Injury and diabetes are both common. In the United States, approximately 27 million injured are treated in emergency departments per year, and 2.5 million are hospitalized. Approximately, 200,000 die as a result of injury (1). Diabetes mellitus (DM) affects an estimated 30 million people, of which 7 million are undiagnosed (2). It is therefore one of the most common comorbidities and increasing in incidence (3).
Around 7–9% of trauma patients admitted to the US trauma centers have a previous diagnosis of DM (4, 5). The impact of DM on outcomes in injured patients is therefore an area of great interest. Early studies focused on the association between presenting hyperglycemia and outcomes, showed that an elevated serum glucose level is associated with higher mortality and infection rates (6–8). However, these studies did not distinguish between hyperglycemia in diabetic patients and stress-induced hyperglycemia. We have previously demonstrated that, compared with nondiabetics with normoglycemia, patients with stress-induced hyperglycemia had a more than 2-fold increase in mortality (5). Diabetic patients with hyperglycemia, in contrast, had a nonsignificant, but near-50% increase in mortality.
Although our understanding of the impact of acute hyperglycemia on admission has improved, data regarding the impact of longer term glycemic control in diabetic trauma patients are still lacking. Hemoglobin A1c (HbA1c) is widely used to assess the overall quality of patients’ glycemic control. In the setting of specific conditions, patients with an HbA1c of more than 7.9% have been shown to be at increased risk of fractures and those with HbA1c of more than 6.5% have also been shown to have higher rates of delayed union (9, 10). Similarly, diabetic patients with an HbA1c of more than 8% have been shown to have delayed wound healing following burns (11). In the general trauma setting, a small study of injured patients admitted to an ICU in India has shown that, on univariate analysis, an HbA1c of more than 6% was associated with nonsignificant increases in a number of complications. Using a logistic regression model, which adjusted for Acute Physiology and Chronic Health Evaluation II score, injury severity score, baseline random blood sugar, and age at admission, the authors showed that patients with an HbA1c of more than 6% had a 4.6 times greater risk of “poor outcome” (a composite of death or discharge against medical advice) than those with an HbA1c of less than 6%. However, the study only included 120 patients, only 29 of whom had an elevated HbA1c, and the confidence intervals were wide (12).
Given the high prevalence of DM, the relationship between glycemic control and outcome in diabetic trauma patients, whether previously known or newly diagnosed, warrants further investigation. We hypothesized that patients with poorer glycemic control have increased rates of morbidity and mortality. The purpose of this study was to evaluate whether diabetic patients with poor glycemic control have worse outcomes and whether there is a demonstrable relationship between the quality of glycemic control and outcome.
MATERIALS AND METHODS
Design, setting, data source, and study population
This is a retrospective study of trauma patients evaluated at University of Alabama at Birmingham (UAB) Hospital, an American College of Surgeons verified Level 1 trauma center, between 2011 and June 2018. Data were obtained from the trauma registry and the electronic health record. To more accurately identify patients with DM, and undiagnosed DM, in particular, admission HbA1c has been part of our routine trauma panel since September 2009. Patients who had no HbA1c levels recorded upon initial presentation to the hospital were excluded from the study population. Hyperglycemia was managed with subcutaneous short- and long-acting insulin. Patients with serum glucose of more than 300 mg/dL, or hyperglycemia refractory to subcutaneous insulin administration, were managed with insulin infusions, in the step-down or ICU setting. The study was approved by the UAB Institutional Review Board for Human Use.
Demographic details, admission physiology, details of injury mechanism, revised trauma scores, injury severity scores, length of hospital and ICU stay, duration of ventilator support, hospital mortality, as well as the occurrence of pneumonia, acute kidney injury, sepsis, and urinary tract infection (UTI) (as defined by the National Trauma Data Bank) were extracted (13).
Patients were stratified into 4 groups, based on their admission HbA1c level, in accordance with definitions used by the American Diabetic Association (ADA) (14). Patients with an HbA1c of less than 5.7% were classified as nondiabetic or having excellent glycemic control. Patients with an HbA1c between 5.7% and 6.4% were classified as having good glycemic control, those with an HbA1c of 6.5% to 8.0% were classified as having moderate glycemic control, and those with an HbA1c of 8.1% or more were classified as having poor glycemic control. In line with ADA guidance, a serum glucose level of less than 140 mg/dL was considered normal (14). Demographic, injury, and clinical characteristics were compared among groups using χ2 tests and analysis of variance for categorical and continuous variables, respectively. A Poisson regression with robust error variance (15) that was adjusted for age, sex, injury severity score, and injury mechanism was used to estimate risk ratios (RRs) and associated 95% confidence intervals (CIs) for the association between HbA1C and the occurrence of pneumonia, acute kidney injury, sepsis, UTI, as defined by the National Trauma Databank. Given the differences in baseline characteristics, a subgroup analysis was performed, based on patients’ age. A threshold of 60 years for dichotomizing age for the subgroup analysis was determined by rounding the weighted average of age for the good, moderate and poor glycemic control groups (57.8 years) to the nearest 10. Models were created in the same methodology utilizing Poisson regression models described above.
The effect of missing HbA1c measurements was evaluated by comparing the baseline characteristics of those patients who had HbA1c measured, and those who did not. In addition, we conducted sensitivity analyses with best and worst-case imputation to determine whether nonrandom missingness could have impacted on outcomes.
The analysis was conducted using SAS version 9.4. A P-value of less than 0.05 was deemed to be statistically significant. All statistical tests were 2-tailed.
In total, 30,406 trauma patients were admitted to UAB Hospital over the duration of the study period; of these, 4272 (14.0%) did not have HbA1c levels evaluated, leaving 26,134 patients who were included in this analysis. There was no difference in regards to demographic, injury, or clinical characteristics between those that did and did not have HbA1c levels evaluated. Of those included, 17,934 patients (68.6%) were classified as nondiabetic or having excellent glycemic control, forming the reference population; 6067 (23.2%) had good glycemic control; 1210 (4.6%) had moderate glycemic control; and 923 (3.5%) had poor glycemic control. Age was youngest for nondiabetics and people with diabetes with excellent glycemic control (P < 0.001) (Table 1). ISS was highest for those with moderate glycemic control (P < 0.001). Patients with moderate or poor glycemic control were more likely to have a blunt injury (P < 0.001). Nondiabetics and people with diabetes with excellent glycemic control additionally had shorter hospital stay (P < 0.001), days in the ICU (P < 0.001), and days on ventilator support (P < 0.001).
Compared with nondiabetics and people with diabetes with excellent glycemic control, the risk of pneumonia was elevated in all 3 other glycemic control groups and was inversely related to the quality of glycemic control, going from a 25% increased risk for those with good glycemic control (RR 1.25, 95% CI 1.07–1.45) to a 53% risk among those with poor glycemic control (RR 1.53, 95% CI 1.13–2.07) (Table 2). The risk of developing acute kidney injury showed a similar pattern, with a 49% increase for those with good glycemic control (RR 1.49, 95% CI 1.15–1.92), a 2.5-fold increase for those with moderate glycemic control (RR 2.53, 95% CI 1.76–3.63), and a more than 3-fold increased risk for those with poor glycemic control (RR 3.20, 95% CI 2.18–4.71). The risk of UTI was elevated in patients with good glycemic control (RR 1.48, 95% CI 1.18–1.86) and those with poor glycemic control (RR 1.83, 95% CI 1.20–2.79), but did not show a clear trend between quality of control and risk. The risk of sepsis, in contrast, did not show any statistically significant difference in patients with poorer glycemic control; however, there was a trend with worse quality of control associated with higher risk of sepsis.
Diabetic patients had an increased risk of mortality, compared with nondiabetics and people with diabetes with excellent glycemic control. The risk again appeared to be related to the quality of glycemic control. Patients with good glycemic control had a 33% higher risk of mortality than nondiabetics and people with diabetes with excellent glycemic control, whereas those with moderate and poor glycemic control had a 51% and 54% higher risk of dying. Sensitivity analyses with best/worst case imputation for patients with missing HbA1c measurements showed no change in direction of effect.
By age, the observed association between glycemic control and pneumonia was observed only for those ≥60 years of age; in particular, compared with nondiabetics and people with diabetes with excellent glycemic control, there was an observed 40% increase in pneumonia risk for good glycemic control (RR 1.40, 95% CI 1.04–1.89), a near 60% increase for those with moderate control (RR 1.56, 95% CI 1.07–2.28), and a 70% increased risk for those with poor control (RR 1.70, 95% CI 1.03–2.81) (Table 3). Conversely, the risk of renal failure was mostly limited to those under 60 years of age with notably strong associations observed for those with moderate (RR 3.61, 95% CI 2.18–5.99) and poor glycemic control (RR 3.52, 95% CI 2.12–5.84). Similarly, the increased risk of UTI was limited to those younger than 60 years of age, particularly those with moderate (RR 2.36, 95% CI 1.35–4.11) and poor glycemic control (RR 1.88, 95% CI 1.04–3.41). Similar associations for mortality were observed for both age groups.
This study demonstrates that trauma patients’ long-term preinjury glycemic control is correlated with increased morbidity and mortality and that there is a “dose-response” relationship between the quality of glycemic control, and outcomes. Worse long-term control is associated with worse outcomes.
These findings are broadly in keeping with those of Lionel et al (12)—to our knowledge, the only other published study of the impact of HbA1c on outcomes in trauma patients. Lionel et al showed that higher HbA1c levels were associated with worse outcomes, but these results were not statistically significant on univariate analysis, and on multivariable logistic regression analysis were associated with wide (albeit statistically significantly different) confidence intervals (12). These findings are probably related to sample size, as our study, incorporating more than 25,000 participants, is substantially larger. Our results are also in keeping with a number of other studies that have shown that trauma patients with a past medical history of diabetes are at a higher risk of infections, including pneumonia, UTI, abscess formation, wound infection, disseminated fungal infection, and acute kidney injury (5, 8, 11, 16–20). A history of diabetes has also been shown to be associated with increased lengths of ICU and hospital stay as well as ventilator support in trauma patients (12, 13, 16, 17, 19, 21). The increased risk of mortality shown in our study is also in keeping with previous studies (21–23).
Our findings have prognostic implications. Routine checking of HbA1c levels as part of admission blood work may be worth considering, particularly since undiagnosed diabetes is common, in the population as a whole, and in trauma patients specifically. Although there is no acute treatment option for poor long-term glycemic control, elevated HbA1c levels can then be used to inform prognostication, and to anticipate, and proactively treat, complications such as infections or acute kidney injury.
This study has limitations. First, this is a retrospective study, and the results could be biased by unmeasured confounders. For example, HbA1c levels can be influenced by anemia, which was not captured as a comorbidity in the study dataset. Second, we were not able to include BMI in our analysis as 23% of the study population was missing BMI because of missing height or weight values; however, models including BMI (and therefore excluding the 23% missing BMI) had similar association estimates. Furthermore, we did not differentiate between patients with insulin-dependent diabetes mellitus and noninsulin-dependent diabetes mellitus, patients with known and occult diabetes, or consider the duration of disease. All of these could in principle influence the likelihood of developing complications. The reliability of HbA1c measurements in type I diabetes has been questioned, but there is now broad agreement that this is not an issue (24). However, this study is strengthened by the fact that it includes a large cohort of patients across a time period of nearly a decade and relies on widely accepted definitions of the quality of glycemic control.
In conclusion, this study adds to our knowledge and understanding of the impact of diabetes and glycemic control on outcomes in trauma patients. In addition to the adverse effects of increased admission hyperglycemia in the trauma population demonstrated in our previous study, current data suggest long-term glycemic levels also play an important role in the development of complications following trauma injury. Lastly, with the observation that admission serum glucose levels are highly correlated with HbA1c levels, it is possible that high admission glucose levels are a marker for poor long-term glycemic control.
1. Centers for Disease Control. Key injury and violence data. Available at Centers for Disease Control and Prevention, National Center for Health Statistics. Underlying Cause of Death 1999–2017 on CDC WONDER Online Database, released December, 2018. Data are from the Multiple Cause of Death Files, 1999–2017, as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program. Accessed at http://wonder.cdc.gov/ucd-icd10.html
(accessed May 12, 2018).
2. Centers for Disease Control and PreventionNational Diabetes Statistics Report, 2017. Atlanta, GA:Centers for Disease Control and Prevention, US Department of Health and Human Services; 2017.
3. Geiss LS, Wang J, Cheng YJ, Thompson TJ, Barker L, Li Y, Albright AL, Gregg EW. Prevalence and incidence trends for diagnosed diabetes among adults aged 20 to 79 years, United States, 1980–2012. JAMA
4. Bosarge PL, Shoultz TH, Griffin RL, Kerby JD. Stress-induced hyperglycemia is associated with higher mortality in severe traumatic brain injury. J Trauma Acute Care Surg
5. Kerby JD, Griffin RL, MacLennan P, Rue LW 3rd. Stress-induced hyperglycemia, not diabetic hyperglycemia, is associated with higher mortality in trauma
. Ann Surg
6. Laird AM, Miller PR, Kilgo PD, Meredith JW, Chang MC. Relationship of early hyperglycemia to mortality in trauma
patients. J Trauma
7. Sung J, Bochicchio GV, Joshi M, Bochicchio K, Tracy K, Scalea TM. Admission hyperglycemia is predictive of outcome in critically ill trauma
patients. J Trauma
8. Yendamuri S, Fulda GJ, Tinkoff GH. Admission hyperglycemia as a prognostic indicator in trauma
. J Trauma
9. Conway BN, Long DM, Figaro MK, May ME. Glycemic control and fracture risk in elderly patients with diabetes. Diabetes Res Clin Pract
10. Liu J, Ludwig T, Ebraheim NA. Effect of the blood HbA1c
level on surgical treatment outcomes of diabetics with ankle fractures. Orthop Surg
11. Schwartz SB, Rothrock M, Barron-Vaya Y, Bendell C, Kamat A, Midgett M, Abshire J, Biebighauser K, Staiano-Coico LF, et al. Impact of diabetes on burn injury: preliminary results from prospective study. J Burn Care Res
12. Lionel KR, John J, Sen N. Glycated hemoglobin A: a predictor of outcome in trauma
admissions to intensive care unit. Indian J Crit Care Med
13. American College of Surgeons. National Trauma
Data Standard Data Dictionary [code book]. Accessed at https://www.facs.org/-/media/files/quality-programs/trauma/ntdb/ntds/data-dictionaries/ntdsdatadictionary2017admissions.ashx
(Accessed May 12, 2018).
14. American Diabetes AssociationDiagnosis and classification of diabetes mellitus
. Diabetes Care
37: (Suppl. 1): S81–S90, 2014.
15. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol
16. Ahmad R, Cherry RA, Lendel I, Mauger DT, Service SL, Texter LJ, Gabbay RA. Increased hospital morbidity among trauma
patients with diabetes mellitus
compared with age- and injury severity score-matched control subjects. Arch Surg
17. Kao LS, Todd SR, Moore FA. The impact of diabetes on outcome in traumatically injured patients: an analysis of the National Trauma
Data Bank. Am J Surg
18. Korim MT, Payne R, Bhatia M. A case-control study of surgical site infection following operative fixation of fractures of the ankle in a large U.K. trauma
unit. Bone Joint J
19. Low ZK, Ng WY, Fook-Chong S, Tan BK, Chong SJ, Hwee J, Tay SM. Comparison of clinical outcomes in diabetic and non-diabetic burns patients in a national burns referral centre in southeast Asia: a 3-year retrospective review. Burns
20. Tebby J, Lecky F, Edwards A, Jenks T, Bouamra O, Dimitriou R, Giannoudis PV. Outcomes of polytrauma patients with diabetes mellitus
. BMC Med
21. Liou DZ, Singer MB, Barmparas G, Harada MY, Mirocha J, Bukur M, Salim A, Ley EJ. Insulin-dependent diabetes and serious trauma
. Eur J Trauma Emerg Surg
22. Lustenberger T, Talving P, Lam L, Inaba K, Bass M, Plurad D, Demetriades D. Effect of diabetes mellitus
on outcome in patients with traumatic brain injury: a national trauma
databank analysis. Brain Inj
23. Selassie AW, Cao Y, Church EC, Saunders LL, Krause J. Accelerated death rate in population-based cohort of persons with traumatic brain injury. J Head Trauma Rehabil
24. Sherwani SI, Khan HA, Ekhzaimy A, Masood A, Sakharkar MK. Significance of HbA1c
test in diagnosis and prognosis of diabetic patients. Biomark Insights