Evaluation of Clinical and Sonographic Features in 55 Children with Tularemia

Introduction

Tularemia is a zoonotic disease caused by the intracellular, Gram-negative bacteria Francisella tularensis (Feigin et al. 2009). The disease is predominant in North America, Europe, and Asia. Transmission of F. tularensis to humans occurs via the bite of infected animals, contact with contaminated animal products, inhalation of aerosol droplets, and ingestion of contaminated food or water (Sandström et al. 1992, Ellis et al. 2002). Tularemia has several clinical forms in humans that are primarily dependent on the route of acquisition. Worldwide, ulceroglandular tularemia is generally the most common form, whereas oropharyngeal tularemia is the most common form in Turkey.

Most studies in the literature are not solely focused on pediatric population. However, there are a few larger series in children. A study from northern Finland that was performed during an outbreak of holarctica-type tularemia in 2007 reported that 50 pediatric patients had the ulceroglandular form (Jounio et al. 2010). A study from Turkey reported that 23 children with tularemia had the oropharyngeal form, whereas four had the glandular form (Kaya et al. 2012).

Cervical, preparotid, axillary, and inguinal lymphadenopathies (LAPs) are common presentations of the oropharyngeal and glandular forms of tularemia (Helvaci et al. 2000, Celebi et al. 2006). Enlarged lymph nodes and abscess formation that can be visualized via ultrasonography (USG) are often described in patients with the oropharyngeal form of tularemia. USG criteria have been described to help differentiate between benign and malign causes of LAP; however, criteria for differentiating between the many causes of benign LAP have yet to be described (Ahuja et al. 1995, Steinkamp et al. 1995). The literature includes only a few studies on USG findings in tularemia. Ulu-Kilic et al. (2013a) described the sonographic features of LAP in 108 tularemia patients from Central Anatolia; the lymph nodes in the tularemia patients were ultrasonographically observed as hypoechoic, round, lacking an echogenic hilus, and tending to exhibit intranodal cystic necrosis. The aim of the present study was to determine the clinical characteristics and sonographic features of LAP and to evaluate the treatment modalities and treatment outcomes in children with tularemia.

Materials and Methods

Demographic data, clinical characteristics, treatment regimens, and ultrasonographic findings in 55 patients diagnosed with tularemia at Dr. Sami Ulus Maternity and Children's Research and Education Hospital, Ankara, Turkey, between December, 2009, and October, 2012, were evaluated. Patient records were reviewed retrospectively in terms of gender, age, city of residence, source of and risk factors for infection, and the most common symptoms. Diagnosis of tularemia was based on a positive history and clinical findings compatible with tularemia, and serological confirmation via the microagglutination test (MAT).

The tularemia MAT was performed at the Public Health Institution of Turkey. Antibody titers ≥1:160 were regarded as tularemia. Ultrasonographic examinations were performed by the same experienced pediatric radiologist using a Logiq 6 ultrasound (US) machine (General Electric Co., Milwaukee, WI) and a 7.5- MHz linear-array small parts transducer. Lymph nodes were classified as belonging to one of the following eight regions, according to their location in the neck, as similarly described by Hajek et al. (1986)—submental; submandibular; parotid; upper cervical (above the hyoid bone and along the common carotid artery and internal jugular vein); middle cervical (between the hyoid bone and cricoid cartilage, along the common carotid artery and internal jugular vein); lower cervical (below the cricoid cartilage); supraclavicular fossa; and posterior triangle.

Inguinal and axillar LAPs were noted in some patients. Assessment of lymph node morphology included nodal shape, size, homogeneity, definition of borders, presence of echogenic hilum, echogenicity, intranodal necrosis, abscess formation, status of surrounding soft tissues, and conglomeration. Lymph node shape was assessed via measurement of the largest and smallest diameters on the same scan, and by calculating the short axis–long axis (S/L) ratio. An S/L ratio <0.5 indicated an oval node, whereas an S/L ratio ≥0.5 indicated a round node. The degree of necrosis within lymph nodes was classified in three stages by the pediatric radiologist, as follows: Stage 1, multiple cortical microabscesses observed as millimetric nodular hypoanechogenic regions located in the cortex of the lymph node; stage 2, cystic regions within the cortex and medulla of the lymph node, and some solid preserved lymph node parenchyma exists; and stage 3, lymph node totally transformed into an abscess.

Patients with an antibody titer of ≥1:160 were treated with gentamicin (7.5 mg/kg per day, three doses, intravenously); gentamicin (intravenously) plus ciprofloxacin (20 mg/kg per day, two doses, per orally); ciprofloxacin per orally or doxycycline (2–4 mg/kg per day per orally) plus gentamicin combination for 14 days. In addition, the time from the onset of symptoms to initiation of appropriate antibiotic treatment was recorded. Treatment failure was defined as an increase in size of the current LAP or appearance of a new LAP, or LAP suppuration despite sufficient medical treatment. Surgical drainage was performed in patients with lymph node fluctuation or treatment failure.

Statistical Package for Social Sciences for Windows, version 16.0 for Windows (SPSS Inc., Chicago, IL) was used for statistical analysis. Descriptive statistics are presented as frequency, percentage, and mean±standard deviation (SD). The level of statistical significance was set at p<0.05.

Results

The study included 55 patients diagnosed as tularemia. In all, 31 (56%) of the patients were male and 24 (44%) were female; mean age was 10.8±4.0 years (range, 2–17 years). Among the patients, 34 (62%) were living in Ankara, 18 (33%) were living in rural areas surrounding Ankara, and three (5%) were living in the central Anatolian Region of Turkey. In all, 20 (36%) of the patients had a history of drinking spring water, contact with animals, and swimming in a lake or stream. In total, 50 (90%) of the patients had oropharyngeal tularemia, four (8%) had glandular tularemia, and one (2%) had oculoglandular tularemia.

The most frequent symptoms were sore throat (67%) and fever (64%), and LAP (100%) was the most common sign, of which submandibular LAPs were the most common. Sonographic measurement of the involved lymph nodes ranged from 14×5 mm to 56×47 mm (mean short axis, 20±9.5 mm; mean long axis, 33±9.2 mm). All involved nodes were hypoechoic. Echotexture of the involved nodes was heterogeneous in 39 (71%) patients. Abscess formation was noted in 36 (65%) patients, of which seven (19%) had multiple cortical microabscesses classified as stage 1. Among the patients with abscess formation, such as cystic regions within the cortex and medulla of the lymph node, 20 (55%) were classified as stage 2. Lymph nodes totally transformed into an abscess classified as stage 3 were observed in nine (26%) patients. Distribution of the involved lymph nodes and their sonographic features are summarized in Table 1.

Among the patients, 50 (90%) had a history of β-lactam antibiotic treatment as initial therapy. All treatment modalities are summarized in Table 2. Treatment failure was observed in 24 (44%) of the 55 patients. There was no significant correlation between treatment regimen and the rate of treatment failure (p>0.05). In all, nine (16%) of the patients did not respond to medical treatment and required surgical intervention; total LAP excision was performed in each case, and histopathological examination showed chronic granulomatous inflammation.

The time from onset of symptoms to initiation of appropriate antibiotic treatment was 1–12 weeks (mean, 4 weeks). The treatment failure rate was significantly higher in the patients in which treatment was given >3 weeks (p<0.05). Increased short and long LAP axis dimensions were strongly associated with treatment failure (p<0.05). The rate of abscess formation was 38% in cases in which initiation of appropriate treatment was ≤3 weeks, versus 62% in those that received treatment in >3 weeks (p<0.05). The relationship between time to treatment initiation and stage of abscess formation in lymph nodes are shown in Table 3.

Discussion

Tularemia is a bacterial zoonosis with diverse clinical manifestations that vary according to the route of transmission (Ellis et al. 2002). The disease is endemic in Turkey. Outbreaks due to ingestion of contaminated water have occurred in the western regions of Turkey and the Marmara and Black Sea coasts, and at the end of 2009 an outbreak of tularemia occurred in Central Anatolia. The main clinical form of these outbreaks was oropharyngeal tularemia (Ulu-Kılıc et al. 2011). In the present study, oropharyngeal tularemia was the most common form, followed by the glandular form. Oropharyngeal tularemia may present with LAPs as the only presenting symptom, which are localized most commonly to the head and neck region. Cervical LAPs were reported as the most frequent physical finding in epidemics in Turkey, as in the present study (Helvaci et al. 2000). Sore throat, fever, and cervical adenitis—the classical triad of oropharyngeal tularemia—were also the most common presenting signs and symptoms in the present study. Patients that are initially misdiagnosed as upper respiratory tract infection, parotitis, acute bacterial lymphadenitis, or dental infections may eventually present with isolated cervical adenitis (Penn 2010).

Enlarged, palpable lymph nodes are common in children with tularemia due to the reactive hyperplasia of lymphatic tissue that is primarily connected with local inflammation. The etiological diagnosis of LAP is based on epidemiological history, serological testing, culturing of the primary infection, and, if needed, histopathological investigation. Enlarged lymph nodes can be detected via various imaging modalities, including USG, computed tomography, and magnetic resonance imaging (Som et al. 2011). USG is often performed for the investigation of LAPs and can aid in the diagnostic work-up. USG is advantageous because it is readily available, fast, highly sensitive (98%), and highly specific (95%) (Khanna et al. 2011).

It was reported that the imaging features of tularemia are nonspecific, but include enlarged lymph nodes with central necrosis (Robson et al. 2000). There are only a few studies on sonographic findings in serologically proven tularemia. The sonographic features of LAPs in tularemia have been described by Ulu-Kilic et al. (2013a) in 108 patients from Central Anatolia. They reported that lymph nodes in tularemia tend to be hypoechoic, round, without an echogenic hilus, and tend to exhibit intranodal cystic necrosis. In the present study, the lymph nodes in the tularemia patients were hypoechoic and tended to be heterogeneous, oval, and without an echogenic hilus. These findings are consistent with those reported by Ulu-Kilic et al. (2013a), except for the shape of the lymph nodes. It is considered that lymph node enlargement with an S/L ratio <0.5 is a strong criterion for benignity (Vassallo et al. 1993). In the present study 60% of the lymph nodes involved were oval, in contrast to Ulu-Kilic et al. (2013a), who reported primarily spherical lymph nodes. Oval lymph nodes, as observed in the present study, might be considered novel; however, the literature contains insufficient data regarding the sonographic appearance of cervical lymph nodes in children with tularemia, which prohibits making a reliable comparison.

To the best our knowledge, the present study is the first to stage the sonographic appearance of LAPs in pediatric tularemia patients. LAP staging was performed in an effort to correlate the clinical findings with the sonographic extent of disease. Although USG is not useful for diagnosing tularemia, it was proposed to have a role in staging and may help identify patients that require surgical drainage.

In cases of tularemia, after the infective organism is introduced into a susceptible host, multiplication occurs locally, with early spread to regional lymph nodes occurring within 48– 96 h (Feigin et al. 2009). LAPs in tularemia patients are histologically divided into three forms, according to the phase of infection. In the early phase, until the end of the first week of infection, an abscess with central necrosis is formed. From the second to sixth week, abscess–granuloma formation occurs, in which there are several small epithelioid granulomas with central necrosis at the cortex and paracortex (Kitamura et al. 1956). According to the present study's USG findings, the LAP with central necrosis and abcess–granuloma formation may be consistent with stages 1 and 2. After the sixth week of infection, necrosis is homogenized and has progressed to caseous necrosis in the center of the granulomatous lesion, and this form may be stage 3, according to USG findings. The sonographic features of involved lymph nodes may also be dependent on the duration of the disease. It was reported that, provided appropriate therapy is started within a week of the onset of disease (i.e., fever), lymph node swelling resolves without further complications; however, when treatment is delayed for >2 weeks, the risk of lymph node suppuration is as high as 30–40% (Tarnvik et al. 2007).

In the present study, treatment outcome was good in patients that started treatment within 3 weeks of the onset of symptoms. When there is a considerable delay in the start of effective tularemia treatment, the lymph nodes continue to enlarge, which can result in abscess formation. The increased short- and long-axis dimensions of enlarged lymph nodes in the present study were strongly correlated with treatment failure (p<0.05), as reported by Ulu-Kılıc et al. (2013a). The rate of abscess formation and the presence of necrosis in LAPs were lower in the present study's patients that received appropriate treatment <3 weeks after the onset of symptoms, as compared to those that received treatment in >3 weeks (p<0.05). In the present study, multiple cortical microabscesses (stage 1) were observed as millimetric nodular hypoanechogenic regions located in the cortex of lymph nodes, which is consistent with lymphoid follicles in the early stages of infection (<3 weeks), and total nodal necrosis (stage 3) was observed in the late stages of infection (>3 weeks). Nodal necrosis occurred more frequently (87.5%) in the patients with treatment failure.

Aminoglycosides, such as streptomycin and gentamicin, are the first-line drugs for the treatment of tularemia (Penn 2010). Tetracycline and chloramphenicol antibiotics are also used to treatment tularemia; because they are bacteriostatic, they are only considered alternatives to aminoglycosides. Fluoroquinolones are also effective against F. tularensisas alternative treatments for tularemia (Kaya et al. 2012). In a recent study, the lowest minimum inhibitory concentration (MIC) values were obtained using fluoroquinolones, and the researchers suggested they could be considered as a first-line treatment for tularemia (Ulu-Kılıç et al. 2013b). Nonetheless, the use of tetracyclines in children aged <8 years and of fluoroquinolones in children aged <18 years is limited in the treatment of tularemia (Feigin et al. 2009). In the present study, there was no difference in the success rate between aminoglycoside treatment and the other regimens. The rate of effectiveness was similar when considering only patients in the early treatment group. Most of the patients in the present study received gentamicin treatment, and combination treatment regimens were used in patients in which gentamicin therapy failed. Orally administered tularemia treatment has become more important, especially in endemic regions. The effectiveness of ciprofloxacin was not evaluated in the present study due to the limited number of patients that received ciprofloxacin monotherapy.

In conclusion, tularemia is a re-emerging disease in Turkey. Epidemiological findings and clinical features may prompt clinicians to consider tularemia in the differential diagnosis of LAPs. Early clinical diagnosis and timely appropriate antibiotic treatment are important for obtaining good treatment outcome. Gentamicin is a drug of choice in the treatment of tularemia in children; however, alternative treatments may sometimes be necessary, especially in cases of delayed diagnosis and treatment. Staging the sonographic findings of LAPs in children with tularemia, even though there are no specific criteria, might be useful for estimating the duration of disease prior to presentation, increasing the likelihood that treatment will be successful, and identifying patients that require surgery.