Oral Peri-operative TIming of Metformin (or) Salsalate to Improve Non-cardiac Surgery Glucose Control

This internal pilot study (n=36) is designed as a 3-arm, parallel, placebo-controlled trial with block randomization. All parties will be blinded.

Patient Recruitment: Patients with T2D on metformin presenting to the Hamilton Health Sciences (HHS) preoperative clinic for elective surgery requiring general anesthetic will be screened for eligibility.

Sample Size: 36 patients over the 1-year study. This sample size will allow for 3 groups of 12, which permits block randomization and precise confidence interval (CI) calculations.

Randomization and Allocation: Patients will be randomized on the day of their pre-operative assessment by our department statistician using a computer-generated table, with an allocation ratio of 1:1:1 in random permuted blocks of 3 or 6. Only the pharmacist and statistician will know the allocation sequence. Concealed medications (over-encapsulated) will be provided to the RC upon confirming the allotment with a unique identification number (UIN), recorded on the patient-specific case record form (CRF).

Application of Intervention: Our intervention involves self-administered metformin (patient's usual dose) vs. salsalate (1.5g) vs. placebo with sips of water 2h prior to surgery. Baseline data (age, height, weight, ASA score, mean arterial BP, and baseline glucose) will be collected by the RC in the pre-operative assessment area, and flash glucose sensors (FreeStyle Libre, Abbot) will be placed on the upper arm as per manufacturer's instructions and readings collected by the RC. Study data will be securely stored in the anesthesia research office, McMaster University. Standard monitors will be used in the operating room, with any additional monitoring left to the anesthesiologist's discretion. Although most anesthetics appear to modify glucose metabolism, general anesthesia will be provided in their preferred manner (including dexamethasone administration for post-operative nausea and vomiting prophylaxis) to improve generalizability. Details of the type of drugs, fluids and blood loss will be recorded. Following surgery, patients will be brought to the PACU for routine monitoring (30-60 min), and additional blood samples will be drawn. Subsequent samples will be collected on the evening of surgery (6-12h) and POD1. Any episodes of hyper- or hypoglycemia will be managed routinely at the treating physician's discretion and details recorded.

Measurements of Inflammation: Blood and urine samples will be collected on the morning of surgery, upon admission to PACU, and will coincide with other tests to minimize needle sticks. Blood samples will be processed at the HHS Core Research Laboratory and divided into 3 serum and 2 EDTA aliquots, barcoded and stored in liquid nitrogen. Urine samples will be divided into 2 barcoded aliquots and stored in liquid nitrogen. This will allow for future analysis and characterization of biochemical pathways related to the surgical stress response. These future analyses will include measurement of insulin and proinsulin since salsalate may modify their endogenous metabolism. Similarly, the investigators eventually plan to measure various inflammatory markers implicated in the development of acute stress-induced insulin resistance (i.e. IL-6, TNFα, and hsCRP), which could provide insight into potential beneficial effects of salsalate or metformin.

Feasibility Outcomes: The primary feasibility outcomes include calculating the recruitment rate (both the percentage of individuals who were approached in the pre-operative clinic that enrolled in the study, and the total number of individuals that were recruited each month (Table 1, supplemental document). The ability to follow study protocol will be reported as the percentage of individuals who were able to complete the study protocol as a fraction of the total number of individuals recruited for i) drug administration and blinding, ii) blood sampling at specified times, and iii) reliability of CGM by comparison with laboratory glucose values. Workload will be calculated per patient, and safety profile assessed by recording adverse events. Together, this may identify pitfalls requiring modification prior to a full RCT.

Statistical Analysis and Feasibility: Reporting and analysis will conform to the CONSORT pilot and feasibility trials guidelines (www.consort-statement.org). Data will be collected using an encrypted online tool (RedCap, Research Electronic Data Capture) hosted at McMaster. To be feasible and continue onto a subsequent RCT the investigators will need to recruit at least 3 patients per month. Additionally, the investigators seek an adherence rate of 88.9% (32/36) to receive study drug, collect more than 90.3% (130/144) of pre-specified blood samples, and ensure 83.3% (30/36) patients have been followed up within 6 weeks of their procedure to document any complications (e.g. SSI). If targets prove difficult, then our protocol will be reevaluated for possible amendments that will not compromise final outcomes. Given this is a pilot, CIs will be calculated for outcomes of interest: rates of hyperglycemia, SSI, and degree of inflammation.

Trial Management: This pilot trial will be conducted at HHS. The trial will enlist support and collaboration from surgeons and fellow anesthesiologists. Study data will be securely stored at the anesthesia research office, McMaster University. An RC will be employed for patient recruitment and data collection. The CRFs for study patients will be randomly cross-checked for completeness and accuracy. All reports of incorrect randomizations, protocol violations or incomplete data will be noted.

Risk Assessment: This trial does not entail any higher risk to the patient than the present standard of care. It is pragmatic, and uses clinically acceptable doses with a safe profile.