Table 6.

The impact of severity of disease measurements and hematocrit levels on mortality riska

Characteristic All-Cause Death Cardiac Death Infectious Death
RR 95% CI RR 95% CI RR 95% CI
1993 Cohort
no. of access procedures
no procedures (ref)1.001.001.00
1 to 3 procedures1.01(0.97 to 1.05)1.00(0.94 to 1.06)1.10(0.98 to 1.23)
≥4 procedures0.93(0.89 to 0.98)0.94(0.87 to 1.01)1.03(0.90 to 1.18)
no. of transfusions
no transfusion (ref)1.001.001.00
1 to 2 pints1.03(0.97 to 1.09)1.05(0.96 to 1.15)1.07(0.91 to 1.25)
≥3 pints1.20(1.13 to 1.27)1.15(1.05 to 1.25)1.19(1.03 to 1.39)
length of stay
no hospital stay (ref)1.001.001.00
1 to 3 days1.23(1.17 to 1.31)1.22(1.12 to 1.32)1.26(1.06 to 1.49)
4 to 10 days1.47(1.41 to 1.55)1.42(1.32 to 1.52)1.62(1.41 to 1.85)
11 to 20 days1.92(1.82 to 2.03)1.85(1.70 to 2.01)2.29(1.97 to 2.66)
>20 days2.57(2.44 to 2.72)2.42(2.23 to 2.63)3.42(2.96 to 3.96)
hematocrit
<27%1.33(1.26 to 1.40)1.25(1.15 to 1.35)1.53(1.33 to 2.75)
27% to <30%1.12(1.08 to 1.17)1.11(1.05 to 1.17)1.13(1.02 to 1.26)
30% to <33% (ref)1.001.001.00
33% to <36%0.96(0.91 to 1.01)0.97(0.90 to 1.05)1.02(0.88 to 1.19)
1992 & 1993
33% to <36% 0.96 (0.92 to 0.99)
  • a The result from the combined cohorts is compared to that from the 1993 cohort only in the 33% to <36% group. The relative risk is the same 0.96, but the confidence interval is significant. This combined analysis was based on 16,000 patient observations compared to 10,129 in 1993 only. Therefore, the hematocrit group 33% to <36% outcome is sensitive to the number of patients and patient observations in the analysis.