Brucellosis as a Cause of Fever of Unknown Origin in Children Admitted to a Tertiary Hospital in the Aegean Region of Turkey

Introduction

I n 1961, fever of unknown origin (FUO) was defined by Petersdorf and Beason as fever over 38.3°C that continues at least for 3 weeks with no diagnosis reached after 1 week of inpatient investigation (Petersdorf and Beeson 1961). The causes of FUO may be classified under five categories: infectious diseases, collagen vascular diseases, neoplasms, miscellaneous diseases, and undiagnosed illnesses. Infections are the most common cause of FUO. Generally, the agents, site of infection, and clinical manifestations are typical but subtle and overlooked, or are associated with an insidious or prolonged course (Long and Edwards 2008). FUO in children arises from different infectious causes in both developed and developing countries. In the past decade, new infectious disease etiologies have emerged as an important cause of FUO in developed countries, such as Epstein-Barr virus (EBV), cytomegalovirus, human immune deficiency virus (HIV), psittacosis, cat-scratch disease, or bartonellosis (Akpede and Akenzua 2001, Pasic et al. 2006, Long and Edwards 2008). Data from developing countries are limited, but malaria, typhoid fever, and tuberculosis usually feature prominently among the infections in these regions. Miscellaneous infections also cause FUO, which include brucellosis, typhus, and cat-scratch disease (Pasic et al. 2006).

Brucellosis is considered to be the most widespread zoonosis worldwide. It also constitutes a public health problem in Turkey. According to the Turkish Ministry of Health records from 1995 to 2002, the number of cases of brucellosis increased from 8184 to 17,584. The morbidity rate of brucellosis in Turkey increased from an average annual incidence of 1/100,000 in the 1980s to 25/100,000 in 2002 (Turkish Ministry of Health 2007). Brucellosis can invade any organ or system and has a wide clinical spectrum, ranging from asymptomatic to life-threatening extents. Fever is the most common reported clinical manifestation of childhood brucellosis (Sachdev et al. 2001).

The aim of this cumulative case study was to determine the role of brucellosis in children with FUO in the Aegean region of Turkey. Therefore, these data represent occurrences in a single facility in the country.

Methods

We reviewed the records of all children admitted to the Department of Pediatric Infectious Diseases, Ege University Medical School, which is a tertiary hospital in the Aegean region of Turkey, from 2003 to 2008 for evaluation of FUO. FUO was defined according to the classical Petersdorf and Beeson's criteria as fever that is higher than 38.3°C (101°F) on several occasions, present for >3 weeks, and has a cause that is still unexplained after 1 week of evaluation in the hospital. Those children with neutropenia (neutrophil count <500/mm³) who developed fever after admission to the hospital, who were HIV positive, or had a previous diagnosis of immunodeficiency were excluded.

The history and the findings of physical examinations were recorded. Initial evaluation of children with FUO during admission included erythrocyte sedimentation rate, C-reactive protein, complete blood count, urinalysis, urine and stool cultures, blood cultures, chest X-ray, tuberculin test, and EBV serology. According to their histories, physical examinations, and laboratory findings in the following days, along with serology for cytomegalovirus, HIV, mycoplasma, Toxoplasma, Brucella, and hepatitis viruses, tests for antinuclear antibodies, rheumatoid factor, thyroid-stimulating hormone, thyroxine, complement 3 (C3), C4, culture for mycobacteria, echocardiography, radiography of sinus, ultrasonography, and computed tomography of abdomen and chest were also done. Invasive procedures such as bone marrow aspiration/biopsy or liver biopsy and laparotomy were performed when a diagnosis was not revealed by the noninvasive procedures. The causes of FUO were classified into five groups: infection, collagen vascular diseases, neoplasm, miscellaneous, and no diagnosis.

Diagnosis of brucellosis was established by serum agglutination test titer 1/160 and/or isolation of Brucella species from the blood. The serum agglutination test was performed according to previously described techniques (Alton et al. 1975). Blood cultures were processed in a pediatric automatic BacTAlert system (Organon Technica). If no growth was detected within the usual 7-day period, incubation was maintained for a total of 21 days, and blind subcultures were performed on trypticase soy-based sheep blood agar and eosin methylene blue agar after 14 and 21 days. Among the inquiries carried out were demographic characteristics, patient and family history, history of consumption of raw milk or milk products or work in animal breeding, and physical examination findings of children with brucellosis presenting as FUO.

We used descriptive statistics (SPSS statistical software version 11.5) to characterize our population. The chi-square test was used for comparing the number of brucellosis cases in other series with our data. Statistical significance was set at p < 0.05.

This study was approved by the hospital's ethical committee.

Results

During the 6-year study period, 92 patients (52 boys and 40 girls) with FUO were followed up. The median age was 5.1 years (7 months to 16 years) and the median duration of fever prior to admission was 39 days (17–61 days). Final diagnosis is shown in Table 1. Fifty-seven of 92 children (61%) received at least one course of antibiotics.

Fifty-eight of the 92 (63%) children with established diagnosis had infectious diseases. Brucellosis was the most frequent infectious cause of FUO, followed by tuberculosis, EBV, and salmonellosis. Brucellosis was diagnosed in 14 children. The median time between admission and definitive diagnosis in infectious diseases group was 11 days. The median age of children in this group was 5.5 years. In this group, 39 children received at least one course of antibiotics.

The median age of children with brucellosis presented as FUO was 5.9 years (9 months to 12 years) (Table 2). The median duration of fever between admission and definitive diagnosis for patients with brucellosis presented as FUO was 26 days. The diagnosis of brucellosis was confirmed by serum agglutination test (SAT) in 11 patients and blood cultures in 3 patients. Previous suboptimal antibiotic courses and negative blood cultures were the main causes for delayed diagnosis of brucellosis. In two of the patients, after discontinuation of antibiotics, repeated blood cultures confirmed the diagnosis of brucellosis. Eight of 14 children with brucellosis presented as FUO had fever as the only complaint. Five had fever and hepatomegaly and/or splenomegaly, whereas only one had fever and arthritis simultaneously (Table 2). Nine patients (64%) were from rural areas and four (36%) from urban areas of western Anatolia in Turkey. The mode of transmission was consumption of unpasteurized milk and milk products. Histories revealed that the parents of eight patients were working in animal breeding. A positive family history for brucellosis was noted in three patients. Demographic features, symptoms, SAT titers, blood culture results, and antimicrobial combinations for children with brucellosis presenting as FUO are summarized in Table 2.

Discussion

FUO is influenced by age, geographic location, experience of the investigator and availability of diagnostic facilities, time of the study, and referral pattern. The general pattern of FUO in our study is similar to previous studies (Pizzo and Lovejoy 1975, Chantada et al. 1994, Akpede and Akenzua 2001, Pasic et al. 2006). Infections worldwide account for about a quarter to a half of cases of FUO in both children and adults, but the proportion of FUO due to infections is higher in children and in developing countries (Akpede and Akenzua 2001).

In older studies, almost half the patients with infections had upper or lower respiratory tract infections exclusive of granulomatous disease. However, Long and Edwards (2008), Pasic et al. (2006), Pizzo and Lovejoy (1975), Lohr and Hendley (1977), McClung (1972), and Jacobs and Schutz (1998) reported that the most common infections were EBV infection, vertebral and pelvic osteomyelitis, bartonellosis, and complicated urinary tract infection in children (McClung 1972, Pizzo and Lovejoy 1975, Lohr and Hendley 1977, Jacobs and Schutz 1998, Pasic et al. 2006, Long and Edwards 2008) and brucelllosis is an uncommon cause in children with FUO. In their series, it was a cause in only <1% of cases (Long and Edwards 2008). In contrast, in a previous study from Turkey, enteric fever was the most frequent cause, followed by brucellosis, in children with FUO (Ciftçi et al 2003). A meta-analysis carried out with 857 adult patients with FUO from 13 articles in Turkey found that tuberculosis was the most frequent infectious cause of FUO. In this analysis, brucellosis was found to be the second infectious cause (Sipahi et al. 2007). In our study, 58 of the 92 (63%) children with established diagnosis had infectious diseases. Brucellosis was the most frequent infectious cause of FUO, followed by tuberculosis, EBV, and salmonellosis. Comparison of the number of brucellosis cases in Argentina, India, Serbia, and Montenegro series with our data revealed that brucellosis was significantly more common in Turkish children with FUO (Chantada et al. 1994, Pasic et al. 2006, Joshi et al. 2008) (chi square, p < 0.05). Our results suggest that the spectrum of infectious diseases causing FUO appears to depend on geographic and social factors.

Brucelllosis is an explicit public health issue in Turkey. Although Brucella abortus is rare in children, in endemic Brucella melitensis areas such as Turkey, children account for 20%–25% of cases (Mantur et al. 2004). The signs and symptoms of brucellosis are not pathognomonic. The majority of illnesses in infected children are various, with mild to moderate severity (Tanir et al. 2009). Fever is the most common symptom of brucellosis, occurring in 80%–100% of cases. FUO is an uncommon initial diagnosis in these patients in nonendemic areas. The primary manifestation of brucellosis is FUO, with spiking fevers lasting for up to 4 weeks (Long and Edwards 2008). Some patients with brucellosis presented as FUO have tender hepatomegaly, splenomegaly, epistaxis, arthralgia, myalgia, and weight loss. The median duration of fever for patients with brucellosis presented as FUO was 26 days. Eight children with brucellosis diagnosed with FUO had fever as the only complaint. Five of them had fever along with hepatomegaly and/or splenomegaly. One of them had fever and arthritis.

The diagnosis of brucellosis is based on potential exposure to unpasteurized milk and milk products, clinical features suggestive of brucellosis, and serological tests with or without positive blood culture. Isolation of Brucella spp. is the hallmark of diagnosis. However, the rate of isolation from blood is generally low. Serologic tests are the main tools of diagnosis of brucellosis. To diagnose acute cases, assays have been used to detect the presence of specific immunoglobulin M antibodies. SAT for Brucella with titers ≥1/160 is suggestive of active infection (Mantur et al. 2004, Tanir et al. 2009). In our cases, the diagnosis of brucellosis was confirmed by SAT in 11 patients and positive blood cultures in 3 patients. Previous suboptimal antibiotic courses and negative blood cultures were the main causes for delayed diagnosis of brucellosis. In two patients, after discontinuation of antibiotics, repeated blood cultures confirmed the diagnosis of brucellosis.

In conclusion, childhood brucellosis remains a public health problem in Turkey. Brucellosis in children has a wide range of clinical features and may be presented with FUO, particularly in endemic areas. Infectious disease remains as a common cause of FUO, and brucellosis is the most frequent infectious cause of FUO in children in Turkey. Although several infectious diseases may lead to FUO, particularly brucellosis should be considered in the differential diagnosis of children admitted with FUO in Turkey.