The diagnostic yield for cytomegalovirus (CMV) polymerase chain reaction (PCR) viral load in bronchoalveolar lavage (BAL) or in plasma to diagnose CMV pneumonia in lung transplant recipients remains uncertain and was investigated in a large cohort of consecutive lung transplant recipients.Methods
Bronchoscopies within the first year of lung transplantation with CMV detectable in BAL by PCR (ie, viral load, ≥273 IU/mL) were included (66 recipients; 145 bronchoscopies); at each bronchoscopy episode, 2 independent experts reviewed clinical and laboratory information to determine whether the patient at that time fulfilled the criteria for CMV pneumonia per current international recommendations. Corresponding plasma CMV PCR viral load determined at time of the bronchoscopy (n = 126) was also studied. Optimal CMV PCR viral load cutoff for CMV pneumonia diagnosis was determined using receiver operating characteristics.Results
CMV was detected in BAL with CMV PCR in 145 episodes, and 34 (23%) of these episodes fulfilled the criteria for CMV pneumonia. The area under the curve-receiver operating characteristics for CMV in BAL was 90% at the optimum cutoff (4545 IU/mL) with a corresponding sensitivity of 91% and specificity of 77% (in plasma the corresponding values were 274 IU/mL, 63% and 76%, respectively).Conclusions
CMV PCR viral load in BAL had a high performance to diagnose CMV pneumonia in lung transplant recipients; plasma CMV viral load did not reliably aid as a diagnostic tool.