Introduction In an observational cohort of patients admitted with AECOPD, thrombocytosis was associated with inpatient and 1-year mortality.¹ We aimed to validate, and explore mechanisms for, this association within our original DECAF cohort (n = 920).²

Methods Admission platelet counts were categorised as low (<150), normal (150–400), or high (>400) x10⁹ cells/mm³ and odds ratios assessed for inpatient and, among those surviving to discharge, 1-year mortality (normal platelet count=reference). For inpatient mortality, platelet category and DECAF indices were included in multivariate logistic regression. The areas under the ROC curves for DECAF and DECAF+Platelets were compared by the method of DeLong. Associations with thrombocytosis were analysed using Mann-Whitney or Fisher’s exact test. Causes of death at 1-year due to respiratory, cardiac or malignant disease were recorded.

Results Thrombocytosis was associated with inpatient (OR 1.83, 95% CI 1.12–3.00, p = 0.016) and 1-year mortality (OR 1.62 95% CI 1.09–2.30, p = 0.017). Thrombocytopenia was associated with inpatient (OR 3.5, 95% CI 1.51–8.12, p = 0.004), but not 1-year mortality (OR 1.81, 95% CI 0.76–4.312.08, p = 0.181). On multivariate analysis, thrombocytosis (OR 1.85, 95% CI 1.03–3.33 p = 0.039) and thrombocytopenia (OR 3.00 95% CI 1.09–8.24 p = 0.033) independently predicted inpatient mortality, but did not improve predictive power of DECAF (AUROC: DECAF=0.86, DECAF+Platelets=0.86; p = 0.93).

Thrombocytosis was associated with a higher white cell count (p<0.001) and eMRCD score (i.e. more breathless when stable; p = 0.001), lower: albumin (p = 0.004), BMI (p = 0.002), FEV1 (p = 0.010), haemoglobin (p<0.001), and a lower proportion of women (p = 0.004), and patients with eosinopenia (<0.05 x 10⁹/l) (p = 0.008), cardiac death (p = 0.044), current smoking (p = 0.046), AF (p = 0.029) and diabetes (p = 0.006). Thrombocytosis was not related to cardiovascular disease, prior exacerbation and readmission rates or LTOT use, admission PaO₂, pH or NIV, or length of stay.

Discussion Thrombocytosis was an independent predictor of both inpatient mortality and, amongst survivors to discharge, 1-year mortality. Thrombocytosis was not associated with cardiovascular disease and the higher 1-year mortality was not due excess cardiovascular or cancer deaths, suggesting that other mechanisms are responsible. Whilst thrombocytosis was not associated with LTOT use or PaO₂, it was associated with other indices of disease severity, including breathlessness and lower FEV1, BMI and albumin level.

References

1. Harrison Thorax 2014

2. Steer Thorax 2012