Abstract
Background
We evaluated the association of admission blood glucose (ABG) and mortality in patients with and without diabetes mellitus (DM) hospitalized for atrial fibrillation (AF).
Hypothesis
Hyperglycemia on admission is a bad prognostic marker in patients with AF.
Methods
Observational data were collected from electronic records of patients age ≥ 18 years hospitalized for AF in 2011–2013. Twelve‐month data were available in all cases. ABG levels were classified as follows: 70 to 110 mg/dL, normal; 111 to 140 mg/dL, mildly elevated; 141 to 199 mg/dL, moderately elevated; ≥200 mg/dL, markedly elevated. Cox proportional hazards model was used to assess overall survival by ABG categories, adjusted for study variables. Primary outcome measure was mortality at end of follow‐up.
Results
The cohort included 1127 patients (45% male; median age, 75 ± 13 years), of whom 331 had DM. Mortality rates by ABG levels were 19% (77/407 patients), normal ABG; 26% (92/353 patients), mildly elevated ABG; 28% (69/244 patients), moderately elevated ABG; and 41% (50/123 patients), markedly elevated ABG. Data were analyzed for the entire cohort following adjustment for age, sex, CHADS2 score, ischemic heart disease, smoking, and alcohol consumption. Compared with normal ABG, the adjusted hazard ratio for mortality was higher in patients with moderately elevated ABG (2.1, 95% confidence interval: 1.19‐7.94, P < 0.05) and markedly elevated ABG (1.6, 95% confidence interval: 1.02‐5.31, P < 0.05).
Conclusions
In patients with and without DM hospitalized for AF, moderately to markedly elevated ABG levels are associated with increased mortality.
Keywords: Admission, Atrial Fibrillation, Diabetes Mellitus, Glucose, Mortality
1. INTRODUCTION
Atrial fibrillation (AF), the most common sustained cardiac arrhythmia, is associated with significant morbidity and mortality.1, 2 AF increases the risk of stroke, and several scoring systems have been developed to stratify the risk. The CHADS2, currently one of the most popular stratification systems, assigns 1 point for the presence of each of the following: congestive heart failure (HF), hypertension (HTN), age ≥ 75 years, and diabetes mellitus (DM), and 2 points for previous stroke or transient ischemic attack. A CHADS2 score of zero signifies low stroke risk; 1 point, moderate risk; and ≥2 points, high risk.3
Elevated admission blood glucose (ABG) levels are often found in patients with an acute illness, with or without DM. They are associated with a poor outcome following hospitalization for trauma,4 surgery,5 or an infectious disease,6, 7, 8, 9 or for cardiovascular conditions of ischemic or hemorrhagic stroke,10 HF,11, 12 pulmonary embolism,13 and acute myocardial infarction (MI).14, 15, 16 Dziewierz et al17 recently reported that patients with acute MI and an elevated ABG level were at increased risk of life‐threatening complications, especially arrhythmias.
As ABG levels are readily available on admission for most hospitalized patients, they might serve as an outcome predictor in patients with AF. The aim of the present study was to evaluate the association of ABG levels with post‐hospitalization all‐cause mortality in patients with and without DM hospitalized for AF.
2. METHODS
2.1. Patients and setting
The present retrospective observational study was conducted at Rabin Medical Center in Israel, a 1300‐bed university‐affiliated tertiary hospital. Most of the patients in its 10 medical wards are admitted through the emergency department. All patient data are recorded in electronic medical charts based on the same database platform used in the community primary‐care facilities. Mortality data are recorded in the hospital's mortality database, updated from the Population Registry of the Israel Ministry of the Interior.
For the present study, retrospective observational data were extracted from the electronic medical records of all patients who were admitted to the medical wards in Rabin Medical Center between January 1, 2011, and December 31, 2013. Inclusion criteria were age ≥ 18 years with a principal discharge diagnosis of AF, and documented ABG levels. We included patients with valvular and nonvalvular AF. We excluded 18 patients because of missing ABG levels (n = 13) or ABG level < 70 mg/dL (n = 5). In patients with multiple admissions, only the first hospital stay was analyzed. Mortality data were collected until June 1, 2015. The study was approved by the institutional review board of Rabin Medical Center.
2.2. Data collection and definitions
The following data were collected from the patients' electronic files: age, laboratory values at admission, presence of hyperlipidemia, HTN, ischemic heart disease, chronic HF, chronic renal failure, cerebrovascular disease, chronic obstructive pulmonary disease, asthma, interstitial lung disease, and inflammatory bowel disease, and CHADS2 score indicating congestive HF, HTN, age ≥ 75 years, DM, and stroke/transient ischemic attack.
DM was defined as a diagnosis of DM in the medical records or use of any oral hypoglycemic agent, glucagon‐like peptide agonist, or insulin at the time of admission. Patients were stratified by the presence/absence of preexisting DM.
ABG level was defined as the blood glucose measurement made closest to the patient's arrival time at the hospital, within the first 24 h of the admission date. For analysis, we used point‐of‐care measurements performed with beside glucometers or serum glucose levels derived from venous samples. ABG levels were divided into 4 categories: 70 to 110 mg/dL, normal; 111 to 140 mg/dL, mildly elevated; 141 to 199 mg/dL, moderately elevated; and ≥200 mg/dL, markedly elevated. The primary outcome measure of the study was all‐cause mortality risk at the end of follow‐up by ABG categories.
2.3. Statistical analysis
The statistical analysis for this article was generated using SAS software version 9.4 (SAS Institute Inc., Cary, NC). Continuous variables are presented as mean ± SD and categorical variables as n (%). The Kaplan–Meier model was used to assess overall survival by ABG categories, and the Cox proportional hazards model was used to assess overall survival by ABG categories after adjustment for age, sex, smoking, alcohol intake, ischemic heart disease, and CHADS2 score. Two‐sided P values <0.05 were considered statistically significant. Because the interaction between ABG, DM, and mortality was not significant (P = 0.19), we ran the Cox model for the entire cohort of patients with and without DM. This analysis proved a significant association between ABG levels and mortality risk at the end of follow up (P < 0.01). We had complete data for all the study variables, other than body mass index and smoking. No imputation for missing data was done because missing at random (MAR) cannot be assumed. Owing to the small number of patients with low ABG levels, we did not analyze the data for this subgroup.
3. RESULTS
3.1. Baseline characteristics
There were a total of 73 796 admissions to the medical wards of Rabin Medical Center during the study period. We identified 1145 patients with a principal discharge diagnosis of AF; following exclusions, the final cohort consisted of 1127 patients (Figure 1). The cohort included 510 male (45%) and 617 female patients of median age 75 ± 13 years (range, 19–99 years); 331 patients (29%) had preexisting DM and 796 did not.
Figure 1.
Flow diagram of patient screening for the study. The electronic files of a tertiary medical center were reviewed for all patients age ≥ 18 years hospitalized for AF from 2011 to 2013. Abbreviations: ABG, admission blood glucose; AF, atrial fibrillation
3.2. Patient characteristics by ABG category
The median ABG level was 120 mg/dL (range, 70–970 mg/dL). Most of the patients had ABG levels of <140 mg/dL: 36% had 70 to 110 mg/dL (normal) and 31% had 111 to 140 mg/dL (mildly elevated). ABG levels of ≥200 mg/dL were documented in 123 patients (11%), most of whom had preexisting DM (Table 1). Patients with an ABG level of ≥200 mg/dL had significantly higher rates of HTN, ischemic heart disease, congestive HF, and cerebrovascular disease than did patients with an ABG level of 70 to 110 mg/dL (Table 2).
Table 1.
Baseline characteristics of the patients by ABG categories
| Baseline Characteristics | ABG Categories | |||
|---|---|---|---|---|
| 70–110 mg/dL, n = 407 | 111–140 mg/dL, n = 353 | 141–199 mg/dL, n = 244 | ≥200 mg/dL, n = 123 | |
| Age, y, mean ± SD (median) | 71 ± 15 (74) | 74 ± 13 (76) | 75 ± 10 (75) | 77 ± 11 (79)a |
| Male sex | 193 (47) | 153 (43) | 100 (41) | 54 (44) |
| Smoking | 34 (9.0) | 23 (7.0) | 22 (10) | 10 (9.0) |
| Alcohol intake | 5 (1.0) | 2 (1.0) | 2 (1.0) | 2 (2.0) |
| BMI, kg/m2, mean ± SD (median) | 27 ± 5 (27) | 29 ± 7 (28) | 28 ± 6 (28) | 30 ± 11 (28)a |
| Comorbidities | ||||
| HTN | 208 (51) | 199 (56) | 169 (69)a | 89 (72)a |
| Hyperlipidemia | 107 (26) | 111 (31) | 75 (31) | 42 (34) |
| Ischemic heart disease | 72 (18) | 90 (26)a | 54 (22) | 46 (37)a |
| CHF | 51 (13) | 65 (18)a | 34 (14) | 31 (25)a |
| Cerebrovascular disease | 24 (6.0) | 31 (9.0) | 30 (12)a | 16 (13)a |
| DM | 45 (11) | 73 (21)a | 119 (49)a | 94 (76)a |
| CHADS2 score, mean ± SD (median) | 1.2 ± 1 (1) | 1.5 ± 1 (2)a | 1.8 ± 1 (2)a | 2.4 ± 1 (3)a |
| Chronic renal failure | 33 (8.0) | 24 (7.0) | 22 (9.0) | 17 (14) |
| Hypothyroidism | 45 (11) | 40 (11) | 35 (14) | 19 (15) |
| Hyperthyroidism | 10 (2.0) | 5 (1.0) | 3 (1.0) | 2 (2.0) |
| IBD | 0 (0.0) | 1 (0.3) | 3 (1.0) | 0 (0.0) |
| COPD | 9 (2.0) | 11 (3.0) | 17 (7.0)a | 4 (3.0) |
| Asthma | 14 (3.0) | 17 (5.0) | 8 (3.0) | 8 (6.0) |
| Interstitial lung disease | 1 (0.3) | 1 (0.3) | 3 (1.0) | 1 (1.0) |
Abbreviations: ABG, admission blood glucose; BMI, body mass index; CHADS2, CHF, HTN, age ≥ 75 years, DM (1 point each), and stroke (2 points); CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; DM, diabetes mellitus; HTN, hypertension; IBD, inflammatory bowel disease; SD, standard deviation.
Data are presented as n (%) unless otherwise noted.
P < 0.05 compared with normal ABG levels (70–110 mg/dL).
Table 2.
Overall mortality at various time points during follow‐up by ABG level in patients with AF
| ABG | No. of Patients (%) | Mortality | |||||
|---|---|---|---|---|---|---|---|
| In‐hospital | 30 Days | 1 Year | 2 Years | 3 Years | End Follow‐up | ||
| 70–110 mg/dL | 407 (28) | 2 (0.5) | 9 (2.0) | 41 (10) | 56 (14) | 71 (17) | 77 (19) |
| 111–140 mg/dL | 353 (31) | 9 (3.0) | 19 (5.0) | 53 (15) | 70 (20) | 83 (24) | 92 (26) |
| 141–199 mg/dL | 244 (22) | 9 (4.0) | 18 (7.0) | 44 (18) | 56 (23) | 63 (26) | 69 (28) |
| ≥200 mg/dL | 123 (11) | 5 (4.0) | 9 (7.0) | 24 (20) | 41 (33) | 48 (39) | 50 (41) |
| Total | 1127 | 25 (2.0) | 55 (5.0) | 162 (14) | 223 (20) | 265 (24) | 288 (26) |
Abbreviations: ABG, admission blood glucose; AF, atrial fibrillation.
Data are presented as n (%).
3.3. Mortality by ABG category
Twelve‐month follow‐up data were available for all patients. The in‐hospital mortality risk was 0.5% in patients with a normal ABG level, 3% in patients with a mildly elevated ABG level, and 4% each in patients with a moderately or markedly elevated ABG level. Table 2 shows the mortality risks at 30 days and 12, 24, and 36 months after hospitalization by ABG level. At the end of follow‐up, the mortality risks for the different ABG categories were as follows: normal, 19% (77/407 patients); mildly elevated, 26% (92/353 patients); moderately elevated, 28% (69/244 patients); and markedly elevated, 41% (50/123 patients). The Kaplan–Meier analysis of patient survival following admission to time of death is shown in Figure 2.
Figure 2.
Kaplan–Meier analysis of patients following admission. Survival was defined as the time from date of admission to date of death. Observations were censored at the end of follow‐up (P < 0.05)
On comparison of the unadjusted hazard ratios (HRs; Cox proportional hazards model) between subgroups for the entire follow‐up period, a statistically significant difference in overall survival was noted between patients with normal ABG levels and patients with ABG levels that were mildly (HR: 1.4, 95% confidence interval [CI]: 1.10‐1.93, P < 0.05), moderately (HR: 1.6, 95% CI: 1.17‐2.24, P < 0.05), and markedly elevated (HR: 2.6, 95% CI: 1.79‐3.65, P < 0.05). Following adjustment for age, sex, CHADS2 score, ischemic heart disease, smoking, and alcohol intake, there was a statistically significant difference in overall survival between patients with normal ABG levels and patients with moderately elevated (aHR: 2.1, 95% CI: 1.19‐7.94, P < 0.05) and markedly elevated ABG levels (aHR: 1.6, 95% CI: 1.02‐5.31, P < 0.05). Mildly elevated ABG levels were not associated with a higher mortality rate than normal ABG levels. Patients with normal or mildly elevated ABG levels had the highest survival rate at the end of follow‐up, whereas patients with markedly elevated ABG levels had the lowest survival rate (P < 0.001).
3.4. Mortality by DM and ABG
As the lack of interaction between ABG, DM, and mortality may be the result of the sample size, we have completed a separate analysis of patients with DM and patients without DM. ABG was associated with mortality at the end of follow‐up in patients without DM (P = 0.006), but not in those with DM (P = 0.29).
In the group of patients without DM, there was a significant difference in mortality risk between patients with normal ABG and moderately elevated ABG (aHR: 1.72, 95% CI: 1.16‐2.55) and markedly elevated ABG (aHR: 2.34, 95% CI: 1.30‐4.20). The difference between normal ABG levels and mildly normal ABG levels did not reach statistical significance.
4. DISCUSSION
The present study was conducted in patients admitted to a general ward with a diagnosis of AF. The results showed that an elevated ABG level was associated with a higher mortality risk than a normal ABG level. The increased mortality risk was independent of age, sex, smoking, alcohol use, and CHADS2 score.
This study provides further evidence of the prognostic importance of testing blood glucose level at hospital admission, in support of previous reports of an association between ABG levels and various cardiovascular diseases. However, our extensive search of the literature yielded no previous studies associating ABG levels with short‐ and long‐term mortality risks in patients hospitalized for AF. Although some of the earlier studies reported low rates of ABG measurement,18 in our center, blood glucose measurement is a routine part of the admission procedure in the general medical wards, and ABG values were available for 99.6% of our patients.
In a study by Saliba et al.,19 all‐cause mortality risk at 1 and 2 years was lower, compared with our results (8.7% vs 14% and 17.4% vs 20%, respectively). Although that study was based on a computerized database of ambulatory patients with a diagnosis of AF, our study is based on hospitalized patients with AF. Hospitalization may indicate an increased mortality risk, both in‐hospital and shortly following discharge, and these patients may differ from ambulatory patients. Furthermore, it seems that the 30‐day mortality risk is mostly responsible for the difference in 1‐year mortality risk between the 2 studies; if we look at the mortality risk between 30 days after discharge and 1 year, the risk in both studies is similar (9% and 8.7%).
The lower mortality risk with normal or mildly elevated ABG levels suggests that ABG level may serve as a surrogate marker for general well‐being and an important factor in determining the short‐ and long‐term prognosis of patients with AF. The increase in glucose levels on admission might be caused by sympathetic activity, which may mediate atrial arrhythmias as well as myocardial injury. This is in line with the high ABG levels found in patients with MI, congestive HF, and stroke,10, 11, 14 and the pathogenesis underlying the increased mortality associated with these conditions is probably similar to that in our cohort. Although DM is certainly a significant risk factor for cardiovascular mortality,20 in our cohort, ABG had prognostic significance irrespective of a previous diagnosis of DM. Therefore, it seems that it was the presence of hyperglycemia, and not of DM, that was responsible for the increased mortality.
4.1. Study limitations
Our study has several limitations. First, we did not address the potential role of treatment for hyperglycemia in reducing hyperglycemia‐associated mortality in patients without DM. Treatment was left to the discretion of the attending physicians. Second, the retrospective study design precluded the establishment of a causal relationship, as elevated ABG levels might be a surrogate marker of diminished health status. Another limitation is lack of data regarding specific cause of mortality. The major strengths of our study are the large cohort and long‐term follow‐up, representing the real‐life scenario of patients admitted to medical wards. Blood glucose measurement is readily available in medical wards and may thus assist, along with other factors, in the risk stratification of patients admitted to these wards with a diagnosis of AF. Whether treatment of hyperglycemia in those with elevated ABG levels improves outcome requires further study.
5. CONCLUSION
Our study showed a significant association between ABG and mortality in patients hospitalized for AF. Further studies are needed to determine the optimal glycemic control in this patient group and to evaluate the impact of glycemic control on outcome.
Conflicts of interest
The authors declare no potential conflicts of interest.
Akirov A, Grossman A, Shochat T, Shimon I. Hyperglycemia on admission and hospitalization outcomes in patients with atrial fibrillation. Clin Cardiol. 2017;40:1123–1128. 10.1002/clc.22801
REFERENCES
- 1. Chao TF, Liu CJ, Tuan TC, et al. Rate‐control treatment and mortality in atrial fibrillation. Circulation. 2015;132:1604–1612. [DOI] [PubMed] [Google Scholar]
- 2. Díez‐Manglano J, Gomes‐Martín J, Al‐Cheikh‐Felices P, et al. Adherence to guidelines and mortality in atrial fibrillation. Int J Cardiol. 2014;176:430–436. [DOI] [PubMed] [Google Scholar]
- 3. Gage BF, Waterman AD, Shannon W, et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001;285:2864–2870. [DOI] [PubMed] [Google Scholar]
- 4. Yendamuri S, Fulda GJ, Tinkoff GH. Admission hyperglycemia as a prognostic indicator in trauma. J Trauma. 2003;55:33–38. [DOI] [PubMed] [Google Scholar]
- 5. Golden SH, Peart‐Vigilance C, Kao WH, et al. Perioperative glycemic control and the risk of infectious complications in a cohort of adults with diabetes. Diabetes Care. 1999;22:1408–1414. [DOI] [PubMed] [Google Scholar]
- 6. Gamble JM, Eurich DT, Marrie TJ, et al. Admission hypoglycemia and increased mortality in patients hospitalized with pneumonia. Am J Med. 2010;123:556.e11–556.e16. [DOI] [PubMed] [Google Scholar]
- 7. Falciglia M, Freyberg RW, Almenoff PL, et al. Hyperglycemia‐related mortality in critcally ill patients varies with admission diagnosis. Crit Care Med. 2009;37:3001–3009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Akirov A, Elis A. The prognostic significance of admission blood glucose levels in patients with urinary tract infection. Diabetes Res Clin Pract. 2016;115:99–105. [DOI] [PubMed] [Google Scholar]
- 9. Akirov A, Shimon I. The prognostic significance of admission blood glucose levels in elderly patients with pneumonia (GAP Study). J Diabetes Complications. 2016;30:845–851. [DOI] [PubMed] [Google Scholar]
- 10. Capes SE, Hunt D, Malmberg K, et al. Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients: a systematic overview. Stroke. 2001;32:2426–2432. [DOI] [PubMed] [Google Scholar]
- 11. Barsheshet A, Garty M, Grossman E, et al. Admission blood glucose level and mortality among hospitalized nondiabetic patients with heart failure. Arch Intern Med. 2006;166:1613–1619. [DOI] [PubMed] [Google Scholar]
- 12. Kosiborod M, Inzucchi SE, Spertus JA, et al. Elevated admission glucose and mortality in elderly patients hospitalized with heart failure. Circulation. 2009;119:1899–1907. [DOI] [PubMed] [Google Scholar]
- 13. Scherz N, Labarère J, Aujesky D, et al. Elevated admission glucose and mortality in patients with acute pulmonary embolism. Diabetes Care. 2012;35:25–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Capes SE, Hunt D, Malmberg K, et al. Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes: a systematic overview. Lancet. 2000;355:773–778. [DOI] [PubMed] [Google Scholar]
- 15. Timmer JR, van der Horst IC, Ottervanger JP, et al; Zwolle Myocardial Infarction Study Group. Prognostic value of admission glucose in non‐diabetic patients with myocardial infarction. Am Heart J. 2004;148:399–404. [DOI] [PubMed] [Google Scholar]
- 16. Yang JH, Song PS, Song YB, et al. Prognostic value of admission blood glucose level in patients with and without diabetes mellitus who sustain ST‐segment elevation myocardial infarction complicated by cardiogenic shock. Crit Care. 2013;17:R218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Dziewierz A, Giszterowicz D, Siudak Z, et al. Admission glucose level and in‐hospital outcomes in diabetic and non‐diabetic patients with acute myocardial infarction. Clin Res Cardiol. 2010;99:715–721. [DOI] [PubMed] [Google Scholar]
- 18. Evans N, Dhatariya K. Assessing the relationship between admission glucose levels, subsequent length of hospital stay, readmission and mortality. Clin Med (Lond). 2012;12:137–139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Saliba W, Barnett‐Griness O, Rennert G. Red cell distribution width and all‐cause mortality in patients with atrial fibrillation: a cohort study. J Arrhythm. 2017;33:56–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Tancredi M, Rosengren A, Svensson AM, et al. Excess mortality among persons with type 2 diabetes. N Engl J Med. 2015;373:1720–1732. [DOI] [PubMed] [Google Scholar]