INTRODUCTION
Acute febrile illness (AFI) is one of the most common reasons for hospitalization and emergency department visits globally.1,2 Febrile illnesses can have a wide range of etiologies, including bacterial, parasitic, and viral infections, as well as systemic conditions.3–5 The ability to correctly diagnose the cause of fever has critical implications for both patient management and public health because accurate diagnosis can ensure that proper treatment is administered in a timely fashion and can limit further spread of infectious diseases.1,2,6,7 However, it is common for the etiologies of AFI in hospitalized patients to remain unidentified or to be misdiagnosed, especially in low- and middle-income countries and settings with limited laboratory capacity, in part due to the wide range of causes of febrile illnesses and similarity of symptoms.2,4,7
Inaccurate diagnosis can lead to the overuse of antimicrobials, contributing to the growing issue of antimicrobial resistance (AMR) in Thailand and Southeast Asia.8,9 In Thailand, over-prescription of antibiotics in health facilities is common, and misuse of antibiotics also occurs due to the availability of antimicrobials over-the-counter in pharmacies.10–14 Additionally, health literacy about antibiotics and AMR is generally low among adults in Thailand,15,16 and rates of inappropriate antibiotic use are high.10,14
Acute febrile illness can present without localized manifestations such as respiratory or diarrheal signs and symptoms, in which case it is referred to as acute undifferentiated febrile illness (AUFI), which can be particularly difficult to diagnose.1,17 In Thailand and Southeast Asia, the numerous infectious etiologies of AUFI include, but are not limited to, dengue virus, chikungunya virus, Zika virus, Salmonella enterica, Burkholderia pseudomallei, Orientia tsutsugamushi (the cause of scrub typhus), pathogenic Leptospira species, and Rickettsia species.2,5,18–23 Due to the high prevalence of dengue fever in the region, AUFI is frequently misdiagnosed as dengue, which can impede the proper management of illness and delay potentially lifesaving treatment.19,24
Several previous studies have investigated clinical and hematological indicators of bacterial versus nonbacterial infections but few have assessed patients’ risk factors and exposure histories or focused specifically on AUFI.6,25–29 The majority of previous studies in Thailand have analyzed clinical and laboratory factors associated with specific disease diagnoses, such as comparing dengue to chikungunya, scrub typhus, or typhoid.30–34 To our knowledge, no studies in Thailand have compared clinical predictors and exposure histories among patients diagnosed with bacterial and nonbacterial febrile illnesses. An improved ability to determine whether an undifferentiated illness is bacterial or nonbacterial in origin can hasten proper management, which may have three primary results: 1) improve patient outcomes, 2) reduce the chance of further transmission of infectious disease in the surrounding community, and 3) reduce the unnecessary use of antimicrobial treatments. This analysis aims to assess sociodemographic, clinical, and risk factors associated with bacterial and nonbacterial infections in AUFI patients in two border provinces of Thailand (Nakhon Phanom, borders Laos; Tak, borders Myanmar, Figure 1).
Study sites in Nakhon Phanom and Take provinces, Thailand.
Citation: The American Journal of Tropical Medicine and Hygiene 111, 3; 10.4269/ajtmh.23-0731
Results from this analysis may help to increase local knowledge of the characteristics of patients presenting to hospitals with bacterial and nonbacterial causes of AUFI. Findings may aid in the proper screening and treatment of fever in Nakhon Phanom and Tak based on demographic, exposure, and clinical risk factors.
MATERIALS AND METHODS
Project design and study population.
From April 2017 to May 2020, a prospective observational study was conducted among patients aged 2–80 years hospitalized with AFI in 12 healthcare facilities in Nakhon Phanom and Tak (five hospitals along the borders) provinces (Supplemental Table 1). Acute febrile illness was defined as fever (temperature ≥38°C upon admission) or history of fever (subjective or measured) with onset ≤7 days before admission. Patients were excluded if they were returning to the hospital for continuation of treatment of fever within 30 days, had been admitted to any hospital in the previous 14 days, or could not read or understand Thai, Lao, Burmese, or Karen languages. Patients with AFI who had evidence of fewer than two clinical respiratory signs or symptoms (defined as sore throat, rhinitis, cough, difficulty breathing, and sputum production) and no evidence of diarrheal disease (defined as a clinical diagnosis of diarrhea or three or more loose or watery stools within the previous 24 hours reported by the patient or guardian) were categorized as having AUFI. This analysis was restricted to patients with AUFI.
Data and specimen collection.
Written informed consent was obtained from all patients aged ≥18 years and from guardians of patients <18 years of age, and assent was obtained from children aged 7–17 years. Project staff conducted interviews and reviewed medical records to collect information on patient demographics, clinical signs and symptoms, exposure history, underlying comorbidities, laboratory results, and clinical course during hospital stays. All AUFI patients had blood and urine samples collected within 24 hours of hospital admission, and participants ≥18 years with any respiratory symptoms also had sputum specimens collected, if possible.
Laboratory diagnostics.
Blood samples were tested for bacterial pathogens using an automated blood culture system (BD BACTECTM FX, BD Franklin Lakes, NJ) and automated identification and susceptibility testing system (BD PhoenixTM, BD). Sputum and urine were cultured for bacterial pathogens using conventional methods. Rapid diagnostic tests (RDT) were used to test urine samples for Streptococcus pneumoniae (Binax NOW® S. pneumoniae Antigen Card, Abbott, Chicago, IL), and blood samples for dengue (SD Bioline Dengue Duo, Abbott), and malaria (Humasis Malaria Pf/Pan Antigen Test, Humasis, York, United Kingdom). In addition, singleplex real-time polymerase chain reaction assays (RT-qPCR, 7500 Real-Time PCR System, Thermo Fisher Scientific, Waltham, MA) were used to detect bacterial and viral pathogens known to cause febrile illness in Thailand as well as specific dengue serotypes. Supplemental Table 2 details the PCR targets for each pathogen.
Definitions.
Only individuals with positive laboratory results were considered for this analysis. The following definitions were used:
- Dengue virus infections—NS1 antigen positive and/or IgM positive by RDT, and/or dengue serotype positive by RT-qPCR
- Malaria infections—RDT positive for Plasmodium falciparum/pan Plasmodium
- Chikungunya virus, Zika virus, pathogenic Leptospira spp., O. tsutsugamushi, Rickettsia rickettsii, and pan-Rickettsia infections—RT-qPCR or qPCR positive
- Escherichia coli, Klebsiella pneumoniae, Streptococcus pyogenes, Streptococcus agalactiae, Staphylococcus aureus, Acinetobacter baumannii, Streptococcus suis, Haemophilus influenzae, Pseudomonas aeruginosa infections: hemoculture, urine or sputum culture, and/or qPCR positive
- S. pneumoniae infections—BINAX RDT positive, hemoculture, urine or sputum culture positive
- B. pseudomallei infections—hemoculture, urine or sputum culture, and/or qPCR positive
- Japanese encephalitis virus infection—IgM positive by ELISA
- Other Streptococcus spp. infection—hemoculture, urine or sputum culture positive
- Other bacterial infection—any other bacterial pathogens detected from hemoculture, urine culture, sputum culture, and/or qPCR
For this analysis, patients with laboratory-confirmed evidence of bacterial or nonbacterial infections:
Bacterial infections included laboratory-positive results for leptospirosis (Leptospira spp.), scrub typhus (O. tsutsugamushi), rickettsiosis (Rickettsia spp.), melioidosis (B. pseudomallei), E. coli, K. pneumoniae, S. pyogenes, S. agalactiae, S. aureus, A. baumannii, S. suis, H. influenzae, P. aeruginosa, other Streptococcus spp., and other bacterial detections.
- a.Patients with laboratory-confirmed multiple bacterial pathogens, in the absence of nonbacterial pathogens, were classified as having bacterial infection.
Nonbacterial infections included laboratory-positive results for dengue virus, chikungunya virus, malaria, Zika virus, and Japanese encephalitis virus.
- a.Patients with laboratory-confirmed multiple nonbacterial pathogens, in the absence of bacterial pathogens, were classified as having nonbacterial infection.
Patients with laboratory-confirmed both bacterial and nonbacterial pathogens were excluded from the analysis.
STATISTICAL ANALYSES
Descriptive analyses were performed for the distribution of pathogens detected and the demographic characteristics of patients with bacterial or nonbacterial infections. Median and interquartile range (IQR) were applied for continuous variables with non-normal distribution. Chi-square tests were performed to compare the proportion of laboratory-confirmed evidence for pathogen between Nakhon Phanom and Tak provinces. Statistical significance was set at P ≤0.05.
Categorical variables that were similar in nature (e.g., contact with cows/contact with pigs/contact with goats/contact with sheep; contact with febrile household member/contact with febrile coworker/contact with febrile neighbor) and showing the same direction of association in bivariate models were combined to reduce the number of predictors included in multivariable models. Comorbidities identified from medical records (including diabetes, hypertension, heart disease, asthma, chronic obstructive pulmonary disease, cancer, HIV, immunodeficiency, history of tuberculosis, active tuberculosis, liver disease, thyroid disease, thalassemia, anemia, chronic renal disease, and other chronic disease) were combined into one variable specifying the presence of any comorbidities.
Simple logistic regressions were conducted to assess the predictive effects of demographic variables, exposure history, symptoms, antibiotic use, and comorbidities on bacterial versus nonbacterial infections in both provinces combined. All variables with bivariate significance of P ≤0.2 were considered for inclusion into a multivariable model. Multicollinearity among predictor variables was assessed through the variance inflation factor (VIF) with a cutoff value of ≤5.35 A multivariable logistic regression model was fit using backward selection, retaining variables with P ≤0.05. Odds ratios (ORs), adjusted ORs (aORs), and their corresponding 95% CIs were calculated for variables included in the multivariable model.
To assess whether findings may have been driven by commonly detected pathogens, bivariate and multivariable logistic regression analyses were conducted for the most detected bacterial and nonbacterial pathogens. We compared characteristics of individuals with any evidence of each common pathogen (regardless of whether multiple bacterial or nonbacterial pathogens were identified) to those with no evidence to the pathogen of interest. Results were compared with the primary multivariable analysis of bacterial versus nonbacterial infections.
Clinical outcomes and severity of illness were assessed using simple logistic regression to examine associations between days of hospitalization, intubation/mechanical ventilation, and discharge status among patients with bacterial versus nonbacterial infections. Associations were further explored in multivariable models including a priori adjustments for age and the presence of comorbidities.
A sensitivity analysis was performed to examine differences between Nakhon Phanom and Tak: bivariate and multivariable logistic regressions were performed for each province separately to assess whether associations appreciably differed by province. An additional sensitivity analysis was conducted by performing multivariable regression stratified by age group.
All data were analyzed using SAS Version 9.4 (Cary, NC).
RESULTS
During the study period, 21,972 patients with AFI were eligible to participate in the study, and 11,274 (51.3%) consented. Among patients who did not consent, 31% (6,832/21,972) refused to participate, and we were not allowed to enroll children who did not have parents available to provide consent, 16% (3,639/21,972) (Figure 2). Overall, 2,913 (25.8%) of consenting patients presented with AUFI (N = 1,881 in Nakhon Phanom and N = 1,032 in Tak), and 1,326 (45.5%) of these patients had specimens with laboratory-confirmed evidence of one or more pathogens. Among individuals with bacterial or nonbacterial infections, 63 (4.8%) tested positive for both bacterial and nonbacterial pathogens and were excluded, resulting in 1,263 individuals included in the analysis (Figure 2).
Participant inclusion chart.
Citation: The American Journal of Tropical Medicine and Hygiene 111, 3; 10.4269/ajtmh.23-0731
Participant characteristics.
Demographic and clinical characteristics of AUFI patients are shown in Table 1. The median age of AUFI patients was 42 years (IQR: 20–60). Of 1,263 patients, the majority (93.9%) were of Thai nationality, and farming was the most common occupation (31.8%). Patients had an average of 2.1 days of fever (SD: 1.7) before presenting to the hospital, and the most common symptoms in addition to fever were fatigue (86.4%), chills (77.0%), and muscle pain (71.6%). A total of 31.8% of patients reported one or more underlying comorbidities, with hypertension (15.4%), diabetes (13.3%), and chronic renal disease (8.6%) being the most frequently reported among participants.
Demographic and clinical characteristics of patients with acute undifferentiated fever, Nakhon Phanom and Tak provinces, Thailand, April 2017–May 2020
| Characteristics | Overall (N = 1,263) | Bacterial Infections (n = 528) | Non-bacterial Infections (n = 735) | Bacterial vs. Nonbacterial Infections | ||||
|---|---|---|---|---|---|---|---|---|
| n | % | n | % | n | % | OR (95% CI) | P-Value | |
| Province of Hospital | ||||||||
| Nakhon Phanom | 720 | 57.0 | 386 | 73.1 | 334 | 45.4 | 3.26 (2.57–4.15) | <0.0001 |
| Tak | 543 | 43.0 | 142 | 26.9 | 401 | 54.6 | – | – |
| Sex | ||||||||
| Male | 620 | 49.1 | 261 | 49.4 | 359 | 48.8 | 0.98 (0.78–1.22) | 0.84 |
| Female | 643 | 50.9 | 267 | 50.6 | 376 | 51.2 | – | – |
| Age, years | <0.0001 | |||||||
| 2–17 | 283 | 22.4 | 49 | 9.3 | 234 | 31.8 | 0.54 (0.38–0.78) | <0.01 |
| 18–49 | 550 | 43.6 | 153 | 29.0 | 397 | 54.0 | Ref. | Ref. |
| ≥50 | 430 | 34.1 | 326 | 61.7 | 104 | 14.2 | 8.13 (6.09–10.86) | <0.0001 |
| Nationality | ||||||||
| Thai | 1187 | 93.9 | 501 | 94.9 | 686 | 93.3 | 1.33 (0.82–2.15) | 0.25 |
| Other | 76 | 6.0 | 27 | 5.1 | 49 | 6.7 | – | – |
| Employment Status | <0.0001 | |||||||
| Employed* | 791 | 62.6 | 358 | 67.8 | 433 | 58.9 | Ref. | Ref. |
| Student/Preschool Student | 311 | 24.6 | 52 | 9.9 | 259 | 35.2 | 0.24 (0.18–0.34) | <0.0001 |
| Unemployed | 161 | 12.8 | 118 | 22.4 | 43 | 5.9 | 3.32 (2.28–4.84) | <0.0001 |
| Year of Study | <0.001 | |||||||
| Year 1 (April 2017–March 2018) | 382 | 30.3 | 183 | 34.7 | 199 | 27.1 | 2.10 (1.60–2.76) | <0.001 |
| Year 2 (April 2018–March 2019) | 345 | 27.3 | 182 | 34.5 | 163 | 22.2 | 2.56 (1.96–3.38) | <0.001 |
| Year 3 (April 2019–May 2020) | 536 | 42.4 | 163 | 30.9 | 373 | 50.8 | Ref. | Ref. |
| Hospital Type | ||||||||
| Provincial | 512 | 40.5 | 212 | 40.2 | 300 | 40.8 | – | – |
| District | 751 | 59.5 | 316 | 59.9 | 435 | 59.2 | 1.03 (0.82–1.29) | 0.81 |
| Contact with Febrile Household Member, Coworker, or Neighbor | 572 | 45.3 | 159 | 30.1 | 413 | 56.2 | 0.34 (0.27–0.43) | <0.0001 |
| Contact with Farm Animal (Cow, Pig, Goat, or Sheep) | 348 | 27.6 | 169 | 32.0 | 179 | 24.4 | 1.46 (1.14–1.88) | <0.01 |
| Contact with Chicken or Duck | 656 | 51.9 | 323 | 61.2 | 333 | 45.3 | 1.90 (1.52–2.39) | <0.0001 |
| Contact with Cats or Dogs | 800 | 63.3 | 341 | 64.6 | 459 | 62.5 | 1.10 (0.87–1.38) | 0.44 |
| Contact with Rodents | 592 | 46.9 | 288 | 54.6 | 304 | 41.4 | 1.70 (1.36–2.13) | <0.0001 |
| Contact with Stray Animal† | 118 | 9.3 | 63 | 11.9 | 55 | 7.5 | 1.68 (1.15–2.45) | <0.01 |
| Bitten by Mosquitoes or Other Insects | 1244 | 98.5 | 516 | 97.7 | 728 | 99.1 | 0.41 (0.16–1.06) | 0.07 |
| Contact with Flood Water, Mud, Ponds, or Rivers | 435 | 34.4 | 188 | 35.6 | 247 | 33.6 | 1.09 (0.86–1.38) | 0.46 |
| Cut or Scraped Self | 179 | 14.2 | 68 | 12.9 | 111 | 15.1 | 0.83 (0.60–1.15) | 0.26 |
| Cut Down Trees, Bushes, Gathered Wood, and/or Cleared Land | 336 | 26.6 | 157 | 29.7 | 179 | 24.4 | 1.31 (1.02–1.69) | 0.03 |
| Ate raw or Undercooked Fish or Pork | 319 | 25.3 | 153 | 29.0 | 166 | 22.6 | 1.40 (1.08–1.81) | 0.01 |
| Visited Forest | 386 | 30.6 | 158 | 29.9 | 228 | 31.0 | 0.95 (0.75–1.21) | 0.68 |
| Visited or Worked with Rubber Trees | 145 | 11.5 | 65 | 12.3 | 80 | 10.9 | 1.15 (0.81–1.63) | 0.43 |
| Walked Outside with no Shoes | 413 | 32.7 | 188 | 35.6 | 225 | 30.6 | 1.25 (0.99–1.59) | 0.06 |
| Antibiotics Taken within 72 hours before Hospital Presentation | ||||||||
| Yes‡ | 152 | 12.0 | 79 | 15.0 | 73 | 9.9 | 1.58 (1.13–2.22) | <0.01 |
| No | 1073 | 85.0 | 436 | 82.6 | 637 | 86.7 | Ref. | Ref. |
| Unknown | 7 | 0.6 | 5 | 1.0 | 2 | 0.3 | – | – |
| Not Asked | 31 | 2.5 | 8 | 1.5 | 23 | 3.1 | – | – |
| Days of Fever Before Hospital Admission | ||||||||
| 0–1 | 425 | 33.7 | 214 | 40.5 | 211 | 28.7 | 1.50 (1.15–1.95) | <0.0001 |
| 2–3 | 483 | 38.2 | 195 | 36.9 | 288 | 39.2 | Ref. | Ref. |
| 4–7 | 355 | 28.1 | 119 | 22.5 | 236 | 32.1 | 0.75 (0.56–0.99) | <0.01 |
| Symptoms | ||||||||
| Cough | 141 | 11.2 | 63 | 11.9 | 78 | 10.6 | 1.14 (0.80–1.62) | 0.46 |
| Rhinitis | 37 | 2.9 | 14 | 2.7 | 23 | 3.1 | 0.84 (0.43–1.65) | 0.62 |
| Sore Throat | 64 | 5.1 | 25 | 4.7 | 39 | 5.3 | 0.89 (0.53–1.49) | 0.65 |
| Shortness of Breath/Difficulty Breathing | 137 | 10.9 | 76 | 14.4 | 61 | 8.3 | 1.89 (1.30–2.66) | <0.01 |
| Nausea or Vomiting | 702 | 55.6 | 253 | 47.9 | 449 | 61.1 | 0.59 (0.47–0.74) | <0.0001 |
| Yellow Eyes or Skin | 66 | 5.2 | 45 | 8.5 | 21 | 2.9 | 3.17 (1.86–5.39) | <0.0001 |
| Headache | 951 | 75.3 | 358 | 67.8 | 593 | 80.7 | 0.50 (0.39–0.65) | <0.0001 |
| Blood in Urine, Stool, or Vomit | 68 | 5.4 | 30 | 5.7 | 38 | 5.2 | 1.11 (0.68–1.81) | 0.69 |
| Muscle Pain | 904 | 71.6 | 360 | 68.2 | 544 | 74.0 | 0.75 (0.59–0.96) | 0.02 |
| Chest Pain | 190 | 15.0 | 93 | 17.6 | 97 | 13.2 | 1.41 (1.03–1.92) | 0.03 |
| Bone or Joint Pain | 550 | 43.6 | 227 | 43.0 | 323 | 44.0 | 0.96 (0.77–1.21) | 0.74 |
| No Appetite | 947 | 75.0 | 378 | 71.6 | 569 | 77.4 | 0.74 (0.57–0.95) | 0.02 |
| Tiredness, No Energy | 1091 | 86.4 | 460 | 87.1 | 631 | 85.9 | 1.12 (0.80–1.55) | 0.52 |
| Seizures | 30 | 2.4 | 14 | 2.7 | 16 | 2.2 | 1.22 (0.59–2.53) | 0.59 |
| Chills | 972 | 77.0 | 421 | 79.7 | 551 | 75.0 | 1.31 (1.00–1.72) | 0.05 |
| Pale or Cold Skin | 429 | 34.0 | 210 | 39.8 | 219 | 29.8 | 1.56 (1.23–1.97) | <0.01 |
| Rash | 239 | 18.9 | 38 | 7.2 | 201 | 27.4 | 0.21 (0.14–0.30) | <0.0001 |
| Bruises | 27 | 2.1 | 11 | 2.1 | 16 | 2.2 | 0.96 (0.44–2.08) | 0.91 |
| Any Existing Comorbidities§ | 402 | 31.8 | 295 | 55.9 | 107 | 14.6 | 7.43 (5.69–9.71) | <0.0001 |
| Current Smoking | 184 | 14.6 | 95 | 18.0 | 89 | 12.1 | 1.59 (1.16–2.18) | <0.01 |
| Current Alcohol Consumption | 250 | 19.8 | 107 | 20.3 | 143 | 19.5 | 1.05 (0.80–1.39) | 0.72 |
– = not included in the analysis; OR = odds ratio; Ref. = reference. Bold value indicates P <0.05.
Used occupations included farmer, healthcare personnel, government/office worker, merchant, monk, housekeeper, and laborer.
Streets or homeless animal.
Of 152 patients, 27 (17.8%) patients aged 2–17 years, 73 (48.0%) aged 18–49 years, and 52 (34.2%) aged ≥50 years.
Comorbidities included diabetes, hypertension, heart disease, asthma, chronic obstructive pulmonary disease, cancer, HIV, immunodeficiency, history of tuberculosis, active tuberculosis, liver disease, thyroid disease, thalassemia, anemia, chronic renal disease, and other chronic disease.
Laboratory results.
Among 1,263 AUFI patients, 528 (41.8%) tested positive for one or more bacterial pathogens, and 735 (58.2%) tested positive for one or more nonbacterial pathogens. In both provinces, E. coli was the most common bacterial pathogen detected, identified in 160 (12.7%) patients, and dengue virus was the most common nonbacterial pathogen detected, identified in 563 (44.6%) patients (Table 2). Thirty-six patients (2.9%) tested positive for multiple bacterial pathogens, and 115 patients (9.1%) tested positive for multiple nonbacterial pathogens (Supplemental Table 3).
Distribution of laboratory-confirmed diagnoses in Nakhon Phanom and Tak provinces, Thailand, April 2017–May 2020
| Pathogen | Nakhon Phanom (n = 720) | Tak (n = 543) | P-Value | Total (N = 1,263) |
|---|---|---|---|---|
| n (%)* | n (%)* | n (%)* | ||
| Bacterial Detections† (N = 528; Nakhon Phanom: n = 386, Tak: n = 142) | ||||
| Orientia tsutsugamushi | 4 (0.6) | 32 (5.9) | <0.0001 | 36 (2.9) |
| Pathogenic Leptospira Species | 44 (6.1) | 27 (5.0) | 0.38 | 71 (5.6) |
| B. pseudomallei | 26 (3.6) | 0 (0) | <0.0001 | 26 (2.1) |
| Rickettsia spp. | 59 (8.2) | 10 (1.8) | <0.0001 | 69 (5.5) |
| Escherichia coli | 119 (16.5) | 41 (7.6) | <0.0001 | 160 (12.7) |
| Krebseilla pneumoniae | 27 (3.8) | 3 (0.6) | <0.01 | 30 (2.4) |
| Streptococcus agalactiae | 13 (1.8) | 1 (0.2) | <0.01 | 14 (1.1) |
| Streptococcus pneumoniae | 9 (1.3) | 9 (1.7) | 0.55 | 18 (1.4) |
| Streptococcus pyogenes | 3 (0.4) | 1 (0.2) | 0.64‡ | 4 (0.32) |
| Streptococcus aureus | 20 (2.8) | 7 (1.3) | 0.07 | 27 (2.1) |
| Acinetobacter baumannii | 4 (0.6) | 0 (0) | 0.14‡ | 4 (0.3) |
| Streptococcus suis | 3 (0.4) | 0 (0) | 0.26‡ | 3 (0.2) |
| Haemophilus influenza | 1 (0.1) | 0 (0) | 1.00‡ | 1 (0.1) |
| Psedomonas aeruginosa | 2 (0.3) | 0 (0) | 0.51‡ | 2 (0.2) |
| Other Streptococcus Species | 12 (1.7) | 3 (0.6) | 0.07 | 15 (1.2) |
| Other Bacteria (Culture Signaled Positive) | 70 (9.7) | 16 (3.0) | <0.0001 | 86 (6.8) |
| Multiple Bacterial Detections | 29 (4.0) | 7 (1.3) | <0.01 | 36 (2.9) |
| Nonbacterial Detections (N = 735; Nakhon Phanom: n = 334, Tak: n = 401) | ||||
| Dengue Virus | 288 (40.0) | 275 (50.6) | <0.01 | 563 (44.6) |
| Chikungunya Virus | 28 (3.9) | 96 (17.7) | <0.0001 | 124 (9.8) |
| Malaria (Plamodium falciparum/Pan Plasmodium) | 0 (0) | 30 (2.4) | <0.0001 | 26 (2.1) |
| Zika Virus | 1 (0.1) | 2 (0.4) | 0.58‡ | 3 (0.2) |
| Japanese Encephalitis Virus§ | 66 (9.2) | 65 (12.0) | 0.11 | 131 (10.4) |
| Multiple Nonbacterial Detections | 49 (6.8) | 66 (12.2) | <0.01 | 115 (9.1) |
Percentages do not add up to 100% because of cases with more than one bacterial or nonbacterial detection.
R. rikettsii were not detected.
Calculated using Fisher’s exact test.
Cross-reactivity between Japanese encephalitis and dengue IgM was not accounted for because both were considered nonbacterial detections.
Demographic characteristics and exposure histories differed for patients with laboratory-confirmed evidence for the three most common pathogens: dengue virus, E. coli, and chikungunya virus. Notably, dengue virus was found in 70.0% (198/283) of children aged 2–17 years, and E. coli was found in 28.6% (123/430) of individuals aged ≥50. Chikungunya virus comprised 17.7% (96/654) of all laboratory-confirmed nonbacterial infection in Tak province, although it was only found during year 3 of the study (Supplemental Table 4).
The proportion of bacterial infections was higher in Nakhon Phanom (53.6%, 386/720) than in Tak (26.2%, 142/543; P <0.0001). Burkholderia pseudomallei, A. baumannii, S. suis, H. influenzae, and P. aeruginosa were only found in Nakhon Phanom, and malaria was only found in Tak. Of patients with bacterial infections, B. pseudomallei, Rickettsia spp., E. coli, K. pneumoniae, S. agalactiae, and other bacteria were significantly more common in Nakhon Phanom (P <0.05), and O. tsutsugamushi was more common in Tak (P <0.0001). Of patients with nonbacterial detections, dengue virus, chikungunya virus, and malaria were more common in Tak (P <0.01; Table 2).
Bivariate analysis.
Factors associated with infections type are shown in Table 1. Age ≥50 years was associated with bacterial infections (P <0.01), and the youngest age group (2–17 years) was associated with the nonbacterial infections (P <0.0001) compared with the 18- to 49-year age group. Compared with individuals who were employed, students and preschoolers had greater odds of nonbacterial infections, and unemployed individuals had greater odds of bacterial infection (P <0.0001). Bacterial and nonbacterial infections did not significantly differ by sex, nationality, or type of hospital (provincial or district).
Bacterial infection was associated with exposure to farm animals, poultry, rodents, and stray animals, as well as a recent history of cutting down trees and eating raw or undercooked fish or pork (P <0.05). Contact with febrile household members, coworkers, or neighbors was associated with nonbacterial infection (P <0.0001).
Bacterial infections were more common among individuals who had taken antibiotics in the 72 hours before presentation to the hospital (P <0.01). Shortness of breath, jaundice, chest pain, chills, pallor, fewer days of fever before hospital admission, presence of comorbidities, and smoking were associated with bacterial infection, whereas nausea and/or vomiting, headache, muscle pain, decreased appetite, and rash were associated with nonbacterial infection (P <0.05).
Multivariable analysis.
Results from the multivariable analysis are shown in Table 3. The VIF for all variables was <5, and multicollinearity was determined to be unlikely to impact the multivariable model. Patients in Nakhon Phanom province had greater odds of bacterial infection than patients in Tak (aOR: 2.82, 95% CI: 2.02–3.93). Bacterial infection was independently associated with age ≥50 years (aOR: 4.18, 95% CI: 2.85–6.14), unemployment (aOR: 1.68, 95% CI: 1.01–2.79), study years 1 (aOR: 2.92, 95% CI: 2.01–4.24) and 2 (aOR: 3.30, 95% CI: 2.25–4.82), contact with farm animals (aOR: 1.82, 95% CI: 1.29–2.57), antibiotic use within 72 hours prior to hospital presentation (aOR: 2.37, 95% CI: 1.50–3.74), jaundice (aOR: 2.31, 95% CI: 1.15–4.46), chest pain (aOR: 1.79, 95% CI: 1.18–2.73), pallor (aOR: 1.70, 95% CI: 1.23–2.35), and existing comorbidities (aOR: 2.77, 95% CI: 1.93–3.96). Contact with febrile individuals (aOR: 0.42, 95% CI: 0.31–0.57), nausea and/or vomiting (aOR: 0.73, 95% CI: 0.54–1.00), muscle pain (aOR: 0.44, 95% CI: 0.31–0.64), and rash (aOR: 0.45, 95% CI: 0.29–0.70) were associated with lower odds of bacterial infection.
Demographic, exposure, and clinical predictors for bacterial versus nonbacterial infections, Nakhon Phanom and Tak provinces, Thailand, April 2017–May 2020
| Predictor | aOR* (95% CI) |
|---|---|
| Province of Hospital | |
| Nakhon Phanom | 2.82 (2.02–3.93) |
| Tak | Ref. |
| Age, years | |
| 2–17 | 1.16 (0.48–2.77) |
| 18–49 | Ref. |
| ≥50 | 4.18 (2.85–6.14) |
| Employment Status | |
| Employed | Ref. |
| Student/Preschool | 0.50 (0.21–1.16) |
| Unemployed | 1.68 (1.01–2.79) |
| Year of Study | |
| Year 1 (April 2017—March 2018) | 2.92 (2.01–4.24) |
| Year 2 (April 2018–March 2019) | 3.30 (2.25–4.82) |
| Year 3 (April 2019–May 2020) | Ref. |
| Contact with Febrile Household Member, Neighbor, or Coworker | 0.42 (0.31–0.57) |
| Contact with Cow, Pig, Goat, or Sheep | 1.82 (1.29–2.57) |
| Antibiotics Taken within 72 hours before Hospital Presentation | |
| Yes | 2.37 (1.50–3.74) |
| No | Ref. |
| Yellow Eyes or Skin | 2.31 (1.15–4.63) |
| Nausea and/or Vomiting | 0.73 (0.54–1.00) |
| Muscle Pain | 0.44 (0.31–0.64) |
| Chest Pain | 1.79 (1.18–2.73) |
| Pallor | 1.70 (1.23–2.35) |
| Rash | 0.45 (0.29–0.70) |
| Any Existing Comorbidities† | 2.77 (1.93–3.96) |
aOR = adjusted odds ratio; Ref. = reference.
Adjusted for all other variables included in the model.
Comorbidities include diabetes, hypertension, heart disease, asthma, COPD, cancer, HIV, immunodeficiency, history of tuberculosis, active tuberculosis, liver disease, thyroid disease, thalassemia, anemia, chronic renal disease, and other chronic disease.
Multivariable results for factors associated with E. coli and dengue virus infections are shown in Table 4. Similar to the overall multivariable analysis, E. coli infection was independently associated with age ≥50 years (aOR: 2.72, 95% CI: 1.68–4.41), pallor (aOR: 1.54, 95% CI: 1.03–2.32), and existing comorbidities (aOR: 2.43, 95% CI: 1.56–3.78); lower odds of E. coli infection were observed among those with contact with febrile individuals (aOR: 0.53, 95% CI: 0.35–0.82). Dengue virus infection was independently associated with contact with febrile individuals (aOR: 1.74, 95% CI: 1.28–2.40), nausea and/or vomiting (aOR: 1.70, 95% CI: 1.23–2.34), hematocrit >40 mg% (aOR: 2.21, 95%CI: 1.56–3.12), white blood cell count <4,000/mm3 (aOR: 6.48, 95% CI: 4.45–9.43), and platelet count <100,000/mm3 (aOR: 2.66, 95% CI: 1.84–3.84). Age ≥50 years (aOR: 0.33, 95% CI: 0.22–0.50), antibiotic use within 72 hours before hospital admission (aOR: 0.65, 95% CI: 0.41–1.05), chest pain (aOR: 0.48, 95% CI: 0.31–0.75), pallor (aOR: 0.66, 95% CI: 0.48–0.92), and existing comorbidities (aOR: 0.47, 95% CI: 0.31–0.69) were associated with lower odds of dengue virus infection. Additional associations between the infection of each pathogen and variables not included in the bacterial versus nonbacterial multivariable analysis are reported in Table 4 (see Supplemental Table 4 for bivariate results).
Factors associated with the most common bacterial (E. coli) and nonbacterial (dengue virus) infections, Nakhon Phanom and Tak provinces, Thailand, April 2017–May 2020
| Characteristics | Dengue Virus (N = 563) | E. coli (N = 160) |
|---|---|---|
| aOR* (95% CI) | aOR* (95% CI) | |
| Female Sex | – | 2.82 (1.82–4.37) |
| Age, years | ||
| 2–17 | 2.08 (1.38–3.12) | 0.06 (0.01–0.48) |
| 18–49 | Ref. | Ref. |
| ≥50 | 0.33 (0.22–0.50) | 2.72 (1.68–4.41) |
| Year of Study | ||
| Year 1 (April 2017–March 2018) | 1.52 (1.04–2.22) | – |
| Year 2 (April 2018–March 2019) | 0.94 (0.64–1.40) | – |
| Year 3 (April 2019–May 2020) | Ref. | – |
| Contact with Febrile Household Member, Coworker, or Neighbor | 1.74 (1.28–2.40) | 0.53 (0.35–0.82) |
| Contact with Cats or Dogs | – | 1.68 (1.10–2.57) |
| Antibiotics Taken within 72 hours before Hospital Presentation† | ||
| Yes | 0.65 (0.41–1.05) | – |
| No | Ref. | – |
| Days of Fever before Hospital Admission | ||
| 0–1 | 1.15 (0.77–1.72) | 1.01 (0.65–1.56) |
| 2–3 | Ref. | Ref. |
| 4–7 | 1.38 (0.95–2.00) | 0.55 (0.30–1.02) |
| Nausea and/or Vomiting | 1.70 (1.23–2.34) | – |
| Headache | 1.51 (1.03–2.21) | 0.53 (0.35–0.83) |
| Chest Pain | 0.48 (0.31–0.75) | – |
| Bone or Joint Pain | 0.66 (0.47–0.92) | – |
| No Appetite | – | 0.57 (0.37–0.88) |
| Chills | – | 1.76 (1.02–3.02) |
| Pale or Cold Skin | 0.66 (0.48–0.92) | 1.54 (1.03–2.32) |
| Rash | 1.40 (0.95–2.07) | – |
| Any Existing Comorbidities‡ | 0.47 (0.31–0.69) | 2.43 (1.56–3.78) |
| Hematocrit >40 mg% | 2.21 (1.56–3.12) | 0.59 (0.32–1.09) |
| White Blood Cell <4,000/mm3 | 6.48 (4.45–9.43) | 0.68 (0.36–1.29) |
| Platelet <100,000/mm3 | 2.66 (1.84–3.84) | 0.47 (0.26–0.86) |
– = not included in multivariable model; aOR = adjusted odds ratio; E. coli = Escherichia coli; OR = odds ratio; Ref. = reference.
Adjusted for all other variables included in the model.
Excluded unknown.
Comorbidities included diabetes, hypertension, heart disease, asthma, chronic obstructive pulmonary disease, cancer, HIV, immunodeficiency, history of tuberculosis, active tuberculosis, liver disease, thyroid disease, thalassemia, anemia, chronic renal disease, and other chronic disease.
Clinical outcomes.
The clinical outcomes of bacterial and nonbacterial infections are shown in Table 5. Patients with bacterial infection had greater odds of more severe outcomes, including longer hospital stays (OR: 3.67, 95% CI: 2.88–4.69) and intubation/mechanical ventilation (OR: 5.15, 95% CI: 1.90–13.97); they also had lower odds of recovery or improvement of their condition at the time of discharge (OR: 0.11, 95% CI: 0.05–0.23). When adjusting for age and comorbidities, bacterial infection was independently associated with longer hospital stays (aOR: 2.75, 95% CI: 2.08–3.64) and lower odds of recovery or improvement (aOR: 0.14, 95% CI: 0.07–0.31).
Associations between clinical outcomes and bacterial vs. nonbacterial infections, Nakhon Phanom and Tak provinces, Thailand, April 2017–May 2020
| Characteristics | Bacterial Infections (N = 528) | Non-Bacterial Infections (N = 735) | OR (95% CI) | aOR (95% CI)* |
|---|---|---|---|---|
| n (%) | n (%) | |||
| No. of Days Hospitalized | ||||
| 0–1 | 37 (7.0) | 48 (6.5) | – | – |
| 2–4 | 223 (42.2) | 526 (71.6) | – | – |
| ≥5 | 268 (50.8) | 161 (21.9) | 3.67 (2.88–4.69)† | 2.75 (2.08–3.64)† |
| Intubation/Mechanical Ventilation | 18 (3.4) | 5 (0.7) | 5.15 (1.90–13.97) | 2.49 (0.82–7.56) |
| Discharge Status | ||||
| Recovery/Improved | 474 (89.9) | 724 (98.8) | 0.11 (0.05–0.23)† | 0.14 (0.07–0.31)† |
| Not Improved | 49 (9.3) | 9 (1.2) | – | – |
| Deceased | 4 (0.8) | 0 (0.00) | – | – |
aOR = adjusted odds ratio; OR = odds ratio.
Adjusted for age and existing comorbidities.
Compared with other categories combined.
Sensitivity analyses.
Multivariable results did not appreciably differ when stratifying by province (Supplemental Table 5). Most associations remained the same as in combined analyses, but there were several differences. In Nakhon Phanom, patients aged 2–17 years had lower odds of bacterial infection compared with those aged 18–49 years (aOR: 0.43, 95% CI: 0.22–0.84). Use of antibiotics within 72 hours before hospitalization did not differ for patients with bacterial and nonbacterial infections. In Tak, a history of visiting the forest within the previous 2 weeks was independently associated with greater odds of bacterial infection (aOR: 1.89, 95% CI: 1.16–3.07), whereas the presence of jaundice, nausea and/or vomiting, and rash did not significantly differ for patients with bacterial and nonbacterial infections. Furthermore, contact with farm animals was not significantly associated with bacterial infection in either province alone.
DISCUSSION
This analysis identified several demographic characteristics, exposures, and clinical indicators that were associated with bacterial and nonbacterial etiologies of AUFI. Our results highlight the importance of considering contextual factors to aid in the diagnosis of AUFI, especially in settings with limited resources and laboratory capacities.
Several demographic characteristics and exposure were found to be associated with bacterial infection, indicating that it may be useful to consider patient characteristics when assessing individuals with AUFI. Notably, we found that older age and the presence of comorbidities were both associated with increased odds of bacterial infection, which is consistent with current knowledge on susceptibility to bacterial infections.36–38 Both dengue and E. coli may have influenced these findings because our independent analyses showed that patients age ≥50 years and those with comorbidities had greater odds of E. coli infection and lower odds of dengue virus infection. In contrast, those with higher hematocrit level and lower white blood cell and platelet counts had a greater odds of dengue infection, consistent with previous literature.23,33,34 Older patients and those who had comorbidities may have an increased susceptibility to bacterial infection, which can inform diagnostics and treatment decisions. When we adjusted for age and comorbidities, patients with bacterial infections tended to have more severe outcomes including longer hospitalization and decreased odds of recovery at the time of hospital discharge. This is consistent with findings in published literature, in which bacterial infections are more likely than nonbacterial infections to result in sepsis and other complications.4,39 However, this analysis did not consider hematological indicators or complications that arose during hospitalization, so reasons for the severity of bacterial infections were not assessed. Proper diagnosis of bacterial pathogens and early appropriate antibiotic treatment may improve patient outcomes.
Our findings also show that individuals who took antibiotics in the 72 hours before hospital presentation had greater odds of bacterial infection. Differing symptoms and symptom severity between bacterial and nonbacterial infections may have influenced the likelihood of taking antibiotics before hospital presentation, but further research should be done to assess drivers of antibiotic use. The high proportion (10%) of participants with nonbacterial infections who took antibiotics before seeking care should not be overlooked. Providers can encourage patients to seek proper clinical and/or laboratory diagnosis before beginning antimicrobial treatment to reduce the chances of not being able to confirm bacterial infections and also because of the substantial burden of AMR in Thailand and the Southeast Asia region.9,40,41
Several signs and symptoms were associated with bacterial versus nonbacterial infections. Patients presenting with jaundice had greater odds of bacterial infection, which may be reflective of current knowledge that several bacterial infections, including leptospirosis and rickettsiosis, can cause jaundice, especially when the infection is severe or has resulted in sepsis.17,39,42 However, diagnostic tests for viral hepatitis were not included in the analysis, and results may have differed if hepatitis was included as a nonbacterial infection in this analysis. We also found that chest pain and pallor were associated with bacterial infections. Although some nonbacterial pathogens can cause chest pain and pallor, our results indicate that bacterial causes may be more likely among febrile patients in Thailand. Considered together with other symptoms, exposure history, and patient characteristics, the presence of signs and symptoms such as jaundice, chest pain, and pallor may prompt healthcare workers to investigate a potential bacterial etiology, possibly leading to quicker diagnosis and treatment.
Although many infectious etiologies can cause nausea and/or vomiting, muscle pain, and rash, we found these symptoms to be associated with nonbacterial infection. The high prevalence of dengue infections in our analysis, as well as the chikungunya outbreak identified in year 3 of the study, are likely drivers of these findings, given that all three signs and symptoms are known to be characteristic of dengue and/or chikungunya infections.43,44 Because of the high prevalence of dengue and chikungunya in Thailand’s border regions,45 the presence of nausea and/or vomiting, muscle pain, or rash alongside fever can prompt healthcare professionals to test for both pathogens before prescribing antibiotics.
Differences by province.
A significantly greater proportion of nonbacterial infections were found in participants from Tak than in those from Nakhon Phanom, which is likely due to its location bordering Myanmar. In Thailand, higher incidences of dengue and malaria are seen in provinces bordering Myanmar than in provinces bordering Laos,24,45 and this was reflected in our findings of higher proportions of both dengue and malaria detections in participants from Tak than in those from Nakhon Phanom. Chikungunya was only identified during year 3 of the study, which corresponds with nationwide dengue and chikungunya outbreaks that were observed in 2019,43,46,47 and may explain why patients enrolled in the third year of the study were significantly less likely to have bacterial infections.
Although there were not many differences between the combined and individual analyses for Nakhon Phanom and Tak, the provinces are different in terms of populations, geography, common types of illnesses. For instance, melioidosis, Rickettsia, K. pneumoniae, and S. agalactiae were predominantly found in Nakhon Phanom; scrub typhus, chikungunya virus, and malaria were predominantly found in Tak. These differences should be considered in conjunction with clinical presentations and epidemiological information, in diagnostic and treatment decisions at the local level.
Limitations.
This analysis is subject to at least four limitations. First, only patients admitted to 12 government hospitals were included in the surveillance activity, and only 51.3% of eligible patients consented to participate; therefore, the results may not be representative of the general population or other geographic areas of Thailand. Patients who choose to visit private hospitals or who only have ease of access to smaller clinics or outpatient departments may have different demographic characteristics or behaviors than those included in our analysis. Second, although testing was conducted for an array of pathogens, it was not done for all possible infectious etiologies. A substantial proportion of eligible patients did not have pathogens detected and were excluded from this analysis, but it is possible that some pathogens were not included in the diagnostic tests that were conducted. Our results may have differed if additional pathogens were included. Third, although we conducted pathogen-specific analyses for the most common pathogens detected in this study, the sample size was not large enough to investigate either the associations between indicators and other pathogens of interest or the effects of multiple bacterial or nonbacterial infections in individuals. Future studies with larger sample sizes would be useful for the comprehensive assessment of the etiologies of AUFI. Finally, there is substantial overlap in the clinical presentations and risk factors for pathogens causing AUFI. Our findings are not meant to serve as diagnostic criteria for bacterial or nonbacterial etiologies, but they may be used to inform targeted laboratory testing (e.g., diagnostic testing focusing on bacterial versus testing focusing on viral or parasitic diseases) and treatment options (e.g., an initial treatment regimen centered on use of empiric antibiotics versus one centered on supportive care that might include empiric use of antiviral or antiparasitic agents) based on the available evidence in febrile persons admitted to hospitals.
CONCLUSION
Accurate diagnosis of the etiologies of AUFI is challenging in Thailand. Our findings may support inferential decision-making for laboratory testing and treatment options for patients who are suspected to have a fever of infectious origin, influencing timely and appropriate treatment. Comprehensive assessment of exposure history, symptoms, and risk factors can aid healthcare professionals in resource-limited settings to narrow down the likely cause of illness. Improvements in the ability to differentiate between bacterial and nonbacterial etiologies of AUFI may help direct clinical and laboratory assessment of patients with AUFI, which could hasten provision of appropriate care and treatment, including appropriate antibiotic use and timely implementation of precautions to reduce onward transmission of infections, thereby reducing AMR and improving the health outcomes of patients with AUFI in Thailand.
Supplemental Materials
ACKNOWLEDGMENTS
We thank Nakhon Phanom and Tak Provincial Health Office, hospital directors, hospital and laboratory staff at Nakhon Phanom Hospital and Mae Sot General Hospital for their assistance in conducting the surveillance activity, as well as hospital directors and staff at district hospitals (Nae Kae, Srisongkram, Renunakhon, Banphang, That Phanom Crown Prince, Plapak hospitals in Nakhon Phanom and Thasongyang, Umphang, Maeramad, and Poppra hospitals in Tak) in both provinces who contributed to data collection and rapid diagnostic testing. We also thank laboratory staff at Bamrasnaradura Infectious Diseases Institute for conducting molecular diagnostic testing. We thank Somsak Thamthitiwat and Sean M. Griffing for their assisting with the protocol development, Pornpak Khunatorn and Thantapat Akarachotpong for IT support, and Juraiporn Ratanodom for administrative assistance.
REFERENCES
- 1.↑
Rhee C , Kharod GA , Schaad N , Furkawa 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.
- 2.↑
Tam P-YI , Obaro SK , Storch G , 2016. Challenges in the etiology and diagnosis of acute febrile illness in children in low- and middle-income countries. J Pediatric Infect Dis Soc 5: 190–205.
- 3.↑
Crump JA , 2014. Time for a comprehensive approach to the syndrome of fever in the tropics. Trans R Soc Trop Med Hyg 108: 61–62.
- 4.↑
Crump JA , Newton PN , Baird SJ , Lubell Y , Holmes KK , Bertozzi S , Bloom BR & Jha P Major Infectious Diseases. 3rd ed. Washington, DC: The International Bank for Reconstruction and Development/The World Bank, 365–385.
- 5.↑
Mayxay M et al., 2013. Causes of non-malarial fever in Laos: A prospective study. Lancet Glob Health 1: e46–e54.
- 6.↑
Kapasi AJ , Dittrich S , Gonzalez IJ , Rodwell TC , 2016. Host biomarkers for distinguishing bacterial from non-bacterial causes of acute febrile illness: A comprehensive review. PLoS One 11: e01600278.
- 7.↑
Gasem MH et al., for INA-RESPOND , 2020. An observational prospective cohort study of the epidemiology of hospitalized patients with acute febrile illness in Indonesia. PLoS Negl Trop Dis 14: e0007927.
- 8.↑
Antimicrobial Resistance Collaborators , 2022. Global burden of bacterial antimicrobial resistance in 2019: A systematic analysis. Lancet 399: 629–655.
- 9.↑
World Health Organization , 2024. Antimicrobial Resistance in South-East Asia. Available at: https://www.who.int/thailand/health-topics/antimicrobial-resistance. Accessed September 23, 2022.
- 10.↑
Khamsarn S , Nampoonsak Y , Busamaro S , Tangkoskul T , Seenama C , Rattanaumpawan P , Boonyasiri A , Thamlikitkul V , 2016. Epidemiology of antibiotic use and antimicrobial resistance in selected communities in Thailand. J Med Assoc Thai 99: 270–275.
- 11.↑
Zellweger RM , Carrique-Mas J , Limmathurotsakul D , Day NP , Thwaites GE , Baker S , 2017. A current perspective on antimicrobial resistance in Southeast Asia. J Antimicrob Chemother 72: 2963–2972.
- 12.↑
Haenssgen MJ , Charoenboon N , Xayavong T , Althaus T , 2020. Precarity and clinical determinants of healthcare-seeking behaviour and antibiotic use in rural Laos and Thailand. BMJ Glob Health 5: e003779.
- 13.↑
Apisarnthanarak A , Danchaivijitr S , Khawcharoenporn T , Limsrivilai J , Warachan B , Bailey TC , Fraser VJ , Thammasart University Antibiotic Management Team , 2006. Effectiveness of education and an antibiotic-control program in a tertiary care hospital in Thailand. Clin Infect Dis 42: 768–775.
- 14.↑
Siltrakool B , Berrou I , Griffiths D , Alghamdi S , 2021. Antibiotics’ use in Thailand: Community pharmacists’ knowledge, attitudes and practices. Antibiotics (Basel) 10: 137.
- 15.↑
Chanvatik S , Kosiyaporn H , Lekagul A , Kaewkhankhaeng W , Vongmongkol V , Thunyahan A , Tangcharoensathien V , 2019. Knowledge and use of antibiotics in Thailand: A 2017 national household survey. PLoS One 14: e0220990.
- 16.↑
Sirijoti K , Hongsranagon P , Havanond P , Pannoi W , 2014. Assessment of knowledge attitudes and practices regarding antibiotic use in Trang province, Thailand. J Health Res 28: 299–307.
- 17.↑
Bhargava A , Ralph R , Chatterjee B , Bottieau E , 2018. Assessment and initial management of acute undifferentiated fever in tropical and subtropical regions. BMJ 363: k4766.
- 18.↑
Acestor N , Cooksey R , Newton PN , Menard D , Guerin PJ , Nakagawa J , Christophel E , Gonzalez IJ , Bell D , 2012. Mapping the aetiology of non-malarial febrile illness in Southeast Asia through a systematic review—Terra Incognita impairing treatment policies. PLoS One 7: e44269.
- 19.↑
Capeding MR et al., 2013. Dengue and other common causes of acute febrile illness in Asia: An active surveillance study in children. PLoS Negl Trop Dis 7: e2331.
- 20.↑
Ellis RD et al., 2006. Causes of fever in adults on the Thai–Myanmar border. Am J Trop Med Hyg 74: 108–113.
- 21.↑
Kasper MR , Blair PJ , Touch S , Sokhal B , Yasuda CY , Williams M , Richards AL , Burgess TH , Wierzba TF , Putnam SD , 2012. Infectious etiologies of acute febrile illness among patients seeking health care in south-central Cambodia. Am J Trop Med Hyg 86: 246–253.
- 22.↑
Leelarasamee A , Chupaprawan C , Chenchittikul M , Udompanthurat S , 2004. Etiologies of acute undifferentiated febrile illness in Thailand. J Med Assoc Thai 87: 464–472.
- 23.↑
Luvira V , Silachamroon U , Piyaphanee W , Lawpoolsri S , Chierakul W , Leaungwutiwong P , Thamwomkuno C , Wattanagoon Y , 2019. Etiologies of acute undifferentiated febrile illness in Bangkok, Thailand. Am J Trop Med Hyg 100: 622–629.
- 24.↑
Polwiang S , 2016. Estimation of dengue infection for travelers in Thailand. Travel Med Infect Dis 14: 398–406.
- 25.↑
Fernandez-Carballo BL , Escadafal C , MacLean E , Kapasi AJ , Dittrich S , 2021. Distinguishing bacterial versus non-bacterial causes of febrile illness—A systematic review of host biomarkers. J Infect 82: 1–10.
- 26.↑
Oved K et al., 2015. A novel host-proteome signature for distinguishing between acute bacterial and viral infections. PLoS One 10: e0120012.
- 27.↑
Chalupa P , Beran O , Herwald H , Kasprikova N , Holub M , 2011. Evaluation of potential biomarkers for the discrimination of bacterial and viral infections. Infection 39: 411–417.
- 28.↑
Dupuy AM et al., Maurice Rapin Institute Biomarkers Group , 2013. Role of biomarkers in the management of antibiotic therapy: An expert panel review: I—Currently available biomarkers for clinical use in acute infections. Ann Intensive Care 3: 22.
- 29.↑
Lubell Y et al., 2015. Performance of C-reactive protein and procalcitonin to distinguish viral from bacterial and malarial causes of fever in Southeast Asia. BMC Infect Dis 15: 511.
- 30.↑
Laoprasopwattana K , Kaewjungwad L , Jarumanokul R , Geater A , 2012. Differential diagnosis of chikungunya, dengue viral infection and other acute febrile illnesses in children. Pediatr Infect Dis J 31: 459–463.
- 31.↑
Laoprasopwattana K , Limpitikul W , Geater A , 2020. Using clinical profiles and complete blood counts to differentiate causes of acute febrile illness during the 2009–11 outbreak of typhoid and chikungunya in an endemic area. J Trop Pediatr 66: 504–510.
- 32.↑
Watt G , Jongsakul K , Chouriyagune C , Paris R , 2003. Differentiating dengue virus infection from scrub typhus in Thai adults with fever. Am J Trop Med Hyg 68: 536–538.
- 33.↑
Assawawiroonhakarn S , Geater A , Laoprasopwattana K , 2017. Using clinical profile and initial laboratory results to differentiate leptospirosis, scrub typhus and dengue viral infections among children with an acute febrile illness. Southeast Asian J Trop Med Public Health 48: 1274–1280.
- 34.↑
Chaloemwong J , Tantiworawit A , Rattanathammethee T , Hantrakool S , Chai-Adisaksopha C , Rattarittamrong E , Norasetthada L , 2018. Useful clinical features and hematological parameters for the diagnosis of dengue infection in patients with acute febrile illness: A retrospective study. BMC Hematol 18: 1–10.
- 35.↑
Craney TA , Surles JG , 2002. Model-dependent variance inflation factor cutoff values. Qual Eng 14: 391–403.
- 37.↑
Esper AM , Moss M , Lewis CA , Nisbet R , Mannino DM , Martin GS , 2006. The role of infection and comorbidity: Factors that influence disparities in sepsis. Crit Care Med 34: 2576–2582.
- 38.↑
Seo H , Sim YS , Min KH , Lee JH , Kim BK , Oh YM , Ra SW , Kim TH , Hwang YI , Park JW , 2022. The relationship between comorbidities and microbiologic findings in patients with acute exacerbation of chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 17: 855–867.
- 40.↑
Sumpradit N et al., 2017. New chapter in tackling antimicrobial resistance in Thailand. BMJ 358: j3415.
- 41.↑
Lim C , Takahashi E , Hongsuwan M , Wuthiekanun V , Thamlikitkul V , Hinjoy S , Day NP , Peacock SJ , Limmathurotsakul D , 2016. Epidemiology and burden of a multidrug-resistant bacterial infection in a developing country. eLife 5: e18082.
- 42.↑
Syhavong B et al., 2010. The infective causes of hepatitis and jaundice amongst hospitalised patients in Vientiane, Laos. Trans R Soc Trop Med Hyg 104: 475–483.
- 43.↑
World Health Organization , 2022. Dengue and Severe Dengue. Available at: https://www.who.int/news-room/fact-sheets/detail/dengue-and-severe-dengue. Accessed June 12, 2023.
- 44.↑
World Health Organization , 2020. Chikungunya. Available at: https://www.who.int/news-room/fact-sheets/detail/chikungunya. Accessed September 23, 2022.
- 45.↑
Xu Z , Bambrick H , Pongsumpun P , Tang IM , Yakob L , Devine G , Frentiu FD , Williams G , Hu W , 2020. Does Bangkok have a central role in the dengue dynamics of Thailand? Parasit Vectors 13: 22.
- 46.↑
Wangsri S , Angskul T , 2019. MOPH tackles dengue fever after 62 deaths reported. ReliefWeb. Available at: https://reliefweb.int/report/thailand/moph-tackles-dengue-fever-after-62-deaths-reported. Accessed June 14, 2024.
- 47.↑
Luangtongwattana N , 2021. Incidence of severe dengue fever and dengue fever with unusual manifestations in patients older than 15 years in Prapokklao Hospital, Chanthaburi province during the 2019 outbreak. J Prapokklao Hosp Clin Med Edu Cent 38: 419–428.