Fine-needle aspiration to improve diagnosis of melioidosis of the head and neck in children: a study from Sarawak, Malaysia

Melioidosis, the infection caused by the environmental saprophyte Burkholderia pseudomallei, is associated with a high case fatality rate [1, 2]. This is due partly to the extensive intrinsic antibiotic resistance of the organism [3], but also to difficulties in case recognition and diagnosis [4, 5], especially in rural settings where the disease commonly occurs [6, 7].

Difficulties in clinical recognition of melioidosis arise mainly due to the varied and non-specific presentation of patients. Patients from different regions can have different disease manifestations [8], and there are also variations between pediatric and adult presentations of the disease [9, 10]. Difficulties in confirming suspected infections further lengthen the diagnostic process. Laboratory investigations, including bacterial culture, rapid antigen detection, direct real-time polymerase chain reaction assays (PCR), and serology all suffer from limited sensitivity or specificity [8, 11, 12]. For example, sensitivity and specificity of PCR on clinical isolates were reported to be 73.2% and 89.2%, respectively [13]. Bacterial culture, although 100% specific, may have a sensitivity of only 60% [11].

In view of and despite the limitations of the various diagnostic modalities, adequate sampling of the site of infection is key in diagnosing melioidosis [14]. The use of fine-needle aspiration to obtain pus and tissue samples from various body sites, including mediastinal lymph nodes, spleen, and prostate, has been reported to have enabled confirmation of melioidosis infections [15‐17]. This minimally invasive technique may be particularly useful for difficult to access sites and when surgical biopsies are not possible or available [18, 19].

Bintulu Hospital is a 302-bed hospital which provides primary and secondary health services to a total population of 256,000, including 72,000 children aged < 15 years, residing in Bintulu Division and Belaga district of Kapit Division in central Sarawak, Malaysian Borneo. Although adult general surgery and otorhinolaryngology services are available, specialized pediatric services are only available in a tertiary referral center located over 300 km away in southern Sarawak.

In 2017, we reported a broad description of melioidosis infections among a total of 42 children presenting to three hospitals in central Sarawak, including Bintulu Hospital [20]. One interesting finding was that cervical lymph node infections were frequently present, documented in 26% of cases. We also found that 24% of all children had a fatal outcome. In comparison, the pediatric case fatality rate was lower in other regions [9, 21]. The higher apparent fatality rate found in Sarawak may in fact suggest that many children with localized or less severe melioidosis had not been diagnosed, possibly related to the difficulties with diagnosis and limited resources. The aim of this study was to provide a detailed description of head and neck melioidosis among children in Sarawak and determine whether fine-needle aspiration of suspected head and neck lesions could facilitate the detection of B. pseudomallei and improve melioidosis diagnosis.

We performed a retrospective descriptive study of all children with melioidosis of the head and neck presenting to Bintulu Hospital in Sarawak, Malaysian Borneo, over a 10-year period from January 2011 to December 2020.

Based on admission body weight, poor nutritional status was diagnosed in children aged ≤ 10 years if weight-for-age was below − 2 z-score using the World Health Organization Child Growth Standards [22], while children aged > 10 years were considered to have poor nutritional status if their body weight was below the 3rd percentile using the Centers for Disease Control and Prevention weight-for-age percentiles [23]. Lymphadenopathy was defined as an enlarged, non-tender lymph node with no obvious signs of inflammation. Lymphadenitis was defined as an enlarged, tender lymph node, with the presence of signs of inflammation. A lymph node abscess was diagnosed based on the presence of fluctuance in an enlarged lymph node. Melioidosis was considered to be localized if a single or multiple organ-site(s) of involvement were confined to the head and neck, in the absence of a positive blood culture and clinical/radiological/microbiological evidence of dissemination to other non-head or neck sites. Disseminated melioidosis was defined as the presence of infection in ≥ 1 discrete body sites other than those found in the head and neck and/or a positive blood culture. Septic shock was defined by the presence of hypotension (systolic blood pressure below the 5th percentile) with evidence of inadequate tissue perfusion unresponsive to fluid replacement [24]. Leukocytosis was defined by an elevated white blood cell count > 14 × 10⁹/L in children < 2 years, > 12 × 10⁹/L in children 2–9 years, and > 10.5 × 10⁹/L in children > 9 years [25]. Neutrophilia was defined as an absolute neutrophil count > 8.5 × 10⁹/L in children < 6 years and > 8.0 × 10⁹/L in children 6 years or older. Monocytosis was defined as an absolute monocyte count > 1.0 × 10⁹/L. Thrombocytosis was defined as a platelet count > 400 × 10⁹/L. A melioidosis-active antibiotic regimen was defined as the use of either ceftazidime or a carbapenem during the intensive phase and oral trimethoprim–sulfamethoxazole during the eradication phase, with amoxicillin–clavulanate or doxycycline used as second-line oral agents when resistance, contraindications, or intolerance to trimethoprim–sulfamethoxazole was present [13].

Before 2016, pus and other samples from the head and neck were mainly collected during incision and drainage procedures using sterile bottles before being transported as quickly as possible to the microbiology laboratory. These specimens were then cultured directly on blood agar (BA), chocolate agar (CA), and MacConkey agar (MAC) (standard methods). From 2016, in an attempt to improve culture-confirmation of melioidosis infections, a fine-needle aspiration technique with inoculation in blood culture bottles (FNA + BCB) was used in addition to standard methods. With this method, fine-needle aspiration of affected lymph nodes and other lesions in the head and neck were performed using a 23-gauge needle, similar to fine-needle aspiration biopsy procedures previously described [26]. Aspirated contents were inoculated directly into pediatric BACTEC blood culture bottles (Becton Dickinson, USA). When minimal (i.e., pus/tissue present only in the hub of the needle) or no pus/tissue was aspirated, a small amount of culture medium from the blood culture bottle was aspirated through the needle into the syringe and flushed back into the bottle. Culture bottles were then incubated in the BACTEC blood culture system (Becton Dickinson, USA). Positive growth was subcultured onto BA, CA, and MAC.

B. pseudomallei was identified with either API^®20NE (BioMérieux, France) or BBL™ Crystal™ Identification Systems (Becton Dickinson, USA). Blood samples collected from patients were subjected to the BACTEC blood culture system.

Statistical analysis was performed using SPSS Statistics 21. Continuous variables were presented using the median and interquartile range (IQR).

This study reports the demography, clinical presentations, laboratory findings, treatment, and outcome of children with melioidosis of the head and neck in Sarawak, Malaysian Borneo. We show that children with head or neck melioidosis in Sarawak have distinct manifestations with involvement predominantly of cervical lymph nodes. In addition, we describe how the use of fine-needle aspiration improves diagnosis of melioidosis in these children.

The head and neck region was a major site of melioidosis infection in Sarawakian children, with melioidosis of the head and neck representing nearly 60% of the melioidosis cases diagnosed between 2011 and 2020. Head and neck infections are a well described feature of pediatric melioidosis; in Cambodia and Thailand, 65% and 33% of children, respectively, are reported to have head or neck involvement [27, 28]. Interestingly, however, head and neck infections are rarely reported in children in northern Australia [6, 9]. In adults, head and neck melioidosis is uncommon [29, 30]. The high prevalence of head or neck infections among children in Southeast Asia is thought to be related to acquisition of B. pseudomallei through the oral/nasal mucosa or conjunctiva, possibly via exposure to contaminated water sources, as opposed to inhalation and ingestion postulated in adults with pneumonia or bacteraemia and skin inoculation among children in northern Australia [31, 32].

In our study, cervical lymph nodes were the most common organ-site involved in the head and neck, and differentiating B. pseudomallei infection from other common causes of cervical lymph node disease, including pyogenic or tuberculous (TB) infection, by clinical, routine laboratory, and even histologic examination was difficult. Cervical lymph node involvement was present in 19 (95%) of our 20 cases and was the primary reason for admission in 80% (16/20) of the cases. These results contrast with findings from Cambodia and Thailand, where salivary glands, and in particular parotid glands, are the most common site of infection [33, 34]. It is unlikely that this discrepancy was due to misdiagnosis of parotitis among children in our study, as none of the children had pre-auricular swelling. Additionally, 2/3 had ultrasound imaging confirming cervical lymph node involvement, although in 25% small parotid abscesses were simultaneously found. It is uncertain whether parotitis is being over-diagnosed in the rest of Southeast Asia, although it remains possible that a true difference in predilection of organ-sites exists. Reasons for this, however, are not yet known. In our study, a broad range of cervical lymph node presentations from unilateral painless adenopathy to bilateral lymph node abscesses were found. Most laboratory findings were non-specific except for monocytosis which occurred in 63%. Whilst monocytosis has been shown to be a useful predictor of TB [35], the utility of this hematological finding for melioidosis diagnosis has not been evaluated. Histological examination showed findings that were either indistinguishable from TB or were non-specific. Thus, a high index of suspicion for melioidosis is required in children presenting with cervical lymph node swelling in melioidosis-endemic regions, and microbiological confirmation is essential.

Fine-needle aspiration, mainly of affected cervical lymph nodes, improved the diagnosis of melioidosis of the head and neck among children in this study, as it facilitated sampling of the site of infection and enabled microbiological confirmation of melioidosis (Additional file 3: Table S2). Indeed, the diagnosis was obtained in 75% of our cases using fine-needle aspiration, and these cases were not diagnosed by any other method. In addition, over 5 times more culture-confirmed melioidosis cases were detected in the second half of the study, mainly attributable to the use of fine-needle aspiration. Before 2016, with the exception of the three cases reported in this study, most diagnoses of head and neck melioidosis in Bintulu Hospital were made based on clinical and serological findings. Blood cultures were usually negative, and unless obvious collections were present, pus/tissue from lesions were rarely sampled as the lack of pediatric surgical/otorhinolaryngology services resulted in a reluctance to perform biopsies/drainage procedures. When drainage procedures were performed (usually late in the course of the illness), pus cultures were frequently negative. In an attempt to narrow the diagnostic gap, FNA + BCB was initiated at the beginning of 2016. Fine-needle aspiration is a simple and safe procedure, which may be performed at the bedside, and has routinely been used in the diagnosis of TB lymphadenitis [36, 37]. Complications of fine-needle aspiration, including bleeding, damage to surrounding structures, fistula creation, or seeding of tissue planes with infection or neoplastic cells, are rare [38]. In contrast, open/excisional biopsies are more invasive and require surgical expertise and general anesthesia. Fine-needle aspiration may thus be a useful option in the resource-limited regions where most melioidosis infections occur.

The use of blood culture bottles for bacterial culture of non-blood specimens has previously been reported. For example, improved yield has been documented in the culture of synovial fluid, pleural fluid, and brain abscess biopsies [39‐41]. A number of reasons may explain the higher rates of B. pseudomallei isolation with the use of blood culture bottles compared to direct plating on agar reported in the present study. Probably most importantly, the use of blood culture bottles enabled detection of B. pseudomallei even when no or minimal pus/tissue was present, using a technique of ‘rinsing’ the biopsy needle and syringe by aspirating contents of the blood culture bottle. This similar technique has been shown to be useful in culture-confirmation of TB lymphadenitis [42]. The ability to isolate B. pseudomallei from infected lymph nodes even when no pus is present, with similar detection times as for large samples of purulent aspirates, suggests a sensitive approach with the potential to enable earlier diagnosis. Other reasons for the improved yield may include the benefits of liquid culture media versus solid culture media and the presence of antibiotic-binding resins in commercial blood culture media [43‐45]. However, it is probable that any suitable enrichment broth (other than blood culture media) would have performed equally well, and this should be evaluated in future studies.

Even though bacteremia was rarely detected among children with melioidosis of the head and neck, disseminated infections were diagnosed in 65% of cases. The spleen was the most common organ affected in disseminated infections; occult splenic abscesses were present in 85% (11/13) of children with disseminated disease and in 58% (11/19) overall. As previously reported, detection of splenic abscesses by abdominal ultrasonography offers a useful and rapid clue to melioidosis diagnosis [46]. In addition, the detection of occult splenic abscesses in children with seemingly localized head and neck melioidosis in this study is also important in view of recent suggestions that localized infections might adequately be treated with oral antibiotics alone [21]. Clearly, a thorough diagnostic work-up to exclude occult dissemination is needed before such a treatment decision can be considered. Alternatively, in melioidosis-endemic regions, the detection of a head or neck focus of infection (e.g., cervical lymph node infection) in a child presenting primarily with infection in another organ-site, for example pneumonia or septic arthritis, may suggest a diagnosis of melioidosis and the need for empirical melioidosis-active antibiotics.

This study has several limitations. Firstly, throat swabs for culture of B. pseudomallei were not performed. Throat swabs have been shown to be useful in the diagnosis of melioidosis and may have been especially beneficial in the present study, as some of the children could have had a pharyngo-cervical syndrome as reported in previous studies [27, 47]. Throat swab culture for B. pseudomallei requires the use of selective media, such as Ashdown’s agar. However, Ashdown’s agar cannot be used in Sarawak as the majority of cases are caused by gentamicin-susceptible B. pseudomallei [48], while other B. pseudomallei selective media and modified Ashdown (with colistin instead of gentamicin) are not readily available. Secondly, this study does not report cases of suspected head and neck melioidosis in which all cultures, including FNA + BCB, were negative and as such cannot inform on sensitivity of the method. This is mainly related to the study methodology and the lack of a gold standard for diagnosis. Another limitation of the study was that no cost-effectiveness calculations were undertaken. These may be important, as melioidosis is largely an affliction of resource-limited regions. Other limitations include the small sample size, the possibility of selection bias related to the monocentric nature of the study, and limitations inherent to the retrospective (bias of information) descriptive (bias of confusion) study design. Despite these limitations, the study manages to provide a fairly comprehensive characterization of pediatric head and neck melioidosis in Sarawak, of which very limited data was previously available.

Nearly 60% of children with melioidosis in Sarawak had head or neck involvement. Cervical lymph nodes were the most common site of infection in the head and neck; salivary and lacrimal gland infections also occurred but were less frequently observed. Fine-needle aspiration, mainly of affected cervical lymph nodes, was extremely useful in enabling microbiological confirmation of melioidosis, allowing the diagnosis of 75% of our cases. Further studies are needed to confirm the sensitivity, specificity and cost-effectiveness of this simple and safe technique, which may be particularly useful in resource-constrained regions.