Table 3.
Thirty-day mortality and corresponding HRs in patients with different preoperative serum albumin concentration undergoing surgery because of colorectal cancer, overall and stratified by type of admission, cancer type, and period of surgery
| HR (95% CI) |
|||||
|---|---|---|---|---|---|
| Number of patients N |
Number of deaths N |
30-Day mortality %* (95% CI) |
Crude | Adjusted† | |
| Colorectal cancer surgery, albumin g/L | |||||
| ≤25 | 401 | 108 | 26.9 (22.9 to 31.6) | 15.89 (11.69 to 21.59) | 7.59 (4.95 to 11.64) |
| 26–30 | 784 | 154 | 19.6 (17.0 to 22.6) | 10.91 (8.18 to 14.56) | 5.19 (3.53 to 7.63) |
| 31–35 | 1742 | 163 | 9.4 (8.1 to 10.8) | 4.92 (3.70 to 6.56) | 2.58 (1.80 to 3.69) |
| 36–40 | 3065 | 149 | 4.9 (4.2 to 5.7) | 2.50 (1.87 to 3.34) | 1.75 (1.25 to 2.45) |
| >40 | 3347 | 66 | 2.0 (1.6 to 2.5) | 1.00 | 1.00 |
| Elective admission‡, albumin g/L | |||||
| ≤25 | 122 | 30 | 24.6 (17.9 to 33.2) | 15.32 (9.84 to 23.83) | 8.08 (4.45 to 14.67) |
| 26–30 | 348 | 62 | 17.8 (14.2 to 22.3) | 10.57 (7.38 to 15.15) | 5.31 (3.27 to 8.62) |
| 31–35 | 1128 | 90 | 8.0 (6.5 to 9.7) | 4.50 (3.23 to 6.26) | 2.56 (1.68 to 3.90) |
| 36–40 | 2544 | 99 | 3.9 (3.2 to 4.7) | 2.15 (1.55 to 2.98) | 1.52 (1.04 to 2.21) |
| >40 | 3118 | 57 | 1.8 (1.4 to 2.4) | 1.00 | 1.00 |
| Non-elective admission‡, albumin g/L | |||||
| ≤25 | 279 | 78 | 28.0 (23.1 to 33.6) | 8.19 (4.11 to 16.33) | 7.78 (3.17 to 19.12) |
| 26–30 | 435 | 92 | 21.2 (17.6 to 25.3) | 5.83 (2.94 to 11.57) | 5.36 (2.23 to 12.88) |
| 31–35 | 612 | 73 | 11.9 (9.6 to 14.8) | 3.14 (1.57 to 6.28) | 2.68 (1.12 to 6.44) |
| 36–40 | 516 | 50 | 9.7 (7.4 to 12.6) | 2.51 (1.23 to 5.10) | 2.32 (0.97 to 5.55) |
| >40 | 227 | 9 | 4.0 (2.1 to 7.5) | 1.00 | 1.00 |
| Age 0–59 years, albumin g/L | |||||
| ≤25 | 48 | 6 | 12.5 (5.8 to 25.7) | 23.03 (7.03 to 75.46) | 7.39 (1.32 to 41.22) |
| 26–30 | 64 | 9 | 14.1 (7.6 to 25.3) | 25.55 (8.56 to 76.24) | 11.84 (2.65 to 52.97) |
| 31–35 | 234 | 9 | 3.9 (2.0 to 7.3) | 6.68 (2.24 to 19.94) | 2.18 (0.51 to 9.35) |
| 36–40 | 549 | 10 | 1.8 (1.0 to 3.4) | 3.15 (1.08 to 9.20) | 2.29 (0.69 to 7.61) |
| >40 | 857 | 5 | 0.6 (0.2 to 1.4) | 1.00 | 1.00 |
| Age 60–69 years, albumin g/L | |||||
| ≤25 | 60 | 9 | 15.0 (8.1 to 26.8) | 15.11 (6.37 to 35.87) | 6.63 (2.13 to 20.61) |
| 26–30 | 170 | 26 | 15.3 (10.7 to 21.6) | 15.39 (7.77 to 30.50) | 4.75 (1.77 to 12.76) |
| 31–35 | 380 | 21 | 5.5 (3.6 to 8.4) | 5.26 (2.59 to 10.69) | 2.55 (1.03 to 6.31) |
| 36–40 | 824 | 17 | 2.1 (1.3 to 3.3) | 1.94 (0.92 to 4.05) | 1.34 (0.57 to 3.15) |
| >40 | 1122 | 12 | 1.1 (0.6 to 1.9) | 1.00 | 1.00 |
| Age 70–79 years, albumin g/L | |||||
| ≤25 | 154 | 43 | 27.9 (21.5 to 35.7) | 11.89 (7.35 to 19.24) | 8.76 (4.40 to 17.44) |
| 26–30 | 280 | 54 | 19.3 (15.1 to 24.4) | 7.71 (4.86 to 12.24) | 5.23 (2.79 to 9.77) |
| 31–35 | 605 | 46 | 7.6 (5.8 to 10.0) | 2.87 (1.78 to 4.61) | 2.12 (1.19 to 3.80) |
| 36–40 | 1064 | 49 | 4.6 (3.5 to 6.1) | 1.70 (1.06 to 2.72) | 1.41 (0.83 to 2.40) |
| >40 | 990 | 27 | 2.7 (1.9 to 4.0) | 1.00 | 1.00 |
| Age 80+ years, albumin g/L | |||||
| ≤25 | 139 | 50 | 36.0 (28.6 to 44.5) | 7.54 (4.56 to 12.45) | 7.09 (3.50 to 14.36) |
| 26–30 | 270 | 65 | 24.1 (19.4 to 29.6) | 4.57 (2.82 to 7.40) | 4.57 (2.40 to 8.68) |
| 31–35 | 523 | 87 | 16.6 (13.7 to 20.1) | 3.03 (1.90 to 4.83) | 2.66 (1.46 to 4.85) |
| 36–40 | 628 | 73 | 11.6 (9.4 to 14.4) | 2.05 (1.28 to 3.31) | 2.17 (1.22 to 3.86) |
| >40 | 378 | 22 | 5.8 (3.9 to 8.7) | 1.00 | 1.00 |
Age and admission type were largely responsible for the change in estimates by adjustment.
*Calculated using the Kaplan-Meier method.
†Mutually adjusted for gender, age (both as a continuous and a categorical variable), type of admission, operation year (calendar year), county, cancer site, cancer stage (excluded patients with stage ‘unknown’), comorbidity level, alcohol-related disease, liver disease, marital status, haemoglobin, sodium, potassium and creatinine (number of observations with complete data=8033).
‡Information on type of admission is missing for some patients. Therefore, the sum of patients with non-elective and elective admissions is not equal to the number of all patients included in the study.