Baseline characteristics

Baseline characteristics are shown in Table 1. 618,694 patients with a mean age of 47.6 (17.9) years were included, of whom 299,159 (48.4%) were men. According to the categorization after the index blood glucose level: 1871 (0.3%) had hypoglycemia, 525,636 (85%) had NGT, 77,442 (13%) had dysglycemia and 13,745 (2%) had hyperglycemia, respectively (Table 1). Patients with hyperglycemia were older and more often male compared to the other groups. The prevalence of hypertension, chronic obstructive pulmonary disease, prior stroke, and peripheral arterial disease was more common in patients with hyperglycemia. Patients with dysglycemia and hyperglycemia had the highest prevalence of atrial fibrillation, coronary heart disease, prior revascularization procedure (CABG and PCI), chronic kidney disease (CDK), and history of malignancy (i.e., any history of malignancy within 2 years prior to index-date), compared to the other groups. Patients with dysglycemia were more often treated with statin therapy, aspirin, P2Y12-inhibitors, angiotensin converting enzyme inhibitors (ACEI)/angiotensin receptor blockers (ARB) and oral anticoagulants (OAC) than the other groups.

Early (30-day) outcomes—event, event rates and risk of mortality, myocardial infarction, stroke, and heart failure due to blood glucose level categorization

During the first 30-days a total of 4780 patients died (0.8%): 80 (4.3%) patients with hypoglycemia, 2452 (0.5%) patients with NGT, 1360 (1.8%) patients with dysglycemia and 888 (6.5%) patients with hyperglycemia, respectively. Event, event rates and HRs are all shown in Table 2.

Table 2 Early (30-day) event, event rates and relative risks for all-cause mortality, cardiovascular mortality, myocardial infarction, stroke and heart failure in 618,694 patients attending the emergency department

Within the first 30-days, after multiple adjustments, patients with hypoglycemia had the highest risk of all-cause mortality HR 10.87 (95% CI 8.69–13.61), followed by patients with hyperglycemia HR 7.04 (95% CI 6.50–7.63), and patients with dysglycemia HR 2.14 (95% CI 1.99–2.29), respectively, compared to patients with NGT. In contrast patients with hyperglycemia had the highest risk of cardiovascular mortality HR 16.43 (95% CI 14.36–18.80) followed by patients with hypoglycemia HR 5.29 (95% CI 2.62–10.67), and patients with dysglycemia HR 3.13 (95% CI 2.74–3.58), compared to patients with NGT (Table 2).

After multiple adjustments, the risk of myocardial infarction, stroke and heart failure was highest among patients with hyperglycemia HR 3.39 (95% CI 3.10–3.70), HR 2.07 (95% CI 1.90–2.26) and HR 1.94 (95% CI 1.68–2.25), respectively, compared to patients with NGT. Corresponding numbers for patients with dysglycemia were HR 1.70 (95% CI 1.60–1.81), HR 1.44 (95% CI 1.37–1.51), and HR 1.21 (95% CI 1.10–1.33), respectively; and for patients with hypoglycemia HR 0.77 (95% CI 0.37–1.62), HR 1.07 (95% CI 0.64–1.77), and HR 0.53 (95% CI 0.20–1.42), respectively, compared to patients with NGT (Table 2).

Long-term outcomes—event, event rate and risk of mortality and myocardial infarction, stroke and heart failure due to blood glucose level categorization

During a mean follow-up time of 3.9 years (maximum 9 years), a total of 44,532 (7.2%) patients died: 214 (11.5%) with hypoglycemia, 31,635 (6.0%) with NGT, 9858 (12.7%) with dysglycemia, and 2825 (20.6%) patients with hyperglycemia, respectively. Events, event rate and HRs of mortality, myocardial infarction, stroke and heart failure between categorized groups are shown in Table 3. The long-term outcome of mortality is illustrated by a Kaplan Meier curve in Fig. 1. Long-term outcome of cardiovascular mortality, myocardial infarction, stroke, and hospitalization of heart failure are further illustrated in Kaplan Meier curves (Additional file 1: Fig. S1).

Table 3 Long term event, event rates and relative risks for all-cause mortality, cardiovascular mortality, myocardial infarction, stroke and heart failure in 618 694 patients attending the emergency department Fig. 1

Crude estimated Kaplan–Meier curve between blood glucose and death (all-cause mortality) in 618,694 patients with previous unknown diabetes categorized into four different groups, i.e., hypoglycemia (< 3.9 mmol/L), normal glucose levels (3.9–7.7 mmol/L), dysglycemia (7.8–11.0 mmol/L) and hyperglycemia (≥ 11.1 mmol/L) according to one random glucose blood level due to visiting emergency department at seven different hospitals in Sweden between 2006 and 2016

After multiple adjustments, the relative risk of all-cause mortality was highest among patients with hypoglycemia HR 2.58 (95% CI 2.26–2.96) followed by patients with hyperglycemia HR 1.69 (95% CI 1.63–1.76) and patients with dysglycemia HR 1.16 (95% CI 1.13–1.19), respectively, compared to the reference category of NGT. After multiple adjustments, the relative risk of cardiovascular mortality between groups was much the same as the relative risk for all-cause mortality (Table 3).

For the secondary outcomes, after multiple adjustments, the risk of myocardial infarction, stroke and heart failure were highest among patients with hyperglycemia HR 2.18 (95% CI 2.04–2.34), HR 1.54 (95% CI 1.44–1.65) and HR 1.49 (95% CI 1.36–1.63), respectively, compared to patients with NGT (Table 2). Corresponding numbers were for patients with dysglycemia HR 1.36 (95% CI 1.31–1.42), HR1.19 (95% CI 1.15–1.24) and HR 1.08 (95% CI 1.03–1.14), respectively; and for patients with hypoglycemia HR 0.89 (0.57–1.37), HR 1.14 (95% CI 0.81–1.60), and 1.02 (95% CI 0.66–1.56), respectively, compared to patients with NGT (Table 3).

After excluding the first 30-day from the analysis the results were much the same as for the main analysis (Additional file 1: Table S2).

Sensitive analysis of early and long-term outcomes and competing risk analysis of long-term outcomes—event, event rate and risk of myocardial infarction, stroke and heart failure due to blood glucose level categorization

As the glucose levels of patients attending the ED may be influenced by acute stress, we finally adjusted our model for a marker of acute stress, i.e., WBC count, in which the relative risks of early and long-term outcomes did not change, supporting that there were no confounding effect from acute stress (Tables 2 and 3). The association between blood glucose levels and the relative risk of cardiovascular events with competing risk of death was also investigated. In a competing risk regression analysis, one could see that the sub distribution HRs for myocardial infarction, stroke and heart failure was not statistically affected after this analysis (Additional file 1: Table S3).

Mortality and cardiovascular event rates related to sex

Event rates and risk of all-cause mortality, cardiovascular mortality, myocardial infarction, stroke and heart failure due to blood glucose level categorization in women and men, respectively, is presented in Additional file 1: Table S4. Age and sex standardized mortality rate for women was in the hypoglycemia group 57.2 (95% CI 37.5–76.9), NGT group 14.2 (95% CI 13.7–14.6), dysglycemia group 19.2 (95% CI 18.0–20.4) and hyperglycemia group 34.8 (95% CI 30.2–39.3), calculated per 1000 person-years, respectively. Corresponding numbers for men was in the hypoglycemia group 54.9 (95% CI 36.0–73.8), NGT group 20.1 (95% CI 19.5–20.6), dysglycemia group 23.5 (95% CI 22.3–24.8) and hyperglycemia group 40.5 (95% CI 36.4–44.5) calculated per 1000 person-years, respectively.