Patients in three hospitals in Sweden who will have knee arthroplasty surgery for the first time are invited to participate in the study, approximately 2000 patients. The fasting glucose value (fP-glucose) is obtained via a blood sample the day after the knee arthroplasty surgery. The sample is analyzed and the glucose value obtained is documented and sent to the Swedish Knee Arthroplasty Register (SKAR). In the SKAR there is information on patient characteristics, diagnosis, prosthesis, anesthetic form and primary and secondary operations, etc. The unique personal numbers of the included patients are submitted to the National Board of Health and Welfare, which matches the cohort with the Patient Register to identify adverse events, re-admission and death in a year after the surgery. SKAR has carried out several interactions with the National Board of Health and Welfare, and possesses clinical expertise as well as registers and biostatistics expertise. Incidence calculation of the proportion that develops elevated glucose levels after surgery and Cox regression for group comparison (elevated glucose vs. non-elevated) taking into account factors within groups such as age, gender, etc. This is a register-based observation study. Since the incidence of elevated glucose value after a knee prosthesis operation is not known, we are planning for a representative selection from three major prosthetic clinics in Sweden. Regarding the secondary purposes, our ability to answer these depends on the presence of elevated glucose levels. The inclusion start in January 2019 and lasts for one year. Incidence calculations can be made as soon as all patients are included, but 90-day data from the patient register can reasonably be completed

Purpose

To reduce the risk of prosthetic joint infection (PJI) after primary knee arthroplasty surgery.

The aims are to study:

• the presence of postoperative hyperglycemia after knee arthroplasty.
• if there is an increased risk of wound disturbance
• if there is an increased risk of PJI.
• if there is an increased risk of other complications, re-admissions and mortality.
• if postoperative hyperglycemia is more common after general than spinal anesthesia.
• if patients developing hyperglycemia has longer hospital stay.

Background

Postoperative hyperglycemia is a consequence of the insulin resistance that may occur in otherwise healthy nondiabetic patients after uncomplicated elective surgery (Thorell et al. 1999). Hyperglycemia has been associated with an increased risk of PJI (Mraovic et al. 2011) and to be a risk factor in many common postoperative complications (van den Berghe et al. 2001).

Knee arthroplasty surgery in Sweden has more than doubled since the millennium with more than 14,000 primary knee arthroplasties in 2016. The risk of revision after a primary knee arthroplasty is the lowest reported in the world (less than 5% after 10 years in total knee arthroplasty) (SKAR 2017). However during the latest 10 years the risk of revision due to infection has increased as compared to the previous 10 years. PJI is an uncommon complication, 1% - 1.5% (SKAR 2017), but devastating for the patient and an expensive complication for the health care and the society.

During 2009-2013 the Prosthetic Related Infections Shall be Stopped project (PRISS) was performed in Sweden. The purpose of that project was to half the risk of PJI in knee and hip arthroplasties by optimizing the routines. All hospitals performing knee and hip arthroplasty surgery in Sweden participated (www.lof.se). The PRISS project resulted in, among other things, documents/recommendations including a compilation and evaluation of the best-known knowledge within four important areas; risk factor and optimizing, prophylactic antibiotics, follow-up and infection registration and optimal operation room environment. These documents are updated every second year or in case of new knowledge. One of the issues that the expert group, within risk factors and optimizing, requested was further research concerning if screening for hyperglycemia in patients with unknown diabetes has an effect on the risk of PJI (www.lof.se).

Postoperative hyperglycemia has been shown to be relatively common (Pili-Floury et al. 2009, Jämsen et al. 2015). Pili-Floury et al. (2009) showed that 29/38 patients without a medical history of diabetes before the surgery became hyperglycemic for the first two days following arthroplasty surgery. A study from a Finnish hospital found that one fourth (47/191) of the hip and knee arthroplasty patients included in the study had pre-diabetes (increased fasting glucose and/or impaired glucose tolerance) and half of these patients (24/47) developed perioperative hyperglycemia (Jämsen et al. 2015).

Prevention and treatment of insulin resistance has been shown to have an effect on morbidity, mortality, and recovery after surgery (Ljungqvist et al. 2007). It has been suggested that insulin resistance can be avoided using epidural analgesia, minimally invasive surgical technique, optimal pain control, and preoperative carbohydrate treatment (Ljungqvist et al. 2007). General anesthesia has become more common in Sweden and was used in 30% of the surgery during 2016 with large variations between hospitals. Minimal invasive surgery is sometimes used in uni-compartmental knee arthroplasty (UKA) but not in total knee arthroplasty (TKA) in Sweden (SKAR 2017). The hospitals use standardized multi-modal pain control including local infiltration analgesia (LIA), while the use of preoperative carbohydrate treatment may vary.

There are no general recommendations in Sweden to measure fasting glucose levels and/or HBA1c either preoperatively or postoperatively in knee arthroplasty patients unless they have diagnosed diabetes. In a survey from the Swedish Knee Arthroplasty Register (SKAR) to all 74 hospitals in Sweden performing knee arthroplasty surgery 2017, only 5 hospitals routinely obtained fasting plasma glucose (fP-Glucose). In these hospitals, patients having an elevated glucose level are submitted to the primary care for further investigation. The number of such patients is unknown.

Work plan

1. Study population All patients operated with a primary knee arthroplasty from 1st of January 2019 to 31st of December 2020 in high volume hospitals will be invited to participate in the study, approximately 2,000 patients. The operation will be registered in the SKAR to which all hospitals performing knee arthroplasty surgery in Sweden report their knee arthroplasties (primaries, revisions and re-operations). The validity of the register has been considered to be high (coverage 100%, completeness 97%, and response rate 99%) (SKAR 2017). All variables registered are shown and described in the variable list (attachment 1).

2. Plasma glucose The pre-, per- and postoperative routines at each hospital will be as usual, except that for a fP-glucose measurement. In the morning the day after surgery, fP-Glucose will be measured, using the ordinary routine for blood samples at each hospital. The glucose level in plasma is reported as millimoles per liter (mmol/L). According to the Nordic Reference Interval Project (NORIP) the reference interval for venous and capillary fP-Glucose is 4.2-6.3 mmol/L, i.e. slightly different from the World Health Organization (WHO) reference value which is <6.1 mmol/L.

   The result of the test will be documented on a form (attachment 2) that includes the patient's personal identification (ID), the date of fP-Glucose measurement, if the patient has diabetes (type 1 or 2) or not and in case of diabetes what treatment (diet, tablets, insulin). The form will be sent to the SKAR and entered in a separate file which can be linked to the SKAR database. Using the SKAR will provide information on the proportion of patients participating in the study as well as for their follow-up with respect to further surgeries or complications.

3. Prosthetic joint infection Revisions and re-operations due to infection or suspected infection as well as other causes after primary knee arthroplasty are registered in the SKAR (attachment 1). Patients treated for infection or suspected infections not requiring surgery are documented in the department's local infection register at each hospital respectively.

4. Other complications, re-admissions and mortality The SKAR has together with the National Board of Health and Welfare examined the ICD10 and the Nordic Medico-Statistical Committee codes (NOMESCO, a classification of surgical procedures) that occur in the National Patient Register (NPR) during admission for, and after knee arthroplasty. This has resulted in a longer list of codes that may represent adverse events when they occur during the hospital stay or in readmissions within 90 days and 1 year of surgery (SKAR). The SKAR will send data on registered patients to the NPR which performs the match. The NPR will check if they during the hospital stay or within 1 year of the surgery received ICD10 and/or procedure codes that correspond to the definition of adverse events.

5. Length of hospital stay The length of hospital stay for each patient is obtained from the electronic record system used by the orthopedic departments at each site.

6. Statistics The primary aim of this register-based, observational, study is to investigate how common postoperative hyperglycemia is after knee arthroplasty. The secondary aim is to investigate if hyperglycemic patients have an increased risk of PJI or adverse events and if hyperglycemia is more common after general than spinal anesthesia.

   The proportion of hyperglycemia patients in Sweden is unknown. The study is to be performed at large arthroplasty units that operate approximately 400-800 knee arthroplasties/year. Assuming that the proportion of patients with hyperglycemia is similar to that in the Finnish study, i.e. 25% (Jämsen et al. 2015), the hospitals together operate 1800 primary knee arthroplasties during one year and that 80% of the patients accept to participate in the study, would result in 450 patients having hyperglycemia. This would provide a reasonable material for evaluation of risks concerning infections and adverse events.

7. Ethics In some hospitals in Sweden, fP-glucose is a routine while in those hospitals included in the planed study it is not. However in the morning the day after surgery the hemoglobin value is controlled routinely and performed at the included hospitals. From the same splinter as hemoglobin is taken in the finger or vein, fP-glucose can be obtained. The splinter in the finger or vein may be painful and experienced as unpleasant by the patients instantly but this additional test will not cause any further discomfort. The patients will be sent written information (attachment 2) about the study together with the invitation for the preoperative visit at each hospital. At the preoperative visit the patients has possibility to further questions if requested. Considering the patients integrity all results will be presented on aggregated level and the material is relatively large and the risk for individual identification may be seen as negligible.

Importance

Currently, we have no knowledge on the prevalence of postoperative hyperglycemia after knee arthroplasty in patients without diagnosed diabetes or its effect on outcome. Prosthetic joint infection is a devastating complication for the patients as well as it is costly for the health care and the society. Should postoperative hyperglycemia be a risk factor in the development of PJI it may be a modifiable risk factor by prevention and/or treatment.

By including patients having a primary knee arthroplasty we will gain knowledge on the proportion of patients without diagnosed diabetes having elevated postoperative glucose values and to evaluate if such patients run an increased risk of PJI or other adverse events.

Further, general anesthesia during arthroplasty surgery has become more popular in Sweden in recent years. However, it has been suggested that general anesthesia may be increase the risk for hyperglycemia after surgery (Ljungqvist et al. 2007). Thus, this study may help us find out if this is the case or not.

Dependent on the findings of the study, a continuation may be needed in order to evaluate the possible benefit of a general preoperative screening in arthroplasty patients.

Additional/Safety

Patients with elevated postoperative fP-Glucose level will be followed-up by the primary care. In case of fP-Glucose levels associated with diabetes the routines of the department will be followed.

Time-frame The collection of data will start January 1st 2019 if all permissions are granted and continue until December 31st 19. During 2019-2020 the patients will be followed in the SKAR as well in medical journals and local infection registers. When the last patient is included we may have the possibility to present the frequency of hyperglycemia after knee arthroplasty surgery in the hospitals included. The data from the NPR for patients participating in the study will probably not be ready until the spring 2020. In the summer 2020 we may have data from the Patient register for complications, readmission and death during the first 90 day after surgery and be able to present a short follow-up. During the spring 2021 we may have the 1 year follow-up from the NPR.

Inclusion Criteria:

• Patients operated on with a primary knee arthroplasty in 3 hospitals in Sweden.
• Approved to participate in the study.

Exclusion Criteria:

• Not approved to participate in the study.

Patients operated on with a primary knee arthroplasty from Three hospitals in Sweden