Table 3

Outcomes of non-selective versus cardioselective β-blocker therapy

Non-selective β blocker versus cardioselective β blocker*
 Unadjusted—full cohort 0.97 (0.70 to 1.35)
 Covariate adjusted—full cohort 0.89 (0.63 to 1.25)
 Covariate and propensity score adjusted—full cohort 0.86 (0.60 to 1.24)
 Matched cohort adjusted for unbalanced covariates§ 0.75 (0.47 to 1.19)
Late ventilation
 Unadjusted—full cohort 1.28 (0.93 to 1.76)
 Covariate adjusted—full cohort 1.29 (0.93 to 1.79)
 Covariate and propensity score adjusted—full cohort 1.24 (0.88 to 1.74)
 Matched cohort adjusted for unbalanced covariates§ 1.06 (0.68 to 1.65)
30-Day all-cause readmission**
 Unadjusted—full cohort 1.35 (1.21 to 1.50)
 Covariate adjusted—full cohort 1.21 (1.08 to 1.36)
 Covariate and propensity score adjusted—full cohort 1.20 (1.06 to 1.35)
 Matched cohort adjusted for unbalanced covariates§ 1.25 (1.08 to 1.44)
  • * Among β-blocker treated cases (n=10 070).

  • Accounting for within-hospital clustering.

  • Covariates included in all models: underlying cardiovascular condition (hypertension, ischaemic heart disease, congestive heart failure), age group, principal diagnosis, admissions in previous year for chronic obstructive pulmonary disease, insurance payor, atrial fibrillation/flutter, pulmonary circulation disease, other neurological disorders, renal failure, metastatic cancer, solid tumour without metastasis, diabetes, weight loss, deficiency anaemias, drug abuse, psychoses, depression, sleep apnoea, pneumonia, other infections, chronic pulmonary heart disease, day 1 or 2 methylxanthine bronchodilators, arterial blood gas, non-invasive ventilation, morphine, loop diuretics, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, antiplatelets, calcium channel blockers, lipid lowering non-statins, and hospital region.

  • § Each non-selective β-blocker-treated case matched on propensity with one cardioselective β-blocker-treated case. 2321 of 2459 (94.4%) non-selective treated cases matched.

  • Unbalanced covariates adjusted for attending physician specialty, valvular disease, sleep apnoea, day 1 or 2 tetracyclines, day 1 or 2 non-invasive ventilation.

  • ** Among 9873 survivors in full cohort and 4545 survivors in matched cohort.