SUMMARY Patients with a clinical diagnosis of tuberculous lymphadenitis (TBLN) undergoing fine needle aspiration (FNA) biopsy at Port Moresby General Hospital (PMGH) were recruited in a pilot study to compare Xpert® MTB/RIF (Xpert) with microscopy and cytology. From a total of 1080 patients attending the FNA clinic during the study period 107 were recruited, of whom 105 were analysed. Xpert detected Mycobacterium tuberculosis in 65/105 subjects (62%), acid-fast bacilli (AFB) were found in 35/105 (33%) and cytology was positive in 59/105 (56%). 3 of 7 samples unsuitable for microscopy and 9 of the 28 cases (32%) initially classified as non-TBLN were Xpert positive. Xpert was comparable to cytology but more sensitive than microscopy. Xpert also detected multidrug-resistant tuberculosis (MDR-TB) TBLN cases. The results demonstrated that FNA samples are suitable for Xpert analysis at PMGH to diagnose TBLN, which has the added advantage of detecting MDR-TB.

Data on the accuracy of Xpert® MTB/RIF (Xpert) assay in detecting TB in lymph node aspirates in Papua New Guinea (PNG) is scanty. This study evaluated Xpert performance in diagnosing tuberculous lymphadenitis (TBLN) using lymph node needle aspirates at the Port Moresby General Hospital (PMGH). The objective of the study was to compare Xpert accuracy to acid fast bacilli (AFB) microscopy, cytomorphology, a composite reference test (CRS) and culture. A total of 107 eligible subjects were recruited out of 1080 clinic attendees. Results showed Xpert detected significantly more cases of TBLN than AFB microscopy (66 vs 35; p=0.001). Compared to AFB microscopy Xpert had a sensitivity of 45.4% (95% CI 33.1-58.1), specificity of 87.8% (95% CI 73.8-95.9), positive predictive value (PPV) of 85.7% (95% CI 71.6-93.4) and negative predictive value (NPV) of 50.0%% (95% CI 43.8-56.1). There was no difference between Xpert and cytomorphology (66 vs 60; p=0.5). Compared to cytomorphology Xpert had a sensitivity of 71.6% (95% CI 58.5-82.5), specificity of 51.1% (95% CI 35.7-66.3), PPV of 66.1% (95% CI 58.2-73.2) and NPV of 57.5% (95% CI 45.2-68.9). There was no difference between Xpert and CRS (66 vs 71; p=0.6). Compared to CRS Xpert had a sensitivity of 76.0% (95% CI 64.4- 85.3), specificity of 66.6% (95% CI 49.0-81.4), PPV of 81.8% (95% CI 73.5-87.9) and NPV of 58.4% (95% CI 46.7-69.4). Culture was completed on 24 subjects with positive isolates in 14 giving a culture yield of 58.3%. Of the 24 subjects, Xpert was positive in 21 subjects. There was no difference between Xpert and culture (21 vs 14; p=0.8). Compared to culture Xpert had a sensitivity of 100.0% (95% CI 76.8-100.0), specificity of 30.0% (95% CI 6.6-65.2), PPV of 66.6% (95% CI 57.1-75) and NPV of 100.0%. The results suggest Xpert is more sensitive than AFB microscopy but comparable to cytomorphology and CRS for TBLN diagnosis in the PNG context. Xpert can be used for diagnosing TBLN at PMGH

The pattern of antibiotics prescribed to patients with peripheral lymphadenopathy was assessed and compared with existing standard treatment guidelines (STGs) in Papua New Guinea (PNG). Information was obtained from patients referred to the Port Moresby General Hospital for fine needle aspiration biopsy by interviewing patients and reviewing patients’ clinic attendance record books and referral letters. Of the 107 patients recruited for the study, 51 (48%) were prescribed antibiotics and, of these, prescription data were obtained from 40 (78%). Amoxycillin, which is recommended as a first-line antibiotic for peripheral lymphadenopathy in PNG STGs, was prescribed in only 18/40 patients (45%). There was also high variability in other antibiotic selection, antibiotic combinations and treatment duration. The results highlight a need for ongoing training on rational antibiotic prescribing.