Fever in Travelers Returning from Malaria-Endemic Areas: Don't Look for Malaria Only
Heli M. Siikamäki, MD, Division of Infectious Diseases, Department of Medicine, Helsinki University Central Hospital, Aurora Hospital, Building 5, 3rd floor, PO Box 348, FIN-00029 HUS, Helsinki, Finland. E-mail: firstname.lastname@example.org
Background. Returning travelers with fever pose challenges for clinicians because of the multitude of diagnostic alternatives. Case data in a Finnish tertiary hospital were analyzed in order to define the causes of fever in returned travelers and to evaluate the current diagnostic approach.
Methods. A retrospective study of patient records comprised 462 febrile adults who, after traveling in malaria-endemic areas, were admitted to the Helsinki University Central Hospital (HUCH) emergency room from 2005 to 2009. These patients were identified through requests for malaria smear.
Results. The most common groups of diagnoses were acute diarrheal disease (126 patients/27%), systemic febrile illness (95/21%), and respiratory illness (69/15%). The most common specific main diagnosis was Campylobacter infection (40/9%). Malaria was diagnosed in 4% (20/462). Blood culture was positive for bacteria in 5% of those tested (21/428). Eight patients were diagnosed with influenza. HIV-antibodies were tested in 174 patients (38%) and proved positive in 3% of them (5/174, 1% of all patients). The cause of fever was noninfectious in 12 (3%), remaining unknown in 116 (25%). Potentially life-threatening illnesses were diagnosed in 118 patients (26%), the strongest risk factors were baseline C-reactive protein (CRP) ≥100 (OR 3.6; 95% CI 2.0–6.4) and platelet count ≤140 (OR 3.8; 95% CI 2.0–7.3). Nine patients (2%) were treated in high dependency or intensive care units; one died of septicemia. Forty-five patients (10%) had more than one diagnosis.
Conclusions. The high proportion of patients with more than one diagnosis proves the importance of careful diagnostics. Every fourth returning traveler with fever had a potentially life-threatening illness. Septicemia was as common as malaria. The proportion of HIV cases exceeded the prevalence in population for which Centers for Disease Control and Prevention, USA (CDC) recommends routine HIV testing. Both blood cultures and HIV tests should be considered in febrile travelers.
Fever is one of the most common reasons for seeking medical help after travel, others being gastrointestinal, respiratory, and skin symptoms.1–6 In several studies malaria is reported as both the most common single reason for travel-related fever without local findings1–3,5,7–9 and the primary cause of death.5,9 In addition to tropical diseases, cosmopolitan infections are frequently diagnosed, and in a minority of cases, noninfectious causes like rheumatic diseases and malignancies are found. Type of traveler1,4–6,9–13 and destination of travel2,3,5,6,8,9 are both associated with the etiology of the fever; a correlation with travelers' country of origin has also been reported.6
The number of foreign leisure trips made by Finnish residents (population 5.3 million) has nearly doubled within the past 10 years (3.6 million in 2009) with an increasing trend in travel to malaria-endemic countries.14 The area most favored by Finns outside Europe and the East Mediterranean region is Asia/Oceania (226,000 trips/yr, including Thailand with 121,000 trips/yr) followed by the Americas (126,000) and Africa (109,000).15
The clinician on call faces a multitude of diagnostic alternatives when examining febrile travelers.16 To define the causes of fever and to evaluate the current diagnostic approach, patient data of travelers returning with fever from tropical or subtropical areas were analyzed in an emergency room of a Finnish tertiary hospital.
A retrospective study was conducted on the medical records of adult travelers returning from tropical or subtropical areas with fever admitted to the emergency room of internal and pulmonary medicine of Helsinki University Central Hospital (HUCH), a tertiary hospital serving 1.4 million inhabitants. To identify retrospectively these patients among the 12,300 patients seen in the emergency room during the study period between January 2005 and March 2009, the request for a malaria smear was used as a search tool.
The current diagnostic guideline and practice in HUCH is to routinely obtain malaria smear, hemoglobin, white blood cell (WBC), platelet count, P-CRP, creatinine, sodium, potassium, liver enzymes, two blood cultures, urine sample, and chest X-ray from patients with unexplained fever returning from a malaria-endemic area. Other tests are chosen by the physician in charge on the basis of the clinical symptoms. Malaria smears were taken from a mean of 20 (range 7–68) patients/mo, altogether 1008 patients (2% of all patients). The first 10 patients of each month were included. Adult patients (≥16 years of age) who had traveled in the tropics or subtropics within a year and had a malaria smear taken because of fever (measured or reported axillar temperature >37.5°C prior to, or at the time of presentation) were included in the study. Altogether 500 patients were collected; 462 patients met the inclusion criteria and were included for the final analysis.
The study protocol was approved by the Department of Internal Medicine of HUCH.
The following information was retrieved from the medical records: demographic data (age, gender, country of birth, country of residence), detailed travel history (travel destination, time and length of stay), reason for travel [holiday/work or study/visiting friends and relatives (VFR)/foreign visitor/recently arrived immigrant/expatriate living in malaria area], symptoms, time of the beginning of the symptoms, chest X-ray and laboratory findings [hemoglobin, P-CRP, WBC, platelet count, creatinine, alanine transaminase (ALAT), HIV test, influenza antigen, blood culture, stool culture and parasites], antimicrobial treatment, diagnoses, and duration of hospitalization.
Diagnoses were recorded at three different time points: (1) the working diagnosis at the emergency room, (2) the discharge diagnosis, and (3) the final diagnosis evaluated at least 1 year after discharge (>1 diagnosis/patient possible on each occasion). Complications and significant underlying diseases were recorded separately.
The final clinical or etiological diagnosis of all patients was defined by the same infectious diseases specialist (H. S.), who had access to all the results. Diagnoses were listed in the order of relevance to the symptoms as judged by the specialist. The diagnoses were coded according to the classification used by GeoSentinel3: a standardized list of 588 possible individual diagnoses categorized under 21 broad syndromes was used. Septicemia was defined as a symptomatic condition with a positive blood culture. Unknown bacterial infection was defined as a clinical picture, C-reactive protein (CRP) (CRP median 136, range 50–275 mg/L), and a timely response to systemic antibiotic therapy, all compatible with bacterial infection. Potentially life-threatening illness was defined as a disease potentially leading to death if left without specific or supportive treatment.
The countries visited were grouped into five regions: Sub-Saharan Africa, Southeast Asia, Central Asia and Indian Subcontinent, South and Central America and the Caribbean, Other (North Africa, West Asia, Northeast Asia), modified from GeoSentinel.3
Chi-square tests, t-tests, and Mann–Whitney tests served to test for differences between the groups. The binary and multinomial logistic regression models served to identify explanatory variables to the outcome variables. Variables that were found to have p value less than 0.2 were included in the multivariable models. To identify independent risk factors, forward and backward selection with Akaike information criteria (AIC) was used. One variable (duration of the trip) had 72 missing values of the 462, and to take that into account in the model, we used multiple imputation with an assumption that the missingness process was missing at random (MAR). The analysis was carried out with SPSS 18.0.2 (SPSS, Inc., Chicago, IL, USA).
The demographic and travel data are presented in Table 1. Information about referral was available for 417 patients; of these, 368 (88%) were referred by a doctor, 4 were brought by ambulance, and 45 were taken in without referral. A total of 249 (54%) patients were hospitalized; for those the median length of hospitalization was 5 days. Ten patients (2%) were referred because of a recent history of being treated for malaria in an endemic area.
Table 1. Patient characteristics and travel data
|Number of travelers||Total||462|
|Age (years)||Median (IQR)*||34.0 (27–44)|
|Duration of travel (days)†||Median (IQR)*||15 (13–30)|
|Geographic region visited (number of travelers)||Sub-Saharan Africa||193 (41.8%)|
| ||South-East Asia||128 (27.7%)|
| ||Central Asia and Indian Subcontinent||93 (20.1%)|
| ||South and Central America and Caribbean||26 (5.6%)|
| ||Other (North Africa, West Asia, Northeast Asia)||21 (4.5%)|
| ||Not known||1|
|Reason for travel||Tourism||296 (64.1%)|
| ||Work/education||80 (17.3%)|
| ||Visiting friends and relatives||31 (6.7%)|
| ||Foreign visitor||16 (3.5%)|
| ||Recently arrived immigrant||9 (1.9%)|
| ||Expatriate||8 (1.7%)|
| ||Not known||22 (4.8%)|
|Time between return from the trip and presentation at the hospital (days)‡||Median (IQR)*||5 (2–16)|
|Duration of symptoms before presentation at the hospital (days)§||Median (IQR)*||3 (2–7)|
The final diagnoses regarded as the main cause of fever, including potentially life-threatening illnesses, are presented in Table 2. An etiological or clinical diagnosis was established in 346 (75%) cases. The discharge diagnosis differed from the working diagnosis in 193 (43%) cases. The final diagnosis was different from the working diagnosis in 256 (55%) and from the discharge diagnosis in 115 (25%) cases. The data below describe the final diagnoses.
Table 2. Final diagnoses
|Acute diarrheal disease, all||126 (27.3%)|
| Gastroenteritis of unknown origin||59|
| Campylobacter infection||40|
| Clostridium difficile infection||1|
|Systemic febrile illness, all||95 (21%)|
| Unknown bacterial infection||8|
| Paratyphoid fever†||3|
| Typhoid fever†||2|
| Intra-abdominal abscess†||1|
| Acute viral infection||6|
| HIV infection†||4|
| Epstein-Barr virus infection||4|
| Herpes simplex infection||3|
| Nephropathia epidemica (Puumala virus infection)||3|
| Viral meningitis||2|
| Cytomegalovirus infection||1|
| HHV-6 infection||1|
| Plasmodium falciparum†||16|
| Plasmodium vivax||2|
| Plasmodium ovale||1|
| Species unknown||1|
| Pulmonary histoplasmosis||1|
|Respiratory illness, all||69 (14.9%)|
| Upper respiratory tract infection||26|
| Legionnaires' disease†||1|
| Pulmonary tuberculosis†||1|
| Pneumocystis jirovecii pneumonia†||1|
|Non-diarrheal gastrointestinal diagnosis, all||13 (2.8%)|
| Viral hepatitis||2|
| Other gastrointestinal diagnoses†,‡||10|
|Genitourinary diagnosis, all||19 (4.1%)|
| Acute urinary tract infection||2|
| Generalized gonococcal infection||1|
|Dermatologic diagnosis, all||12 (2.6%)|
| Skin infection||3|
| Anaphylactic reaction†||1|
|Other specified noninfectious diagnosis, all||12 (2.6%)|
| Collagen disease||7|
| Non-Hodgkin lymphoma†||1|
| Subacute thyroiditis||1|
| Mefloquine intolerance||1|
|No specific diagnosis||116 (25.1%)|
The most common main groups of diagnosis were acute diarrheal disease (126/27%), systemic febrile illness (95/21%), and respiratory illness (69/15%). Campylobacter was the most common specific cause of acute diarrheal disease and the most common single specific diagnosis.
Malaria was diagnosed in 20 patients, 8 of whom were VFRs. Plasmodium falciparum was the causative pathogen in 16 cases; in four of them the disease was complicated and required intensive care treatment.
Blood cultures were obtained from 428 (93%) of the patients and were positive for bacteria in 21 (5%) of these (Salmonella species 5, Escherichia coli 3, Salmonella paratyphi 3, Salmonella typhi 2, Staphylococcus aureus 2, Burkholderia pseudomallei 1, Klebsiella pneumoniae 1, Shigella sonnei 1, Streptococcus pyogenes 1, Streptococcus viridians 1, Pseudomonas aeruginosa 1).
Nasal swabs for influenza A and B antigen were taken from 47 patients (10% of all), including 20 of the 111 meeting the criteria of influenza-like illness (respiratory symptoms, fever >38.5°C); the test was found positive in 7 patients (15% of those tested).
HIV test was taken from 174 patients and repeated in 17 patients. A new HIV diagnosis was established in five patients (5/174, 3% of those tested).
More than one specific diagnosis was established in 45 (10%) patients: 41 patients had two and 4 had three separate diagnoses. The most common group of additional diagnoses was acute diarrheal disease (20/49 diagnoses), followed by respiratory (9/49) and systemic febrile illness (6/49, including 2 Epstein-Barr, 1 dengue, 1 HIV, 1 Herpes simplex virus infection, and 1 viral meningitis), genitourinary (4/49), dermatologic (3/49), and non-diarrheal gastrointestinal disease (3/49), and noninfectious diagnoses (4/49).
Travel Destination and Diagnoses
Patients returning from Central Asia and the Indian Subcontinent had acute diarrheal disease more frequently (38/93, 41%) than travelers from other areas (88/369, 24%) (p = 0.002). Most of the malaria (18/20) and all rickettsiosis cases (6) came from Sub-Saharan Africa, and most dengue cases from Asia (9/14). Rare severe diseases acquired in Asia were diagnosed: two cases of melioidosis and one case each of leptospirosis, hepatitis E, and pulmonary histoplasmosis.
Potentially Life-Threatening Illnesses and Their Risk Factors
Potentially life-threatening illness was diagnosed in 118 (26%) patients. Nine (2%) needed treatment in high dependency or intensive care units (four with P. falciparum malaria, two septicemia, two pneumonia, one leptospirosis). Significant complications developed in 19 patients (4%). One patient died of P. aeruginosa septicemia.
In the multivariate model, potentially life-threatening illness was associated with older age (≥40 years, OR 2.3, 95% CI 1.4–3.8), having a baseline CRP value ≥100 (OR 3.6, 95% CI 2.0–6.4), platelet count ≤140 (OR 3.8, 95% CI 2.0–7.2), and a white blood cell count ≥8 (OR 2.0, 95% CI 1.2–3.5). Patients with gastrointestinal symptoms were less likely to be diagnosed with a life-threatening illness (OR 0.4, 95% CI 0.2–0.6). There was no independent association between life-threatening illness and region of birth, duration of travel, muscle or joint symptoms, or urinary tract symptoms. Risk factors for malaria and septicaemia as compared to other final diagnoses are presented in Table 3.
Table 3. Independent risk factors for malaria and septicaemia
|Continent of birth|| ||0.000|| ||0.156|
| Other than Africa||1.0|| ||1.0|| |
| Africa||22.7 (4.7–110.4)|| ||2.4 (0.7–8.1)|| |
|Destination|| ||0.006|| ||0.640|
| Other than Africa||1.0|| ||1.0|| |
| Africa||15.0 (2.2–101.7)|| ||1.3 (0.5–3.4)|
|Muscle or joint symptoms|| ||0.112|| ||0.034|
| No||1.0|| ||1.0|| |
| Yes||3.3 (0.8–14.7)|| ||2.7 (1.1–7.0)|| |
|Headache|| ||0.086|| ||0.394|
| No||1.0|| ||1.0|| |
| Yes||3.7 (0.8-16.4)|| ||0.67 (0.24–1.8)|| |
|CRP*|| ||0.002|| ||0.005|
| <100 mg/L||1.0|| ||1.0|| |
| ≥ 100 mg/L||10.6 (2.3–47.6)|| ||3.9 (1.5–10.3)|| |
|Platelet count*|| ||0.000|| ||0.405|
| ≥140 × 109/L||1.0|| ||1.0|| |
| <140 × 109/L||41.0 (8.1–208.9)|| ||1.7 (0.5–5.5)|
The present data, while confirming several findings of previous studies, provide additional information useful in the diagnostic approach to returning travelers with fever.
Patient Selection and Limitations
To retrospectively identify returned travelers with fever, requests for malaria smear were considered an accurate approach: doctors on duty are aware of the national recommendation to request a malaria smear from all febrile travelers who have returned from malaria-endemic areas. The first 10 patients each month were included to ensure even distribution throughout the year. Although the most common destination of Finnish tourists is Thailand, patients in the present study most commonly had visited Sub-Saharan Africa.
The classification of potentially life-threatening illnesses was created by the study group as a tool to evaluate if the selection of patients referred to tertiary care was accurate. The classification is naturally ambiguous but a rather strict definition was preferred.
Those included were not representative of all febrile travelers, but patients referred to a tertiary hospital. Accordingly, the proportion of those with a potentially life-threatening illness was high. High dependency or intensive care treatment was needed for 2%, consistent with the findings of Bottieau and colleagues.9 Hospitalization proved more common (54%) than in other reports (26%–27%),5,9 which may partly be explained by the national guidelines advising to observe febrile travelers with strong suspicion of malaria until a sufficient number of malaria smears has been collected. The median length of hospitalization (5 days) in our study was similar to that in other reports (4–5 d).8,9
The final diagnosis differed from the working diagnosis in 55%, and from the discharge diagnosis in 25%. Besides reflecting variation between clinicians in interpretation and coding, this difference is likely to be explained by the fact that complete laboratory results, especially stool cultures and antibody tests, are often not yet available at the time of discharge.
Notably, in our study, every 10th patient had more than one diagnosis, similar to a previous report9 which stresses the importance of thoroughness in diagnosing travelers with fever.
The present data were collected before the onset of the influenza A (H1N1) pandemic in 2009. Nasal swabs for influenza A and B antigen were taken only in 18% of cases that met the criteria of influenza-like illness. These data are consistent with previous studies, suggesting influenza to be under-diagnosed in travelers.17 The pandemic increased the use of rapid diagnostic tests, hopefully not only temporarily.
HIV infection was diagnosed in 3% of those tested, 1% of all patients. Similar proportions of HIV cases have been found in another study on febrile returning travelers.9 Despite the widely recognized possibility of negative test at the early course of acute HIV infection, the test was repeated later only in 17 cases. There are studies on testing HIV in selected groups of returning travelers,18–20 but this group has not been systematically tested. In populations where the prevalence of HIV is >0.1%, Centers for Disease Control and Prevention, USA (CDC) recommend offering routine HIV testing for everyone in contact with health care.21 Our results suggest that travelers are a high-risk group for HIV infection; therefore, routine HIV testing should be recommended for all travelers with fever.
Potentially Life-Threatening Illnesses and Risk Factors
When examining returning traveler with fever, the most important task is to recognize potentially life-threatening infections. In other studies, malaria has been reported as the most common reason for fever without localized symptoms in returning travelers1–3,5,7–9; in most investigations septicemia has not been reported.1–3,5,8 In the study of Antinori 2004,7 blood culture was taken from 56% of febrile returning travelers and found positive in 10% of them. In Bottieau's report (2006),9 the diagnosis was made by blood culture in 2% of all patients. In our study, blood cultures were taken from 93%, of which septicemia was detected in 5%. The high proportion of septicemia may reflect the selection of our patients, most of whom had been referred to the tertiary hospital after initial contact within primary or secondary care. In our study mortality was 0.2% (1/462) which corresponds to other reports (0.2%–1.2%).4,5,9 In other studies malaria has been the main cause of death5,9; in our study there were no malaria-related deaths.
Risk factors for tropical diseases have been examined by Bottieau and colleagues22; we focused on risk factors for malaria and septicaemia, and found differences between them. Several independent risk factors were listed for malaria patients: they were more likely to have traveled and/or to be born in Africa, had CRP levels >100 mg/L and platelet counts <140x109/L. These findings are in line with other studies.23,24 Septicemia, in contrast, proved more difficult to predict; the only independent risk factors recognized were CRP level >100 mg/L, and muscle or joint symptoms. The equal proportion of septicaemia and malaria cases testifies to the importance of blood cultures in the examination of febrile travelers and suggests a low threshold for empiric antimicrobial therapy.
Every fourth patient had a diagnosis classified as a potentially life-threatening illness, further emphasizing the importance of rapidity when evaluating returning travelers with fever. In the multivariate model, several factors were independently associated with this heterogeneous group of conditions. Two predictors were found in the history of the patient (age >40, absence of gastrointestinal symptoms), one in physical examination (dermatological symptoms), and three in laboratory tests (high CRP, low platelet, and high leukocyte counts). However, none of the individual variables or combinations of variables could be used to exclude severe diagnosis. This highlights the importance of thorough history and careful examination as well as follow-up of all febrile travelers.
As travels to tropical and subtropical areas are increasing in number, there will be more travelers returning with fever. The high proportion of patients with more than one diagnosis urges clinicians to thoroughness in examining these patients. The diagnostic approach of taking both malaria smears and blood cultures from patients returning with fever from the tropics and subtropics is justified in a tertiary hospital. We also recommend that HIV tests should be taken routinely from febrile travelers and influenza tests from those fulfilling the criteria for influenza-like illness.
We thank Associate Professor Sakari Jokiranta, and the personnel of HUSLAB for help in identifying the patients.
This study was supported by the Finnish Society for Study on Infectious Diseases.
Declaration of Interests
The authors state they have no conflicts of interest to declare.