Abstract
Purpose
This study aimed to evaluate effects of perioperative dexamethasone on hospital length of stay (LOS) and glycemic control for patients with type 2 diabetes mellitus undergoing total hip arthroplasty (THA).
Methods
We performed retrospective case review of THA performed in adults (≥18 years old) with type 2 diabetes at Springfield Memorial Hospital (Springfield, IL) immediately before (2013), during (2014), and after (2015) publication of consensus guidelines for use of perioperative dexamethasone. Hospital LOS was the primary endpoint. Capillary blood glucose by hospital day, proportion of patients treated with insulin, and median insulin dose by hospital day were secondary endpoints.
Results
A total of 209 patients were included: 109 not dosed with dexamethasone (“no dexamethasone”), and 100 treated with perioperative dexamethasone. The most common dose of dexamethasone was 4 mg (63% of patients). Mean (95% CI) reduction in adjusted hospital LOS for dexamethasone-treated patients, compared to controls, was −2.8 (−3.7 to −1.9) days for all patients, −1.6 (−2.7 to −0.5) days for those with arthritis as the indication for THA, and −4.0 (−5.9 to −2.1) days for those with fracture as indication for THA (P<0.001 for all). Glycemic control measured by median capillary blood glucose was no different or slightly better in the dexamethasone group than the no dexamethasone group, except for postoperative day 1 among patients treated with insulin prior to surgery.
Conclusions
Perioperative dexamethasone significantly reduces hospital LOS for patients with type 2 diabetes undergoing THA, with modest effects on hyperglycemia.
Keywords: type 2 diabetes mellitus, hip arthroplasty, orthopedic surgery, glycemic control, dexamethasone, perioperative management
Total hip arthroplasty (THA) is one of the most frequently performed surgeries in the United States, with 450,000–500,000 procedures annually and more than 800,000 predicted by the year 2030.1,2 The prevalence of diabetes mellitus (DM) among patients undergoing THA is significant and increasing. In the Function and Outcomes Research for Comparative Effectiveness in Total Joint Replacement (FORCE-TJR) study, nearly 12% of patients undergoing THA in the United States had preexisting DM,3 and the proportion of Medicare beneficiaries undergoing THA with known DM doubled from 7.1% to 15.5% in the interval 1991–2008.4 The high and growing prevalence of DM makes it an important comorbidity in the management of THA patients.
Significant nausea occurs in more than 20% of patients and vomiting in approximately 15% of patients following joint arthroplasty,5 and nearly 90% of patients receive opioids for pain management following THA.6 However, since publication of consensus guidelines in 2014,7 high-dose preoperative dexamethasone, a potent synthetic glucocorticoid, has been increasingly used to mitigate postoperative nausea, vomiting and severe pain requiring opioid analgesics. Preoperative dexamethasone has been demonstrated to significantly improve nausea and vomiting, reduce dosing of narcotics, and shorten hospital length of stay (LOS), though most studies have focused on nondiabetic patients.8–13
Glucocorticoids exacerbate hyperglycemia in patients with DM chiefly by increasing insulin resistance,14 and worsening glycemic control after THA might offset some or all of the benefits of preoperative dexamethasone for diabetic patients. Studies in the peer-reviewed literature are mixed, with some showing no significant effect of dexamethasone on glycemic control9,15 while others demonstrate hyperglycemia is more difficult to manage intraoperatively and measurably higher after surgery.16,17
A recent large, single-center study of patients undergoing orthopedic procedures that included a relatively large proportion of individuals with DM (nearly 25%) showed reduced hospital LOS for all subjects (ie, with or without DM) who had received perioperative dexamethasone, with measurable increases in mean peak glucose limited to the first 24 hours of hospital admission.18 However, this study did not assess how factors such as prehospital glycemic control or outpatient diabetes management impact risk of hyperglycemia after dexamethasone, and the study’s time frame (January 2000–August 2016) potentially meant that changes in surgical approaches and prostheses could have had an impact on clinical outcomes.19–21 Herein, we examine the impact of perioperative dexamethasone on glycemic control and hospital LOS for patients with type 2 DM undergoing THA admitted to a single center during 2013–2015, a period that includes the year before, year of, and year following publication of consensus guidelines for use of perioperative glucocorticoids.7
METHODS
Study Design and Patient Selection
This retrospective study utilized the electronic medical record (EMR) of Springfield Memorial Hospital (Springfield, IL) to identify patients with type 2 DM who underwent THA within the time frame of January 2013–December 2015 and compare outcomes of those who received perioperative dexamethasone with those who did not (ie, the control group). THA patients were chosen for evaluation because THA is the second most common surgery at Springfield Memorial after cardiothoracic procedures, providing a sizable population of patients with diabetes undergoing a single surgery. Comparison of the first year after routine use of perioperative dexamethasone to the years before implementation of these consensus guidelines was conducted to minimize the potential effects of changes in surgical approach and hardware on patient outcomes. The study was approved by the Springfield Committee for Research Involving Human Subjects, the local institutional review board for Memorial Health, and the Southern Illinois University School of Medicine.
Patients ≥18 years of age with coincident International Classification of Diseases, Ninth Revision (ICD-9) codes for THA (V43.64) and type 2 DM (250.x0 and 250.x2) during the same hospital admission were eligible for inclusion in the study; both diagnoses were confirmed by manual chart review. Patients were excluded if they had a diagnosis of type 1 DM, were treated with any glucocorticoid before surgery, continued treatment with glucocorticoids in hospital after pulse-dose dexamethasone at start of surgery, or were managed by Southern Illinois University’s hospital diabetes team. Diabetes team-managed patients were excluded due to the small number (n=10) during the study time frame and to make results more generalizable, since most patients with type 2 DM undergoing THA at Springfield Memorial are managed by either surgical or general medicine teams.
Charts were manually reviewed to collect patient demographic data, including age, sex, weight, key chronic health conditions (coronary heart disease, hypertension, and stage 2 or higher chronic kidney disease [CKD]), mode of preoperative diabetes management, hemoglobin A1c (HbA1c) during or immediately prior to hospital admission, and indication for THA. Weight was recorded as a surrogate for obesity defined by body mass index because height was not routinely reported. Increasing age, female sex, obesity, coronary heart disease, and CKD have been identified as risk factors for prolonged hospital LOS following THA, and hypertension is the most common comorbidity among patients undergoing THA.22,23 Clinical data collected by manual chart review included dose of dexamethasone administered if received, capillary blood glucose (CBG) measurements, postsurgical insulin dose requirement, and hospital LOS.
A flowchart showing selection of patients is presented in Figure 1.
Figure 1.
Flowchart of patient selection. DM2, type 2 diabetes mellitus; THA, total hip arthroplasty.
Statistical Analysis
The primary endpoint was hospital LOS. Multiple measures of glycemic control were evaluated as secondary endpoints, including mean and median CBG for the entire hospital admission and by hospital day, proportion of patients dosed with insulin by hospital day, and total insulin dose administered by hospital day.
Based on LOS (3 ± 1 days) for hip arthroplasty patients in the evaluated time frame of 2013–2015, a sample size of 152 patients (76 in each group) was required to detect a 0.5-day difference in LOS between groups at a β of 0.8 and α of 0.05.18 Ultimately, our study sample exceeded this minimum, as we analyzed data from all eligible patients to increase power and reduce risk for error. A 152-patient sample also was deemed adequate to detect a significant difference of 20 mg/dL in mean CBG between groups; standard practice at Springfield Memorial is to check CBG levels in patients with DM at meals, bedtime, and 0200.
Continuous variables were expressed as mean ± standard deviation if normally distributed or median with interquartile range (IQR) if skewed and compared with parametric or nonparametric t-tests as appropriate. Categorical variables were expressed as proportion and compared using Fisher’s exact test or chi-squared test as indicated. Difference in hospital LOS between the dexamethasone-treated and untreated groups was further evaluated by linear mixed-model regression to control for other factors that might affect the outcome, including year of surgery, age, sex, weight, coronary heart disease, hypertension, CKD, mode of outpatient DM management, HbA1c, and THA indication. Two-way analysis of variance (ANOVA) was performed to evaluate the effects of treatment (dexamethasone vs no dexamethasone) and time (hospital day) on glycemic control. CBG measurements were log-transformed for the ANOVA because the raw data were not normally distributed. Statistical analyses were performed using Prism 8 (GraphPad Software), ProStat 5.5 (Poly Software International, Inc.), and Stata 17 (StataCorp LLC) software.
RESULTS
A total of 209 patients were included in the study. There were 109 patients who were not dosed with dexamethasone (control group) and 100 patients who received perioperative dexamethasone (dexamethasone group). All patients underwent general anesthesia and received a single dose of parenteral dexamethasone shortly before anesthesia induction. The majority (63%) of dexamethasone group patients received 4 mg of dexamethasone; mean dexamethasone dose was 5.8 ± 2.5 mg (median: 4, range: 4–12).
Baseline patient characteristics are presented in Table 1. Both groups were well matched for age, weight, sex distribution, preoperative HbA1c, mode of preoperative diabetes management, and coronary heart disease. Patients in the dexamethasone group were more likely to have stage 2 or higher CKD than control patients. A very high proportion of patients in both groups were diagnosed with hypertension, though prevalence was higher in the control group. Perioperative dexamethasone was more likely to be dosed for patients with arthritis or an unspecified indication for THA rather than fracture. As expected, most control group patients underwent THA in 2013 and 2014, while dexamethasone group patients had almost all their surgeries in 2014 and 2015.
Table 1.
Baseline Patient Characteristics
| Parameter | Control group (n=109) | Dexamethasone group (n=100) | P |
|---|---|---|---|
| Median age (IQR) | 73 (65–80) years | 69 (60–79) years | 0.07 |
| Mean weight ± SD | 93.4 ± 22.6 kg | 96.0 ± 23.0 kg | 0.51 |
| Median hemoglobin A1c (IQR) | 6.9% (6.4%–7.3%) | 6.8% (6.3%–7.2%) | 0.46 |
| Female sex | 61.5% | 55.0% | 0.40 |
| Coronary heart disease | 43.1% | 37.0% | 0.40 |
| CKD stage 2 or higher | 37.6% | 52.0% | 0.04 |
| Hypertension | 92.7% | 83.0% | 0.03 |
| Diabetes treatment | 0.27 | ||
| Insulin | 20.2% | 14.1% | |
| Oral medication | 58.7% | 65.7% | |
| Diet | 21.1% | 20.2% | |
| Indication for hip arthroplasty | <0.01 | ||
| Fracture | 37.9% | 14.0% | |
| Arthritis or unspecified | 65.1% | 86.0% | |
| Year of surgery | <0.01 | ||
| 2013 | 65.1% | 1.0% | |
| 2014 | 26.6% | 42.0% | |
| 2015 | 8.3% | 57.0% |
CKD, chronic kidney disease; IQR, interquartile range; SD, standard deviation.
Unadjusted hospital LOS for patients treated with dexamethasone was significantly shorter than for control patients (2.8 ± 1.7 days vs 4.6 ± 3.1 days; P<0.001), and the difference in LOS remained significant after adjusting for relevant covariates by linear mixed-model regression (Table 2). Adjusted LOS was 2.8 days shorter for dexamethasone-treated patients than control patients, and LOS remained shorter with dexamethasone treatment regardless of indication for THA. The magnitude of improvement in adjusted LOS was greater for patients with fracture as the indication for THA (−4.0 days) than patients with arthritis as the indication for surgery (−1.6 days). There was no difference in LOS by year of surgery for all patients, regardless of diabetes status, undergoing THA at Springfield Memorial in 2013–2015 (Figure 2), mitigating against an effect of year of surgery on the LOS result. Adjusted mean LOS for the control (n=109) and dexamethasone (n=100) groups were 5.1 (95% CI: 4.6–5.6) and 2.3 (95% CI: 1.7–2.8) days, respectively. Unadjusted mean LOS for all THA patients by year of surgery was 3.1 days (95% CI: 2.9–3.3) in 2013 (n=449), 3.2 days (95% CI: 2.9–3.5) in 2014 (n=480), and 2.9 days (95% CI: 2.7–3.1) in 2015 (n=416).
Table 2.
Adjusted Difference in Hospital Length of Stay* by Indication for Total Hip Arthroplasty
| Indication | Patient group | Regression coefficient for length of stay | 95% CI | P |
|---|---|---|---|---|
| All | Control | Reference | ||
| Dexamethasone | −2.8 days | −3.7 to −1.9 days | <0.001 | |
| Arthritis | Control | Reference | ||
| Dexamethasone | −1.6 days | −2.7 to −0.5 days | <0.001 | |
| Fracture | Control | Reference | ||
| Dexamethasone | −4.0 days | −5.9 to −2.1 days | <0.001 |
Adjusted from linear mixed-model regression with covariates including year of surgery, age, sex, weight, coronary heart disease, chronic kidney disease, hypertension, hemoglobin A1c, and mode of outpatient diabetes management.
Figure 2.
Adjusted mean hospital length of stay (LOS) and 95% CI for patients with type 2 diabetes mellitus (DM2) in the control and dexamethasone (Dex) groups, and unadjusted mean LOS for all patients undergoing total hip arthroplasty (THA) at Springfield Memorial Hospital in 2013–2015. Adjusted mean hospital LOS for the control (n=109) and Dex (n=100) groups was 5.1 days (4.6–5.6) and 2.3 days (1.7–2.8), respectively. Unadjusted mean LOS for all THA patients by year of surgery was 3.1 days (2.9–3.3) in 2013 (n=449), 3.2 days (2.9–3.5) in 2014 (n=480), and 2.9 days (2.7–3.1) in 2015 (n=416). Adjusted hospital LOS was determined by linear mixed-model regression, with covariates including year of surgery, age, sex, weight, coronary heart disease, chronic kidney disease, hypertension, hemoglobin A1c, mode of outpatient diabetes management, and indication for THA. *P<0.001 for the comparison between the control and Dex groups.
The impact of perioperative dexamethasone on glycemic control during hospitalization is presented in Table 3, and the effect of dexamethasone by mode of diabetes treatment before THA is presented in Figure 3. Overall, median CBG for patients in the no dexamethasone group was actually higher than for dexamethasone-treated patients (160 mg/dL [IQR: 130–202] based on a total of 1409 recorded measurements vs 152 mg/dL [IQR: 123–192] based on 1046 measurements; P<0.001). There were no differences in glycemic control on admission or postoperative day (POD) 1, when adverse effects of dexamethasone on blood glucose would be anticipated due to its approximately 24-hour duration of activity. Dexamethasone only raised blood glucose significantly in the subset of patients treated with insulin prior to hospital admission, and the effect was limited to POD 1 (median: 223 mg/dL [IQR: 158–273] based on 55 CBG measurements vs median: 184 mg/dL [IQR: 138–218, n=112] based on 112 CBG measurements; P=0.002). Median CBG for insulin-treated patients receiving dexamethasone was actually lower than controls on POD 2 and 3, perhaps due to lower numbers of dexamethasone-treated patients in hospital on these days.
Table 3.
Median CBG by Hospital Day for All Patients in Each Treatment Group
| Hospital day | No dexamethasone group, median CBG (IQR); N | Dexamethasone group, median CBG (IQR); N | P |
|---|---|---|---|
| Admission | 152 (120–202) mg/dL; 347 | 147 (119–191) mg/dL; 274 | 0.39 |
| POD 1 | 161 (134–201) mg/dL; 452 | 162 (130–215) mg/dL; 339 | 0.61 |
| POD 2 | 168 (139–208) mg/dL; 362 | 153 (131–185) mg/dL; 265 | <0.001 |
| POD 3 | 154 (125–199) mg/dL; 253 | 136 (113–170) mg/dL; 168 | <0.001 |
CBG, capillary blood glucose; IQR, interquartile range, N, number of CBG measurements; POD, postoperative day.
Figure 3.
Median (interquartile range [IQR]) capillary blood glucose (CBG) by hospital day for controls (purple symbols and lines) and dexamethasone-treated patients (pink symbols and lines). A: Median CBG by hospital day for patients with type 2 diabetes managed with insulin prior to total hip arthroplasty, with comparisons by nonparametric t-test. Median CBG on hospital day 1 was significantly higher for insulin-treated patients receiving dexamethasone than insulin-treated controls (223 mg/dL [IQR: 158–273] per 55 total measurements] vs 184 mg/dL [IQR: 138–218] per 112 total measurements; P=0.002). B: Median CBG by hospital day for patients managed with diet or oral hypoglycemic agents (OHA), with comparisons by nonparametric t-test. Median CBG on hospital day 2 was slightly higher in the controls than the dexamethasone-treatment group (164 mg/dL [IQR: 135–200] per 265 total measurements vs 153 mg/dL [IQR: 131–184] per 226 total measurements; P=0.016).
For patients managed by lifestyle changes (eg, diet and exercise) or with non-insulin therapeutics, median CBG remained fairly stable in both the control and dexamethasone groups from surgery through POD 3 and below 180 mg/dL on all hospital days. Median CBG on POD 2 was slightly higher in the control than the dexamethasone treatment group (median: 164 mg/dL [IQR: 135–200] based on 265 measurements vs median: 153 mg/dL [IQR: 131–184] based on 226 measurements; P=0.016). Two-way ANOVA revealed a significant interaction between the effects of treatment group (dexamethasone or no dexamethasone) and hospital day on glycemic control (F[3,7]=6.263; P<0.0001), with simple main effects analysis showing a significant effect of hospital day on blood glucose (P<0.0001). There was no significant effect of treatment group on glycemic control in simple main effects analysis (P=0.09).
Insulin dose requirements were similar for the dexamethasone and control groups. There were no significant differences in the proportion of patients receiving at least 1 dose of insulin from day of surgery through POD 3. The dexamethasone group did not receive more insulin than control patients, and controls actually received a modestly higher median dose of insulin on POD 3 that was statistically but not clinically significant (control group: 12.0 units [IQR: 6.0–28.5]; dexamethasone group: 10.0 units [IQR: 4.0–22.3); P=0.04). Almost all patients were managed only with correction factor (“sliding scale”) insulin while in hospital.
DISCUSSION
To the best of our knowledge, this is the first study to evaluate the effect of perioperative dexamethasone on hospital LOS in a cohort consisting entirely of patients with type 2 DM undergoing THA. Perioperative dexamethasone significantly reduced hospital LOS for these patients regardless of the indication for THA. The impact of dexamethasone on glycemic control was minor and limited to higher blood sugars among patients treated with insulin prior to surgery on POD 1 only. Our results indicate that perioperative dexamethasone is of clinical benefit to patients with type 2 DM undergoing THA and that dexamethasone should not be withheld due to concerns about adverse impact on blood glucose levels after surgery.
Though older studies evaluating the effects of methylprednisolone11 or prednisone and hydrocortisone24 failed to show an impact of perioperative glucocorticoids on hospital LOS, recent randomized controlled trials of dexamethasone in the management of nondiabetic total joint arthroplasty patients have demonstrated reduced LOS. Backes et al evaluated the effects of 1 10-mg dose of dexamethasone before surgery or 2 10-mg doses of dexamethasone (1 dose before, the other after surgery) for patients undergoing THA or total knee arthroplasty and found reduced LOS for both dexamethasone groups and greater improvement in LOS for the two-dose dexamethasone regimen.13 Dissanayake et al compared 2 8-mg dexamethasone doses to placebo in patients undergoing either THA or total knee arthroplasty and documented a 1-day improvement in median LOS for THA patients that did not quite reach statistical significance; there was no change in LOS for total knee arthroplasty patients.25 In a study including only THA patients, Lei et al compared 2 or 3 doses of 10 mg dexamethasone to placebo and demonstrated a dose-related improvement in hospital LOS.12 Though there are few studies including patients with diabetes, the recent retrospective study by Herbst et al included 1187 patients with diabetes undergoing total joint arthroplasty and found comparable reductions in median hospital LOS for patients with or without diabetes.18
Our study confirmed and extended Herbst and colleagues’ recent finding that perioperative dexamethasone reduces hospital LOS for patient with diabetes.18 We evaluated a well-circumscribed time interval 1 year before, the year of, and 1 year after publication of guidelines on use of perioperative dexamethasone to limit the likelihood that changes in THA technique or joint hardware would affect the results. The indication for THA was known for all patients, and though the magnitude of reduction in LOS was larger for patients with osteoarthritis than those with hip fractures, both groups of patients experienced significant reductions in LOS when treated with perioperative dexamethasone. Moreover, the dexamethasone regimen was known for all steroid-treated patients, and our study results indicated that LOS is reduced significantly by a single perioperative dose of dexamethasone in patients with type 2 DM. While the improvement in adjusted hospital LOS for this study (2.8 days) is the largest reported to date for patients with or without diabetes, this outcome conforms to the results of recent studies that the use of dexamethasone in management of patients undergoing THA significantly reduces hospital LOS.12,13,18
The effects of perioperative dexamethasone on glycemic control were modest in this study and limited to POD 1 for patients treated with insulin prior to THA. Median CBG was actually better on POD 2 and POD 3 for insulin-requiring patients who received dexamethasone than those who did not and on POD 2 for patients receiving dexamethasone and managed with diet or non-insulin therapeutics prior to surgery. This pattern was similar to the glycemic control findings by Herbst et al.18 Results of 2-way ANOVA in this study indicated that hospital day, but not treatment with dexamethasone, was the factor that had a significant effect on glycemic control. In a study of more than 2000 patients undergoing total joint arthroplasty at Oschner Clinic from 2011 to 2015,15 657 had a preoperative diagnosis of diabetes, and there was no difference in mean blood glucose level regardless of dexamethasone treatment status on any POD. Insulin dose requirements in our study also were unaffected by treatment with dexamethasone. Though the findings conflict with expectations due to the known adverse effects of glucocorticoids on hyperglycemia,14 the results of this study and other recent, larger studies failed to show a significant or persistent effect of perioperative dexamethasone on glycemic control other than on POD 1 and, in our study, only patients managed with insulin prior to surgery. The typical dose of dexamethasone (4 mg) in this study was significantly lower than most others, and this could account for the limited effect of dexamethasone on postsurgical blood sugars.
As noted, a strength of this study was its narrow time frame (3 years) spanning immediately before, during, and after publication of guidelines regarding use of perioperative glucocorticoids, thereby limiting changes in THA care. There was no change in mean LOS for all patients undergoing THA at Springfield Memorial by year during the study, indicating that improvement in LOS for the dexamethasone group reflects a benefit of treatment with dexamethasone rather than an unrelated change in surgical or postsurgical management. Study patients were a well-defined cohort of individuals with type 2 DM, and knowledge of patients’ treatment regimens allowed the effects of perioperative dexamethasone on hospital hyperglycemia to be evaluated as a function of ambulatory diabetes management. Importantly, preoperative glycemic control in both the control and dexamethasone groups was representative of typical patients undergoing joint arthroplasty; for example, in a previous multicenter study of 1645 patients with type 2 DM undergoing either total knee arthroplasty (n=1004) or THA (n=641), mean preoperative HbA1c was 6.6% and nearly two-thirds of patients had a preoperative HbA1c below 7%.26 Finally, we evaluated the effects of dexamethasone on total daily insulin dose by hospital day in addition to daily mean CBG. Surprisingly, mean daily insulin dose was not increased for dexamethasone-treated patients on either the day of surgery or POD 1, when dexamethasone activity would be anticipated to affect blood glucose levels.
Limitations
Results should be considered within the context of several study limitations. Although care was taken to include all variables that might significantly influence the results, data were collected retrospectively and may be affected by unmeasured confounding factors. Both patient groups were generally well matched, but some clinical characteristics (eg, prevalence of hypertension and stage 2 or higher CKD) differed, and the proportion of patients undergoing THA for hip fracture was higher in the control group. Guidelines for perioperative glucocorticoids are silent on use in patients with hypertension and CKD, and the changes in prevalence of hypertension and CKD were in opposite directions, making it unlikely these characteristics had a coordinated effect on patient outcomes. All baseline patient characteristics, including hypertension and CKD, were included in the linear mixed-effects model regression to determine adjusted hospital LOS for controls and dexamethasone-treated patients, accounting for possible effects of baseline characteristics on hospital LOS. It does not appear that the differences in prevalence of hypertension and CKD between the control and dexamethasone groups were due to practice-based selection bias, and the large difference in adjusted hospital LOS between groups makes it unlikely these patient characteristics significantly affected LOS. Although arthritis was a more frequent indication for THA among patients treated with dexamethasone, there was a significant reduction in hospital LOS regardless of indication for THA. The magnitude of LOS improvement was smaller for patients with arthritis as the indication for THA than fracture, a result that would be anticipated to actually reduce the overall improvement in LOS for the dexamethasone group. Weight was used as a proxy measure for obesity defined by body mass index; however, this may have been suboptimal to control for the effects of obesity on hospital LOS.
The patient population for our study was much smaller than other recent retrospective studies,15,18 but the smaller study population allowed for consideration of factors (eg, THA indication, mode of preoperative diabetes management) that might be impractical to determine in a study with thousands of patients. Also, our hospital LOS and glycemic control results are generally consistent with those reported by the larger studies. The timing and amount of dexamethasone in this study is considerably different than recent randomized controlled trials, and our results may not be relevant to patients with type 2 DM treated with higher or more frequent doses of dexamethasone. Most patients in this study had well-controlled diabetes as suggested by a mean HbA1c of <7% in each treatment group, and while this is consistent with preoperative glycemic control reported in other joint arthroplasty studies, the effects of perioperative dexamethasone on hospital LOS and postoperative glucose levels in this particular sample may not extend to patients with uncontrolled hyperglycemia prior to surgery.
Lastly, the impact of perioperative dexamethasone on postoperative pain or nausea was not measured, and the study design’s retrospective nature does not allow a determination of causality for results.
CONCLUSIONS
This study demonstrated that perioperative dexamethasone is associated with a significant reduction in hospital length of stay for patients with type 2 diabetes undergoing total hip arthroplasty for treatment of osteoarthritis or fracture. The adverse effect of dexamethasone on hospital glycemic control is limited to patients treated with insulin prior to surgery and only on postoperative day 1. Insulin-treated patients may benefit from endocrinology or diabetes team management of hyperglycemia in the hospital. Our findings indicate that patients with type 2 diabetes and favorable preoperative glycemic control benefit from perioperative dexamethasone, like their nondiabetic peers, and that the effects of a single perioperative dose of dexamethasone on postsurgical glycemic control are modest.
Patient-Friendly Recap
Dexamethasone can be administered during surgical procedures to relieve subsequent pain and nausea, but this anti-inflammatory steroid also increases insulin resistance.
Authors studied how use of dexamethasone during a total hip replacement affected surgery outcomes and glycemic levels in patients with type 2 diabetes.
The patients who received dexamethasone had a significantly shorter hospital stay, on average, while maintaining similar glycemic control as those not receiving this steroid during surgery.
While patients with diabetes can safely receive perioperative dexamethasone – and thus experience less vomiting – those taking insulin may benefit from having their postoperative care managed by diabetes specialists.
Acknowledgments
We thank the staff of the Health Informatics Office of Springfield Memorial Hospital for their assistance with data collection.
Footnotes
Author Contributions
Study design: Williams, Ansari, Jakoby. Data acquisition or analysis: all authors, Manuscript drafting: Williams, Jakoby. Critical revision: Botchway, Jakoby.
Conflicts of Interest
None.
References
- 1.Sloan M, Premkuma A, Sheth NP. Projected volume of primary total joint arthroplasty in the U.S., 2014 to 2030. J Bone Joint Surg Am. 2018;100:1455–60. doi: 10.2106/JBJS.17.01617. [DOI] [PubMed] [Google Scholar]
- 2.Singh JA, Yu S, Chen L, Cleveland JD. Rates of total joint replacement in the United States: future projections to 2020–2040 using the National Inpatient Sample. J Rheumatol. 2019;46:1134–40. doi: 10.3899/jrheum.170990. [DOI] [PubMed] [Google Scholar]
- 3.Franklin PD, Miozzari H, Christofilopoulos P, Hoffmeyer P, Ayers DC, Lübbeke A. Important patient characteristics differ prior to total knee arthroplasty and total hip arthroplasty between Switzerland and the United States. BMC Musculoskelet Disord. 2017;18(1):14. doi: 10.1186/s12891-016-1372-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cram P, Lu X, Kaboli PJ, et al. Clinical characteristics and outcomes of Medicare patients undergoing total hip arthroplasty, 1991–2008. JAMA. 2011;305:1560–7. doi: 10.1001/jama.2011.478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wang Y, Yang Q, Lin J, et al. Risk factors of postoperative nausea and vomiting after total hip arthroplasty or total knee arthroplasty: a retrospective study. Ann Transl Med. 2020;8(17):1088. doi: 10.21037/atm-20-5486. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Shah R, Kuo YF, Westra J, Lin YL, Raji MA. Opioid use and pain control after total hip and knee arthroplasty in the US, 2014 to 2017. JAMA Netw Open. 2020;3(7):e2011972. doi: 10.1001/jamanetworkopen.2020.11972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gan TJ, Diemunsch P, Habib AS, et al. Consensus guidelines for the management of postoperative nausea and vomiting. Anesth Analg. 2014;118:85–113. doi: 10.1213/ANE.0000000000000002. [DOI] [PubMed] [Google Scholar]
- 8.Yue C, Wei R, Liu Y. Perioperative systemic steroid for rapid recovery in total knee and hip arthroplasty: a systematic review and meta-analysis of randomized trials. J Orthop Surg Res. 2017;12(1):100. doi: 10.1186/s13018-017-0601-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Nurok M, Cheng J, Romeo GR, Vecino SM, Fields KG, YaDeau JT. Dexamethasone and perioperative blood glucose in patients undergoing total joint arthroplasty: a retrospective study. J Clin Anesth. 2017;37:116–22. doi: 10.1016/j.jclinane.2016.11.012. [DOI] [PubMed] [Google Scholar]
- 10.Meng J, Li L. The efficiency and safety of dexamethasone for pain control in total joint arthroplasty: a meta-analysis of randomized controlled trials. Medicine (Baltimore) 2017;96(24):e7126. doi: 10.1097/MD.0000000000007126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lunn TH, Kehlet H. Perioperative glucocorticoids in hip and knee surgery – benefit vs. harm? A review of randomized clinical trials. Acta Anaestheiol Scand. 2013;57:823–34. doi: 10.1111/aas.12115. [DOI] [PubMed] [Google Scholar]
- 12.Lei Y, Huang Q, Xu B, Zhang S, Cao G, Pei F. Multiple low-dose dexamethasone further improves clinical outcomes following total hip arthroplasty. J Arthroplasty. 2018;33:1426–31. doi: 10.1016/j.arth.2017.11.057. [DOI] [PubMed] [Google Scholar]
- 13.Backes JR, Bentley JC, Politi JR, Chambers BT. Dexamethasone reduces length of hospitalization and improves postoperative pain and nausea after total joint arthroplasty: a prospective, randomized controlled trial. J Arthroplasty. 2013;28(8 Suppl):11–7. doi: 10.1016/j.arth.2013.05.041. [DOI] [PubMed] [Google Scholar]
- 14.Suh S, Park MK. Glucocorticoid-induced diabetes mellitus: an important but overlooked problem. Endocrinol Metab (Seoul) 2017;32:180–9. doi: 10.3803/EnM.2017.32.2.180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Godshaw BM, Mehl AE, Shaffer JG, Meyer MS, Thomas LC, Chimento GF. The effects of peri-operative dexamethasone on patients undergoing total hip or knee arthroplasty: Is it safe for diabetics? J Arthroplasty. 2019;34:645–9. doi: 10.1016/j.arth.2018.12.014. [DOI] [PubMed] [Google Scholar]
- 16.Wasfie T, Tabatabai A, Hedni R, Hollern L, Barber K, Shapiro B. Effect of intra-operative single dose of dexamethasone for control of post-operative nausea and vomiting on the control of glucose levels in diabetic patients. Am J Surg. 2018;215:488–90. doi: 10.1016/j.amjsurg.2017.11.019. [DOI] [PubMed] [Google Scholar]
- 17.Tien M, Gan TJ, Dhakal I, et al. The effect of anti-emetic doses of dexamethasone on postoperative blood glucose levels in non-diabetic and diabetic patients: a prospective randomised controlled study. Anaesthesia. 2016;71:1037–43. doi: 10.1111/anae.13544. [DOI] [PubMed] [Google Scholar]
- 18.Herbst RA, Telford OT, Hunting J, et al. The effects of perioperative dexamethasone on glycemic control and postoperative outcomes. Endocr Pract. 2020;26:218–25. doi: 10.4158/EP-2019-0252. [DOI] [PubMed] [Google Scholar]
- 19.Shon WY, Park BY, Rajsankar NR, Park PS, Im JT, Yun HH. Total hip arthroplasty: past, present, and future. What has been achieved? Hip Pelvis. 2019;31:179–89. doi: 10.5371/hp.2019.31.4.179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Zagra L. Advances in hip arthroplasty surgery: What is justified? EFORT Open Rev. 2017;2:171–8. doi: 10.1302/2058-5241.2.170008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Lewis PL, Graves SE, de Steiger RN, et al. Does knee prosthesis survivorship improve when implant designs change? Findings from the Australian Orthopedic Association National Joint Replacement Registry. Clin Orthop Relat Res. 2020;478:1156–72. doi: 10.1097/CORR.0000000000001229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sibia US, MacDonald JH, King PJ. Predictors of hospital length of stay in an enhanced recovery after surgery program for primary total hip arthroplasty. J Arthroplasty. 2016;31:2119–23. doi: 10.1016/j.arth.2016.02.060. [DOI] [PubMed] [Google Scholar]
- 23.Papalia R, Zampogna B, Torre G, et al. Preoperative and perioperative predictors of length of hospital stay after primary total hip arthroplasty – our experience on 743 cases. J Clin Med. 2021;10(21):5053. doi: 10.3390/jcm10215053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sculco PK, McLawhown AS, Desai N, Su EP, Padgett DE, Jules-Elysee K. The effect of perioperative corticosteroids in total hip arthroplasty: a prospective double-blind placebo-controlled pilot study. J Arthroplasty. 2016;31:1208–12. doi: 10.1016/j.arth.2015.11.011. [DOI] [PubMed] [Google Scholar]
- 25.Dissanayake R, Du HN, Robertson IK, Ogden K, Wiltshire K, Mulford JS. Does dexamethasone reduce hospital readiness for discharge, pain, nausea, and early patient satisfaction in hip and knee arthroplasty? A randomized, controlled trial. J Arthroplasty. 2018;33:3429–36. doi: 10.1016/j.arth.2018.07.013. [DOI] [PubMed] [Google Scholar]
- 26.Tarabichi M, Shohat N, Kheir MM, et al. Determining the threshold for HbA1c as a predictor for adverse outcomes after total joint arthroplasty: a multicenter, retrospective study. J Arthroplasty. 2017;32:S263–S267e1. doi: 10.1016/j.arth.2017.04.065. [DOI] [PubMed] [Google Scholar]