2. Methods and Patients
This was a single tertiary referral center retrospective cohort study.
The study was approved by the local institutional review board
(0403-21-RMB-D). Inclusion criteria were patients aged 18 years old or
older, who underwent tonsillectomy or tonsillotomy (with or without
uvulopalatopharyngoplasty [UPPP]) between 2015-2022. Patients’
medical records were reviewed.
Post-tonsillectomy bleeding was defined as any instance of fresh blood
discharge through the mouth, either reported by the patient or seen by
medical stuff. This included any bleeding ranging from a few drops of
blood to major hemorrhage with hemodynamic compromise. We decided on
this broad definition in order to minimize the chance for interobserver
variability on declaring PTB.
All patients were admitted after the operation for at least 2 days.
Post-operative pain was assessed by VAS (0 to 10) by the department
nurse at least once in an 8-hour shift and recorded in the medical
records.
After discharge all patients were scheduled an outpatient clinic visit 2
weeks after the operation. During that visit the patients were
questioned regarding episodes of PTB (even if they occurred in another
hospital) and the answers were documented in the medical record.
Descriptive analysis of quantitative variables was demonstrated using
mean, standard deviation and range. Categorial variables were described
with prevalence and percentage. Quantitative groups were compared by
independent sample t-test or Wilcoxon rank sum test (according to sample
size and distribution). Qualitative data were compared using Chi-square
test or Fisher’s exact test. Multi-variable logistic regression model
was used to test correlation between the independent variables and PTB
(variables that were correlated to PTB in a statistically significant
manner in the univariable analysis). Co-linearity was excluded. Results
are shown as adjusted odds ratio. Power of the correlation between
qualitative variables was tested using Cramer’s coefficient. A p-value
less than 0.05 was considered statistically significant.
Statistical analysis was performed using SPSS © 24.0 software for
Windows.
Transparent reporting of a multivariable prediction model for individual
prognosis or diagnosis (TRIPOD) guidelines has been followed in this
study.
There are no sources of funding to disclose for this study.
The authors declare they did not have any conflicts of interest.