Introduction Small intestinal bacterial overgrowth (SIBO) is probably the most common cause for chronic gastrointestinal (GI) symptoms following cancer treatments. There is no diagnostic gold standard. We assessed whether the glucose hydrogen methane breath test has greater value than the hydrogen breath test alone and whether a duodenal (D2) aspirate improves the diagnostic yield.

Methods Patients in a cancer centre referred for potential SIBO. Breath hydrogen (H2) and methane (CH4) were measured in parts/million (ppm) using a QuinTron BreathTracker DP at baseline, then—after 75 g oral glucose—at 20 min time points up to 3 h or until positive. Positive test: fasting H₂ ≥20 or CH₄ ≥10 ppm or a rise in H₂ ≥12 or CH₄ ≥6 ppm. Some patients also had duodenal aspirates collected endoscopically. Positive result: >10⁴ colony forming units/ml. Patients with positive tests were treated with antibiotics. Results were assessed retrospectively.

Results 126 patients, 66 males and 60 females, median group age 61 years (range 35–86), treated for gynaecological (30%), upper GI (25%), lower GI (9%), urological (28%), other cancers (10%—myeloma, ependymoma, bronchial, breast and lymphoma) were referred. Daily troublesome GI symptoms included flatulence (85%), borborygmi (63%), belching (54%), bloating (52%), abdominal pain (50%), steatorrhoea (38%), nausea/vomiting (23%) and diarrhoea (19%). 60 (48%) had a positive breath test—5% H2, 10% CH4 and 33% both gases. 21 (17%) of 86 patients with D2 aspirates tested positive for Streptococcus (n=6), E. Coli (3), Candida (3), Klebsiella (2), Enterococcus (2), Pseudomonas (2), Neisseria (1), Aeromonas (1) and Stenotrophomonas (1). 17 (33%) had negative breath tests. Six of these had positive D2 aspirates with 3 (50%) responsive to antibiotics, 11 had negative aspirates but a 73% antibiotic response rate. 24 (46%) tested positive for both gases. 33% of these had a positive D2 aspirate with a 75% response rate to antibiotics and 67% negative aspirates but 81% response. Six patients (12%) were positive only for H2 (all had negative aspirates), 1 responded to antibiotics. 5 (10%) tested positive for CH4 only, of which 1 had a positive D2 aspirate with a response to antibiotics and 4 (80%) had negative aspirate but 2 responded.

Conclusion Methane breath testing identifies 10% more patients with SIBO compared to the glucose hydrogen breath test alone. D2 aspirate increases the detection rate by 12%. A trial of antibiotics, with other tests negative, benefits 15% of patients. All current diagnostic methods are flawed. Better diagnostic tests for SIBO are required.

Competing interests None declared.