Skip to main content
. 2023 Jan 2;13:52. doi: 10.1038/s41598-022-26476-5

Table 3.

The event numbers and the estimated subdistribution hazard ratios for different types of GI bleeding using a multivariable regression model.

Before matching After matching
HD (n = 81,116) PD (n = 6298) aSHR (95% CI) HD (n = 6296) PD (n = 6296) aSHR (95% CI)
Overall bleeding, no. (%) 20,970 (25.85) 815 (12.94) 1.23** (1.15–1.33) 950 (15.09) 815 (12.94) 1.13* (1.03–1.24)
    Upper GI bleeding 14,150 (17.44) 567 (9.00) 1.23** (1.12–1.34) 653 (10.37) 567 (9.01) 1.11 (0.99–1.25)
    Lower GI bleeding 1231 (1.52) 40 (0.64) 1.49* (1.08–2.06) 51 (0.81) 40 (0.64) 1.26 (0.83–1.91)
    Unspecified bleeding 5589 (6.89) 208 (3.30) 1.24* (1.09–1.42) 246 (3.90) 208 (3.30) 1.18 (0.99–1.40)

aSHR Adjusted subdistribution hazard ratio; CI Confidence interval.

aHRs were adjusted for age, sex, selected comorbidities (diabetes mellitus, hypertension, coronary artery disease, peripheral vascular disease, heart failure, stroke, chronic obstructive pulmonary disease, hyperlipidemia, rheumatological disease) and medications (Aspirin, Antiplatelet agent, NSAID, Cox-2 selective inhibitors, Corticosteroids, Selective Serotonin Reuptake Inhibitors, Anticoagulants, Gastroprotective agents, Aldosterone antagonists, Calcium channel blocker, Nitrates).

*p value < 0.05.

**p value < 0.001.