"Steroid Hyperglycemia" Is Not Benign
Year: 2009
Abstract Number: 1047-P
Authors: PANKAJ SHAH, NAIFA L. BUSAIDY, SHITIJ KAPOOR, STEVEN I. SHERMAN, VICTOR R. LAVIS
Institutions: Rochester, MN, Houston, TX
Results: To explore the clinical significance of hyperglycemia in glucocorticoid treated patients (GCC: ≥30 mg/day hydrocortisone or equivalent) we analyzed 3940 inpatients over 3 mo in a cancer center.
1628 (41.3%) of 3940 got GCC during >=1 hospitalization. GCC group was younger (53.4±17.1 vs. 57.1±16.1 yr. p<0.001), more males (54.6% vs. 47.9%, p<0.01), but similar ethnicity to the no GCC group.
Out of those tested for glucose 401(28%) of 1447 in GCC group and 237 (18%) of 1338 in no GCC group had sustained significant hyperglycemia (SSH, bld glucose [BG] ≥200 mg/dl on ≥2 days). Fewer patients with SSH received antidiabetic therapy in the GCC group than in the no GCC group (74.4% vs. 87.3%, p<0.001), but more (p <0.001) received insulin vs. oral agents. Average BG were comparable in SSH on GCC or no GCC (171.9±41.8 vs. 170.7±35.7 mg/dl), more with GCC than no GCC had ≥ 1 BG ≥300mg/dl (42.3% vs. 30.8%, OR 1.65 (95% CI 1.21-2.24)).
Within the GCC group, hyperglycemia was associated with longer hospital stay, more frequent doubling of creatinine, infections requiring IV antibiotics, and more in-hospital and long term mortality (table & figure) as compared to no SSH. Adjusted for treatment of hyperglycemia, the OR for in-hospital mortality with SSH was 2.8 (1.7-4.4) vs 1.9 (0.6-6.2) for no GCC. | Hospital stay (d) | Doubling of creatinine | Infection | Hospital Death |
| No GCC No SSH | 6.6±6.0 | 21/1258 (1.7%) | 213/1325 (16.1%) | 27/1101 (2.5%) |
| No GCC SSH | 11.3±14.5 | 12/ 261 (4.6%) | 69/ 276 (25.0%) | 6/ 237 (2.5%) |
| p<0.00001 | 2.8 (1.4-5.8)^ | 1.7 (1.3-2.4)^ | 1.0 (0.4-2.5)^ |
| GCC No SSH | 8.2±8.4 | 59/1773 (3.3%) | 357/1851 (19.3%) | 75/1046 (7.2%) |
| GCC SSH | 12.4±14.7 | 52/ 695 (7.5%) | 195/ 704 (27.7%) | 81/ 401 (20.2%) |
| p<0.00001 | 2.3 (1.6-3.4)^ | 1.6 (1.3-2.0)^ | 3.3 (2.3-4.6)^ |
The association of poor outcomes with hyperglycemia is at least as strong in GCC-treated patients as in those not treated with GCC.[figure1]