Introduction
Hyperglycaemia is frequent among hospital patients and is related to increased complications, poorer outcomes and increased mortality.1,2 We examined the relationship between hyperglycaemia on admission with morbidity and mortality. By reviewing hospital admissions during different seasons, we studied seasonal variation in the relationship of hyperglycaemia to mortality in view of the known seasonal variation in blood glucose levels.3,4
Material and methods
We retrospectively examined the records of 1,132 hospital admissions. Hyperglycaemia was defined as an admission random glucose level of above 11.0 mmol/L. For statistical analyses, we used the Mann–Whitney U test, complemented by Spearman's rank correlation and chi-squared tests with a significance level of p=0.05.
Results and discussion
Hyperglycaemia was present in 14.1% of patients admitted to the hospital, of whom, 3.9% had no previously documented history of diabetes. Patients with new-onset hyperglycaemia on admission, had a significantly higher mortality rate than previously diagnosed diabetes (43.3% vs 17.9%; p=0.006). Logistic regression showed that plasma glucose on admission was independently associated with increased 1-year mortality (odds ratio 1.035; p=0.034). Table 1 shows the continuous variables significantly related with mortality at 1 year, while Table 2 shows the relationship of diabetes history and of plasma glucose on admission related with mortality and re-hospitalisation at 1 year.
Table 1.
Continuous variables significantly related with mortality at 1 year (analysed using Mann–Whitney U test)
| Mortality, yes, median (IQR) | Mortality, no, median (IQR) | p-value | |
|---|---|---|---|
| Age on admission, years | 80.0 (71.0–86.0) | 69.0 (49.0–80.0) | <0.001 |
| Length of stay in hospital, days | 6.0 (3.0–13.0) | 3.0 (2.0–6.0) | <0.001 |
| Plasma glucose on admission, mmol/L | 7.0 (5.8–9.8) | 6.4 (5.5–8.43) | 0.005 |
| Serum creatinine, μmol/L | 94.0 (70–142.8) | 81.0 (65–105.5) | <0.001 |
| Serum urea, mmol/L | 9.2 (6.7–14.9) | 6.5 (4.8–9.3) | <0.001 |
| Estimated GFR, mL/min/1.73m2 | 60.0 (36.0–88.5) | 80.0 (56.5–103.0) | <0.001 |
| Serum sodium, mmol/L | 138.0 (135.0–140.0) | 139.0 (136.0–141.0) | 0.005 |
| Serum chloride, mmol/L | 98.0 (94.2–101.6) | 100.2 (97.1–102.6) | <0.001 |
| Serum osmolality, mOsm/kg | 302.2 (294.6–314.0) | 300.8 (295.0–307.0) | 0.020 |
GFR = glomerular filtration rate; IQR = interquartile range.
Table 2.
Diabetes history and blood glucose on admission related with mortality and re-hospitalisation at 1 year
| Plasma glucose on admission | Mortality, n (%) | p-value | Re-hospitalised, n (%) | p-value |
|---|---|---|---|---|
| Patient with known diabetes | ||||
| <11 mmol/L | 36 (22.8) | 0.006 | 90 (57.0) | 0.003 |
| >11 mmol/L | 14 (17.9) | 0.006 | 44 (56.4) | 0.003 |
| Patient not known to have diabetes | ||||
| <11 mmol/L | 90 (18.1) | 0.006 | 217 (43.6) | 0.003 |
| >11 mmol/L | 13 (43.3) | 0.006 | 10 (33.3) | 0.003 |
Hyperglycaemia at admission was also an independent predictor of increased length of stay (p≤0.001). A longer inpatient course was associated with an increased 90-day and 1-year mortality (p<0.001 for both). After adjusting for confounding variables, admission plasma glucose and length of stay remained significant predictors of 1-year mortality outcome (p=0.034 and p=0.003, respectively). In April 2019, a higher 90-day and 1-year mortality rate were noted (p=0.001 and p=0.015, respectively), together with a higher proportion of patients admitted with new-onset hyperglycaemia (p=0.012). In January 2019, more previously known patients with diabetes were admitted with hyperglycaemia (p=0.012). Younger patients were more likely to be admitted in the summer. When comparing mortality at 90 days and 1 year, April 2019 had a higher mortality rate than expected (p=0.001 and p=0.015, respectively). Interestingly, this was not observed in April 2020.
Conclusion
Our results indicate that plasma glucose is an important prognostic marker and may indicate a more severe illness. We recommend that these patients are highlighted with a greater level of care. A glycosylated haemoglobin level taken at admission in cases of new-onset hyperglycaemia can aid differentiation between stress hyperglycaemia and undiagnosed diabetes.
References
- 1.Umpierrez GE, Hellman R, Korytkowski MT, et al. Management of hyperglycemia in hospitalized patients in non-critical care setting: an endocrine society clinical practice guideline. J Clin Endocrinol Metab 2012;97:16–38. [DOI] [PubMed] [Google Scholar]
- 2.Evans NR, Dhatariya KK. Assessing the relationship between admission glucose levels, subsequent length of hospital stay, readmission and mortality. Clin Med 2012;12:137–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kershenbaum A, Tarabeia J, Stein N, Lavi I, Rennert G. Unraveling seasonality in population averages: an examination of seasonal variation in glucose levels in diabetes patients using a large population-based data set. Chronobiol Int 2011;28:352–60. [DOI] [PubMed] [Google Scholar]
- 4.Gikas A, Sotiropoulos A, Pastromas V, et al. Seasonal variation in fasting glucose and HbA1c in patients with type 2 diabetes. Prim Care Diabetes 2009;3:111–4. [DOI] [PubMed] [Google Scholar]