Evaluation of microscopy and first response™ malaria Ag PAN/Pf RDT compared to qPCR and hemoglobin levels in pregnant women at delivery in Busia County, Western Kenya

Abstract

Background

The World Health Organization (WHO) recommends parasitological confirmation before treating suspected malaria cases. However, diagnosing malaria in pregnancy (MiP) remains challenging due to parasite sequestration in the placenta, which results in low peripheral parasite density often undetectable by point-of-care diagnostic tools. Studies have reported conflicting findings on the relationship between ABO blood groups, anaemia, and MiP. This study investigated the performance of microscopy and the First Response™ Malaria Ag PAN/Pf rapid diagnostic test (mRDT) compared to quantitative polymerase chain reaction (qPCR) in diagnosing MiP. It also explored the associations between MiP, ABO blood groups, and anaemia.

Methods

This cross-sectional study enrolled 236 women delivering at Sio Port and Port Victoria Sub-County Hospitals, Busia County, Kenya (April 2023–March 2024). Peripheral blood was tested for malaria using microscopy, mRDT, and qPCR while placental blood was analyzed using microscopy and qPCR. ABO blood groups and haemoglobin levels were determined from peripheral samples.

Results

Peripheral malaria prevalence was 7.2% (16/223), 8.1% (18/223), and 9.0% (20/223) as determined by microscopy, mRDT, and qPCR, respectively. Placental malaria prevalence was 7.2% (17/236) by microscopy and 15.7% (37/236) by qPCR. When compared with peripheral blood qPCR, mRDT demonstrated higher sensitivity (90% (95% CI: 0.77–1.0)) but lower specificity (80% (95% CI: 0.74–0.85)) than microscopy (80% (95% CI: 0.63–0.97) sensitivity; 89% (95% CI: 0.85–0.93) specificity). Plasmodium falciparum predominated in placental (58.8%, 10/17) and peripheral (81.3%, 13/16) samples by microscopy. Mean haemoglobin ± SD was 10.5 ± 1.45 g/dL, with 69.9% (95% CI 65.2–71.6) anaemia prevalence (Hb ≤ 11 g/dL). Blood group O + was most common among the participants (43.2%, 102/236). No significant associations were observed between MiP and ABO blood groups or anaemia (P > 0.05).

Conclusion

mRDTs demonstrated superior performance compared to microscopy in detecting malaria in pregnancy, although diagnostic limitations remain. Anaemia remained highly prevalent reflecting broader maternal health issues. These findings emphasize the need for integrated maternal health strategies during antenatal care in malaria endemic regions.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-026-13577-z.

Background

Every year, approximately 125 million women are at risk for contracting malaria worldwide [1]. According to the 2024 World Health Organization (WHO) Malaria Report, there were an estimated 36 million pregnancies of which 12.4 million (34%) were affected by malaria in 33 moderate-to-high-transmission countries in the WHO African Region [2]. In sub-Saharan Africa, where the malaria burden is highest, pregnant women are at the highest risk of MiP in the first trimester and Plasmodium falciparum is responsible for up to 10,000 malaria-related maternal deaths annually primarily due to severe anaemia, and approximately 200,000 infants die [3]. MiP contributes to poor pregnancy outcomes including higher rates of miscarriage, intrauterine death, premature delivery, low-birth- weight neonates, and perinatal mortality. In addition, low birth weight (LBW) linked to MiP contributes to about 11% of newborn fatalities [4, 5]. In Kenya, data from the Ministry of Health indicate that 6.3% of women attending their initial antenatal care visit had MiP in 2019 [6].

Pregnant women susceptibility to MiP declines with increasing gravidity due to the gradual acquisition of antibodies that block Plasmodium falciparum adhesion to placental receptor chondroitin sulfate A (CSA) [7, 8]; nevertheless, the impact of MiP extends beyond health, imposing significant socio-economic and financial burdens on affected populations [9, 10]. To address this, the WHO recommends universal diagnosis and treatment of malaria cases, as well as preventive measures such as intermittent preventive treatment in pregnancy (IPTp) with sulphadoxine-pyrimethamine (SP) and the use of insecticide-treated mosquito nets (ITNs) in endemic areas [11]. Despite these strategies, malaria remains a significant public health burden, particularly among pregnant women and children under five years of age in Kenya [8, 12].

Diagnosing MiP remains a challenge. First, P. falciparum-infected erythrocytes sequester in the placenta, resulting in low peripheral parasite density levels that often cannot be detected by point-of-care diagnostic tools. Secondly, scarcity of resources in many malaria-endemic areas limits access to more advanced diagnostic methods such as PCR [13]. The WHO currently recommends microscopy and mRDTs as primary diagnostic tools. However, microscopy often fails to detect submicroscopic parasite density (low levels of parasites in blood), particularly in placental malaria cases [14]. While mRDT offer an alternative [15], they are qualitative and prone to false-positives [16, 17]. Additionally, increasing reports of P. falciparum histidine rich protein 2/3 (pfhrp2 and/or pfhrp3) gene deletions pose a growing challenge, as these mutations allow parasites to evade detection by most HRP2-based mRDTs [18]. Quantitative polymerase chain reaction (qPCR) offers a more sensitive molecular method for detecting P. falciparum genetic material (parasite DNA) in patient samples, though it is expensive and requires specialized training and infrastructure [19].

Previous studies have reported an association between ABO blood group and anaemia in the context of malaria. Specifically, blood group O has been shown to confer a protective effect against severe Plasmodium falciparum malaria compared to non-O blood groups [20–22]. However, other studies have not found a significant association between ABO blood group or related clinical outcomes [23, 24]. These inconsistent findings highlight the need for further research to clarify the role of ABO blood group in malaria susceptibility and disease severity. Anaemia is a significant public health concern during pregnancy, affecting about 50% of pregnant women in malaria endemic areas of Africa [25]. It has multiple causes, including nutritional deficiencies, red blood cell disorders (such as sickle cell and thalassaemia), and infections like schistosomiasis, HIV, hookworm, and malaria [26]. Malaria alone accounts for about 25% of severe anaemia in pregnant women in malaria endemic areas [27]. However, in Busia County, a malaria endemic region, there is limited data on the prevalence of malaria in pregnancy (MiP), the diagnostic performance (sensitivity and specificity) of microscopy and mRDTs in this population, and the relationship between uncomplicated malaria, ABO blood group and anaemia.

This study aimed to evaluate the diagnostic performance (sensitivity and specificity) of microscopy and mRDTs using peripheral blood samples, compared to qPCR, among pregnant women at delivery in Busia County, Western Kenya. In addition, the study compared malaria detection in peripheral and placental blood samples using microscopy and qPCR to assess their utility in diagnosing MiP. It also determined the prevalence of MiP among delivering women and examined its association with ABO blood group and maternal hemoglobin levels. These findings will help inform improved detection, timely treatment, and prevention strategies for MiP, thereby supporting better maternal and fetal health outcomes.

Methods

Study design, site and population

This cross-sectional, prospective study was conducted between April 2023 and March 2024 at Sio Port and Port Victoria, Level-4 sub-County hospitals in Busia County, western Kenya. The two hospitals are located in a Lake endemic malaria ecological zone [28]. They offer fundamental medical care and are the primary referral facilities in the region. This is a hot, dry and low-lying rural area and the population is mainly engaged in subsistence farming and fishing. The areas have two rainy seasons: “long rains” (April–July) and “short rains” (September–November) with an annual rainfall between 250 and 1000 mm [29]. Pregnant women presenting for delivery at the hospital were screened for malaria using microscopy. Participants with confirmed malaria or a history of fever (temperature > 37.5 °C) were included in the analysis after providing written informed consent, with individuals below 18 years providing written assent prior to inclusion. Pregnant women with pre-existing severe illness, debilitating diseases, a history of drug use that could compromise the immune system, mental illness, or those in critical condition requiring emergency care were excluded from the study. Additionally, participants who refused to participate, or had a retained placenta or morbidly adherent placenta, were also excluded.

Sample size determination and sampling technique

Fisher’s algorithm was used to calculate the required sample size for the study, incorporating an estimated prevalence from the most recent relevant study and assuming a simple random sampling approach [30]. This study used a predicted prevalence (P) of 19%, an estimate of the overall prevalence of malaria in the Lake Victoria endemic region according to the Malaria Indicator Survey 2020 [31].

Where;

N is the determined sample size.

Z = Z-statistic with a 0.05 alpha (1.96).

P = Prevalence estimates based on the most recent prior research (0.19)

Q = 1 – P (1-0.19)

L = the permitted margin of error (0.05)

Therefore;

Therefore, a minimum of 236 pregnant women were required for the study. This sample was equally distributed across the two study sites, with 118 participants enrolled from Port Victoria and 118 from Sio Port.

Socio‑demographic and clinical data collection

A well-structured questionnaire was used to collect information on obstetric and sociodemographic variables. Obstetric data included gravidity and gestational age while sociodemographic data included age, level of education, occupation, marital status, and place of residence. Additional data was collected on the history of malaria infection and treatment, as well as any history of fever (Additional file 2).

About 5 mL of whole placental blood was collected in ethylenediaminetetraacetic acid (EDTA) vacutainer tubes (Becton-Dickinson, Franklin Lakes, NJ, USA) from the enrolled study participants immediately after delivery for malaria parasite screening by microscopy, and qPCR assays. About 7 µL of whole placental blood was used to prepare thin and thick blood smears for microscopy as described [32]. Additionally, about 500 µL of peripheral blood was collected from a finger prick used for preparing thin and thick smears for microscopy as described [33], for mRDT [34], for ABO blood grouping [35] and for haemoglobin determination [6], and the remaining blood was collected as dried blood spot (DBS) on FTA filter paper (Whatman Inc., Bound Brook, NJ) for molecular analysis [36]. Of the total 236 participants enrolled, peripheral DBS samples were successfully collected from 223. Thirteen participants were excluded due to missing DBS cards at the time of collection. Thirteen participants lacked peripheral DBS samples but had complete demographic and placental blood data and were therefore included in demographic and placental analyses but excluded from peripheral molecular analyses. Whole blood samples were transported to KEMRI-Centre for Global Health Research (CGHR) laboratory under cold chain (4 °C) for processing and analysis. DBS samples were stored at -20 °C in airtight bags containing desiccant and humidity indicator cards until DNA extraction.

Malaria diagnosis

Rapid diagnostic testing

Rapid diagnostic testing was carried out according to the manufacturer’s instructions using the First Response^™ Malaria PAN (test detects all human malaria species)/Pf diagnostic kit (Access Bio Inc. USA). RDTs were performed on peripheral blood samples only, as the manufacturer’s validation and performance evaluations are based on peripheral blood [37]. Briefly, 5µL of peripheral whole blood was dispensed into the sample well using a calibrated dropper, followed by the addition of buffer to the buffer well. The test device was placed on a level surface for 15 min, as previously described [38, 39], after which the results were read. Three bands (Control, PAN, and Pf) indicated P. falciparum infection with possible additional non-falciparum species. Two bands (Control and Pf) indicated P. falciparum infection only, while two bands (Control and PAN) indicated non-falciparum malaria. The appearance of the Control band alone indicated no detectable Plasmodium antigen (malaria negative) while tests in which the control band failed to appear were considered invalid and required repetition [39].

Diagnosis by microscopy

Thick and thin blood smears from the placental and peripheral blood were prepared as previously described [33, 40]. Briefly, thin smears were fixed in absolute methanol, whereas thick smears were spread into a thick layer without fixation. After air drying, both thin and thick smears were stained with 10% Giemsa at pH 7.2 for 15 min, as previously described [40], and examined under oil immersion. Parasites were detected and quantified by expert microscopists as previously described. Microscopy has a detection limit of 50–100 parasites/µL under optimal conditions. Slides were declared negative after examination of 100 high-power fields [41, 42]. A second independent expert microscopist read the slides, blinded to the RDT and the first reader’s results. In case of discrepancies, between the first and second microscopists, a third microscopist resolved the discrepancy.

DNA extraction, Plasmodium parasites detection, and speciation

DNA was extracted once per sample from 236 whole placental blood samples and 223 dry blood spots using the QIAamp® DNA mini kit (QIAGEN, Inc., Germany) following the manufacturer’s instructions with the 13 participants missing DBS excluded from peripheral molecular analyses [43]. To minimize the risk of DNA contamination when cutting sample discs from DBS cards, the punching device was reused but cleaned between each sample with 70% ethanol to remove residual DNA and cellular material [44]. The concentration and purity of extracted DNA was assessed by a Nanodrop, ND 1000 Spectrophotometer (Thermo Scientific™, US) [45]. Plasmodium parasites genus was detected using Quant Studio 6 Flex (Applied Biosystems Inc, Foster City, CA) as previously described [46]. A negative control (nuclease free water) and positive control (3D7 genomic DNA) were included for each assay for results validation. Additionally, Ribonuclease P (RNase P), a human housekeeping gene, was employed to QC the DNA extraction process. Species of the positive Plasmodium parasites were subsequently determined using a set of primers and probes as earlier described [46]. Infections with more than one Plasmodium species present were defined as mixed malaria infections. Samples with Ct values ≤ 40 with a sigmoidal shape amplification curve were considered positive [17, 34]. Each sample was analyzed in duplicate for Plasmodium detection and species identification by qPCR. A sample was considered positive if at least one replicate showed amplification.

Haemoglobin determination

Maternal haemoglobin level at delivery was measured once per sample using the portable Mission plus® HB machine (ACON Laboratories, Inc., San Diego, California, USA). The device was routinely calibrated and internal quality controls were performed to ensure accuracy. About 10 µL of peripheral blood was transferred using a capillary tube into a single-use test strip fitted in an Hb meter for measuring haemoglobin, and results were read after 15 s. Anaemia status was assessed using haemoglobin concentrations according to WHO guidelines. Haemoglobin levels were classified as follows: severe anaemia < 7.0 g/dL, moderate anaemia 7.0–9.9 g/dL, mild anaemia 10.0–10.9 g/dL, and anaemia < 11.0 g/dL [47].

ABO blood typing

ABO blood group typing was by forward typing on whole blood samples, and all tests were conducted in duplicate as described [48] according to the manufacturer’s instructions using a commercially available kit (Accucare, Lab Care Diagnostics, India). Briefly, on a spotlessly clean glass slide, one drop of each participant’s whole peripheral blood was placed in three distinct spots. A drop of anti-A, anti-B, and Rhesus factor (D) antisera (from the test kit) was applied to each of the three spots on the glass slide. The antisera and blood cells were mixed using an applicator stick to achieve homogeneity. The blood group was scored after visually observing the mixture for evidence of agglutination within 2 min. The blood typing results were interpreted as described [22].

Statistical analysis

Descriptive statistics, including means, standard deviations, frequencies, and percentages, were used to summarize the data. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), along with their corresponding 95% confidence intervals (CIs), were calculated to assess the diagnostic performance of microscopy and RDTs compared to qPCR. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using standard diagnostic accuracy definitions as described [49, 50]. Agreement between microscopy, RDT, and qPCR results was assessed using Cohen’s Kappa coefficient. Kappa values were interpreted according to standard benchmarks: k ≤ 0 as indicating no agreement and 0.01–0.20 as none to slight, 0.21–0.40 as fair, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.00 as almost perfect agreement [51]. Exact binomial confidence intervals were used to estimate the CIs for sensitivity, specificity, PPV, and NPV.

Correlation analyses were conducted to examine the relationship between maternal haemoglobin levels and parasite density. A logistic regression analysis was performed to evaluate the association between various maternal characteristics and malaria infection status. The Wilcoxon rank-sum test was applied to compare parasite density across different blood types.

All p-values were two-sided, and a P-value < 0.05 was considered statistically significant. Analyses were performed using R statistical software version 4.1.1.

Results

Demographic characteristics of the study participants

The participants’ age ranged between 15 and 38 years, with a mean ± SD of 24.6 ± 5.6 years. The majority 33.9% (80/236) were in the 21–25 years age group. The mean gestational age [SD] was 38.4 weeks [1.36]. Most participants were primigravida 42.4% (100/236), while nearly equal proportions were multigravida 29.7% (70/236) and secundigravida 28.0% (66/236). A majority 69.5% (164/236) were married, and 54.2% (128/ 236) had more than 8 years of formal schooling, and out of these, 16.1% (38/236) were house wives, 28.8% (68/236) farmers, 22.9% (54/236) traders, 9.3% (22/236) civil servants and 22.9% (54/236) were other groups. Overall, 233 participants (98.7%) reported using insecticide-treated nets (ITNs). Among them, 42.4% (100/236) were primigravida, 28.0% (66/236) were secundigravida, and 28.3% (67/236) were multigravida. The demographic characteristics are summarized in Table 1.

Table 1.

Sociodemographic and obstetric characteristics of the study participants (N = 236)

Study variables	Number (%)
Study sites
Port Victoria	118 (50%)
Sio Port	118 (50%)
Age (Years)
Mean [SD]	24.6 [5.6]
Gestation at delivery (weeks)
Mean [SD]	38.4 [1.36]
Gravidity
Primigravida	100 (42.4%)
Secundigravida	66 (28.0%)
Multigravida	70 (29.7%)
Marital status
Married	164 (69.5%)
Single	72 (30.5%)
Education level
Primary	108 (45.8%)
Secondary	98 (41.5%)
Tertiary	30 (12.7%)
Occupation
House wife	38 (16.1%)
Farmer	68 (28.8%)
Traders	54 (22.9%)
Civil servant	22 (9.3%)
Others	54 (22.9%)
ITN usage	233 (98.7%)
Primigravida	100 (42.4%)
Secundigravida	66 (28%)
Multigravida	67 (28.3%)

Primigravida refers to a woman pregnant for the first time; secundigravida refers to a woman in her second pregnancy; and multigravida refers to women who have been pregnant three or more times

Proportion of malaria infection and species among pregnant women

All slides were examined by two experts with an agreement of 96% (kappa value = 0.864, 95% CI 0.810–0.899). The proportion of malaria in peripheral blood was 9.0% (95% CI: 16–25) by qPCR. Screening of 223 peripheral blood samples depicted RDT prevalence of 8.1% (18/223), PCR 9.0% (20/223) and microscopy 7.2% (16/223). Additionally, out of the 236 placental blood samples screened for malaria, 15.7% (37/236) and 7.2% (17/236) were positive by qPCR and microscopy, respectively. The overall prevalence of maternal malaria by qPCR was 20.2% (45/223), whereas microscopy detected 10.8% (24/223). Twelve (12) participants had positive infections in both peripheral and placental samples infections by qPCR. Similarly, microscopy detected nine (9) participants positive in both peripheral and placental blood films, who were counted once in the calculation of overall prevalence. However, 17 participants were infected and detected exclusively in placental samples by qPCR (Table 2).

Table 2.

Proportion of malaria by different diagnostic methods

Study Participants
	Positive			Negative
	n (%)	95% CI	P-value	n (%)	95%CI	P-value
Peripheral blood ( N  = 223)
Microscopy	16 (7.2)	12–26	< 0.001*	207 (92.8)	197–214	0.679
RDT	18 (8.1)	14–28		205 (91.9)	194–211
qPCR	20 (9.0)	16–25		203 (91.0)	199–209
Placental blood ( N  = 236)
Microscopy	17 (7.2)	10–22	0.987	219 (92.8)	210–226	0.578
qPCR	37 (15.7)	20–42		199 (84.3)	194–214
Peripheral and Placental blood ( N  = 223)
Microscopy	24 (10.8)	6.7–14.8	< 0.001*	199 (89.2)	83.9–94.5	0.457
qPCR	45 (20.2)	15.1–25.9		178 (79.8)	74.5–85.1

N- Overall sample size, n- number of participants in each category, RDT-rapid diagnostic test, qPCR- quantitative Polymerase Chain Reaction. Placental Plasmodium infection was diagnosed among 236 delivered women whereas peripheral Plasmodium infection was diagnosed in 223 delivered women (13 samples were excluded from peripheral molecular analyses due to missing DBS cards at the time of collection). Overall maternal malaria prevalence was calculated using women with available peripheral samples (n = 223) *Statistically significant

Malaria speciation using all the three diagnostic tools identified P. falciparum as the predominant species in both peripheral 81.3% (13/16) and placental 58.8% (10/17) blood samples by microscopy. In peripheral blood, it was detected in 81.3% (13/16) samples by microscopy, 38.9% (7/18) by RDT, and 100% (20/20) by qPCR. Mixed infection of P.falciparum and P. malariae was detected in 18.8% (3/16) by microscopy. Additionally, RDT identified 61.1% (11/18) mixed-species infections in peripheral blood. Similarly, in placental blood samples, P. falciparum was the most common species, with 58.8% (10/17) detected by microscopy and 97.3% (36/37) by qPCR. One placental blood sample had mixed infection of P. falciparum, P. malariae, P. ovalewallikeri and P. ovalecurtisi (Additional file 1, Table 1).

Distribution of malaria parasite density in peripheral and placental blood by microscopy

Low parasite density was recorded in 12.5% (2/16) of positive peripheral blood films and 52.9% (9/17) of positive placental blood films. Moderate parasite density was observed in 43.8% (7/16) of peripheral blood and 11.8% (2/17) of placental blood films. In contrast, high parasite density was found in 43.8% (7/16) of peripheral blood and 35.3% (6/17) of placental blood films (Table 3).

Table 3.

Distribution of malaria parasite densities of peripheral blood samples as determined by microscopy

Parasite density (/µL)
	Low (< 1000/µL) N (%)	Moderate (1000-4,999/µL) N (%)	High (≥ 5,000/µL) N (%)
Peripheral Blood (N = 16)
	2 (12.5)	7 (43.8)	7 (43.8)
Placental Blood (N = 17)
	9 (52.9)	2 (11.8)	6 (35.3)
Gravidity (Peripheral blood)
Primigravida (n = 10)	2 (20.0)	4 (40.0)	4 (40.0)
Secundigravida (n = 4)	0 (0)	2 (50.0)	2 (50.0)
Multigravida (n = 2)	0 (0)	1 (50.0)	1 (50.0)
Gravidity (Placental blood)
Primigravida (n = 12)	6 (50.0)	4 (33.3)	2 (16.7)
Secundigravida (n = 3)	0 (0)	2 (66.7)	1 (33.3)
Multigravida (n = 2)	0 0)	1 (50.0)	1 (50.0)

Primigravida refers to a woman pregnant for the first time; secundigravida refers to a woman in her second pregnancy; and multigravida refers to women who have been pregnant three or more times

Parasite density varied by gravidity in both peripheral and placental blood. In peripheral blood, (n = 16) were positively infected by microscopy. Out of which, (10/16) were primigravida with 20.0% low, 40.0% moderate, and 40.0% high density infections. Secundigravida (4/16) had 50.0% moderate and 50.0% high density infections. Multigravida (2/16) also showed 50.0% moderate and 50.0% high-density infections. In placental blood (n = 17) were positively infected by microscopy. Out of which, primigravida (12/17) were predominantly associated with low-density infections (50.0%), followed by moderate (33.3%) and high (16.7%) densities. Among secundigravida (3/17), infections were mainly moderate (66.7%) and high (33.3%). Multigravida (2/17) showed equal proportions of moderate (50.0%) and high (50.0%) parasite densities (Table 3).

Performance of microscopy and RDT compared to qPCR in diagnosing malaria in peripheral blood

When compared with qPCR, RDT demonstrated higher sensitivity 90.0 (95% CI: 0.77–1.0) than microscopy 80.0 (95% CI: 0.63–0.97). However, microscopy showed greater specificity 89.0 (95% CI: 0.85–0.93) compared with RDT 80.0 (95% CI: 0.74–0.85). The positive predictive value was higher for microscopy 42.0 (95% CI: 0.26–0.58) than RDT 31.0 (95% CI: 0.19–0.42), while both tests demonstrated very high negative predictive values; microscopy: 98.0 (95% CI: 0.96–0.99); RDT: 99.0 (95% CI: 0.97–1.00). Overall diagnostic accuracy was higher for microscopy 88.0 (95% CI: 0.83–0.92) than RDT 81.0 (95% CI: 0.75–0.86). Agreement with qPCR was moderate for microscopy (κ = 0.48) and fair for RDT (κ = 0.36) (Table 4).

Table 4.

Diagnostic performance of RDT and microscopy in diagnosing malaria in pregnancy in comparison with qPCR in peripheral blood (N = 223)

Metric	Microscopy (95% CI)	RDT (95% CI)
True Positives (TP)	16	18
False Positives (FP)	22	41
True Negatives (TN)	181	162
False Negatives (FN)	4	2
Apparent Prevalence	0.17 (0.12–0.22)	0.26 (0.20–0.32)
True Prevalence	0.09 (0.05–0.13)	0.09 (0.05–0.13)
Sensitivity	0.80 (0.63–0.97)	0.90 (0.77-1.0)
Specificity	0.89 (0.85–0.93)	0.80 (0.74–0.85)
Positive Predictive Value	0.42 (0.26–0.58)	0.31 (0.19–0.42)
Negative Predictive Value	0.98 (0.96–0.99)	0.99 (0.97-1.00)
Positive Likelihood Ratio	7.32 (4.63–11.57)	4.43 (3.25–6.09)
Negative Likelihood Ratio	0.22 (0.10–0.56)	0.13 (0.04–0.49)
Accuracy	0.88 (0.83–0.92)	0.81 (0.75–0.86)
Kappa Statistic	0.48 (Moderate)	0.36 (Fair)

N = 223, Values in parentheses represent 95% confidence intervals (95% CI)

Accuracy of microscopy and RDT compared to qPCR in diagnosing malaria in peripheral blood of women with different gravidities

Both microscopy and RDT demonstrated similar sensitivity across gravidity groups, with values of 85.7% and 100% in primigravida, 66.7% and 66.7% in secundigravida, and 66.7% and 66.7% in multigravida, respectively. However, the specificity of the two methods varied by gravidity. In primigravida women, microscopy showed higher specificity compared with RDT (86.2% vs. 73.6%). A similar trend was observed among secundigravida women, where microscopy demonstrated higher specificity than RDT (94.9% vs. 83.1%). In multigravida women, specificity was comparable between microscopy and RDT (87.7% vs. 86.0%). Overall, microscopy tended to show higher specificity than RDT across most gravidity groups, while the sensitivity of the two diagnostic methods remained largely comparable (Table 5).

Table 5.

Diagnostic performance of microscopy and RDT compared to qPCR across different gravidity levels (N = 223)

Diagnostic Test	Gravidity Group	TP	FP	TN	FN	Sensitivity % (95% CI)	Specificity % (95% CI)	PPV % (95% CI)	NPV % (95% CI)
Microscopy	Multigravida	2	7	50	1	66.7 (9.4–99.2)	87.7 (76.3–94.9)	22.2 (2.8–60.0)	98.0 (89.6–100)
	Primigravida	12	12	75	2	85.7 (57.2–98.2)	86.2 (77.1–92.7)	50.0 (29.1–70.9)	97.4 (90.9–99.7)
	Secundigravida	2	3	56	1	66.7 (9.4–99.2)	94.9 (85.9–98.9)	40.0 (5.3–85.3)	98.2 (90.6–100)
RDT	Multigravida	2	8	49	1	66.7 (9.4–99.2)	86.0 (74.2–93.7)	20.0 (2.5–55.6)	98.0 (89.4–99.9)
	Primigravida	14	23	64	0	100 (76.8–100)	73.6 (63.0-82.4)	37.8 (22.5–55.2)	100 (94.4–100.0)
	Secundigravida	2	10	49	1	66.7 (9.4–99.2)	83.1 (71.0-91.6)	16.7 (2.1–48.4)	98.0 (89.4–99.9)

n = 223, TP: true positives; TN: true negatives; FP: false positives; FN: false negatives; PPV: positive predictive value; NPV: negative predictive value; PLR: positive likelihood ratio; NLR: negative likelihood ratio. Primigravida refers to a woman pregnant for the first time; secundigravida refers to a woman in her second pregnancy; and multigravida refers to women who have been pregnant three or more times

Association between anaemia and malaria infection among pregnant women presenting for delivery in peripheral blood

This study revealed a mean haemoglobin ± SD of 10.5 ± 1.45 g/dL (95% CI 10.3–10.7) and an anaemia prevalence of 69.9% (95% CI 65.2–71.6). The majority of participants, 39.0% (92/236) had moderate anaemia (haemoglobin 7–9.9 g/dL) followed by mild anaemia 29.2% (69/236) (haemoglobin 10–10.9 g/dL). Only a few pregnant women, 1.7% (4/236) had severe anaemia (haemoglobin < 7 g/dL) whereas 30.1% (71/236) were non-anaemic (haemoglobin > 11 g/dL). There was no significant association between haemoglobin levels (P = 0.25, R = 0.19) as well as anaemia (P = 0.43) and malaria infection among pregnant women (Fig. 1A and B, respectively).

Fig. 1.

Association between maternal log- transformed hemoglobin levels (A) and anaemia status (B) with log-transformed parasite density in peripheral blood (N = 223). Both horizontal and vertical axes are presented on the base-10 logarithmic scale in Fig. A and vertical axis in Fig. B, respectively)

Association between maternal blood group, anaemia, age, gravidity, maternal BMI with malaria infection in peripheral blood

Blood group O + was the most prevalent 43.2% (102/236) among the enrolled individuals, followed by B + 22.9% (54/236), A + 19.9% (47/236), and AB + 14.0% (33/236), respectively. Among malaria-positive individuals detected by qPCR in peripheral blood, O + was the most common blood group 45.0% (9/20), followed by A + 30.0% (6/20), B + 20.0% (4/20), and AB + 5.0% (1/20) (Additional file 1, Fig. 1). Overall, none of the analyzed variables demonstrated a statistically significant association with malaria infection status (all p > 0.05) (Table 6).

Table 6.

Association between maternal blood group, anaemia, age, gravidity, maternal BMI and malaria infection status in peripheral blood samples (n = 223)

Characteristic	OR	95% CI	p-value
Blood group
A+	—	—
AB+	0.19	0.01–1.37	0.2
B+	0.50	0.11–2.16	0.4
O+	0.56	0.16–2.05	0.4
Anaemia
Anaemic	—	—
Healthy	1.14	0.37–3.42	0.8
Age (Years)	0.90	0.76–1.04	0.2
Gravidity
Multigravida	—	—
Primigravida	2.61	0.43–23.1	0.3
Secundigravida	1.05	0.15–9.06	> 0.9
Maternal BMI	0.97	0.79–1.17	0.7

n = 223, CI: Confidence Interval, OR: Odds Ratio, BMI: Body mass index, Primigravida refers to a woman pregnant for the first time; secundigravida refers to a woman in her second pregnancy; and multigravida refers to women who have been pregnant three or more times

Discussion

Efficient malaria control in endemic regions relies on accurate diagnosis and timely treatment of Plasmodium species [52–54]. Misidentification can lead to unnecessary treatments, contributing to drug resistance, which complicates malaria control efforts [55]. To improve diagnostic accuracy, the WHO recommends microscopy and RDTs for routine use at health care points in malaria endemic countries [52, 56]. This study aimed to assess the diagnostic performance of microscopy and First Response™ RDT compared to qPCR, and to explore associations between malaria, anaemia, and ABO blood groups among pregnant women.

The proportion of participants with malaria was 9.0% by qPCR, higher than reports from Ethiopia (2.8%) and India (5.4%) [57, 58], but lower than in Malawi (19.6%) and Burkina Faso (18.1%) [59, 60]. While our study participants reported high usage of long-lasting insecticide nets (LLINs) at 98.7%, the proportion of participants with malaria remained relatively high. This could be due to several factors, including variations in local transmission intensity, insecticide resistance, mosquito biting behavior, or inconsistent use of nets during peak exposure times. For comparison, LLINs usage in Ethiopia was only 39.9% [61], and slightly lower than levels reported in South Sudan (89%) and Mali (83.7%), the highest across 11 East African countries analyzed in a meta-analysis [62]. These findings suggest that although LLINs coverage is crucial, additional preventive measures and monitoring of vector control effectiveness are needed to further reduce malaria risk.

Interestingly, the number of malaria cases in our study appeared to decline with increasing gravidity, consistent with previous observations in endemic regions [17, 63]. Primigravida are generally more susceptible to malaria due to the lack of acquired immunity against placental-binding P. falciparum variants (VAR2CSA). With successive pregnancies, women develop antibodies that reduce placental parasite sequestration, lowering both infection prevalence and parasite density [7, 64]. Behavioral factors, such as more consistent use of LLINs or adherence to intermittent preventive treatment, may also contribute to the observed decline in multigravida [65].

The detection of malaria parasites in placental blood (15.7% by qPCR and 7.2% by microscopy) underscores the persistence of submicroscopic infections at delivery, which may go undetected by conventional diagnostic methods [66]. The higher proportion of malaria infections detected by qPCR compared to microscopy reflects the greater sensitivity of molecular assays in identifying low-density peripheral infections, which may be associated with placental parasite sequestration [67]. Although peripheral blood microscopy identified a proportion of infections, qPCR detected additional cases, particularly from placental samples, indicating that reliance on peripheral microscopy alone may underestimate malaria in pregnancy due to placental parasite sequestration [68]. Parasite density distribution differed between placental and peripheral blood with placental blood showing a higher proportion of low-density infections, and peripheral blood a slightly higher proportion of moderate and high-density infections. This pattern likely reflects the sequestration of infected erythrocytes within the placental intervillous spaces, a defining characteristic of placental malaria pathogenesis [17, 69]. These findings are consistent with studies from Kenya and other endemic regions where submicroscopic placental malaria has been linked to plasmodium parasite sequestration and adverse pregnancy outcomes such as maternal anaemia, low birth weight, and preterm delivery [67, 70, 71].

Compared to qPCR, RDTs showed higher sensitivity (90%) but lower specificity (80%), while microscopy had lower sensitivity (80%) but higher specificity (89%). These results align with previous studies in Uganda and Nigeria [38, 72], though other reports show lower RDT performance [17]. The relatively high detection rate of RDTs may partly be explained by the persistence of HRP2 antigen in the bloodstream following parasite clearance, as reported in previous studies [73]. However, HRP2 antigen deletions were not directly assessed in this study.

In comparison to qPCR, microscopy and RDT exhibited false positive rates of 9.9% (22/223) and 18.4% (41/223), respectively. This findings are consistent with reports from Ethiopia [74], and Tanzania [75] where RDT and microscopy similarly showed discordance and false positivity relative to molecular assays. Such discrepancies likely stem from regional differences in transmission patterns and diagnostic practices. For RDTs, prolonged HRP2 antigen circulation and cross-reactivity with non-falciparum species may explain the false positives [76, 77]. Microscopy, as a diagnostic method, may occasionally result in misidentification of parasites as artifacts, particularly at low parasite densities, even when performed by experienced microscopists [56, 78, 79].

False negatives are a significant problem as the inability to treat infected individuals may result in ongoing disease transmission, heightened morbidity, and increased death [54]. Although the proportion of false negatives was low in this study (1.8% for microscopy, and 0.9% for RDT), they remain a concern. Submicroscopic infections and low parasite density, especially common in pregnancy, may explain these findings [56, 80]. Submicroscopic infections may persist for several months without exhibiting typical symptoms and require active detection and targeted treatment. These results are consistent with studies in Tanzania [81], but contrast with reports from Ghana and Nigeria showing higher false negative rates [47, 67].

Plasmodium falciparum accounted for 81.3% of cases, consistent with patterns seen in Ethiopia [82], and subtropical and tropical regions [83]. This reinforces the importance of prioritizing P. falciparum in preventive strategies for pregnant women, given its potential for severe maternal and fetal complications. Our finding emphasizes the need for aggressive prevention and control of malaria that relies on timely detection and treatment, especially among pregnant women, to prevent the severe form of the disease and its maternal-child adverse effects. Regional differences in species distribution, such as lower P. falciparum prevalence elsewhere [17, 84], likely reflect varying ecological and transmission factors.

Although placental histology is considered the gold standard for detecting sequestered P. falciparum parasites in the placenta, it was not performed in this study. Instead, placental blood microscopy and qPCR were used as indicators for placental infection, which may underestimate true placental sequestration. Future studies should incorporate placental tissue histology to provide a more comprehensive assessment of malaria in pregnancy.

Anaemia was highly prevalent (69.9%) among the study participants, higher than previous reports from other regions in Kwale County and Pumwani maternity hospital, Kenya, which reported a prevalence of 60% and 57%, respectively [6, 85] and lower than those in previous studies in southwest Ethiopia (23.5%) and northwest Ethiopia (16.5%) [72, 86]. Malaria and anaemia are key factors that can impair the immune system of pregnant women, and when combined, they serve as major risk factors and adverse pregnancy outcomes [8]. The absence of a significant association between anaemia and malaria (P > 0.05) in the current study suggests that other factors such as; nutritional deficiencies, co-infections, or genetic disorders, may play a greater role in this population [6, 87]. This highlights the need for broader screening and nutritional interventions beyond malaria control. This study did not assess other contributors to anaemia, such as nutritional deficiencies, geophagy, and genetic conditions such as sickle cell disease, or any chronic infections. The absence of these data limits our ability to attribute anaemia specifically to malaria or fully characterize its multifactorial causes in this population. Nevertheless, anaemia classification was based on laboratory-confirmed haemoglobin measurements, enhancing reliability of the findings. Also, hemoglobin measurements were performed once per sample, which may introduce measurement variability. Future studies should consider duplicate measurements to improve precision.

ABO blood group determines the individual’s susceptibility to malaria. While some literature suggests that blood groups A, B, and AB confer greater susceptibility to P. falciparum than group O [22, 88, 89]. The underlying mechanism remains poorly understood. Some studies have attributed this to the likelihood of the non-O blood group to form rosettes, a condition typically marked by elevated P. falciparum parasite density during malaria infection [20]. Although more malaria cases were observed among participants with blood group O, this likely reflects the high proportion of blood group O in the cohort rather than increased susceptibility. No statistically significant association was found between malaria infection and blood group type (P > 0.05). This contrasts with findings from a study in Mumbai, India, where individuals with blood group O were found to be at higher risk of malaria infection [88]. Differences in transmission intensity, genetic background, or methodology may account for these contradicting findings [90]. Conversely, a study conducted in Douala, Cameroon, reported that individuals with blood group O were less likely to contract malaria compared to those with blood groups A and B [91]. These mixed findings may be influenced by immunological, genetic, and immune traits as well as environmental factors, underscoring the complexity of host susceptibility [90–92]. Although all forward typing tests were conducted in duplicate, the lack of reverse typing could have resulted in minor misclassification of ABO groups, potentially affecting downstream analyses dependent on precise blood group. Future studies should incorporate both forward and reverse typing to improve accuracy.

Conclusion

This study found that RDTs were more sensitive than microscopy in detecting malaria in pregnancy in Busia, although both methods showed limitations. Using both diagnostics in tandem may improve detection accuracy and minimize false negatives in routine clinical practices. In resource-limited settings, ensuring the quality of diagnostic tools through quality-assured RDTs and strengthened microscopy capacity is essential for effective malaria diagnosis and control.

Malaria in pregnancy was predominantly caused by Plasmodium falciparum and was most common among primigravida, underscoring the need for targeted screening and preventive interventions during antenatal care. The high prevalence of anaemia, despite limited association with malaria or blood groups, may point to broader maternal health challenges such as nutritional deficiencies and co-infections.

These findings highlight the importance of integrating malaria diagnosis within the routine ANC visits and anaemia management during pregnancy in Busia County, Kenya. Strengthening these targeted interventions can help reduce maternal malaria burden and improve pregnancy outcomes.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

ACT

Artemisinin Combined Therapy

BMI

Body Mass Index

CGHR

Centre for Global Health Research

DNA

Deoxyribonucleic Acid

IES

Infected Erythrocytes

IPTp-SP

Intermittent Preventive Treatment with Sulphadoxine-Pyrimethamine

IRC

Institutional Review Committee

IVPB

Inter-villous Placental Blood

IVS

Inter-villous Space

ITN

Insecticide Treated Nets

KEMRI

Kenya Medical Research Institute

LLINs

Long Lasting Insecticide Nets

MiP

Malaria in Pregnancy

MOH

Ministry of Health

mRDT

Malaria Rapid Diagnostic Test

NPV

Negative Predictive Value

NLR

Negative Likely Ratio

PM

Placental Malaria

PPV

Positive Predictive Value

PLR

Positive Likely Ratio

qPCR

Quantitative Polymerase Chain Reaction

QC

Quality Control

SERU

Scientific and Ethics Review Unit

WHO

World Health Organization

Author contributions

DK: Conceptualization of the idea, laboratory examination, collected data, data analysis and interpretation, drafted the original manuscript. DMW: Laboratory examination, Manuscript review, data curation, and interpretation. GI, RM, EC, and SAC: Data collection and data curation. MG, JMO, JLB, and AHI: Manuscript review and data analysis and interpretation. BNO: Conceptualization, study design, data analysis, supervision, and manuscript preparation. All authors reviewed and approved the final manuscript.

Funding

The Bill and Melinda Gates Foundation supported this work through The African Academy of Sciences (grant number APTI − 18 − 05).

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

This study sought approval from Egerton University Ethics Review Committee (approval No: EUREC/APP/180/2022), and the National Commission for Science, Technology, and Innovation (License No. NACOSTI/P/22/18216). The study was permitted by the County Director of Public Health, Busia County. All the respondents were asked for their informed consent before enrolling in the study. The participants’ identifiable information was anonymized throughout the survey, and confidentiality was maintained. Pregnant women who were infected with Plasmodium parasites were treated according to the current Ministry of Health national treatment guidelines. The research was done in compliance with ethical principles as outlined by Declaration of Helsinki for medical research involving human participants, including research using identifiable human data.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.