Cytomegalovirus viral load as predictor of the clinical course of hypoxic pneumonia in children

BACKGROUND: Children under 1 year of age with hypoxic pneumonia regularly have concurrent cytomegalovirus (CMV) viremia. In these children, the diagnosis of CMV-associated pneumonia and the prediction of an outcome are difficult. It is unclear whether quantification of blood CMV viral load (CMV-VL) can predict outcomes in these children. METHODS: This was a retrospective study including children (1–12 months of age), with detectable CMV-VL and hypoxic pneumonia admitted to the paediatric intensive care unit of Tygerberg Hospital, Cape Town, South Africa between 1 January 2014 and 31 December 2015. Clinical, radiological and biochemical data were collected. RESULTS: Of the 87 participants included (median age: 3.9 months, IQR 2.2–4.8), 35 were (40%) born prematurely. The median weight-for-age Z -score was –2.68 (IQR –3.0 to –0.83); 37 (43%) were severely underweight for age; 27 (31%) were HIV-positive, 3 were on ART. The median CMV-VL was log 4.0 (IQR 3.3–4.79); CMV high was defined as CMV-VL > median; CMV-VL < median was classified as CMV low . Overall survival was 90%; 12 (15.4%) remained oxygen-dependent at Day 28 post-admission. There was no difference in survival, 24-h post-admission ratio of arterial oxygen partial pressure to fractional inspired oxygen (P a O 2 :F i O 2 ), oxygen dependence or ventilation duration between CMV low and CMV high . High-frequency oscillation ventilation duration was longer ( P = 0.005) and Pneumocystis jirovecii (PJP) co-infection more frequent ( P = 0.018) in CMV high . CONCLUSION: CMV-VL is unable to predict the clinical outcome in children with hypoxic pneumonia. Specific treatment for CMV should be considered in all children at risk of CMV-associated pneumonia with detectable CMV-VL.