More than 34 million people in the United States have diabetes mellitus (DM) and use hemoglobin A1c (A1c) as a critical laboratory marker for glycemic control.1,2 Although a diagnosis of DM is an established risk factor for a multitude of perioperative complications in orthopaedic surgery, uncontrolled DM and/or a diagnosis of DM with associated end-organ manifestations is a more consistent predictor of poor outcomes.3–10 Specifically, hemoglobin A1c as a marker of glycemic control in orthopaedic patients with DM has been evaluated as a predictor of adverse postoperative outcomes, such as 2-year mortality, 90-day hospital readmission, surgical site infection (SSI), revision surgery, functional scores, delayed union, and poor radiographic outcomes.3–6,8

Prior research involving DM, glycemic control, and complications in orthopaedic patients has primarily focused on elective procedures. Given the concern for increased risk of postoperative complications with markedly elevated A1c levels, most orthopaedic surgeons proceed with caution when considering elective surgical treatment in patients with uncontrolled DM. By contrast, traumatic orthopaedic injuries and their associated surgical indications may dictate surgical treatment even if poor glycemic control is present. Therefore, fracture surgery may offer further insight into the effect of glycemic control on postoperative outcomes in orthopaedic patients. The limited data concerning A1c and outcomes in patients undergoing orthopaedic surgery for a fracture vary in the definitions of postoperative complications and use institution-specific definitions of SSI.5,7,8,10

Thus, we aimed to evaluate A1c as a predictor of postoperative SSI in patients with orthopaedic trauma using a consensus definition of SSI in fracture surgery.11 Furthermore, we sought to investigate if patients with A1c levels consistent with a diagnosis of DM who undergo orthopaedic trauma surgery experience a higher rate of SSI compared with patients with A1c levels below the cutoff for diagnosis of DM who undergo orthopaedic trauma surgery. We hypothesized that elevated A1c values would be markedly associated with the rate of postoperative SSI in patients with orthopaedic trauma, and patients with A1c levels above 6.5 undergoing orthopaedic trauma surgery would experience a markedly higher rate of postoperative SSI compared with patients with A1c levels below 6.5 undergoing orthopaedic trauma surgery.

Methods

The study aims were addressed through a retrospective chart review at a single level one trauma center. After Institutional Review Board approval, the electronic medical record was queried for patients aged 18 years or older treated surgically for an acute fracture by a fellowship-trained orthopaedic trauma surgeon at our institution from July 1, 2012, to December 1, 2020, with a laboratory value for hemoglobin A1c available within 3 months of their surgery. Patients were excluded if they underwent surgery for a diagnosis other than fracture or if they were initially treated for their fracture at an outside hospital. The primary outcome of interest was SSI as defined by ‘Fracture related infection: A consensus on definition from an international expert group,’ by Metsemakers et al11 Criteria confirmatory and suggestive of a diagnosis of SSI were identified in the medical record by review of the treating orthopaedic surgeon's documentation and available laboratory, culture, and pathology results (Table 1). The suggestive criteria of radiological signs were omitted from review because of anticipated difficulty with consistent valid retrospective identification within the medical record. An A1c value of 6.5 was selected as a point of evaluation within the study given the established criteria for the diagnosis of DM is defined by a hemoglobin A1c value ≥ 6.5.12 The diagnosis of grossly uncontrolled diabetes was defined by a hemoglobin A1c value ≥ 10.1,13 A1c values within 3 months of fracture surgery were used because of the acceptance of A1c as the standard measure of a patient's glycemic control over the preceding 3 months.2

Table 1 - Factors Confirmatory and Suggestive of Fracture Related Infection as Defined by Metsemakers et al Confirmatory criteria of FRI Suggestive criteria of FRI Fistula or sinus tract Clinical signs: Pain, redness, swelling, fever Purulent drainage or presence of pus Radiological signs: Bone lysis, implant loosening, sequestration, periosteal bone formation Phenotypically indistinguishable pathogens identified by culture from at least two separate deep tissue/implant specimens New-onset joint effusion Presence of microorganisms in deep tissue specimens confirmed by histopathological examination Elevated serum inflammatory markers (ESR, CRP, WBC) Persistent, increasing, or new-onset wound drainage Pathogenic organism identified by culture from a single deep tissue/implant specimen

FRI = Fracture-related infection

The rate of positive factors suggestive of SSI for the cohort was calculated. The rate of positive factors confirmatory of SSI for the cohort was calculated. A receiver operating characteristic curve (ROC) was calculated using hemoglobin A1c as a predictor for both confirmatory and suggestive SSI criteria. SSI in patients with hemoglobin A1c values above and below the cutoff for diagnosis of DM (A1c < 6.5) was analyzed using the Pearson Chi-Square test. SSI in patients with hemoglobin A1c values above and below the cutoff for diagnosis of grossly uncontrolled DM (A1c < 10) was analyzed using the Pearson Chi-Square test. A P-value <0.05 was used to establish statistical significance. Open fracture is a known risk factor for postoperative fracture-related infection in orthopaedic trauma patients.14,15 In light of this fact and its potential effect on study results, the patient cohort was divided based on fracture type (open versus closed) and analyzed in the same manner as described above. A total of 925 patients were evaluated in the patient cohort based on inclusion and exclusion criteria. The study population is further described in Table 2.

Patient Characteristic Number of Patients Sex  Male 440  Female 485 Closed fracture 730 Open fracture 195 Comorbidities  Diabetes mellitus 521  Coronary artery disease 134  Congestive heart failure 114  Chronic kidney disease 101  Peripheral vascular disease 48  Hypertension 630 Average BMI 32.01 (range 13.1-95.8)
Results

A total of 925 patients met criteria for the final analysis. There were signs suggestive of SSI in 19.6% of patients (182/925). There were signs confirmatory of SSI in 11.9% of patients (110/925). To evaluate A1c as a tool to predict SSI in patients with orthopaedic trauma based on the current consensus definition of SSI in fracture surgery, a ROC was calculated using hemoglobin A1c as a predictor for SSI criteria (Figures 1 and 2). The ROC for signs suggestive of SSI demonstrated an area under the curve (AUC) of 0.535 (Table 3). The ROC for signs confirmatory of SSI demonstrated an AUC of 0.539 (Table 3).

Figure 1:

Graph showing ROC evaluating A1c as a predictor of criteria suggestive of surgical site infection. ROC = receiver operating characteristic curve

Figure 2:

Graph showing ROC evaluating A1c as a predictor of criteria confirmatory of surgical site infection. ROC = receiver operating characteristic curve

Table 3 - Area Under the Curve for Receiver Operating Characteristic Curve Using A1c as a Predictor of Surgical Site Infection Criteria Signs Suggestive of SSI AUC Signs Confirmatory of SSI AUC 0.535 0.539

AUC = area under the curve

In patients with A1c levels below the threshold for diagnosis of DM (<6.5), signs suggestive of SSI were identified in 18.3% of patients (100/547). This was similar to the rate of signs suggestive of SSI in patients with A1c levels consistent with a diagnosis of DM (82/378, 21.7%) (P = 0.199). In patients with A1c levels below the threshold for diagnosis of DM (<6.5), signs confirmatory of SSI were identified in 11.0% of patients (60/547). This was similar to the rate of signs confirmatory of SSI in patients with A1c levels consistent with a diagnosis of DM (50/378, 13.2%) (P = 0.297). In patients with completely uncontrolled DM (A1c > 10), signs suggestive of SSI were identified in 22.5% of patients (16/71). Signs suggestive of SSI were identified in 19.4% of patients (166/854) with A1c values < 10 (P = 0.528). In patients with completely uncontrolled DM (A1c > 10), signs confirmatory of SSI were identified in 14.1% of patients (10/71). Signs confirmatory of SSI were identified in 11.7% of patients (100/854) with A1c values < 10 (P = 0.552).

To further characterize the utility of A1c as a predictor of SSI criteria in this patient cohort, ROCs were calculated on patients with closed fractures and A1c values < 6.5, > 6.5, < 10, and > 10. Table 4 summarizes the AUC for each ROC calculated for the closed fracture cohort. Similarly, ROCs were calculated on patients with open fractures and A1c values < 6.5, > 6.5, < 10, and > 10. Table 4 also summarizes the AUC for each ROC calculated for the open fracture cohort.

Table 4 - Area Under the Curve Values for Closed and Open Fracture Receiver Operating Characteristic Curves A1c Value ROC AUC for Signs Suggestive of SSI in Closed Fractures ROC AUC for Signs Confirmatory of SSI in Closed Fractures ROC AUC for Signs Suggestive of SSI in Open Fractures ROC AUC for Signs Confirmatory of SSI in Open Fractures <6.5 0.500 0.438 0.598 0.700 >6.5 0.520 0.560 0.473 0.437 <10 0.535 0.523 0.528 0.564 >10 0.448 0.327 0.679 0.444

AUC = area under the curve, ROC = receiver operating characteristic curve

The rate of signs suggestive and confirmatory of SSI in the group of patients with closed fractures was calculated (Table 5). In patients with closed fractures and A1c levels below the threshold for diagnosis of DM (<6.5), signs suggestive of SSI were identified in 13.5% of patients (59/436). This was similar to the rate of signs suggestive of SSI in patients with closed fractures and A1c levels consistent with a diagnosis of DM (48/294, 16.3%) (P = 0.295). In patients with closed fractures and A1c levels below the threshold for diagnosis of DM (<6.5), signs confirmatory of SSI were identified in 8.0% of patients (35/436). This was similar to the rate of signs confirmatory of SSI in patients with closed fractures and A1c levels consistent with a diagnosis of DM (28/294, 9.5% (P = 0.480). In patients with closed fractures and completely uncontrolled DM (A1c > 10), signs suggestive of SSI were identified in 17.2% of patients (10/58). Signs suggestive of SSI were identified in 14.4% of patients (97/672) with closed fractures and A1c values < 10 (P = 0.561). In patients with closed fractures and completely uncontrolled DM (A1c > 10), signs confirmatory of SSI were identified in 10.3% of patients (6/58). Signs confirmatory of SSI were identified in 8.5% of patients (57/672) with closed fractures and A1c values < 10 (P = 0.627).

Table 5 - Rates of Signs of SSI in Closed Fractures A1c Value Rate of Signs Suggestive of SSI P for Signs Suggestive of SSI Rate of Signs Confirmatory of SSI P for Signs Confirmatory of SSI <6.5 59/436 = 13.5% 0.295 35/436 = 8.0% 0.480 >6.5 48/294 = 16.3% 28/294 = 9.5% <10 97/672 = 14.4% 0.561 57/672 = 8.5% 0.627 >10 10/58 = 17.2% 6/58 = 10.3%

SSI = surgical site infection

The rate of signs suggestive and confirmatory of SSI in the group of patients with open fractures was calculated (Table 6). In patients with open fractures and A1c levels below the threshold for diagnosis of DM (<6.5), signs suggestive of SSI were identified in 36.4% of patients (40/110). This was similar to the rate of signs suggestive of SSI in patients with open fractures and A1c levels consistent with a diagnosis of DM (34/85, 40.0%%) (P = 0.604). In patients with open fractures and A1c levels below the threshold for diagnosis of DM (<6.5), signs confirmatory of SSI were identified in 22.7% of patients (25/110). This was similar to the rate of signs confirmatory of SSI in patients with open fractures and A1c levels consistent with a diagnosis of DM (22/85, 25.9%) (P = 0.609). In patients with open fractures and completely uncontrolled DM (A1c > 10), signs suggestive of SSI were identified in 46.2% of patients (6/13). Signs suggestive of SSI were identified in 37.4% of patients (68/182) with open fractures and A1c values < 10 (P = 0.528). In patients with open fractures and completely uncontrolled DM (A1c > 10), signs confirmatory of SSI were identified in 30.8% of patients (4/13). Signs confirmatory of SSI were identified in 23.6% of patients (43/182) with open fractures and A1c values < 10 (P = 0.561).

Table 6 - Rate of Signs of SSI in Open Fractures A1c Value Rate of Signs Suggestive of SSI P Rate of Signs Confirmatory of SSI P <6.5 40/110 = 36.4% 0.604 25/110 = 22.7% 0.609 >6.5 34/85 = 40.0% 22/85 = 25.9% <10 68/182 = 37.4% 0.528 43/182 = 23.6% 0.561 >10 6/13 = 46.2% 4/13 = 30.8%

SSI = surgical site infection

Discussion

A large portion of the US population have DM.1,2 Previous research focusing on elective procedures has demonstrated numerous postoperative complications in patients with DM and poor glycemic control, measured by A1c value, undergoing orthopaedic surgery.3–10 Limited data exist in the current literature regarding A1c and outcomes in patients with orthopaedic trauma.5,7,8,10 Fracture surgery may provide further insight into the effect of glycemic control on postoperative complications given the appropriate hesitancy of orthopaedic surgeons to pursue elective surgery in patients with poor DM management. The available studies evaluating A1c as it relates to orthopaedic trauma surgery use institution-specific definitions of SSI.5,7,8,10 To our knowledge, this is the first study investigating A1c as a predictor of SSI in patients with orthopaedic trauma using the consensus definition established by Metsemakers et al.11

We found A1c was not a useful tool to predict SSI in this cohort of 925 patients with orthopaedic trauma based on the current consensus definition of SSI in fracture surgery (Figures 1, 2 and Table 3). Despite numerous other studies demonstrating a connection between A1c and outcomes in orthopaedic surgery patients, we found no association between A1c values and postoperative SSI after fracture surgery. We found no statistical significance between the rate of signs suggestive or confirmatory of SSI in patients with A1c values consistent with a diagnosis of DM (A1c > 6.5) and those with A1c values below the cutoff for diagnosis of DM (A1c < 6.5). In addition, no statistical significance was identified between the rate of signs suggestive or confirmatory of SSI in patients with completely uncontrolled DM (A1c > 10) compared with patients with A1c values < 10. Similar statistically insignificant findings between the rate of signs suggestive or confirmatory of SSI were maintained when the patient cohort was divided into closed fractures and open fractures and analyzed based on the A1c values of < 6.5 compared with > 6.5 and A1c < 10 compared with A1c > 10 (Tables 5 and 6). When ROCs were calculated using hemoglobin A1c as a predictor for SSI criteria in closed and open fractures separately, results demonstrated that A1c was not a useful predictor of SSI criteria with the exception of confirmatory signs of SSI in patients with open fractures and A1c < 6.5 (Table 4). With a ROC analysis showing an AUC of approximately 0.5 and no pertinent results seen in using A1c as a continuous variable or using the commonly accepted threshold of 6.5 and an additional excessive value of 10 in categorical testing, A1c was not reliably predictive of SSI in orthopaedic trauma surgery based on the consensus definition of SSI in fracture surgery. Given the unexpected nature of these findings and rejection of the null hypothesis, a power analysis was performed using the accepted A1c value for defining the diagnosis of DM at 6.5 which demonstrated a 90% power to detect a difference in SSI rates if one existed in our data. For the A1c value defining the diagnosis of grossly uncontrolled DM at 10, power analysis demonstrated a 60% power to detect a difference in SSI rates if one existed in our data.

This study has several limitations. As a retrospective review, it is unknown how many total patients were treated for an acute fracture by an orthopaedic traumatologist at our institution during the study time frame and were not included because of lack of an A1c value or information pertaining to SSI. This is in contrast to a prospective study in which patients could be followed until study completion or the primary outcome of concern of SSI. It is also unknown why each patient had laboratory work obtained for their A1c value, but studying a cohort undergoing such investigation could have selected for patients perceived to be at a higher risk for diagnosis of DM and its associated complications by the ordering provider. In addition, inclusion of patients whose A1c value was obtained postoperatively could have selected for a population of patients who underwent laboratory investigation after a diagnosis of SSI was either established or suspected by providers in the postoperative period and, therefore, bias results. To evaluate these potential confounding biases, statistical analysis was performed on a subset patient cohort with a laboratory value for hemoglobin A1c obtained within 3 months preoperatively or 2 weeks postoperatively to eliminate any patients with an A1c value obtained after a potential diagnosis of SSI. The review revealed that approximately 90% of the study cohort had their A1c value obtained preoperatively or in the perioperative period < 2 weeks postoperatively. Statistical analysis of this smaller cohort (N = 830) demonstrated very similar results as the study cohort with absolutely no association between A1c and signs suggestive or confirmatory of SSI. A potential secondary limitation of this study cohort is the broad inclusion of all surgically treated acute fractures which met the other inclusion criteria. This group includes heterogeneous fracture types with diverse surgical treatment options and variable baseline risk of SSI before consideration of patient related factors. This limitation may decrease the generalizability of the results at the benefit of increased confidence in the rejection of the null hypothesis.

Despite extensive research in orthopaedics involving DM and glycemic control, numerous avenues for future investigation continue to exist. Although A1c provides information regarding patient's blood glucose levels over a 3-month time span, blood glucose management in the acute perioperative period may affect SSI and long-term outcomes.4,8 A1c is the only one parameter used in the management of patients with DM. Future studies could evaluate the potential predictive nature of DM-associated end-organ complications or other comorbidities on postoperative complications after orthopaedic trauma surgery in this patient cohort. The risk of SSI is of high importance when considering fracture healing and the presence of orthopaedic implant; however, it is the only one potential postoperative complication in patients with orthopaedic trauma which may be affected by the presence of DM and poor glycemic control. In addition, future studies should continue to use and expand on the consensus definition of fracture-related infection. Future validation of confirmatory and suggestive criteria of fracture-related infection by prospective data could affect treatment algorithms for SSI in patients with orthopaedic trauma.

Although existing literature has demonstrated an association with postoperative infection in orthopaedic surgery patients who have DM and elevated hemoglobin A1c values, in this cohort of patients with orthopaedic trauma, hemoglobin A1c was not a valuable tool to predict postoperative SSI, and no statistical significance was identified between the rate of SSI in patients with hemoglobin A1c values above and below the cutoff for diagnosis of DM (A1c > 6.5) or in those patients with completely uncontrolled DM represented by a hemoglobin A1c value > 10. Given these findings, routine A1c monitoring is not a reliable predictor of SSI criteria in patients with orthopaedic trauma based on the current consensus definition of SSI in fracture surgery.

Acknowledgments

We thank Josny Thimothee, MD, and Lauren Pole, BS, for their contributions to data collection.

References 1. Centers for Disease Control and Prevention: National diabetes statistical report, 2020, Available at: https://www.cdc.gov/diabetes/data/statistics-report/index.html, Accessed November 13, 2020. 2. Radin MS: Pitfalls in hemoglobin A1c measurement: When results may Be misleading. J Gen Intern Med 2014;29:388-394. 3. Cancienne JM, Werner BC, Browne JA: Is there a threshold value of hemoglobin A1c that predicts risk of infection following primary total hip arthroplasty? J Arthroplasty 2017;32:S236-S240. 4. Chrastil J, Anderson MB, Stevens V, Anand R, Peters CL, Pelt CE: Is hemoglobin A1c or perioperative hyperglycemia predictive of periprosthetic joint infection or death following primary total joint arthroplasty? J Arthroplasty 2015;30:1197-1202. 5. Liu J, Ludwig T, Ebraheim NA: Effect of the blood HbA1c level on surgical treatment outcomes of diabetics with ankle fractures. Orthop Surg 2013;5:203-208. 6. McElvany MD, Chan PH, Prentice HA, Paxton EW, Dillon MT, Navarro RA: Diabetes disease severity was not associated with risk of deep infection or revision after shoulder arthroplasty. Clin Orthop Relat Res 2019;477:1358-1369. 7. Regan DK, Manoli A, Hutzler L, Konda SR, Egol KA: Impact of diabetes mellitus on surgical quality measures after ankle fracture surgery: Implications for "Value-Based" compensation and "pay for performance". J Orthop Trauma 2015;29:e483-e486. 8. Reich MS, Fernandez I, Mishra A, Kafchinski L, Adler A, Nguyen MP: Diabetic control predicts surgical site infection risk in orthopaedic trauma patients. J Orthop Trauma 2019;33:514-517. 9. Richards JE, Kauffmann RM, Zuckerman SL, Obremskey WT, May AK: Relationship of hyperglycemia and surgical-site infection in orthopaedic surgery. J Bone Joint Surg Am 2012;94:1181-1186. 10. Schmidt T, Simske NM, Audet MA, Benedick A, Kim CY, Vallier HA: Effects of diabetes mellitus on functional outcomes and complications after torsional ankle fracture. J Am Acad Orthop Surg 2020;28:661-670. 11. Metsemakers WJ, Morgenstern M, McNally MA, et al.: Fracture-related infection: A consensus on definition from an international expert group. Injury 2018;49:505-510. 12. Centers for Disease Control and Prevention: All about your A1c, Available at: https://www.cdc.gov/diabetes/managing/managing-blood-sugar/a1c.html, Accessed July 19, 2022. 13. Centers for Medicare and Medicaid Services: Comprehensive diabetes care: Hemoglobin A1c (HbA1c) poor control (>9.0%): Ages 18 to 75, Available at: https://www.medicaid.gov/state-overviews/scorecard/comprehensive-diabetes-care/index.htmlAccessed July, 19 2022. 14. Gustilo RB, Merkow RL, Templeman D: The management of open fractures. J Bone Joint Surg Am 1990;72:299-304. 15. Garner MR, Sethuraman SA, Schade MA, Boateng H: Antibiotic prophylaxis in open fractures: Evidence, evolving issues, and recommendations. J Am Acad Orthop Surg 2020;28:309-315.