Detection of Helicobacter pylori in mechanically ventilated patients: The LARA-¹³C-urea breath test and serology

Objective: Of critically ill patients in intensive care 0.6-9% will develop ulcerations in the stomach with overt bleeding. Recently, the presence of antibodies against Helicobacter pylori was found to be associated with an increased rate of upper gastro-intestinal bleeding in critically ill patients. The exact role of H. pylori in the pathophysiology of stress ulceration is not known, but is currently being investigated without prior validation of diagnostic tests in this specific patient population. Therefore, we studied the accuracy of the commonly used diagnostic tests for H. pylori in mechanically ventilated patients. Design: Prospective study comparing the Laser Assisted Ratio Analyser-¹³C-urea breath test (LARA¹³C-UBT) in ambulant and mechanically ventilated patients. Antibodies against H. pylori were detected in the same patients. Setting: A 20 bed mixed medical, surgical and cardiac ICU in a major teaching hospital in the centre of Amsterdam. Subjects: One hundred consecutive patients admitted for elective cardiac surgery. Interventions: A pre-operative ambulant LARA-¹³C-UBT was performed to detect H. pylori and was used as the 'gold standard'. Post-operatively, a second LARA-¹³C-UBT was performed on the same patient in the ICU during mechanical ventilation. Serum antibodies against H. pylori were also determined in the ambulant and mechanically ventilated state. Endpoints: The LARA-¹³C-UBT and serology in mechanically ventilated state were compared to the LARA¹³C-UBT and serology in ambulant state. Measurements and main results: The LARA-¹³C-UBT during mechanical ventilation reached a sensitivity of 94% and a specificity of 92% compared to the ambulant test. Positive predictive value was 88% and negative predictive value was 96%. The mean serum H. pylori antibody titre of all patients decreased from 4.09 U/l to 3.34 U/l (16%) postoperatively (p < 0.0001), in all probability due to blood loss and haemodilution, which resulted in three false negative tests post-operatively. Six false positive pre-operative serological tests could be related to antibiotic use in the previous year. Sensitivity of post-operative serology was 72% and specificity 70%. Conclusions: The LARA-¹³C-UBT is an accurate and noninvasive method to detect H. pylori in mechanically ventilated patients and is easy to carry out. Serum antibody testing is less accurate due to blood loss, haemodilution and previous antibiotic use and should therefore be interpreted with caution.