Secondary Logo

Journal Logo

Cohort Study

Day of surgery capillary glucose predictability of complications and length of stay for total knee arthroplasty patients with diabetes: a retrospective cohort study

Robin, Alex DSSa,; Gautreau, Sylvia PhDb; MacSween, Mary C. MDb; Cartier, Louis-Jacques MDb

Author Information
doi: 10.1097/SR9.0000000000000023
  • Open

Abstract

Diabetes mellitus is a well-established risk factor for increased length of stay (LOS) and complications in orthopedic surgeries1. Patients with diabetes develop more complications following orthopedic surgeries than those without diabetes—including surgical site infection, pulmonary embolism, acute kidney injury, and myocardial infarction2–5. With the prevalence of this disease predicted to continue to increase and total knee arthroplasty (TKA) being the second most common surgery in the country, it is important for orthopedic surgeons to consider diabetes in their plan of risk reduction for adverse postoperative events6,7.

Hyperglycemia causes an impaired immune response and decreased fibrinolysis leading to a host predisposed to infections and thrombotic events8,9. As diabetes is a dynamic disease, preoperative evaluation traditionally assesses both long-term diabetes control (A1C) and short-term control (random glucose, day of surgery capillary blood glucose). However, little consensus in surgery exists on the application of preoperative A1C and random glucose, and, notably of the day of surgery capillary blood glucose10–12. Between different national associations, there is no consistent optimal perioperative glucose target.

At our center, patients with diabetes undergo a A1C and a random glucose 1 to 6 weeks before surgery. Patients over 50 years also undergo random glucose. Capillary glycemia is routinely done on the day of surgery upon patient arrival. While measures of glycemia are routine in preoperative workups, a better understanding of its interpretation would be invaluable.

Marked hyperglycemia delaying surgery is often an empiric decision by the surgeon and anesthetist due to a lack of consistency from various authorities. Further research is required to lead to more specific guidance.

The purpose of this study is to ascertain if preoperative capillary blood glucose is associated with adverse events or longer LOS, in patients with diabetes undergoing TKA. Our study examines the predictive postoperative effect of these preoperative laboratory values (A1C, random glucose, day of surgery glucose) in 226 patients with diabetes who underwent TKA.

Methods

This was a single-centered retrospective cohort study in patients over 45 years, with diabetes (type 1 or 2) who underwent elective, primary TKA at The Moncton Hospital, between April 2015 and April 2019, identified using our local Discharge Abstract Database. Patients who underwent 2 unique TKA operations (on different knees) were considered separately, while patients who underwent revisional or bilateral TKA were excluded. Patients without prior documented diabetes in medical records and who did not meet Diabetes Canada Clinical Practice Guidelines were excluded13.

Using electronic medical record and charts, preoperative A1C and random blood glucose (within 2 mo of the surgery), as well as the preoperative day of surgery point-of-care (POC) glucose, were collected. A provincial POC glucose digital database was also consulted. Other data collected included: sex, region, comorbidities for Charlson Comorbidity Index, body mass index, indication for surgery, smoking status, medical management of diabetes, time of day of surgery, hospital LOS and complications14. Complications arising within 90 days postoperative included urinary tract infection, acute kidney injury, wound infection, hemorrhage, congestive heart failure, and venous thromboembolism (VTE) (deep vein thrombosis, pulmonary embolism).

Receiver operating characteristic (ROC) analyses were used to determine whether blood glucose readings predict complications and LOS. To facilitate the ROC analyses, complications and LOS were coded dichotomously—as yes/no and as 3 days or less versus 4 days or greater, respectively. The area under the receiver operating characteristic curves (AUCs) were reported with their corresponding 95% confidence intervals (CIs). We considered a P-value >0.05 statistically significant.

Results

Demographics

As per Tables 1 and 2, a total of 226 patients were included in our analysis, 38.2% male, mean body mass index 36.9 kg/m2. Overall, 27.2% (60) of our patients were on insulin therapy, while only 5 of 226 patients had type 1 diabetes. The mean Charlson Comorbidity Index, which predicts the 10-year survival rate in patients with multiple comorbidities, was 4.4±1.8 points, out of a maximum of 37 points.

Table 1 - Patient demographics, continuous variables.
Mean±SD Range (Minimum–Maximum) Median 25th Percentile 75th Percentile Interquartile Range
Age (y) 67.3±8.4 45.0 (42.0–87.0) 67.0 62.0 73.0 11.0
BMI 36.9±7.4 42.3 (19.9–62.2) 35.2 31.8 42.0 10.2
Years of diabetes diagnosis 10.7±6.7 56.0 (1.0–57.0) 10.0 6.0 15.0 9.0
Insulin units* 93.6±65.3 263 (10.0–273.0) 71.0 43.0 139.0 96.0
CCI score 4.4±1.8 12.0 (1.0–13.0) 4.0 3.0 5.5 2.5
ASA 3.1±0.6
*For patients on insulin therapy.
ASA indicates American Society of Anesthesiologists; BMI, body mass index; CCI, Charlson Comorbidity Index.

Table 2 - Patient demographics, categorical variables.
Frequency, n (%)
Sex
 Male 86 (38.2)
 Female 139 (61.8)
Smoking
 Current 14 (6.2)
 Former 103 (45.8)
 Never 108 (48.0)
Place of residence
 In-province 195 (86.7)
 Out-of-province 30 (13.3)
Diabetes control
 Diet-controlled 27 (12.3)
 1 antihyperglycemic agent 81 (36.8)
 ≥2 antihyperglycemic agents 52 (23.6)
 Insulin±antihyperglycemic agent 60 (27.2)
Anemia
 Yes 28 (12.6)
 No 195 (87.4)

Postoperative considerations

As in Table 3, the average LOS was 4.5 days with a 3-day median; the 2020 Canadian Joint Replacement Registry data does not report a subgroup with diabetes for LOS, but including patients with and without diabetes showed a 2.9-day average with a 2-day median15. This difference was expected as patients without diabetes are known to generally have shorter LOS than patients with diabetes16.

Table 3 - Study data.
Mean±SD Range (Minimum–Maximum) Median 25th Percentile 75th Percentile Interquartile Range
Length of stay (d) 4.5±9.4 135.0 (1.0–136.0) 3.0 2.0 4.0 2.0
Day of surgery capillary glycemia 8.2±2.2 14.9 (4.2–19.1) 7.8 6.7 9.2 2.5
Preoperative A1C 6.9±1.0 7.1 (4.2–11.3) 6.7 6.2 7.4 1.2
Preoperative random glucose 8.0±3.2 24.9 (2.8–27.7) 7.2 6.0 8.9 2.9

Notably, one 72-year-old patient in our study had a prolonged LOS at 136 days. Of note, capillary blood glucose morning of surgery was 16.8 mmol/L, preoperative random glucose 27.7 mmol/L, and A1C 6.2%. This patient did not have any complications fitting our criteria but had a history of severe obesity, long-standing diabetes with chronic kidney disease, and normocytic anemia (hemoglobin=94). A1C was likely falsely low due to chronic kidney disease and anemia.

Table 4 shows that 24 patients had complications; most commonly infection (14 patients), followed by VTE (3 patients) and AKI (3 patients). One patient had both infection and VTE.

Table 4 - Complications.
Complications Frequency, n (%)
Yes 25 (11.0)
 Infection 14 (6.2)
 Venous thromboembolism 4 (1.8)
 Hemorrhage 3 (1.3)
 Congestive heart failure 1 (0.4)
 Cerebrovascular accident 1 (0.4)
No 202 (89.0)

Correlation statistics

ROC curve was used to determine a correlation between day of surgery POC glucose with postoperative LOS and complications. AUC values of .56, .64, and .71 have been described in the literature as small, medium, and large effect sizes, respectively17. The ROC curve (Figs. 1, 2) for the prediction of day of surgery POC glycemia on LOS after TKA was nonsignificant (P=0.063). The AUC was 0.578 (95% CI: 0.491–0.664).

Figure 1
Figure 1:
Receiver operating characteristic (ROC) curve for point-of-care glycemia prediction of hospital length of stay following total knee arthroplasties. The diagonal line (in green) indicates the point of a chance prediction. ROC curve indicated in blue.
Figure 2
Figure 2:
Receiver operating characteristic (ROC) curve for point-of-care glycemia prediction of complications following total knee arthroplasties. The diagonal line (in green) indicates the point of a chance prediction. ROC curve indicated in blue.

Similarly, the ROC curve for the day of surgery POC blood glucose for postoperative complications was 0.564 (95% CI: 0.426–0.701; P=0.319), which was nonsignificant.

Our secondary analysis, with ROC curves for preoperative A1C and random glucose, did not differ in chance. For A1C compared with an outcome of complications and LOS, the AUC was 0.610 (95% CI: 0.475–0.744) with a P-value of 0.092 and 0.543 (95% CI: 0.462–0.625) with a P-value of 0.277, respectively. For preoperative random glucose, the AUC for an outcome of complications was 0.570 (95% CI: 0.451–0.688) with a P-value of 0.264, and the AUC for LOS was 0.505 (95% CI: 0.423–0.587) with a P-value of 0.901.

Discussion/conclusion

To our knowledge, this is the first study to evaluate immediate preoperative capillary glucose levels and the impact on complications in a cohort of elective orthopedic surgery patients. Although the practice of such monitoring is firmly entrenched in preoperative care, little data exists to guide management. There are conflicting recommendations on how to proceed in a patient with significant hyperglycemia on the day of surgery, despite diabetes prevalence continuing to rise18,19.

Our data indicate that the random blood glucose and A1C done in the weeks before surgery are appropriate indicators of glycemia on presentation to the hospital. These are thus valid targets to avoid significant hyperglycemia on the day of surgery, to attenuate delays and inefficient use of resources.

Although there was a trend for significance, our study demonstrated no statistical relation between POC day of surgery glucose with post-TKA LOS or complications. Despite perioperative glycemic control being known to be an accurate predictor of general surgical complications, no studies have been able to reproduce this for orthopedic surgeries20,21.

Of particular interest are the minority of patients who present with significant hyperglycemia on the day of surgery. Such patients are at risk of cancellation and rebooking of their procedure with its consequent inefficiencies. We identified a trend in these patients showing a significantly higher preoperative A1C and random glucose. In our study, patients presenting with a glycemia ≥10 mmol/L on the day of surgery had an average A1C of 8±1.16% versus 6.76±0.89% in the <10 mmol/L subgroups. Similarly, patients with a glycemia ≥10 mmol/L on day of surgery had random glucose of 10.98±4.70 versus 7.41±2.29 mmol/L in standard routine preoperative bloodwork. This information could be useful for risk stratification during preoperative assessment and reinforces that physicians should be alert in patients with random glucose and A1Cs in this range, on routine testing.

Our secondary analysis also showed no established relationship with negative outcomes and preoperative A1C or random glucose. We did demonstrate that random glucose and A1C are predictors of day-of-operation glycemia. Demographic background, including age, Charlson Comorbidity Index, region (rural vs. urban setting), and sex were not significant predictors.

A minority of our patients had LOS exceedingly greater than our mean of 4.5 days. 5 patients stayed ≥10 days with the longest staying 136 postoperative days, mainly due to coexisting comorbidities. Of note, 4 of the 5 had capillary glycemia >10 mmol/L before surgery with an average age of 74.2 years. This validates the utility of preoperative assessment to achieve satisfactory glycemia.

We suspect that numerous institution-specific protocols contributed to the overall acceptable day of surgery glucose control in our population. Diabetes Canada recommends a perioperative glycemic target between 5.0 and 10.0 mmol/L for noncardiovascular surgeries in patients with known diabetes; 83.5% of our patients met their recommended range22. Standardized testing likely flags poorly controlled diabetes which encourage more aggressive preoperative glycemic control or delay surgery when necessary. In addition, our surgical wards have a postoperative subcutaneous insulin clinical order set, which may have adequately corrected any postoperative glycemia with suboptimal control before admission. Our protocols for medication adjustment before surgery also appear to effectively avoid preoperative hypoglycemia, as we had no hypoglycemia (lowest POC glycemia was 4.2 mmol/L).

Limitations of our study are the single-center design and relatively small sample size of 226, with relatively satisfactory glycemic control. It is possible a larger study including more patients with suboptimal control could demonstrate statistical significance. Of note, we did have 30 patients from out-of-province. Due to regional databases, 5 patients had no laboratory evidence in our electronic medical record, although they did have diet-controlled diabetes based on chart diagnosis.

Due to the need to input dichotomous values for our ROC stats, the outcome classification of 3 days or less versus 4 days or greater was maybe not the proper cutoff. The rationale for this particular grouping for LOS was based on previous LOS data from our institution, as the Canadian Joint Registry does not have specific data on people with diabetes. Similarly, as ROC requires dichotomous values, it likely underestimated the effect for a patient with both infection and VTE.

Finally, perhaps there truly is no relation between preoperative POC day of surgery glucose with LOS or complications in patients who underwent primary TKA operations. One large study equally found this ambiguous result for patients undergoing noncardiac surgery23. It may also be that acute hyperglycemia may be a poor predictive biomarker as it is vulnerable to rapid fluctuation.

On the basis of this, we propose a need for further large-scale research of preoperative diabetes management, specifically in orthopedics. We identified modifiable predictors (A1C and random glucose), that should be used as targets for physicians to prevent patients from presenting with out-of-range glycemia day of the procedure.

Ethical approval

Ethical considerations were approved by the Research Ethics Board of Horizon Health Network (NB, Canada) on May 27, 9 2019, under #100259-3705.

Sources of funding

None.

Author contribution

A.R.: data collection, data interpretation, manuscript writing. S.G.: study design and proposal, data interpretation, manuscript review. M.C.M.: study supervision, data interpretation, manuscript review. L.-J.C.: data interpretation, manuscript review.

Conflicts of interest disclosure

The authors declare that they have no financial conflict of interest with regard to the content of this report.

Research registration unique identifying number (UIN)

This study was registered to the Research Registry, under the number researchregistry6528.

Guarantor

Alex Robin.

Acknowledgments

A special thanks to Donaldo D. Canales, MA, and Andrew Flewelling, PhD, of HHN Research Services for statistical analysis and expertise. The authors also thank all the wonderful staff at the Moncton Hospital, notably in the diabetes clinic, medical records department, and laboratory for making this project enjoyable.

References

1. Marchant MH, Viens NA, Cook C, et al. The impact of glycemic control and diabetes mellitus on perioperative outcomes after total joint arthroplasty. J Bone Joint Surg Am 2009;91:1621–9.
2. Mraovic B, Hipszer BR, Epstein RH, et al. Preadmission hyperglycemia is an independent risk factor for in-hospital symptomatic pulmonary embolism after major orthopedic surgery. J Arthroplasty 2010;25:64–70.
3. Fisichella L, Fenga D, Rosa MA. Surgical site infection in orthopaedic surgery: Correlation between age, diabetes, smoke and surgical risk. Folia Med (Plovdiv) 2014;56:259–63.
4. Oladeji LO, Raley JA, Menendez ME, et al. Risk factors for in-hospital myocardial infarction after shoulder arthroplasty. Am J Orthop 2015;44:E142–7.
5. Kateros K, Doulgerakis C, Galanakos SP, et al. Analysis of kidney dysfunction in orthopaedic patients. BMC Nephrology 2012;13:101.
6. International Diabetes Federation. Facts & figures; 2020. Available at: https://www.idf.org/aboutdiabetes/what-is-diabetes/facts-figures.html. Accessed January 20, 2020.
7. Canadian Institute for Health Information. Inpatient hospitalization, surgery and newborn statistics, 2018–2019; 2020. Available at: https://www.cihi.ca/en/hospital-stays-in-canada. Accessed October 23, 2020.
8. Mraovic B, Suh D, Jacovides C, et al. Perioperative hyperglycemia and postoperative infection after lower limb arthroplasty. J Diab Sci Technol 2011;5:412–8.
9. Turina M, Fry DE, Polk HC. Acute hyperglycemia and the innate immune system: clinical, cellular, and molecular aspects. Crit Care Med 2005;33:1624–33.
10. Kremers HM, Lewallen LW, Mabry TM, et al. Diabetes mellitus, hyperglycemia, hemoglobin a1c and the risk of prosthetic joint infections in total hip and knee arthroplasty. J Arthroplasty 2015;30:439–443.
11. Akiboye F, Rayman G. Management of hyperglycemia and diabetes in orthopedic surgery. Curr Diab Rep 2017;17:1–11.
12. Sudhakaran S, Surani SR. Guidelines for perioperative management of the diabetic patient. Surg Res Pract 2015;2015:284063.
13. American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diab Care 2012;35(suppl 1):S64–71.
14. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83.
15. Canadian Institute for Health Information. Hip and knee replacements in Canada: CJRR annual report; 2020. Available at: https://www.cihi.ca/en/hip-and-knee-replacements-in-canada-cjrr-annual-report. Accessed September, 2020.
16. 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.
17. Rice ME, Harris GT. Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law Human Behav 2005;29:615–20.
18. Dogra P. Diabetic Perioperative Management. Treasure Island, FL: StatPearls; 2020. Available at: https://www.ncbi.nlm.nih.gov/books/NBK540965/. Accessed October 23, 2020.
19. Saklayen MG. The global epidemic of the metabolic syndrome. Curr Hypertens Rep 2018;20:12.
20. Kwon S, Thompson R, Dellinger P, et al. Importance of perioperative glycemic control in general surgery: a report from the Surgical Care and Outcomes Assessment Program. Ann Surg 2013;257:8–14.
21. Kang Z-Q, Huo J-L, Zhai X-J. Effects of perioperative tight glycemic control on postoperative outcomes: a meta-analysis. Endocr Connect 2018;7:R316–27.
22. Jeon CY, Furuya EY, Berman MF, et al. The role of pre-operative and post-operative glucose control in surgical-site infections and mortality. PLoS One 2012;7:e45616.
23. Abdelmalak BB, Knittel J, Abdelmalak JB, et al. Preoperative blood glucose concentrations and postoperative outcomes after elective non-cardiac surgery: an observational study. Br J Anaesth 2014;112:79–88.
Keywords:

Diabetes; Perioperative; Orthopedics

Copyright © 2021 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of IJS Publishing Group Ltd.