The annual number of patients undergoing total hip arthroplasty (THA) and total knee arthroplasty (TKA) has been steadily increasing throughout the past decade, with a concomitant increase in number of comorbidities.1
One of the reasons for the increase in comorbidity is the increasing prevalence of diabetes,2 likely due to increased age and prevalence of obesity in the general population and increased lifespan among diabetics in most Western countries.3
Diabetes is a generally acknowledged risk factor for postoperative mortality and morbidity in many types of surgeries.4,5 This also applies to major joint arthroplasty,6–8 although a few recent studies have been unable to demonstrate such association.9,10
Furthermore, pre- and perioperative hyperglycemia are independent predictors of postoperative morbidity after noncardiac surgery5 and specifically after major joint arthroplasty.6,11,12 This finding has led to suggestions of aggressive postoperative glycemic control,13 despite concerns about the increasing risk of hypoglycemia.14 Consequently, specific considerations regarding the perioperative continuum of care in the diabetic patient14,15 and increased focus on perioperative blood glucose management are recommended.16,17
No previous studies of THA and TKA in diabetics have been done in a standardized fast-track setting with optimized perioperative care, including spinal anesthesia, opioid-sparing multimodal analgesia, early mobilization, and discharge to home, which improved outcome after surgery.18 Because the fast-track approach has also been shown to reduce postoperative insulin resistance,19 there is a need to clarify the role of preoperative diabetes in fast-track surgery20 and to elucidate which types of postoperative morbidities occur in these potential high-risk patients.21
Using prospectively collected data, we investigated the association between diabetes and length of hospital stay (LOS) > 4 days, 30- and 90-day readmissions, and mortality in 8055 unselected patients undergoing primary elective unilateral THA and TKA with a standardized fast-track approach. We also studied whether there was an association between the intensity of antihyperglycemic treatment (insulin, oral antihyperglycemics, and diet treatment) and LOS > 4 days and readmissions. Finally, we estimated the number of surgical type 2 diabetics needed for 1 additional case of LOS > 4 days and readmissions (the adjusted number needed to harm [NNH]),22,23 thereby providing a more clinically oriented measure of the specific influence of diabetes on postoperative outcomes after fast-track THA and TKA.
METHODS
All data were collected from 7 departments between February 1, 2010, and November 30, 2012. No approval was needed from the Regional Ethics Committee because this was an analytic observational cohort study using existing prospectively collected data. Permission was acquired from the Danish Data Protection Agency and the Danish National Board of Health to review and store deidentified medical records without informed authorization of release of protected health information.
Preoperative Data
From February 1, 2010, all patients undergoing elective primary THA or TKA at hospitals participating in the Lundbeck Foundation Centre for Hip and Knee Replacement Collaboration have been completing a simple preoperative questionnaire on various patient characteristics. These include body mass index (BMI), living alone/with others, preoperative use of walking aids (yes/no), and known comorbidity such as type 1 and type 2 diabetes, pharmacologically treated cardiac disease, hypercholesterolemia, and hypertension. Staff is available for completing the questionnaire, and completeness of data is >95% of all procedures performed at the participating departments.24 Data are stored in the Lundbeck Foundation Centre database, which is registered as an ongoing study registry on preoperative risk factors on ClinicalTrials.gov (ID: NCT01515670).
For this study, further information on specific pharmacologic antihyperglycemic treatment 6 months before surgery was acquired through the Danish National Database of Reimbursed Prescriptions, which registers all prescriptions with reimbursements dispensed at Danish pharmacies.25 In Denmark, all antihyperglycemic medications are reimbursed, and the use of prescription databases has been shown to have a positive predictive value of 98% for type 2 diabetics using oral antihyperglycemics.26 If patients reported type 2 diabetes but had no prescriptions in the Danish National Database of Reimbursed Prescriptions, they were considered as having “diet treatment.” Thirty-six patients didn’t report diabetes but had prescriptions on antihyperglycemic drugs and were thus classified as diabetics.
Perioperative Treatment
The participating departments all used a similar perioperative fast-track approach, including spinal anesthesia, opioid-sparing multimodal analgesia, minimal use of drains and catheters, mobilization on day of surgery, and functional discharge criteria.27 Median LOS of 2 to 3 days until discharge to home has been reported in a smaller cohort included in this study.24
Preoperative evaluation of diabetic status (HbA1C or fasting glucose) was at the discretion of the consulting anesthesiologist. Diabetic patients were typically scheduled for surgery early in the day and intraoperative use of 50 to 100 mL/h of IV glucose (5%)-insulin-potassium or glucose (5%)-insulin was standard in all departments. Intraoperative blood glucose was monitored each hour followed by appropriate adjustments of infusion rate of glucose and insulin. An intraoperative blood glucose >5 and <10–14 mmol/L was considered acceptable in all departments. All patients received fast-track care postoperatively regardless of diabetic status and were encouraged to eat and drink as soon as possible, with glucose infusion in diabetics being discontinued after the first meal. Postoperative blood glucose monitoring and thresholds for additional insulin treatment were according to local guidelines. In 1 department, a specific diabetes team consisting of 2 specialized nurses supported by a diabetologist was available.
Postoperative Data
Information on LOS, transfers between departments, and readmissions was obtained through the Danish National Health Registry. Danish hospitals have to report to the Danish National Health Registry in order to receive reimbursement, allowing complete follow-up on readmissions regardless of location.28 LOS was defined as number of nights in hospital after surgery until discharge to own home, including transfer to other hospitals or departments (e.g., intensive care units, general surgical or medical wards and in-hospital rehabilitation units). An LOS > 4 days was chosen as the cutoff for “prolonged” hospitalization because we have previously established that >80% of patients are discharged within 4 days, suggesting that patients with longer LOS have had complications.24 In case of transfer between hospitals/departments, LOS > 4 days, or 90-day readmissions, discharge forms and medical charts were consulted to determine primary reasons for these occurrences. Causes of in-hospital mortality were evaluated using the complete medical charts and autopsy reports if available. Causes of mortality outside hospital were evaluated using death certificates and/or autopsy reports if available.
Finally, we constructed a composite outcome of “diabetes-related” morbidity, defined as the occurrence of specific postoperative morbidity more frequently occurring in diabetics than in nondiabetics4,6–8,29 or due to common diabetes-induced organ dysfunctions, and resulting in LOS > 4 days or readmissions. This included any case of cardiac arrhythmia, acute congestive heart failure, myocardial infarction, prosthetic or wound infections, renal insufficiency, cerebral attacks (transient ischemic attacks or cerebral stroke), pneumonia, and urinary tract infection causing LOS > 4 days or readmissions, as well as sepsis and dysregulated blood glucose causing LOS > 4 days and “other infections” causing readmission. All categories of outcome variables were predefined before data analysis (Table 1).
;)
Table 1: Definitions of Complications
Study Outcomes
Our primary aim was to investigate the association of diabetes and LOS > 4 days, readmissions, and mortality within 90 days postoperatively, both overall and related to antihyperglycemic treatment regimens.
Our second aim was to determine the relationship between preexisting diabetes and its treatment regimen and the occurrence of “diabetes-related morbidity” postoperatively, with an adjusted NNH22,23 for both LOS > 4 days, 30- and 90-day readmissions, and “diabetes-related comorbidity.” The adjusted NNH is the number of surgical type 2 diabetics needed for 1 more case of LOS > 4 days, readmission, or diabetes-related morbidity related specifically to the presence of diabetes while controlling for other confounders.22,23
During data preparation, we found that there were too few patients with type 1 diabetes to include them in the main analyses. Crude results for these few patients are therefore reported separately.
Statistics
Distribution of data was assessed using Kolmogorov-Smirnov with Lilliefors correction. Nonparametric data are reported as medians with interquartile ranges (IQR). Mann-Whitney U test was used for unadjusted analysis of nonparametric data, while χ2 test and Fisher exact test were used for categorical data when appropriate. To adjust for baseline differences between type 2 diabetics and nondiabetics, we calculated propensity scores30 using logistic regression with type 2 diabetes as outcome and age, sex, BMI, use of walking aids, operated joint, living alone/in institution/with others, pharmacologically treated cardiac disease, treated hypertension, and treated hypercholesterolemia as covariates. Age and BMI were entered as continuous covariates, while 2 separate variables were created regarding living alone or in institution using living with others as reference. The remaining variables were dichotomized (present/not present) with type 2 diabetes treatment categorized as insulin (with/without oral treatment), oral antihyperglycemic drugs only, and diet only (no pharmacologic treatment) versus nondiabetics for subanalyses.
Of 8804 consecutive surgical patients, we initially excluded 36 (0.4%) due to missing information on diabetic status. Of the remaining 8768 procedures with completed preoperative questionnaires and information on diabetes, 43 (0.5%) procedures were in type 1 diabetics and 957 were in type 2 diabetics (10.9%). The percentage of type 2 diabetics was 9.3%–12.7% of all procedures in the different departments. It was possible to calculate propensity scores for 890 (93.0%) procedures in type 2 diabetics and for 7414 (95.4%) in nondiabetics. The C-statistic for the propensity model was 0.82, indicating a satisfactory predictive power. Trimming of the data set was performed to further improve accuracy of the model by excluding 249 procedures (3.0%) for which there was no propensity score overlap between groups.31 Thus, the final study population consisted of 890 (11.0%) procedures in type 2 diabetics and 7165 (89.0%) in nondiabetics.
Analysis on outcomes was done using multiple logistic regressions, adjusting for patient characteristics and for propensity score as a continuous variable in a “doubly robust” model.30 Department of surgery (dichotomized into those with significantly more readmissions than the department with the fewest versus others) was also included in the multiple regression analysis to adjust for potential logistical and population differences not otherwise accounted for in the standardized perioperative setup. Only the first readmission was included for regression analysis. We also tested for relevant interactions using pairwise separate models and applying the Wald test statistics. Results are reported as adjusted odds ratios (OR) with 95% confidence intervals (95% CI) and a 2-tailed P-value of <0.05 considered significant. The adjusted NNH for type 2 diabetics with 95% CI was calculated based on the OR.23
Statistical analysis was performed using SPSS v. 20 (IBM Corp., Armonk, NY), with the adjusted NNH macro available from http://www.rbsd.de/softw.html.
RESULTS
Primary Outcome
Type 2 diabetics were older, had more comorbidity than nondiabetics, and were most often treated with oral antihyperglycemics (Table 2).
Table 2: Patient Characteristics
Although more type 2 diabetics (11.3%) than nondiabetics (8.1%) had LOS > 4 days (unadjusted P = 0.001; Fig. 1, A), there was no association between LOS > 4 days and type 2 diabetes when adjusting for covariates. This was regardless of antihyperglycemic treatment (Table 3). When comparing the causes of LOS > 4 days, renal, urologic, and cerebral complications were significantly more frequent in type 2 diabetics (Table 4).
Figure 1: Unadjusted incidence and 95% confidence intervals for study outcomes. LOS = length of hospital stay; T2D = type 2 diabetics. *P < 0.05.
Table 3: Multiple Logistic Regression Analysis of the Diabetes Characteristics Associated with Primary and Secondary Outcomes and Adjusted Number Needed to Harm (NNH)a
Table 4: Types of Complications
Readmissions within 30 and 90 days were both more frequent in type 2 diabetics (6.5% and 10.2%) than in nondiabetics (5.5% and 7.8%) (Fig. 1, B and C). However, after adjusting for covariates, no association between type 2 diabetes and 30- or 90-day readmissions was found (Table 3). Renal and urologic complications reoccurred more frequently in type 2 diabetics (Table 4).
Finally, 90-day all-cause mortality was 0.3% in both type 2 diabetics (n = 3) and in nondiabetics (n = 23).
Secondary Outcomes, Diabetes-Related Morbidity, and Adjusted Number Needed to Harm
The composite outcome of diabetes-related morbidity occurred in 6.0% of type 2 diabetics and in 3.6% of nondiabetics. (Fig. 1, D). After covariate adjustment, only insulin-treated type 2 diabetes was found to be associated with diabetes-related morbidity (Table 3).
Because there was no significant association between type 2 diabetes and study outcomes, infinity was included in all estimates of the adjusted NNH (Table 3).
Type 1 Diabetics
In the 43 excluded type 1 diabetics, median LOS was 3 days (IQR: 2–4), with 9 (20.9%) patients having LOS > 4 days, and 30- and 90-day readmission rates of 11.6% and 18.6%, respectively. There were 4 patients (9.3%) with diabetes-related morbidity and 1 death due to cancer within 90 days.
DISCUSSION
This study reveals several new aspects about type 2 diabetes as a preoperative risk factor in fast-track total THA and TKA. First, although there were about 3% more type 2 diabetics with LOS > 4 days and 90-day readmissions, this was not associated with type 2 diabetes after adjusting for other clinical covariates. We also found that only insulin-treated type 2 diabetes was associated with specific diabetes-related morbidity. Finally, we estimated that approximately 80 patients with type 2 diabetes need surgery in order for 1 additional case of LOS > 4 days, increasing to at least 100 patients for readmissions and diabetes-related morbidity.
The influence of diabetes has not been investigated in fast-track surgery, despite the fact that fast-track methodology has been shown to decrease postoperative morbidity and time to discharge,24,32,33 probably by reducing surgical stress, organ dysfunction, and pain and thereby enhancing recovery.18,34,35 In this study, we did not find type 2 diabetes to be independently associated with LOS > 4 days or readmissions, regardless of stratifying by different antihyperglycemic treatments. One possible explanation may be that the fast-track approach with an optimized perioperative setup, including the intraoperative period (type of anesthesia and opioid-sparing analgesia) and postoperative recovery (with early mobilization and adequate multimodal analgesia), is different from the standard of care used in previous studies on diabetic surgical patients.4,7,36 The fast-track approach may reduce the detrimental effects of postoperative stress and organ dysfunction, thereby limiting metabolic derangement such as hyperglycemia caused by pain-induced insulin resistance,19,37 which may be particularly important in diabetics. However, it should be considered that other recent studies without specific fast-track protocols have also found diabetes to be unrelated to various postoperative complications.9,10,38 Notably, the guidelines for perioperative management of diabetic patients in all departments in our study were similar to recent National Health Service guidelines from the United Kingdom,15 although only 1 department had an inpatient specialist team available.
In diabetics, the combination of related comorbidity (i.e., vascular and renal disease), postoperative insulin resistance, and hyperglycemia may be responsible for a higher risk and fewer routine discharges after surgery,15,17,39 including total joint arthroplasty.7,36 However, age in itself and need of walking aids may also be relevant because they have been shown to influence outcomes after fast-track THA and TKA.24 This assumption is further illustrated by the high point estimates and inclusion of infinity in the CIs of the adjusted NNH. Thus, a very large number of patients with type 2 diabetes would be needed to demonstrate the effects of further interventions specifically targeting the diabetes. This may be of importance when considering the cost–benefit of further preoperative optimization of antihyperglycemic treatment and perioperative interventions intended to further reduce perioperative blood glucose.40,41
Apart from type 2 diabetes not being associated with LOS > 4 days or readmissions, we were also unable to find any association with postoperative morbidity previously associated with diabetes,6–8,29,36,42 except for those treated with insulin. Insulin treatment is usually indicated in patients with longer duration of type 2 diabetes and often more comorbidity than in patients treated only with lifestyle changes or oral antihyperglycemic drugs. This may account for the increased risk of diabetes-related morbidity in these patients. However, the composite secondary outcome was intended to evaluate whether any lack of association between our primary outcomes and type 2 diabetes would be due to lack of differentiation between causes of morbidity. Consequently, no conclusions should be drawn regarding the specific importance of individual outcome components because these may differ in frequency and severity.43
In contrast to previous studies, where presence of diabetes has been based on diagnostic codes only7,8,36 and often with no information on antihyperglycemic treatments,4,10 the strength of our study was the use of prospectively collected data intended for this type of research, combined with extraction of the specific pharmacologic treatment of diabetes from the Danish National Database of Reimbursed Prescriptions database,25 allowing detailed information about the treatment of hyperglycemia. To our knowledge, the only 2 other studies using a similar approach found an increased risk of periprosthetic joint infection in diabetics.6,29 However, antihyperglycemic treatment was not found related to infections in the study by Jämsen et al.,6 while Pedersen et al.29 surprisingly found that patients with diabetes for less than 5 years were at higher risk of revisions due to infection after THA compared with other diabetics. The differences between these studies and ours can be explained both by different follow-up periods and definitions on outcomes but likely also by the differences in perioperative setup and consequently longer LOS during their study periods.33,44 Notably, we did not find diet-treated type 2 diabetes to be associated with LOS > 4 days or readmissions, even without adjusting for comorbidity. Considering that these patients often have shorter duration of diabetes and are well controlled in relation to hyperglycemia, this may not be unexpected.45
Limitations
There are some caveats to consider regarding our study. First, we do not know if the 118 patients answering yes to having type 2 diabetes but without prescriptions dispensed within the past 6 months actually had type 2 diabetes. However, patient- versus physician-reported diabetes has substantial agreement in older patients having THA and TKA,46 and in only 0.5% of patients reporting to be nondiabetics did we find prescriptions for antihyperglycemic treatment, supporting this finding.
Regarding specific perioperative treatment of diabetics, we can only report on the general guidelines of the participating departments because no patient-specific data were available. These reflected the current recommendations that additional preoperative testing and medical consultations should depend on clinical evaluation rather than be used routinely,47 but we have no data on the consistency of preoperative evaluation or adequacy of glucose control, which could have influenced the results. In this context, the presence of a specialized diabetes team in only 1 department could also be a potential confounder. However, we believe the influence of these factors would be limited because these are highly standardized, dedicated, elective fast-track departments where exemptions from standard would be rare.
Regarding follow-up, the Danish National Health Registry records all readmissions, evaluation of causes of LOS > 4 days and readmissions depended on the medical charts and discharge papers. Thus, an inaccurate description would lead to misclassification. However, we believe that our approach is far preferable compared with the singular use of administrative data because these are known to be less reliable.48,49 Also, LOS can only be considered a surrogate marker for successful recovery after surgery, which is why we investigated the causes of LOS > 4 days. Thus, our data allow analysis of the reasons for prolonged LOS. However, because this is a descriptive study it can only provide associations, not causality.
The same applies to the multiple regression analysis, where factors not included in the model may have influenced the results.50 Furthermore, limitations in the syntax used for calculating the adjusted ORs and NNH necessitated that age and BMI be used as continuous instead of categorical variables, which may have been preferable.
Finally, although there was a risk of selection bias, this has been shown to be nonexistent in the Lundbeck Foundation Centre database,24 which gives a reliable picture of the “everyday” patient at the participating large university and community hospitals. In this context, the value of large prospective cohort studies when analyzing complex outcomes is well recognized.51,52 Nonetheless, it is prudent to limit our conclusions to the concept of fast-track THA/TKA and not to other procedures in general. Thus, further procedure-specific studies will be needed to investigate whether our results can be reproduced within other specialties using a fast-track approach.
CONCLUSIONS
In conclusion, this large prospective study in fast-track THA and TKA found no association between type 2 diabetes per se and impaired postoperative outcome. Further studies of perioperative morbidity in diabetics are needed in other types of fast-track surgeries to establish whether our findings apply to all fast-track procedures or only THA and TKA.
APPENDIX
The following are members of the Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative Group: Kjeld Soeballe, Department of Orthopedics, Aarhus, Denmark; Torben B. Hansen, Department of Orthopedics, Holstebro, Denmark; Henrik Husted, Orthopedic Department, Hvidovre, Denmark; Mogens B. Laursen, Orthopedic Division, Aalborg, Denmark; Lars T. Hansen, MD, Orthopedic Department, Grindsted, Denmark; Per Kjærsgaard-Andersen, Department of Orthopedics, Vejle, Denmark; and Søren Solgaard, Department of Orthopedics, Gentofte, Denmark.
DISCLOSURES
Name: Christoffer C. Jørgensen, MD.
Contribution: This author contributed to the original idea and design of the study, completed data collection and data analysis, wrote the first draft of the manuscript, and takes direct responsibility for the integrity of the manuscript.
Attestation: Christoffer C. Jørgensen reviewed the original data and data analysis and is the archival author.
Name: Sten Madsbad, MD, DMSci.
Contribution: This author contributed to the design of the study, analysis of data, and critically revised the manuscript.
Attestation: Sten Madsbad attests to the final version of the manuscript.
Name: Henrik Kehlet, MD, PhD.
Contribution: This author contributed to the original idea and design of the study, data analysis, and critically revised the manuscript.
Attestation: Henrik Kehlet reviewed the study data and analysis and attests to the final version of the manuscript.
This manuscript was handled by: Franklin Dexter, MD, PhD.
ACKNOWLEDGMENTS
The Danish National Database of Reimbursed Prescriptions database is supported by a grant from the Clinical Epidemiological Research Foundation and Aarhus University. We thank Professor Lars Pedersen for help with data extraction from the database.
REFERENCES
1. Kirksey M, Chiu YL, Ma Y, Della Valle AG, Poultsides L, Gerner P, Memtsoudis SG. Trends in in-hospital major morbidity and mortality after total joint arthroplasty: United States 1998-2008. Anesth Analg. 2012;115:321–7
2. Danish National Board of Health.
http://www.sst.dk/publ/tidsskrifter/nyetal/pdf/2009/01_09.pdf. Accessed November 30, 2012
3. Støvring H, Andersen M, Beck-Nielsen H, Green A, Vach W. Rising prevalence of diabetes: evidence from a Danish pharmaco-epidemiological database. Lancet. 2003;362:537–8
4. Bower WF, Jin L, Underwood MJ, Lee JF, Lee KF, Lam YH, Ng SK, Vlantis AC, Lai PB. Overt diabetes mellitus adversely affects surgical outcomes of noncardiovascular patients. Surgery. 2010;147:670–5
5. Frisch A, Chandra P, Smiley D, Peng L, Rizzo M, Gatcliffe C, Hudson M, Mendoza J, Johnson R, Lin E, Umpierrez GE. Prevalence and clinical outcome of hyperglycemia in the perioperative period in noncardiac surgery. Diabetes Care. 2010;33:1783–8
6. Jämsen E, Nevalainen P, Eskelinen A, Huotari K, Kalliovalkama J, Moilanen T. Obesity, diabetes, and preoperative hyperglycemia as predictors of periprosthetic joint infection: a single-center analysis of 7181 primary hip and knee replacements for osteoarthritis. J Bone Joint Surg Am. 2012;94:e101
7. Bolognesi MP, Marchant MH Jr, Viens NA, Cook C, Pietrobon R, Vail TP. The impact of diabetes on perioperative patient outcomes after total hip and total knee arthroplasty in the United States. J Arthroplasty. 2008;23:92–8
8. Galat DD, McGovern SC, Larson DR, Harrington JR, Hanssen AD, Clarke HD. Surgical treatment of early wound complications following primary total knee arthroplasty. J Bone Joint Surg Am. 2009;91:48–54
9. Adams AL, Paxton EW, Wang JQ, Johnson ES, Bayliss EA, Ferrara A, Nakasato C, Bini SA, Namba RS. Surgical outcomes of total knee replacement according to diabetes status and glycemic control, 2001 to 2009. J Bone Joint Surg Am. 2013;95:481–7
10. Namba RS, Inacio MC, Paxton EW. Risk factors associated with surgical site infection in 30,491 primary total hip replacements. J Bone Joint Surg Br. 2012;94:1330–8
11. 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;94:1181–6
12. Martinez-Huedo MA, Villanueva M, de Andres AL, Hernandez-Barrera V, Carrasco-Garrido P, Gil A, Martinez D, Jiménez-Garcia R. Trends 2001 to 2008 in incidence and immediate postoperative outcomes for major joint replacement among Spanish adults suffering diabetes. Eur J Orthop Surg Traumatol. 2013;23:53–9
13. Ata A, Valerian BT, Lee EC, Bestle SL, Elmendorf SL, Stain SC. The effect of diabetes mellitus on surgical site infections after colorectal and noncolorectal general surgical operations. Am Surg. 2010;76:697–702
14. Rizvi AA, Chillag SA, Chillag KJ. Perioperative management of diabetes and hyperglycemia in patients undergoing orthopaedic surgery. J Am Acad Orthop Surg. 2010;18:426–35
15. Dhatariya K, Levy N, Kilvert A, Watson B, Cousins D, Flanagan D, Hilton L, Jairam C, Leyden K, Lipp A, Lobo D, Sinclair-Hammersley M, Rayman GJoint British Diabetes Societies. . NHS Diabetes guideline for the perioperative management of the adult patient with diabetes. Diabet Med. 2012;29:420–33
16. Rice MJ, Pitkin AD, Coursin DB. Review article: glucose measurement in the operating room: more complicated than it seems. Anesth Analg. 2010;110:1056–65
17. Kwon S, Thompson R, Dellinger P, Yanez D, Farrohki E, Flum D. Importance of perioperative glycemic control in general surgery: a report from the Surgical Care and Outcomes Assessment Program. Ann Surg. 2013;257:8–14
18. Kehlet H, Wilmore DW. Evidence-based surgical care and the evolution of fast-track surgery. Ann Surg. 2008;248:189–98
19. Ljungqvist O. Jonathan E. Rhoads lecture 2011: Insulin resistance and enhanced recovery after surgery. JPEN J Parenter Enteral Nutr. 2012;36:389–98
20. Kehlet H. Fast-track hip and knee arthroplasty. Lancet. 2013;381:1600–2
21. Kehlet H, Mythen M. Why is the surgical high-risk patient still at risk? Br J Anaesth. 2011;106:289–91
22. Altman DG. Confidence intervals for the number needed to treat. BMJ. 1998;317:1309–12
23. Bender R, Kuss O, Hildebrandt M, Gehrmann U. Estimating adjusted NNT measures in logistic regression analysis. Stat Med. 2007;26:5586–95
24. Jørgensen CC, Kehlet HLundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative Group. . Role of patient characteristics for fast-track hip and knee arthroplasty. Br J Anaesth. 2013;110:972–80
25. Johannesdottir SA, Horváth-Puhó E, Ehrenstein V, Schmidt M, Pedersen L, Sørensen HT. Existing data sources for clinical epidemiology: the Danish National Database of Reimbursed Prescriptions. Clin Epidemiol. 2012;4:303–13
26. Drivsholm TB, Frederiksen K, de Fine Olivarius N, Ødegaard B, Kristensen JK. The prevalence of diabetes in Denmark. Development of a method for a registry-based assessment [in Danish]. Ugeskr Laeger. 2003;165:2887–91
27. Husted H, Solgaard S, Hansen TB, Søballe K, Kehlet H. Care principles at four fast-track arthroplasty departments in Denmark. Dan Med Bull. 2010;57:A4166
28. Andersen TF, Madsen M, Jørgensen J, Mellemkjoer L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull. 1999;46:263–8
29. Pedersen AB, Mehnert F, Johnsen SP, Sørensen HT. Risk of revision of a total hip replacement in patients with diabetes mellitus: a population-based follow up study. J Bone Joint Surg Br. 2010;92:929–34
30. Stürmer T, Joshi M, Glynn RJ, Avorn J, Rothman KJ, Schneeweiss S. A review of the application of propensity score methods yielded increasing use, advantages in specific settings, but not substantially different estimates compared with conventional multivariable methods. J Clin Epidemiol. 2006;59:437–47
31. Stürmer T, Rothman KJ, Avorn J, Glynn RJ. Treatment effects in the presence of unmeasured confounding: dealing with observations in the tails of the propensity score distribution–a simulation study. Am J Epidemiol. 2010;172:843–54
32. Malviya A, Martin K, Harper I, Muller SD, Emmerson KP, Partington PF, Reed MR. Enhanced recovery program for hip and knee replacement reduces death rate. Acta Orthop. 2011;82:577–81
33. Husted H, Jensen CM, Solgaard S, Kehlet H. Reduced length of stay following hip and knee arthroplasty in Denmark 2000-2009: from research to implementation. Arch Orthop Trauma Surg. 2012;132:101–4
34. Adamina M, Kehlet H, Tomlinson GA, Senagore AJ, Delaney CP. Enhanced recovery pathways optimize health outcomes and resource utilization: a meta-analysis of randomized controlled trials in colorectal surgery. Surgery. 2011;149:830–40
35. Ansari D, Gianotti L, Schröder J, Andersson R. Fast-track surgery: procedure-specific aspects and future direction. Langenbecks Arch Surg. 2013;398:29–37
36. Marchant MH Jr, Viens NA, Cook C, Vail TP, Bolognesi MP. The impact of glycemic control and diabetes mellitus on perioperative outcomes after total joint arthroplasty. J Bone Joint Surg Am. 2009;91:1621–9
37. Yang DJ, Zhang S, He WL, Chen HY, Cai SR, Chen CQ, Song XM, Cui J, Ma JP, Zhang CH, He YL. Fast track surgery accelerates the recovery of postoperative insulin sensitivity. Chin Med J (Engl). 2012;125:3261–5
38. Davis C, Tait G, Carroll J, Wijeysundera DN, Beattie WS. The Revised Cardiac Risk Index in the new millennium: a single-centre prospective cohort re-evaluation of the original variables in 9,519 consecutive elective surgical patients. Can J Anesth. 2013;60:855–63
39. Saucedo JM, Marecek GS, Wanke TR, Lee J, Stulberg SD, Puri L. Understanding readmission after primary total hip and knee arthroplasty: who’s at risk? J Arthroplasty. 2013;29:256–60
40. Webster NR, Galley HF. Does strict glucose control improve outcome? Br J Anaesth. 2009;103:331–4
41. Cryer PE. Hypoglycaemia: the limiting factor in the glycaemic management of the critically ill? Diabetologia. 2006;49:1722–5
42. Iorio R, Williams KM, Marcantonio AJ, Specht LM, Tilzey JF, Healy WL. Diabetes mellitus, hemoglobin A1C, and the incidence of total joint arthroplasty infection. J Arthroplasty. 2012;27:726–9.e1
43. Sessler DI, Devereaux PJ. Emerging trends in clinical trial design. Anesth Analg. 2013;116:258–61
44. Mäkelä KT, Peltola M, Sund R, Malmivaara A, Häkkinen U, Remes V. Regional and hospital variance in performance of total hip and knee replacements: a national population-based study. Ann Med. 2011;43(Suppl 1):S31–8
45. Inzucchi SE, Bergenstal RM, Buse JB, Diamant M, Ferrannini E, Nauck M, Peters AL, Tsapas A, Wender R, Matthews DRAmerican Diabetes Association (ADA); European Association for the Study of Diabetes (EASD). . Management of hyperglycemia in type 2 diabetes: a patient-centered approach: position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care. 2012;35:1364–79
46. Gad BV, Higuera CA, Klika AK, Elsharkawy KA, Barsoum WK. Validity of patient-reported comorbidities before total knee and hip arthroplasty in patients older than 65 years. J Arthroplasty. 2012;27:1750–6.e1
47. Apfelbaum JL, Connis RT, Nickinovich DG, Pasternak LR, Arens JF, Caplan RA, Connis RT, Fleisher LA, Flowerdew R, Gold BS, Mayhew JF, Nickinovich DG, Rice LJ, Roizen MF, Twersky RS. Practice advisory for preanesthesia evaluation: an updated report by the American Society of Anesthesiologists Task Force on Preanesthesia Evaluation. Anesthesiology. 2012;116:522–38
48. Severinsen MT, Kristensen SR, Overvad K, Dethlefsen C, Tjønneland A, Johnsen SP. Venous thromboembolism discharge diagnoses in the Danish National Patient Registry should be used with caution. J Clin Epidemiol. 2010;63:223–8
49. Keeney JA, Adelani MA, Nunley RM, Clohisy JC, Barrack RL. Assessing readmission databases: how reliable is the information? J Arthroplasty. 2012;27:72–6.e1–2
50. Wallace AW. Risk of acute myocardial infarction in patients with total hip or knee replacement: comment on “Timing of acute myocardial infarction in patients undergoing total hip or knee replacement: a nationwide cohort study.” Arch Intern Med. 2012;172:1235–6
51. Merkow RP, Ko CY. Evidence-based medicine in surgery: the importance of both experimental and observational study designs. JAMA. 2011;306:436–7
52. Berwick DM. The science of improvement. JAMA. 2008;299:1182–4
