Trends in Acute Pulmonary Embolism Admission Rates and Mortality Outcomes in Australia, 2002-2003 to 2017-2018: A Retrospective Cohort Study.

BACKGROUND: Contemporary Australian epidemiological data on acute pulmonary embolism (PE) are lacking. OBJECTIVES: To determine the admission rates of acute PE in Australia, and to assess the temporal trends in short- and medium-term mortality following acute PE. METHODS: Retrospective population-linkage study of all New South Wales residents admitted with a primary diagnosis of PE between January 1, 2002 and December 31, 2018 using data from the Centre for Health Record Linkage databases. Main outcome measures included temporal trends in total PE admissions and all-cause mortality at prespecified time points up to 1 year, stratified by gender. RESULTS: There were 61,607 total PE admissions between 2002 and 2018 (mean ± standard deviation: 3,624 ± 429 admissions per annum; 50.42 ± 3.70 admissions per 100,000 persons per annum). The mean admission rate per annum was higher for females than for males (54.85 ± 3.65 vs. 44.91 ± 4.34 admissions per 100,000 persons per annum, respectively) and remained relatively stable for both genders throughout the study period. The main study cohort, limited to index PE admission only, comprised 46,382 persons (mean age: 64.6 ± 17.3 years; 44.4% males). The cumulative in-hospital, 30-day, 3-month, and 1-year mortality rates were 3.7, 5.6, 9.6, and 16.8%, respectively. When compared with 2002 as the reference year, there was a significant reduction in in-hospital (odds ratio [OR] = 0.34; 95% confidence interval [CI] = 0.25-0.46), 30-day (OR = 0.58, 95% CI = 0.46-0.73), and 1-year (hazard ratio = 0.74, 95% CI = 0.66-0.84) (all p < 0.001) mortality risk by 2017 after adjusting for age, gender, and relevant confounders. The survival improvements were seen in both genders and were greater for females than for males. CONCLUSION: Mortality following PE has improved with reductions observed in both short- and medium-term follow-ups between 2002 and 2018 with greater reductions in females despite their higher admission rates over time.

Incidence and Impact of Hypoalbuminaemia on Outcomes Following Acute Pulmonary Embolism.

BACKGROUND: Clinical features and outcomes of patients with hypoalbuminaemia in acute pulmonary embolism (PE) have never been studied. The present study investigated the incidence and determined the prognostic significance of hypoalbuminaemia in patients with confirmed acute PE. METHODS: From a dedicated tertiary-referral centre database involving 1,426 consecutive patients admitted with confirmed PE (2000-2012), 1,032 patients had serum albumin assessed on admission (day-1). Patients were stratified into hypoalbuminaemia (<35 g/L) or normal serum albumin (≥35 g/L). Multivariable logistic and Cox proportional-hazards regression methods were used to assess 30-day and 90-day all-cause mortality. RESULTS: Hypoalbuminaemia was present in 160 (15.5%) patients at day-1 and was associated with higher mean (±SD) heart rate (94.4 ± 21.8 vs 87.8 ± 21.5 bpm), lower systolic blood pressure (131.0 ± 24.7 vs 142.3 ± 24.7 mmHg), lower arterial oxyhaemoglobin saturation (93.3 ± 6.1% vs 95.6 ± 4.0%), lower day-1 serum sodium (137.0 ± 4.7 vs 138.8 ± 3.8 mmol/L) and haemoglobin levels (114.4 ± 20.7 vs 131.5 ± 18.7 g/L). Patients with hypoalbuminaemia had higher incidence of malignancy (44.4% vs 18.8%) and chronic renal disease (9.4% vs 5.2%), and at admission were less likely to be taking aspirin/clopidogrel (19.3% vs 27.7%) and more likely to be using enoxaparin (6.7% vs 3.0%). During a mean follow-up of 5.0 ± 4.0 years, patients with hypoalbuminaemia had higher 30-day (16.3% vs 3.6%) and 90-day (26.3% vs 6.2%) mortality. Multivariable analyses showed hypoalbuminaemia independently predicted both 30-day (odds ratio 2.57, 95% confidence interval [CI] 1.03-6.41) and 90-day (hazard ratio 2.42 95% CI 1.38-4.22) mortality. CONCLUSION: Hypoalbuminaemia is an independent predictor of mortality following PE and may improve risk stratification of patients in risk prediction models.