Baseline and follow-up characteristics were summarized using frequencies and percentages for categorical variables. Mean ± standard deviation, median, minimum, and maximum were used to summarize continuous variables. Baseline characteristics were compared using a chi-squared test for categorical variables, Student t test for normally distributed variables, and the Mann-Whitney U test for skewed variables. (See table 1, Supplemental Digital Content 1, which displays the baseline characteristics, http://links.lww.com/PRSGO/C425 .)
Linear regression with inverse probability weighting was used to compare the change in preoperative pain at most recent follow-up for the Avive PUCM and control groups. Change in pain from pre-revision surgery to most recent follow-up was also categorized as (A) worsened or no significant improvement if there was less than two points improvement in VAS pain, or (B) clinically important difference if VAS pain improved greater than or equal to two points. (See table 2, Supplemental Digital Content 2 , which displays the follow-up characteristics, http://links.lww.com/PRSGO/C426 .) Logistic regression with inverse probability weighting was used to compare postrevision satisfaction between groups. A type 1 error level of 5% was used for all statistical tests, and all analyses were conducted in Stata 17 (Supplemental Digital Content 2 , http://links.lww.com/PRSGO/C426 ).
Subanalyses were then conducted by the nerve treated. There were 44 Avive PUCM subjects and 49 control subjects included in the median nerve subanalysis. The ulnar nerve subanalysis included 30 Avive PUCM subjects and 49 control subjects. Twelve subjects in the Avive revision group had radial revision, but no subjects in the control group had radial revision; so a subanalysis was not conducted for radial nerves. Subjects with both median nerve revision and ulnar nerve revision (n = 5) were included in both subanalyses. Baseline and follow-up characteristics were compared using the same methods described above for the entire cohort.
Linear regression with inverse probability weighting was used to compare the change in preoperative pain at most recent follow-up for the Avive PUCM and control groups. Change in pain from pre-revision surgery to most recent follow-up was also categorized as (A) worsened or no significant improvement if there was less than two points improvement in VAS pain, or (B) clinically important difference if VAS pain improved greater than or equal to two points. (
Subanalyses were then conducted by the nerve treated. There were 44 Avive PUCM subjects and 49 control subjects included in the median nerve subanalysis. The ulnar nerve subanalysis included 30 Avive PUCM subjects and 49 control subjects. Twelve subjects in the Avive revision group had radial revision, but no subjects in the control group had radial revision; so a subanalysis was not conducted for radial nerves. Subjects with both median nerve revision and ulnar nerve revision (n = 5) were included in both subanalyses. Baseline and follow-up characteristics were compared using the same methods described above for the entire cohort.
