in the three analyzed experimental sites could be explained by the fact that some sites contain more larval breeding sites than others. In addition, the diversity of parasite density in malaria population could because of population exposition as well as preventing techniques regarding mosquito bites that differ from one site to another.
Furthermore, ROC analysis showed that the blood smear and/or thick drop technique for the classical malaria diagnosis process performs poorly in terms of sensitivity and accuracy with respect to parasite density. Of note, performance of classical malaria diagnosis process increase when patients are clustered in (i) young and (ii) adult depending on different threshold of parasite density. These results confirm the versatility of malaria classical diagnostic technique based solely on parasitemia, as previously mentioned. In addition, several studies shown that the performance of malaria diagnostic technique based on parasite density, in terms of sensitivity and reliability, depends on microscopist experience as well as to parasitemia level of infected subject 2. These results are in line with those published by 3.
The Tukey test supported a low rate of parasite density among patients in Port-Bouët, by contrast Anonkoua-Kouté and Ayamé experimental sites. It is noteworthy to underline that, Port-Bouët experimental site exhibited aging malaria population in comparison to the other two sites i.e. Ayamé and Anonkoua-Kouté. These results suggest parasite density decreasing with aging. Of note, Pearson correlation analysis confirmed this evidence by exhibiting significant negative correlation between parasite density and patients’ age. As previously suggested, the performance of malaria classical diagnosis procedure increase by introducing parasite density threshold basing on patients age. Taking together, we clearly shown the relationship between malaria classical diagnosis process and patient aging. In other words, parasite density in malaria patients displayed a significant negative correlation with age anthropomorphic parameter. We estimated mathematical relationship between parasite density and patients’ age by a linear regression model suggesting aging as an excellent predictor of malaria patients’ parasite density. Indeed, 17 showed in their studies on the threshold value of parasite density in malaria diagnosis that parasite density decreased with patient age increasing. 18 also showed that the P. plasmodium parasite density threshold, likely to cause a febrile attack decreased significantly with age. Of note, ROC curve analysis showed a high performance of the malaria molecular diagnostic technique in contrast to the classical diagnostic technique based on the thick drop and/or blood smear at all the experimental sites. Indeed, performance of malaria molecular diagnosis procedure is age free as opposite to malaria classical diagnosis. The malaria molecular diagnostic process, although having given excellent results with blood, suggested saliva and urine respectively as potential alternative biological fluids for molecular analysis. Previous studies shown traces of Plasmodium DNA in the saliva and urine of malaria patients [19-21] 19. In addition, studies conducted by 10 suggested urine and salivary DNA extracts as alternative sampling to blood in malaria molecular diagnosis procedure. However, the same studies showed a good performance of blood and saliva respectively for malaria molecular diagnosis compared to urine. These results are similar to those of 22 who observed that molecular detection of plasmodial DNA in malaria subjects in urine was less sensitive than in saliva. This would be related to the small amount of DNA template present in urine compared to saliva and blood 10, 20 showed that parasitic DNA average in blood was respectively 600 and 2500 times higher than in saliva and urine samples 20. Studies have also shown that saliva, with or without blood contamination, is more effective than urine samples. Consequently, these non-blood samples have the potential to be used as non-invasive samples for the molecular diagnosis of malaria 23. Our study exhibited very high performance of malaria molecular diagnosis process on the discrimination of pfcrt and pfdhfr molecular biomarkers mutation in blood sample of malaria patients from all three experimental sites by opposite to pfk13 biomarker. 10 showed pfdhfr and pfcrt genes mutation detected in blood as satisfactory biomarkers for characterizing a malaria population. Taking together, the present analysis confirmed pfcrt and pfdhfr molecular biomarkers mutation in blood as the best molecular parameter for predicting malaria infections. Considering as a whole, malaria molecular diagnostic technique performances would therefore depend on the typology of the biological sample as well as to molecular biomarkers used. Is it noteworthy to underline that malaria molecular diagnosis has been executed on patient with high level of parasite density (true positive patients). Interestingly, rigorous selection of malaria patients’ contribute in improving malaria molecular as well as malaria conventional diagnosis performances confirming the versatility of this diagnostic technique.
Malaria diagnosis is a crucial process for patients managing. The reliability of the thick smear/blood smear technique for malaria diagnosis has been the subject of much debate in the biomedical community. This study revealed versatility of the performance of malaria conventional diagnostic processes that depend on patient age, by contrast to malaria molecular diagnosis approach that exhibited high performance in revealing pfdhfr and pfcrt genes mutation in malaria patients’ blood and saliva samples as opposite to pfk13 biomarker in the same biological fluids as well as in urine.
OD, EAER and NDD wrote the paper. NDD and EAER performed statistical analysis. OD and NDD proposed the work protocol as well as organized paper tables and figures. NDD supervised computational statistical analysis. All Authors read and approved paper final version.
The study was conducted in accordance with the Declaration of Helsinki and approval was received from the National Ethics and Research Committee (CNER) of the Ministry of Health in Côte d'Ivoire. After appropriate information and explanations, the adult participants and the parents or legal guardians of all children wishing to participate in the study gave their written consent prior to sampling.
The authors report no conflicts of interest in this work.
The authors express their deep gratitude to Professor Mireille DOSSO (Director of the Pasteur Institute of Côte d'Ivoire) who allowed them to use the equipment of the molecular biology platform of the Pasteur Institute of Côte d'Ivoire to perform the PCR tests. The authors would also like to thank the staff at the study sites (Anonkoua-kouté, Port-Bouët and Ayamé) for their efforts and cooperation in recruiting patients and collecting samples.
Supplementary Table 1. Pearson correlation test between patients anthropomorphic parameters i.e. age, weight, temperature, and parasite density in Anonkoua-Kouté experimental site.
|
Supplementary Table 2. Pearson correlation test between patients anthropomorphic parameters i.e. age, weight, temperature, and parasite density in Ayamé experimental site.
|
Supplementary Table 3. Pearson correlation test between patients anthropomorphic parameters i.e. age, weight, temperature, and parasite density in Port-Bouet experimental site.
|
| [1] | SPILF. Management and prevention of imported malaria. Available at: https://www.infectiologie.com/UserFiles/File/spilf/recos/2017. Accessed 26 April 2021. | ||
| In article | |||
| [2] | Siala E., Abdallah R.B., Bouratbine A., Aoun K. Actualités du diagnostic biologique du paludisme, Revue Tunisienne d'Infectiologie, 2010, 4: 5-9. | ||
| In article | |||
| [3] | WHO. The microscopic examination. Available at https://www.who.int/malaria/areas/diagnosis/microscopy/fr/. Accessed on 05 May 2018. | ||
| In article | |||
| [4] | WHO. Diagnosis of malaria. Available at https://www.who.int/malaria/areas/diagnosis/ en/. Accessed on 07 May 2021. | ||
| In article | |||
| [5] | Mvumbi DM, Boreux R, Mvumbi GL, Bobanga TL.. Situakibanza H N T. Melin P. Kayembe JMN, DeMol P, Hayete MP. Diagnostic du paludisme sévère : comparison de la technique de PCR en temps réel versus microscopie, à Kinshasa, République Démocratique du Congo. Annales Africaines de Medecine, Vol. 8, °N 3, Juin 2015. | ||
| In article | |||
| [6] | Adja A.M., N'goran E.K., Koudou B.G., Dia I., Kengne P., Fontenille D., Chandre F. Contribution of Anopheles funestus, An. gambiae and An. nili (Diptera: Culicidae) to the perennial malaria transmission in the southern and western forest areas of Côte d'Ivoire. Annals of Tropical Medicine and Parasitology, 2011, 105, 13-24. | ||
| In article | View Article PubMed | ||
| [7] | Miguel R.H.; Coura J.R., Samudio F., Suárez-mutis M.C. Evaluation of three different DNA extraction methods from blood samples collected in dried filter paper in Plasmodium subpatent infections from the Amazon region in Brazil. Revista do Instituto de Medicina Tropical de São Paulo, 2013, 55(3): 205-8. | ||
| In article | View Article PubMed | ||
| [8] | Kain KC, Lanar DE. Determination of genetic variation within Plasmodium falciparum by using enzymatically amplified DNA from filter paper disks impregnated with whole blood. Journal of Clinical Microbiology, 1991, 29: 1171-1174. | ||
| In article | View Article PubMed | ||
| [9] | Plowe CV, Wellems TE. Molecular approaches to the spreading problem of drug resistant malaria. Advance in Experimental Medicine Biology, 1995, 390: 197-209. | ||
| In article | View Article PubMed | ||
| [10] | Olefongo D., Ako A.B., Dago D.N., Coulibaly B., Ngazoa-Kakou S., Toure A.O., Djaman J.A. Comparative analysis of genomic DNA amplification yield for Plasmodium falciparum extracted from urine, saliva and blood. Journal of Parasitology and Vector Biology, 2017, 9 (7): pp. 95-105. | ||
| In article | |||
| [11] | Dago D. N., Kablan G. A. J., Alui Konan A., Lallié Hermann D., Dagnogo D., et al. (2021). Normality Assessment of Several Quantitative Data Transformation Procedures. Biostatistics and Biometrics Open Access Journal, 2021, 10(3): 555786. | ||
| In article | View Article | ||
| [12] | NMCP (2018). National guide to the biological diagnosis of malaria. Available at: https://www.pnlp.sn/wp-content/uploads/2018/02. Accessed 24 April 2021. | ||
| In article | |||
| [13] | Dago D.N., Diarrassouba N., Touré A., Lallié H.D., N'Goran K.E., Ouattara H., Kouadio J., Coulibaly A. Whole screening analysis discerning recurrently diagnosed diseases in a Northern locality of Côte d'Ivoire. International Journal of Development Research; 2017, 7 (11): 16598-16604. | ||
| In article | |||
| [14] | Yalamoussa, T., Noel, D.D., Michel, L. Y. and Hervé, K.K. (2017). Screening of phytosanitary practices in vegetable growth activities northern of Côte d’Ivoire. International Journal of Recent Scientific Research, 2017, 8(6), pp. 17396-17402. | ||
| In article | |||
| [15] | Guehi Z.E., Dago D.N., Tehoua L., Yangni A.K.H. Assessment of the obstetrical and anthropometric parameters of women in labor and their newborns at a tribal health center in northern of Cote d’Ivoire. International Journal of Current Research, 2019, 11, (01), 858-864. | ||
| In article | |||
| [16] | Noel, D. D., Olefongo, D., Lazare, T., Wagniman, S., Koffi, N. G. B. S., Joel, K. K., Florent, K. A., Eboulé, A. E. R., Guehi, Z. E., and Yangni-Angaté, K. H. Assessment of recurrently diagnosed diseases dynamism at Korhogo General Hospital in Northern Côte d’Ivoire from 2014 to 2018. International Journal of Medicine and Medical Sciences, 2022, 14(1), 1-19. | ||
| In article | View Article | ||
| [17] | ChippauxJ.P., Akogbeto M.,MassougbodjiA.,Adjagba J. Mesure de la parasitémie palustre et évaluation du seuil pathogène en région de forte transmission permanente. Cotonou, Benin, 1991, 11p. | ||
| In article | |||
| [18] | Richard A., Lallemant M., Trape J.F., Carnevale P. and Mouchet J. Le paludisme dans la région forestière du Mayombe, République populaire du Congo. II. Parasitological observations. Annales de la Société Belge de Médecine Tropicale, 1998, 68, 4: 305-316. | ||
| In article | |||
| [19] | Mharakurwa S., Simoloka C., Thuma P.E., Shiff C.J., Sullivan D.J. PCR detection of Plasmodium falciparum in human urine and saliva samples. Malaria Journal, 2006, 5: 103. | ||
| In article | View Article PubMed | ||
| [20] | Nwakanma D.C., Gomez-Escobar N., Walther M., Crozier S., Dubovsky F., Malkin E., Locke E., Conway D.J. Quantitative detection of Plasmodium falciparum DNA in saliva, blood and urine. JournalofInfectiousDiseases, 2009, 199(11): 1567-74. | ||
| In article | View Article PubMed | ||
| [21] | Putaporntip C., Buppan P., Jongwutiwes S. Improved performance with saliva and urine as alternative DNA sources for malaria diagnosis by mitochondrial DNA based PCR assays. Clinical Microbiology and Infection, 2011, 17: 1484-91. | ||
| In article | View Article PubMed | ||
| [22] | Kwannan Nantavisai Malaria detection using non-blood samples. Songklanakarin Journal of Science and Technology, 2014, 634 36 (6): 633-641. | ||
| In article | |||
| [23] | Aninagyei E., Abraham J., Atiiga P., Antwi S.D., Bamfo S., Acheampong D.O. Evaluating the potential of using urine and saliva specimens for malaria diagnosis in suspected patients in Ghana. Malaria Journal, 2020, 19 (1): 349. | ||
| In article | View Article PubMed | ||
Published with license by Science and Education Publishing, Copyright © 2023 Dagnogo Oléfongo, Dago Dougba Noel, Eboulé Ago Eliane Rebecca, Koffi N'Guessan Bénédicte Sonia and Djaman Allico Joseph
This work is licensed under a Creative Commons Attribution 4.0 International License. To view a copy of this license, visit
https://creativecommons.org/licenses/by/4.0/
| [1] | SPILF. Management and prevention of imported malaria. Available at: https://www.infectiologie.com/UserFiles/File/spilf/recos/2017. Accessed 26 April 2021. | ||
| In article | |||
| [2] | Siala E., Abdallah R.B., Bouratbine A., Aoun K. Actualités du diagnostic biologique du paludisme, Revue Tunisienne d'Infectiologie, 2010, 4: 5-9. | ||
| In article | |||
| [3] | WHO. The microscopic examination. Available at https://www.who.int/malaria/areas/diagnosis/microscopy/fr/. Accessed on 05 May 2018. | ||
| In article | |||
| [4] | WHO. Diagnosis of malaria. Available at https://www.who.int/malaria/areas/diagnosis/ en/. Accessed on 07 May 2021. | ||
| In article | |||
| [5] | Mvumbi DM, Boreux R, Mvumbi GL, Bobanga TL.. Situakibanza H N T. Melin P. Kayembe JMN, DeMol P, Hayete MP. Diagnostic du paludisme sévère : comparison de la technique de PCR en temps réel versus microscopie, à Kinshasa, République Démocratique du Congo. Annales Africaines de Medecine, Vol. 8, °N 3, Juin 2015. | ||
| In article | |||
| [6] | Adja A.M., N'goran E.K., Koudou B.G., Dia I., Kengne P., Fontenille D., Chandre F. Contribution of Anopheles funestus, An. gambiae and An. nili (Diptera: Culicidae) to the perennial malaria transmission in the southern and western forest areas of Côte d'Ivoire. Annals of Tropical Medicine and Parasitology, 2011, 105, 13-24. | ||
| In article | View Article PubMed | ||
| [7] | Miguel R.H.; Coura J.R., Samudio F., Suárez-mutis M.C. Evaluation of three different DNA extraction methods from blood samples collected in dried filter paper in Plasmodium subpatent infections from the Amazon region in Brazil. Revista do Instituto de Medicina Tropical de São Paulo, 2013, 55(3): 205-8. | ||
| In article | View Article PubMed | ||
| [8] | Kain KC, Lanar DE. Determination of genetic variation within Plasmodium falciparum by using enzymatically amplified DNA from filter paper disks impregnated with whole blood. Journal of Clinical Microbiology, 1991, 29: 1171-1174. | ||
| In article | View Article PubMed | ||
| [9] | Plowe CV, Wellems TE. Molecular approaches to the spreading problem of drug resistant malaria. Advance in Experimental Medicine Biology, 1995, 390: 197-209. | ||
| In article | View Article PubMed | ||
| [10] | Olefongo D., Ako A.B., Dago D.N., Coulibaly B., Ngazoa-Kakou S., Toure A.O., Djaman J.A. Comparative analysis of genomic DNA amplification yield for Plasmodium falciparum extracted from urine, saliva and blood. Journal of Parasitology and Vector Biology, 2017, 9 (7): pp. 95-105. | ||
| In article | |||
| [11] | Dago D. N., Kablan G. A. J., Alui Konan A., Lallié Hermann D., Dagnogo D., et al. (2021). Normality Assessment of Several Quantitative Data Transformation Procedures. Biostatistics and Biometrics Open Access Journal, 2021, 10(3): 555786. | ||
| In article | View Article | ||
| [12] | NMCP (2018). National guide to the biological diagnosis of malaria. Available at: https://www.pnlp.sn/wp-content/uploads/2018/02. Accessed 24 April 2021. | ||
| In article | |||
| [13] | Dago D.N., Diarrassouba N., Touré A., Lallié H.D., N'Goran K.E., Ouattara H., Kouadio J., Coulibaly A. Whole screening analysis discerning recurrently diagnosed diseases in a Northern locality of Côte d'Ivoire. International Journal of Development Research; 2017, 7 (11): 16598-16604. | ||
| In article | |||
| [14] | Yalamoussa, T., Noel, D.D., Michel, L. Y. and Hervé, K.K. (2017). Screening of phytosanitary practices in vegetable growth activities northern of Côte d’Ivoire. International Journal of Recent Scientific Research, 2017, 8(6), pp. 17396-17402. | ||
| In article | |||
| [15] | Guehi Z.E., Dago D.N., Tehoua L., Yangni A.K.H. Assessment of the obstetrical and anthropometric parameters of women in labor and their newborns at a tribal health center in northern of Cote d’Ivoire. International Journal of Current Research, 2019, 11, (01), 858-864. | ||
| In article | |||
| [16] | Noel, D. D., Olefongo, D., Lazare, T., Wagniman, S., Koffi, N. G. B. S., Joel, K. K., Florent, K. A., Eboulé, A. E. R., Guehi, Z. E., and Yangni-Angaté, K. H. Assessment of recurrently diagnosed diseases dynamism at Korhogo General Hospital in Northern Côte d’Ivoire from 2014 to 2018. International Journal of Medicine and Medical Sciences, 2022, 14(1), 1-19. | ||
| In article | View Article | ||
| [17] | ChippauxJ.P., Akogbeto M.,MassougbodjiA.,Adjagba J. Mesure de la parasitémie palustre et évaluation du seuil pathogène en région de forte transmission permanente. Cotonou, Benin, 1991, 11p. | ||
| In article | |||
| [18] | Richard A., Lallemant M., Trape J.F., Carnevale P. and Mouchet J. Le paludisme dans la région forestière du Mayombe, République populaire du Congo. II. Parasitological observations. Annales de la Société Belge de Médecine Tropicale, 1998, 68, 4: 305-316. | ||
| In article | |||
| [19] | Mharakurwa S., Simoloka C., Thuma P.E., Shiff C.J., Sullivan D.J. PCR detection of Plasmodium falciparum in human urine and saliva samples. Malaria Journal, 2006, 5: 103. | ||
| In article | View Article PubMed | ||
| [20] | Nwakanma D.C., Gomez-Escobar N., Walther M., Crozier S., Dubovsky F., Malkin E., Locke E., Conway D.J. Quantitative detection of Plasmodium falciparum DNA in saliva, blood and urine. JournalofInfectiousDiseases, 2009, 199(11): 1567-74. | ||
| In article | View Article PubMed | ||
| [21] | Putaporntip C., Buppan P., Jongwutiwes S. Improved performance with saliva and urine as alternative DNA sources for malaria diagnosis by mitochondrial DNA based PCR assays. Clinical Microbiology and Infection, 2011, 17: 1484-91. | ||
| In article | View Article PubMed | ||
| [22] | Kwannan Nantavisai Malaria detection using non-blood samples. Songklanakarin Journal of Science and Technology, 2014, 634 36 (6): 633-641. | ||
| In article | |||
| [23] | Aninagyei E., Abraham J., Atiiga P., Antwi S.D., Bamfo S., Acheampong D.O. Evaluating the potential of using urine and saliva specimens for malaria diagnosis in suspected patients in Ghana. Malaria Journal, 2020, 19 (1): 349. | ||
| In article | View Article PubMed | ||
