INTRODUCTION
Malaria often presents with nonspecific symptoms and can be difficult to distinguish clinically owing to limited diagnostic capabilities and resources in low- to middle-income countries.1 Unfortunately, the overdiagnosis of malaria and the underdiagnosis of other etiologies may be attributed to the similarities between nonspecific acute febrile illness (AFI) syndromes.2
Improved control measures have significantly decreased the worldwide incidence of malaria,3 and in 2010 the WHO recommended antimalarial treatment of AFI only after parasitological confirmation.4 However, in the African sub-region, malaria remains the leading diagnosis given to patients with fever5 because the true etiology of fever is often difficult to ascertain in these resource-limited regions. In Ghana, which is considered a high malaria transmission area, there is little epidemiologic data regarding the possible causes of non-malarial fevers. Once malaria is suspected, clinicians typically do not pursue other potential etiologies, even in circumstances where coinfections could be present. Interestingly, one study addressing the underestimation of non-malaria fevers reported that about 60% of febrile illnesses were caused by different pathogens, even in highly endemic malarial zones.6 Another recent study found that 5% of children under the age of 5 years have bacterial bloodstream infections, caused most frequently by non-typhoidal Salmonella and Salmonella typhi.7
Bacterial diseases such as Q fever and leptospirosis are known to cause AFI syndromes and have been reported in humans, livestock, and rodents in Ghana,8–10 but they have not been well studied in Ghana. Although these bacteria and reservoirs are present in Ghana, these pathogens require testing not typically available at medical treatment facilities and could represent a portion of undiagnosed AFIs. Understanding the prevalence of malaria coinfection and other possible causes of febrile infections should allow healthcare providers to improve diagnostic certainty and guide treatment regimens for their patients,11 especially considering that coinfections often manifest with greater severity and higher mortality.8 Therefore, this study aimed to identify the potential zoonotic or bacterial etiologies of AFIs in Ghana and determine their coinfections with malaria.
MATERIALS AND METHODS
Study design.
This was a cross-sectional study that enrolled febrile patients from inpatient and outpatient settings in Ghana. The study was conducted from July 2015 to December 2019 at the 37 Military Hospital located in Accra and three military clinics in Sekondi-Takoradi. The clinics included the Western Naval Command Medical Center in Sekondi, as well as the 2 Medical Reception Station (2MRS Armed Forces clinic) and the Air Force Medical Center in Takoradi. These clinics in the Western Region of Ghana serve primarily the military and their families, but also civilians.
Recruitment of study participants.
Patients between 1 month and 65 years old presenting with a documented or self-reported fever (≥ 38°C or history of fever within the last 24 hours) without a localizing etiology (i.e., otitis media, meningitis, diarrheal syndromes) were eligible for enrollment. All patients with AFIs seeking medical care at the medical centers were screened for eligibility to participate in the study. The informed consent process was completed for all eligible individuals. For participants less than 18 years old, both child assent and parental permission were obtained. Questionnaires were administered to capture demographic and clinical data.
Ethical considerations.
Ethical approval was obtained from the institutional review boards of the Ghana Health Service (GHS-ERC 14/05/13), Ghana Armed Forces (GAF-IRB 002/13), Noguchi Memorial Institute for Medical Research (NMIMR-IRB 126/12-13), and the U.S. Naval Medical Research Center (NMRC-IRB NAMRU3.2014.0003). The study protocol was approved by the aforementioned institutional review board in compliance with all applicable federal regulations governing the protection of human subjects.
Sample collection and processing.
Blood samples were collected via needle prick (1 month to 4 years) or venipuncture (5 years and above) from all participants. The volume of blood collected was age dependent but did not exceed 10 mL. At the study sites, a specimen of 1 to 10 mL of blood was collected from study participants based on age group (Figure 1) and put in different collection tubes/bottles based on the test to be performed. About 2 to 5 mL of blood was put in a sera separation tube, and samples were centrifuged at 1,500 × g for 15 minutes to obtain sera. One to three milliliters was placed in the BACTEC bottles for bacteria culture, along with two to three drops (approximately 60 μL) of EMJH medium for leptospiral culture. At each site, all blood samples were tested using the First Response Malaria Ag PLDH/HRP2 Combo rapid test kit (Premier Medical Corporation Private Limited, Gujarat, India) to diagnose malaria. As recommended by the manufacturer, 5 μL of whole blood was transferred into the sample collection well, and about 60 μL of assay buffer was dropped into the assay buffer well. The result was read 20 minutes after the specimen and buffer were added to the test device. A sample was interpreted as positive when two color bands appeared at Control line “C” and the test line “P.f” or “PAN.” The kit has a sensitivity and specificity of 100% and 98.1%, respectively (as stated in the packet insert). Sera, whole blood and BACTEC samples were then transferred to the Naval Medical Research Unit-3 (NAMRU-3) laboratory at Noguchi Memorial Institute for Medical Research located in Accra, Ghana, for bacterial culture, serology, or polymerase chain reaction (PCR) (Figure 1). Follow-up blood samples were obtained 2 to 6 weeks after the primary sample was acquired for serological analysis.
A flowchart of the study evaluating 399 acute febrile patients with respect to age. This figure shows the workflow for the number of participants sampled for the different tests based on their age group. PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
A flowchart of the study evaluating 399 acute febrile patients with respect to age. This figure shows the workflow for the number of participants sampled for the different tests based on their age group. PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
A flowchart of the study evaluating 399 acute febrile patients with respect to age. This figure shows the workflow for the number of participants sampled for the different tests based on their age group. PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Serology.
An ELISA was performed using commercially available kits12,13 as follows: The Human Leptospira Antibody (IgM and IgG) ELISA kit (Cusabio, Wuhan, China) was used for the qualitative determination of human Leptospira antibody concentrations in serum. All reagents in the kit and frozen serum samples were brought to room temperature and concentrated wash buffer (20×) was diluted to 1×. The serum samples were first diluted with the sample diluent and absorbent as instructed by manufacturers before they were added into the microtube strips (eight strips), which were pre-coated with Leptospira antigens. Sample dilution preparation for the IgM and IgG ELISA was done by making a 40-fold dilution with the sample diluent. This was achieved by adding 10 μL of sample in 390 μL of sample diluent. For IgM, 100 μL of diluted sample was added to 40 μL of sample absorbent to complete the sample dilution process. All assays were run with positive, negative, and reference (only for IgM) sera, which were provided in the kit. The controls and each sample were read using a microplate reader (BioTek EL 808 microplate reader, Winooski, Vermont, USA) at a wavelength of 450 nm. As recommended by manufacturers, to calculate the human Leptospira antibody (IgM and IgG), the optical density (OD) value of the samples was compared with that of the controls. An OD value less than 0.2 was considered negative, whereas an OD value greater than or equal to 0.5 was considered positive. Samples were retested when their OD value was between 0.2 and 0.5. The assay has a high sensitivity and excellent specificity for the detection of human Leptospira antibodies (IgM and IgG).
A calculated standard unit of < 9 was considered negative and > 11 was considered positive. The kit has a sensitivity and specificity of > 90.
Bacterial culture.
BACTEC bottles containing blood specimen were incubated overnight for 18 to 24 hours at 35–37°C. The cultured blood sample was collected aseptically by disinfecting the surface of the BACTEC bottle with 70% alcohol prep-pad and using a sterile 1-mL syringe with a thin needle and dropping it on the surface of blood, chocolate, and MacConkey agar and allowing it to dry. The plates were then streaked using a sterile 1-μL inoculating loop and incubated at 35–37°C for 18 to 24 hours. Grown bacterial isolates were identified by colony morphology, Gram staining, biochemical tests (catalase, oxidase, and coagulase), and Analytical Profile index (API 20E from bioMerieux, Inc., Marcy-l’Étoile, France) based on the manufacturer’s instructions and standard procedures.14 Blood culture bottles with no bacterial growth after 7 days were subcultured on blood, chocolate, and MacConkey agar (to aid the growth of slow-growing bacteria) before being reported as negative. Leptospirosis culture was performed by inoculating two to three drops of blood in EMJH medium and kept at 37°C in the dark.
DNA extraction.
The DNA was isolated from blood embedded on a Whatman protein saver card (GE Healthcare Ltd, Chicago. Illinois) using the (Qiagen, Hilden, Germany) blood DNA mini kit following manufacturer’s instructions.
Nucleic acid amplification.
Real-time PCR targeting the IS1111a region of C. burnetii was used to detect C. burnetii in collected samples.15 Lig1/Lig2 primers were used to amplify the conserved amino terminus region of the ligA and B genes of Leptospira. Conventional PCR (Applied Biosystems Machine 9700, Norwalk, CT) was performed using the PCR core kit (New England Biolab, Ipswich, MA) and primers as described by Palaniappan et al.16
Statistical methods.
All statistical analyses were conducted using R 3.3.1 software package. The χ2 and Fisher’s exact tests were used to determine significant associations between the presenting symptoms and laboratory test results (ELISA and culture) stratified by malaria status.
Multiple correspondence analysis (MCA) was used to identify data patterns and associations between variables to provide a graphical visualization between active variables (symptoms and sociodemographic characteristics) and supplementary variables (infection status). The infection status consisted of three levels, namely bacterial infection, malaria, and coinfection (bacterial and malaria infections), which were used as the dependent variables. The number of dimensions maintained in the final model was determined using the scree plot.
RESULTS
A total of 399 participants were enrolled; 222 (55.6%) were male and 177 (44.4%) were female (Supplemental Table 1). Less than half of the patients (48.2%) were younger than 18 years, and 29.8% were younger than 5 years (Supplemental Table 1). In addition to fever, other reported symptoms are listed in Supplemental Table 1, with the most common being headache (319, 79.9%), chills (313, 78.5%), joint pain (166, 41.6%), and muscle pain (151, 37.8%). The commonest clinical diagnoses were malaria (283, 70.9%) and respiratory tract infection (65, 16.3%).
All samples (N = 399) were tested for malaria using rapid diagnostic test (RDT) and for leptospirosis and Q fever using PCR. Serological testing was done on 280 samples using serology (participants > 5 years old), and culture was performed on 220 samples (participants > 10 years old because of blood volume limitation) (Figure 1). Malaria RDT was positive in 275/399 participants (68.9%) (Table 1, Figure 2). All PCR results were negative.
Sociodemographic characteristics, clinical presentation, and coinfection status of malaria-positive patients
| Characteristics | Total | Malaria, N = 275 (68.9%) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Sample tested | Leptospirosis* | P value‡ | Q fever* | P value | Sample tested | Culture positive | P value | ||
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | ||||
| Gender | |||||||||
| Male | 158 (567.5) | 122 (59.2) | 6 (6.0) | 0.601 | 11 (9.0) | 0.223 | 100 (61.0) | 5 (0.5) | 0.406 |
| Female | 117 (42.6) | 84 (40.8) | 5 (4.9) | 13 (15.5) | 64 (39.0) | 1 (1.6) | |||
| Age (years)† | |||||||||
| < 5 | 69 (25.1) | – | – | 0.052 | – | 0.401 | – | – | – |
| 5–10 | 42 (15.3) | 42 (20.4) | 0 (0.0) | 8 (19.1) | – | – | – | ||
| 11–17 | 25 (9.1) | 25 (12.1) | 3 (12.0) | 2 (8.0) | 25 (15.2) | 2 (8.0) | 0.937 | ||
| 18–29 | 76 (27.6) | 76 (36.9) | 3 (4.0) | 7 (9.2) | 76 (46.3) | 3 (3.9) | |||
| 30–39 | 32 (11.6) | 32 (15.5) | 4 (12.5) | 5 (15.6) | 32 (19.5) | 1 (3.1) | |||
| ≥ 40 | 31 (11.3) | 31 (15.0) | 1 (3.2) | 2 (6.5) | 31 (18.9) | 0 (0.0) | |||
| Military/Civilian | |||||||||
| Military | 43 (15.6) | 43 (20.9) | 3 (7.0) | 0.702 | 4 (9.3) | 0.790 | 43 (26.2) | 2 (4.7) | 0.653 |
| Civilian | 232 (84.4) | 163 (79.1) | 8 (4.9) | 20 (12.3) | 121 (73.8) | 4 (3.3) | |||
| Season | |||||||||
| Dry | 114 (41.5) | 88 (42.7) | 6 (6.8) | 0.534 | 7 (8.0) | 0.153 | 76 (46.3) | 3 (4.0) | 1.000 |
| Rainy | 161 (58.6) | 118 (57.3) | 5 (4.2) | 17 (14.4) | 88 (53.7) | 3 (3.4) | |||
| Symptoms | |||||||||
| Fever | |||||||||
| Current | 261 (94.9) | 193 (93.7) | 10 (5.2) | 0.662 | 23 (11.9) | 1.000 | 152 (92.7) | 6 (3.1) | 1.000 |
| History of fever | 14 (5.1) | 13 (6.3) | 1 (7.7) | 1 (7.7) | 12 (7.3) | 0 (0.0) | – | ||
| Chills | |||||||||
| Yes | 224 (81.5) | 174 (84.5) | 8 (4.6) | 0.377 | 21 (10.9) | 1.000 | 142 (86.6) | 6 (4.2) | 1.000 |
| No | 48 (17.5) | 30 (14.6) | 3 (10.0) | 3 (10.0) | 20 (12.2) | 0 (0.0) | |||
| Unknown | 3 (1.1) | 2 (1.0) | 0 (0.0) | 0 (0.0) | 2 (1.2) | 0 (0.0) | |||
| Headache | |||||||||
| Yes | 228 (82.9) | 189 (91.7) | 9 (4.8) | 0.505 | 21 (11.1) | 0.475 | 151 (92.1) | 5 (3.3) | 0.396 |
| No | 47 (17.1) | 17 (8.3) | 2 (11.8) | 3 (17.7) | 13 (7.9) | 1 (7.7) | |||
| Retro-orbital pain | |||||||||
| Yes | 54 (19.6) | 49 (23.8) | 3 (6.1) | 0.164 | 4 (8.2) | 0.200 | 41 (25.0) | 2 (4.9) | 0.640 |
| No | 217 (78.9) | 157 (76.2) | 8 (5.1) | 20 (11.5) | 123 (75.0) | 4 (3.3) | |||
| Unknown | 4 (1.5) | 0 (0.0) | – | – | – | – | – | – | – |
| Conjunctivitis | |||||||||
| Yes | 19 (6.9) | 16 (7.8) | 0 (0.0) | 0.738 | 0 (0.0) | 0.286 | 14 (8.5) | 1 (7.1) | 0.420 |
| No | 252 (91.6) | 190 (92.2) | 11 (5.8) | 24 (11.6) | 150 (91.5) | 5 (3.3) | |||
| Unknown | 4 (1.5) | 0 (0.0) | – | – | – | – | – | – | – |
| Sore throat | |||||||||
| Yes | 58 (21.1) | 53 (25.7) | 0 (0.0) | 0.467 | 5 (9.4) | 0.758 | 39 (23.8) | 1 (2.6) | 1.000 |
| No | 211 (76.7) | 152 (73.8) | 11 (10.0) | 19 (12.5) | 124 (7.6) | 5 (4.0) | |||
| Unknown | 6 (2.2) | 1 (0.5) | 0 (0.0) | – | 0 (0.0) | 1 (0.6) | 0 (0.0) | ||
| Rash | |||||||||
| Yes | 15 (5.4) | 11 (5.3) | 0 (0.0) | 0.329 | 1 (9.1) | 1.000 | 6 (3.7) | 0 (0.0) | 1.000 |
| No | 259 (94.2) | 195 (94.7) | 11 (5.6) | 23 (11.8) | 158 (96.3) | 6 (3.1) | |||
| Unknown | 1 (0.4) | 0 (0.0) | |||||||
| Muscle pain | |||||||||
| Yes | 108 (39.3) | 103 (50.0) | 2 (1.94) | 0.173 | 10 (9.7) | 0.485 | 93 (56.7) | 4 (4.3) | 0.712 |
| No | 154 (56.0) | 101 (49.0) | 0 (0.0) | 14 (13.9) | 70 (42.7) | 2 (2.9) | |||
| Unknown | 13 (4.7) | 2 (1.0) | 9 (8.9) | – | 0 (0.0) | 1 (0.6) | 0 (0.0) | ||
| Joint pain | |||||||||
| Yes | 126 (45.8) | 124 (60.2) | 6 (4.8) | 0.319 | 15 (12.1) | 1.000 | 113 (68.9) | 4 (3.5) | 1.000 |
| No | 135 (49.1) | 78 (37.9) | 5 (6.4) | 9 (11.5) | 48 (29.3) | 2 (4.2) | |||
| Unknown | 14 (5.1) | 4 (1.9) | 0 (0.0) | 0 (0.0) | 3 (1.8) | 0 (0.0) | |||
Dry season: November to March and rainy season: April to October.
ELISA (IgM+/IgG−).
Samples for bacteria culture were not obtained from 42 children from 5 years to 10 years.
P value was obtained using χ2 or Fisher’s exact test. The P value determines the association between the infection status (positive and negative) with the social demographic characteristics and clinical symptoms.
Stratification of the different laboratory assays performed on the various sample types. The chart shows the number of participants enrolled and the different tests performed (malaria, bacterial culture, PCR, and serology). PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Stratification of the different laboratory assays performed on the various sample types. The chart shows the number of participants enrolled and the different tests performed (malaria, bacterial culture, PCR, and serology). PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Stratification of the different laboratory assays performed on the various sample types. The chart shows the number of participants enrolled and the different tests performed (malaria, bacterial culture, PCR, and serology). PCR = polymerase chain reaction.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Overall, of the 280 serum samples tested, 10.7% and 5.0% were seropositive (IgM+, IgG) for Q fever and leptospirosis, respectively (Figure 3). Of the 220 samples cultured, approximately 8 (4.0%) were positive for a bacterial pathogen including 5 (2.3%) for S. typhi and 1 (0.5%) each for Salmonella spp., Staphylococcus aureus, and Bacillus cereus (Table 2, Figure 3). Malaria coinfection occurred with leptospirosis, Q fever, or bacteria in 11/206 (5.3%), 24/206 (11.7%), and 6/164 (3.7%), respectively (Table 3, Figure 2). Among the participants with negative malaria tests, 74 serology and 56 blood culture samples were collected. Of these, 3/74 (4.1%) and 6/74 (8.1%) were seropositive for leptospirosis and Q fever, whereas 2/56 (3.6%) blood cultures were positive (Tables 2 and 3, Figure 2). A total of 50 participants provided paired serum samples, resulting in six and seven acute infections (IgM+, IgG−) of Q fever and leptospirosis, respectively. However, only three and one (IgM+, IgG−) remained for Q fever and leptospirosis, respectively, after follow-up (Table 4). Polymerase chain reaction testing of 280 serum samples did not yield any bacterial DNA.
Distribution of bacterial pathogens identified by culture testing is shown for cultured bacteria (N = 220). This bar chart shows the proportion of AFI patients confirmed to be exposed to cultured bacterial infection. AFI = acute febrile illness.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Distribution of bacterial pathogens identified by culture testing is shown for cultured bacteria (N = 220). This bar chart shows the proportion of AFI patients confirmed to be exposed to cultured bacterial infection. AFI = acute febrile illness.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Distribution of bacterial pathogens identified by culture testing is shown for cultured bacteria (N = 220). This bar chart shows the proportion of AFI patients confirmed to be exposed to cultured bacterial infection. AFI = acute febrile illness.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Sociodemographic characteristics, clinical presentation of malaria-negative patients with bacterial pathogen
| Characteristics | Total | Non-Malaria, N = 124 (31.1%) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Sample tested | Leptospirosis* | P value‡ | Q fever* | P value | Sample tested | Culture positive | P value | ||
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | ||||
| Gender | |||||||||
| Male | 64 (51.6) | 42 (56.8) | 0 (0.0) | 0.077 | 1 (2.4) | 0.079 | 34 (60.7) | 1(2.9) | 1.000 |
| Female | 60 (48.4) | 32 (43.2) | 3 (9.4) | 5 (15.6) | 22 (39.3) | 1(4.6) | |||
| Age (years)† | |||||||||
| < 5 | 50 (40.3) | – | – | – | – | – | – | – | – |
| 5–10 | 18 (14.5) | 18 (24.3) | 1 (5.6) | – | 4 (22.2) | – | – | – | – |
| 11–17 | 3 (2.4) | 3 (4.1) | 0 (0.0) | 0.381 | 0 (0.0) | 0.223 | 3 (5.4) | 0 (0.0) | 0.630 |
| 18–29 | 30 (24.2) | 30 (40.5) | 0 (0.0) | 1 (3.3) | – | 30 (53.6) | 2 (6.7) | – | |
| 30–39 | 19 (15.3) | 19 (25.7) | 2 (10.5) | 1 (5.3) | 19 (33.9) | 0 (0.0) | |||
| ≥ 40 | 4 (3.2) | 4 (5.4) | 0 (0.0) | 0 (0.0) | 4 (7.1) | 0 (0.0) | |||
| Military/Civilian | |||||||||
| Military | 16 (12.9) | 16 (21.6) | 0 (0.0) | 1.000 | 0 (0.0) | 0.329 | 16 (28.6) | 0 (0.0) | 1.000 |
| Civilian | 108 (87.1) | 58 (78.4) | 3 (5.2) | 6 (10.34) | 40 (71.40 | 2 (5.0) | |||
| Season | |||||||||
| Dry | 34 (27.4) | 21 (15.5) | 1 (4.8) | 1.000 | 0 (0.0) | 0.175) | 17 (30.4) | 1 (5.9) | 0.519 |
| Rainy | 90 (72.6) | 53 (84.5) | 2 (3.8) | 6 (1.3) | 39 (69.6) | 1 (2.6) | |||
| Symptoms | |||||||||
| Fever | |||||||||
| Current | 116 (93.6) | 69 (93.2) | 2 (2.9) | 0.192 | 6 (8.7) | 1.000 | 51 (91.1) | 2 (3.9) | 1.000 |
| History of fever | 8 (6.5) | 5 (6.7) | 1 (20.0) | 0 (0.0) | 5 (8.9) | 0 (0.0) | |||
| Chills | |||||||||
| Yes | 89 (71.8) | 59 (79.7) | 1 (1.7) | 0.103 | 3 (5.1) | 0.093 | 47 (83.9) | 1 (2.1) | 0.298 |
| No | 35 (28.3) | 15 (20.3) | 2 (13.3) | 3 (20.0) | 9 (16.1) | 1 (11.1) | |||
| Headache | |||||||||
| Yes | 91 (73.4) | 66 (89.2) | 3 (4.6) | 1.000 | 5 (7.6) | 0.510 | 51 (91.1) | 0 (0.0) | 0.006 |
| No | 33 (26.6) | 8 (10.8) | 0 (0.0) | 1 (12.5) | 5 (8.9) | 2 (40.0) | |||
| Retro-orbital pain | |||||||||
| Yes | 21 (16.9) | 17 (23.0) | 0 (0.0) | 1.000 | 0 (0.0) | 0.380 | 15 (26.8) | 0 (0.0) | 1.000 |
| No | 101 (81.5) | 56 (75.7) | 3 (5.4) | 6 (10.7) | 41 (73.2) | 2 (3.6) | |||
| Unknown | 2 (1.6) | 1 (1.4) | 0 (0.0) | 0 (0.0) | 0 (0.0) | 0 (0.0) | |||
| Conjunctivitis | |||||||||
| Yes | 13 (10.5) | 8 (10.8) | 0 (0.0) | 1.000 | 0 (0.0) | 1.000 | 7 (12.5) | 0 (0.0) | 1.000 |
| No | 110 (88.7) | 65 (87.8) | 3 (4.6) | 6 (9.2) | 48 (85.7) | 2 (4.2) | |||
| Unknown | 1 (0.8) | 1 (1.4) | 0 (0.0) | – | 0 (0.0) | 1 (1.8) | 0 (0.0) | ||
| Sore throat | |||||||||
| Yes | 32 (25.8) | 27 (36.5) | 2 (7.4) | 0.550 | 3 (11.1) | 0.662 | 19 (33.9) | 1 (5.3) | 1.000 |
| No | 90 (72.6) | 47 (63.5) | 1 (2.1) | 3 (6.4) | 37 (66.1) | 1 (2.7) | |||
| Unknown | 2 (1.6) | – | – | – | – | – | – | – | – |
| Rash | |||||||||
| Yes | 12 (9.7) | 4 (5.4) | 0 (0.0) | 1.000 | 0 (0.0) | 1.000 | 4 (7.1) | 0 (0.0) | 1.000 |
| No | 112 (90.3) | 70 (94.6) | 3 (4.3) | 6 (8.6) | 52 (92.9) | 2 (2.9) | |||
| Muscle pain | |||||||||
| Yes | 43 (34.7) | 40 (54.1) | 1 (2.5) | 0.591 | 1 (2.5) | 0.088 | 36 (64.3) | 1 (2.8) | 1.000 |
| No | 79 (63.7) | 34 (45.9) | 2 (5.9) | 5 (14.7) | 20 (35.7) | 1 (5.0) | |||
| Unknown | 2 (1.6) | 0 (0.0) | – | – | – | – | – | – | – |
| Joint pain | |||||||||
| Yes | 40 (32.3) | 38 (51.4) | 1 (2.6) | 0.610 | 1 (2.6) | 0.103 | 33 (58.9) | 1 (3.0) | 1.000 |
| No | 82 (66.1) | 36 (48.6) | 2 (5.6) | 5 (13.9) | 23 (41.1) | 1 (4.4) | |||
| Unknown | 2 (1.6) | 0 (0.0) | – | – | – | – | – | – | – |
Dry season: November to March and rainy season: April to October.
ELISA (IgM+/IgG−).
Samples for bacterial culture were not obtained from 18 children from 5 to 10 years of age.
P value was obtained using χ2 or Fisher’s exact test. The P value determines the association between the infection status (positive and negative) with the social demographic characteristics and clinical symptoms.
Comparison of coinfections in the malaria positives and negatives with bacterial infection
| Bacterial infection | Total n (%) | Malaria | ||
|---|---|---|---|---|
| Positive | Negative | P value | ||
| n (%) | n (%) | |||
| Leptospirosis | – | – | – | 0.663 |
| Positive | 14 (5.0) | 11 (5.3) | 3 (4.1) | – |
| Negative | 266 (95.0) | 195 (94.7) | 71 (95.5) | – |
| Q fever | – | – | – | 0.398 |
| Positive | 30 (10.7) | 24 (11.7) | 6 (8.1) | – |
| Negative | 250 (89.3) | 182 (88.4) | 68 (91.9) | – |
| Culture* | – | – | – | 0.976 |
| Positive | 8 (3.6) | 6 (3.7) | 2 (3.6) | – |
| Negative | 212 (96.4) | 158 (96.3) | 54 (96.4) | – |
Sixty participants from 5 to 10 years of age did not provide samples for bacterial culture.
Coxiella burnetii and Leptospira spp. detection in paired sera
| C. burnetii | Leptospira spp. | ||||
|---|---|---|---|---|---|
| Antibodies | Initial (n) | Follow-up | Antibodies | Initial (n) | Follow-up (n) |
| IgM(+)IgG(−) | 3 | 3 | IgM(−)IgG(−) | 5 | 5 |
| IgM(+)IgG(−) | 3 | 3 | IgM(+)IgG(−) | 1 | 1 |
| IgM(+)IgG(−) | 0 | 0 | IgM(−)IgG(+) | 1 | 1 |
| IgM(−)IgG(−) | 9 | 9 | IgM(+)IgG(−) | 0 | 0 |
| IgM(−)IgG(+) | 0 | 0 | IgM(−)IgG(−) | 1 | 1 |
| IgM(−)IgG(−) | 35 | 35 | IgM(−)IgG(−) | 42 | 42 |
The first three MCA dimensions explain approximately 35% of the variability among febrile patients with an infection (Supplemental Table 2, Supplemental Figure 1). Variables defining the first dimension were mainly associated with or without sore throat and joint and muscle pain. The most discriminant variables for dimension 2 were age, with or without muscle and joint pain. Variables that were mainly associated with dimension 3 were age, with or without headache, retro-orbital pain, and conjunctivitis.
The clinical presentation of fever without any other symptoms did not correlate specifically to any of the infections based upon its close proximity to the origin in the MCA analysis, suggesting it cannot be used to distinguish between malaria, bacterial infections, and coinfections (Figure 4), although coinfection was associated with headaches and sore throat. The presence of retro-orbital pain, conjunctivitis, joint pain, and muscle pain correlated to malaria alone rather than to malaria with bacterial coinfection (Figures 5 and 6). The MCA analysis confirmed data from Table 2 that bacterial infection did not correlate with headaches (Figure 5). In addition, bacterial infection correlated with rashes but not with retro-orbital, muscle, or joint pain, conjunctivitis, and chills (Figures 5 and 6). The risk of a febrile patient having an infection was not strongly correlated with age, sex, chills or fever (Supplemental Table 2) as compared to sore throat, muscle pain, and joint pain.
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) fever and chills (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) fever and chills (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) fever and chills (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) conjunctivitis, retro-orbital pain, and headache (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) conjunctivitis, retro-orbital pain, and headache (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 (Dim 1, Dim 2) with the active variables (yes/no) conjunctivitis, retro-orbital pain, and headache (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 with the active variables (yes/no) rashes, muscle pain, and joint pain (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 with the active variables (yes/no) rashes, muscle pain, and joint pain (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
Multiple corresponding analysis (MCA) plot in R of dimensions 1 and 2 with the active variables (yes/no) rashes, muscle pain, and joint pain (red) vs. infection as a supplementary variable (green). The plot represents a two-dimensional graph (dimensions 1 and 2) that indicates the significant impact of the active and supplementary variables on the dimensions.
Citation: The American Journal of Tropical Medicine and Hygiene 109, 5; 10.4269/ajtmh.23-0099
DISCUSSION
Although the WHO reports a global decline in malaria since 2010, it remains a significant problem in sub-Saharan Africa. Despite the positive impact of presidents malaria initiative (PMI) and other malaria initiatives, Ghana—a country with one of the highest burdens of malaria—reported an increased incidence of malaria cases in 2018.17 Therefore, it was not surprising that malaria was highly prevalent (68.9%) in this study of AFI patients. Malaria parasite and bacterial bloodstream coinfections were common among patients with severe febrile illness in a review of low- and middle-income countries by Prasad et al.18 In Ghana, malaria coinfections with bacteremia have been reported in children,7,19 but very little information exists about malaria with other zoonotic bacteria. Our results showed 11.7% Q fever, 5.3% leptospirosis, and 3.7% bacterial pathogen coinfection rates with malaria. This study revealed the presence of some bacterial zoonotic and bloodstream infections as contributors to AFIs in our hospitals. A systematic review of diverse zoonoses causing febrile illness in malaria-endemic regions revealed that the majority of the contributing causes of fever were due to bacterial infections.5 We also recorded 4.1% leptospirosis, 8.1% C. burnetii, and 3.6% bacteremia in patients who were malaria negative, identifying these pathogens as contributing to the burden of nonspecific febrile cases in Ghana.
The overall seropositivity against Q fever antibody was 10.7%, similar to 9.6% IgM seropositivity reported in a febrile study in Mali.20 Other reports of seropositivity of < 8% have been recorded in Chad, the Ivory Coast, and Tanzania.21 In Ghana, Q fever infection in febrile patients was scarcely detected with some reports in children under 5 years old.8 The 5% seropositivity of leptospirosis observed in this study was lower than the 11.5% reported in a study among febrile patients in Mozambique.22 Our result is also in the range of 0–33% reported in a review of leptospirosis in sub-Saharan Africa by de Vries et al.23 This study was conducted in two urban cities in Ghana (Accra and Sekondi-Takoradi); hence, our outcome could justify the observation that leptospirosis, which is a well-known rural disease, could also occur in urban settings, which often comprise poor urban slums,24 but at slightly lower rates.
Seroconversion of the initial and follow-up serum samples was seen in both Q fever and leptospirosis infections. Nine serum samples that were initially negative for acute testing were found to be positive (IgM + IgG−) during follow-up testing for Q fever, adding up to the three serum samples that were positive in the initial testing. Likewise, for leptospirosis, seven serum samples were positive (IgM+ IgG−) for the initial testing, one serum sample seroconverted from IgM+ IgG− to IgM− IgG+, with one serum sample remaining IgM− IgG+, confirming that IgG can remain in the bloodstream for a long time. This finding confirms the need for paired sera testing because it is known to take about 7 to 15 days or more after the onset of illness for seroconversion25 Immunoglobulin M antibody for these infections takes a few weeks to develop during acute infection.
Coxiella burnetii and Leptospira DNA were not detected by PCR; however, antibodies were detected by ELISA. Even though serology alone may not prove infection with Q fever and leptospirosis, it has been documented that as antibodies build, nucleic acid detection decreases to levels below the PCR detection limit.26 This may explain some of the discordant results between nucleic acid amplification and ELISA results.
Bacterial bloodstream infection and bacterial zoonoses have been reported as major causes of fever.5 For 220 samples cultured, approximately 2.3% of S. typhi and 0.5% of Salmonella spp., S. aureus, and B. cereus each were identified (Figure 3). Salmonella typhi is known as one of the more frequently isolated blood-borne pathogens in Ghana.27 Most studies have shown high seroprevalence of enteric fever using the Widal tests, which have a high cross-reactivity rate.28 The low prevalence seen in this study is similar to that of some studies reported in Ethiopia,28 Nigeria,29 and Ghana.30 This could be attributed to irregular shedding of the bacteria or to the majority of patients being from urban areas with improved access to clean water and healthcare. Globally S. aureus is known to cause bacteremia, sepsis, and skin infections31 in hospital or community-acquired settings.32 We found a relatively low percentage (0.5%) of S. aureus isolates in this study, similar to a Ghanaian study that reported an isolation rate of 0.6% among febrile patients in rural hospitals.33 Bacillus cereus is commonly associated with gastrointestinal infection; however, it has increasingly been reported as one of the potentially important pathogens causing nosocomial bloodstream infections.34,35 Our study shows culture-proven bacterial coinfections occur, but are rare.
None of the presenting signs and symptoms were significantly associated with a particular clinical disease with use of Fisher’s exact or χ2 test. This demonstrates the nonspecific presentation of malaria and its overlap with the pathogens responsible for AFI.18,36,37 Therefore, signs and symptoms alone are unlikely to discern the cause of AFIs based solely on clinical presentation, highlighting the need for enhanced diagnostic assays.38,39 However, the MCA helped to visualize probable associations among symptoms, single pathogen infections, and coinfections. In our study, we found that febrile patients presenting with retro-orbital pain, conjunctivitis, and joint and muscle pain were more likely to have a malaria infection alone as opposed to malaria with a bacterial coinfection. The MCA proved productive, showing possible links between symptoms and infection. This could be considered an important additional analytical tool when associations of categorical variables are analyzed.
Our findings suggest that patients 5 years of age and older irrespective of having muscle and joint pain correlated with malaria infection rather than malarial-bacterial coinfection. Njeru and others40 mentioned attempts to predict Q fever, malaria, and typhoid infections based on clinical signs and symptoms modeling. However, like our study, most symptoms were not suggestive of a particular infectious etiology and therefore could not be used to increase the pretest probability of any particular infection. This highlights the challenges of identifying the causes of AFIs solely based on clinical presentation in austere setting. It also emphasizes the importance of improving diagnostics to better characterize the distribution of AFI-causing pathogens. In this study, a febrile person’s age and gender, together with experiencing chills or fever, did not strongly correlate with the infections we studied (Supplemental Table 2), unlike a study in Nigeria that reported the highest prevalence of Salmonella infections was found in young adults.29
Our limited findings in this austere environment demonstrate the need for more comprehensive diagnostic modalities—ones that complement the findings of the clinical exam—to better understand the multiple etiologies contributing to AFIs in tropical environments such as Ghana and West Africa. Continuous febrile illness surveillance can help inform local epidemiology and febrile illness management and will be essential for the detection of disease outbreaks.41
LIMITATIONS
This study was subject to limitations that are associated with using commercial ELISA kits and malaria RDT kits. These are prone to cross-reactivity among pathogens. Acquiring follow-up samples for paired testing was a challenge, as mentioned in other febrile illness research.7,36 Another important limitation to our study was the low rate of follow-up, which was mostly attributed to early patient recovery, inability to take time off work, and other unforeseen circumstances. It is understood that children younger than 5 years may be more prone to malarial-bacterial coinfections; however, because of ethical concerns, we could not draw large volumes of blood from this group. Thus, not all tests were performed on their specimens. Another limitation was the site-by-site variation in sample size due to the smaller number of participants enrolled from Accra (n = 71) compared with Takoradi (n = 328).
CONCLUSIONS
Although positive rapid diagnostic tests for malaria are common among Ghanaians from urban settings presenting with fever, about 10% of participants have evidence suggesting at least one other bacterial cause of AFI. Therefore, additional etiologies for fever should be considered when patients present with AFI, even in the setting of a positive malaria test. In general, clinical signs and symptoms were not significantly associated with the presence of coinfections, but retro-orbital pain, conjunctivitis, joint pain, and muscle pain were more likely among those who had a positive malaria rapid diagnostic test and lacked evidence of coinfections.
This study highlights that bloodstream infections, leptospirosis, and Q fever could contribute to AFIs in Ghana. We report that coinfections with these pathogens and malaria could range from 3–11% among febrile patients in Ghana. We also found that fever without any other symptoms could not be used to accurately distinguish between malaria, bacterial infections, and coinfections. Symptoms associated with malaria included retro-orbital pain, conjunctivitis, and joint and muscle pain, whereas malaria coinfection with Q fever or leptospirosis was associated with headaches and sore throat. Laboratory-confirmed bacterial diagnosis of AFIs should be improved because clinical presentation alone is too similar to malaria to discriminate the presence of other coinfections.
Supplemental Materials
ACKNOWLEDGMENTS
We thank the staff of NAMRU-3 Ghana Detachment and the Parasitology Department of the Noguchi Memorial Institute for Medical Research, especially George Boateng-Safo, Kevin Yartey-Yartey, Julian Adinkrah, Francis Adjei, Ronald E. Bentil, and Jeannette Bentum for their support and efforts. We are also grateful to all personnel at the participating sites.
REFERENCES
- 1.↑
Berkley JA, Maitland K, Mwangi I, Ngetsa C, Mwarumba S, Lowe BS, Newton CRJC, Marsh K, Scott JAG, English M, 2005. Use of clinical syndromes to target antibiotic prescribing in seriously ill children in malaria endemic area: observational study. BMJ 330: 995–999.
- 2.↑
Evans JA, Adusei A, Timmann C, May J, Mack D, Agbenyega T, Horstmann RD, Frimpong E, 2004. High mortality of infant bacteraemia clinically indistinguishable from severe malaria. Q J Med 97: 591–597.
- 3.↑
Crump JA, Newton PN, Baird SJ, Lubell Y, 2017. Febrile illness in adolescents and adults. Disease Control Priorities, Third Edition (Volume 6): Major Infectious Diseases. Washington, DC: The World Bank, 365–385.
- 4.↑
Nadjm B et al., 2010. WHO guidelines for antimicrobial treatment in children admitted to hospital in an area of intense Plasmodium falciparum transmission: prospective study. BMJ 340: 848.
- 5.↑
Maze MJ, Bassat Q, Feasey NA, Mandomando I, Musicha P, Crump JA, 2018. The epidemiology of febrile illness in sub-Saharan Africa: implications for diagnosis and management. Clin Microbiol Infect 24: 808–814.
- 6.↑
Dalrymple U, Cameron E, Bhatt S, Weiss DJ, Gupta S, Gething PW, 2017. Quantifying the contribution of Plasmodium falciparum malaria to febrile illness amongst African children. eLife 6: 1–17.
- 7.↑
Hogan B et al., 2018. Malaria coinfections in febrile pediatric inpatients: a hospital-based study from Ghana. Clin Infect Dis 66: 1838–1845.
- 8.↑
Kobbe R, Kramme S, Kreuels B, Adjei S, Kreuzberg C, Panning M, Adjei O, Fleischer B, May J, 2008. Q fever in young children, Ghana. Emerg Infect Dis 14: 344–346.
- 9.↑
Folitse RD, Opoku-Agyemang T, Amemor E, Opoku ED, Bentum KE, Emikpe BO, 2019. Serological evidence of Coxiella burnetii infection in slaughtered sheep and goats at Kumasi Abattoir, Ghana. J Immunoassay Immunochem 41: 152–157.
- 10.↑
Johnson SAM, Kaneene JB, Asare-Dompreh K, Tasiame W, Mensah IG, Afakye K, Simpson SV, Addo K, 2019. Seroprevalence of Q fever in cattle, sheep and goats in the Volta region of Ghana. Vet Med Sci 5: 402–411.
- 11.↑
Herlihy JM, D’Acremont V, Burgess DCH, Hamer DH, 2016. Diagnosis and treatment of the febrile child. Disease Control Priorities, Third Edition (Volume 2): Reproductive, Maternal, Newborn, and Child Health. Washington, DC: The World Bank, 137–161.
- 12.↑
Cusabio Technologies Human Leptospira antibody [IgM] ELISA Kit Catalog Number. CSB-E05026h. Wuhan, Hubei Province, China.
- 14.↑
Swanson EC, Collins MT, 1980. Use of the API 20E system to identify veterinary Enterobacteriaceae. J Clin Microbiol 12: 10–14.
- 15.↑
Klee SR, Ellerbrok H, Tyczka J, Franz T, Appel B, 2006. Evaluation of a real-time PCR assay to detect Coxiella burnetii .Ann N Y Acad Sci 1078: 563–565.
- 16.↑
Palaniappan RUM et al., 2005. Evaluation of lig-based conventional and real time PCR for the detection of pathogenic leptospires. Mol Cell Probes 19: 111–117.
- 17.↑
World Health Organization , 2019. The “World Malaria Report 2019” at a Glance. Available at: https://www.who.int/news-room/feature-stories/detail/world-malaria-report-2019. Accessed.
- 18.↑
Prasad N, Murdoch DR, Reyburn H, Crump JA, 2015. Etiology of severe febrile illness in low- and middle-income countries: a systematic review. PLoS One 10: e0127962.
- 19.↑
Church J, Maitland K, 2014. Invasive bacterial co-infection in African children with Plasmodium falciparum malaria: a systematic review. BMC Med 12: 1–17.
- 20.↑
Steinmann P, Bonfoh B, Peter O, Schelling E, Traore M, Zinsstag J, 2005. Seroprevalence of Q-fever in febrile individuals in Mali. Trop Med Int Health 10: 612–617.
- 21.↑
Vanderburg S, Rubach MP, Halliday JEB, Cleaveland S, Reddy EA, Crump JA, 2014. Epidemiology of Coxiella burnetii infection in Africa: a OneHealth systematic review. PLoS Negl Trop Dis 8: e2787.
- 22.↑
Ribeiro P et al., 2017. Seroepidemiology of leptospirosis among febrile patients in a rapidly growing suburban slum and a flood-vulnerable rural district in Mozambique, 2012–2014: implications for the management of fever. Int J Infect Dis 64: 50–57.
- 23.↑
de Vries SG, Visser BJ, Nagel IM, Goris MGA, Hartskeerl RA, Grobusch MP, 2014. Leptospirosis in Sub-Saharan Africa: a systematic review. Int J Infect Dis 28: 47–64.
- 24.↑
Wasinski B, Dutkiewicz J, 2013. Leptospirosis-current risk factors connected with human activity and the environment. Ann Agric Environ Med. 20: 239–244. Available at: agro.icm.edu.pl.
- 25.↑
Anderson A et al., 2013. Diagnosis and management of Q fever - United States, 2013: recommendations from CDC and the Q Fever Working Group. MMWR Recomm Rep 62: 1–29.
- 26.↑
Schneeberger PM, Hermans MHA, van Hannen EJ, Schellekens JJA, Leenders ACAP, Wever PC, 2010. Real-time PCR with serum samples is indispensable for early diagnosis of acute Q fever. Clin Vaccine Immunol 17: 286–290.
- 27.↑
Andoh LA, Ahmed S, Olsen JE, Obiri-Danso K, Newman MJ, Opintan JA, Barco L, Dalsgaard A, 2017. Prevalence and characterization of Salmonella among humans in Ghana. Trop Med Health 45: 3.
- 28.↑
Deksissa T, Gebremedhin EZ, 2019. A cross-sectional study of enteric fever among febrile patients at Ambo Hospital: prevalence, risk factors, comparison of Widal test and stool culture and antimicrobials susceptibility pattern of isolates. BMC Infect Dis 19: 288.
- 29.↑
Ishaleku D, Sar T, Houmsou R, 2011. Salmonella typhi infection in a tertiary institution in Nasarawa State, Nigeria. Asian Pac J Trop Med 4: 137–139.
- 30.↑
Labi AK, Obeng-Nkrumah N, Addison NO, Donkor ES, 2014. Salmonella blood stream infections in a tertiary care setting in Ghana. BMC Infect Dis 14: 1–10.
- 31.↑
Shittu A, Oyedara O, Abegunrin F, Okon K, Raji A, Taiwo S, Ogunsola F, Onyedibe K, Elisha G, 2012. Characterization of methicillin-susceptible and -resistant staphylococci in the clinical setting: a multicentre study in Nigeria. BMC Infect Dis 12: 1–10.
- 32.↑
Youssef D, Molony K. & Enany S Frontiers in Staphylococcus aureus. London, United Kingdom: InTechOpen.
- 33.↑
Dekker D et al., 2016. Antibiotic resistance and clonal diversity of invasive Staphylococcus aureus in the rural Ashanti Region, Ghana. BMC Infect Dis 16: 720.
- 35.↑
Ikeda M, Yagihara Y, Tatsuno K, Okazaki M, Okugawa S, Moriya K, 2015. Clinical characteristics and antimicrobial susceptibility of Bacillus cereus blood stream infections. Ann Clin Microbiol Antimicrob 14: 43.
- 36.↑
Rhee C, Kharod GA, Schaad N, Furukawa NW, Vora NM, Blaney DD, Crump JA, Clarke KR, 2019. Global knowledge gaps in acute febrile illness etiologic investigations: a scoping review. PLoS Negl Trop Dis 13: e0007792.
- 37.↑
Crump JA, Kirk MD, 2015. Estimating the burden of febrile illnesses. PLoS Negl Trop Dis 9: e0004040.
- 38.↑
Roddy P et al., 2019. Quantifying the incidence of severe-febrile- illness hospital admissions in sub-Saharan Africa. PLoS One 14: 30220371.
- 39.↑
Maze MJ, Bassat Q, Feasey NA, Mandomando I, Musicha P, Crump JA, 2019. The epidemiology of febrile illness in sub-Saharan Africa: implications for diagnosis and management. Clin Microbiol Infect 24: 808–814.
- 40.↑
Njeru J, Henning K, Pletz MW, Heller R, Forstner C, Kariuki S, Fèvre EM, Neubauer H, 2016. Febrile patients admitted to remote hospitals in Northeastern Kenya: seroprevalence, risk factors and a clinical prediction tool for Q-Fever. BMC Infect Dis 16: 244.
- 41.↑
Halliday JEB et al., 2020. Zoonotic causes of febrile illness in malaria endemic countries: a systematic review. Lancet Infect Dis 20: e27–e37.