Risk of Hematoma After Ketorolac Use in Reduction Mammoplasty.

The electronic health record databases of St. Joseph's Healthcare Hamilton and Hamilton Health Sciences will be searched, using the coding system, for all cases of reduction mammoplasty with post--operative hematoma as a complication. Only hematomas which required return to the operating room for evacuation will be included (i.e. those treated with observation or needle aspiration will be excluded).

Cases are defined as patients suffering a hematoma post reduction mammoplasty will be identified through hospital coding, and matched with controls based on the following criteria: age, body mass index (BMI), pre-existing hypertension, and institution. Controls are defined as patients who had uncomplicated reduction mammoplasty, and will be randomly pulled through retrospective chart review within the same databases, in the same time period (±5 years).

All patient data will be de--identified, and all devices containing patient information will remain on hospital grounds, and be encrypted according to our hospital protocols. The following data will be extracted from patient charts: age, date of surgery, body mass index, surgeon, anesthetist, volume of breast tissue removed per breast, comorbidities, platelet count, coagulation profile, ketorolac and opioid use including dose, timing, frequency, duration, and route of administration of both, time to discharge including length of stay if overnight, comorbidities (including smoking status), American Society of Anesthesiology score, and post--operative usage of drains.

Matching of cases and controls will be performed by the primary investigator (JB), and at the time of matching, the only visible patient information will be the de-identified patient identification (ID) and the four matching variables (ketorolac exposure was not visible).

The frequency of ketorolac exposure between cases and controls will then be reviewed, and the odds ratio (OR) of exposure to ketorolac, with 95% confidence intervals will be calculated. A two by two table will be generated from our data. Since the incidence of hematoma formation after reduction mammoplasty is very low, the odds ratio will closely approximate the relative risk. To evaluate the statistical significance of the ORs, a level of significance of 0.05 will be used. Independent samples t-test will be used to compare continuous demographic data, and Chi-square analysis will be used for ordinal or categorical data between cases and controls.

Sample size was calculated using the POWER program, described by Dupont and Plummer in 1990. A sample size of 25 cases matched with 25 controls (total n = 50) was required from the following data:

• p0 = 0.32 (approximate rate of ketorolac exposure amongst breast reduction patients from previous literature from Cawthorn et al.)
• α = 0.05 (chance of type I error)
• P = 0.8 (power)
• m = 1 (matching ratio1:1)
• Ψ = 3.6 (estimated risk ratio of developing hematoma post breast reduction if exposed to ketorolac, Cawthorn et al.)
• Φ = 0.2 (measurable difference)