The study underwent ethical review in our institution, and written consent was obtained by the parents of infants that participated. The study was a cohort design with 2 study groups. The first group included infants circumcised using the Mogen clamp and combined analgesics (group 1). The analgesics administered before the circumcision included 0.5 mL of 80-mg/mL acetaminophen (administered orally 45 minutes before the clinic appointment); 1 or 2 g of EMLA on the foreskin and abdomen (applied 60 minutes prior to the procedure); and 2 subcutaneous injections of 1.1 mL of 0.5% lidocaine at the 10:30 and 1:30 positions (i.e., DPNB) after EMLA was removed and 10 minutes before the circumcision.The acetaminophen was administered to infants by their parents at home. They were given an information pamphlet and spoke to an office nurse before the scheduled appointment. Infants were last fed at home before the procedure. On arrival to the clinic, EMLA was applied by an office nurse. Part of the dose was placed on the penis and the remainder on a Tegaderm dressing. Then the penis was extended upward and gently pressed on the abdomen and the dressing was placed over the penis and taped to the abdomen. The cream and dressing were removed after 1 hour. During the infant’s stay in the waiting room, a 3 X 3-cm folded gauze embedded with three fourths of a teaspoon of sugar and dipped in grape juice was placed in his mouth. The gauze was replaced by a fresh one 3 different times: while in the waiting room, before the DPNB injection, and before the circumcision.This integrated approach was developed by one of us (N.P.) and is currently in use in the clinical setting. All infants in group 1 were restrained on a circumcision board and circumcised using the analgesics and surgical technique described above (referred to as the Pollock procedure) by one of us (N.P.).The second group included infants circumcised using the Gomco clamp and EMLA (group 2). One gram of EMLA was applied for 60 to 80 minutes prior to the procedure. The EMLA was applied using a similar technique as for group 1. Infants were not offered anything to suck on. Group 2 was taken from our previous double-blind, randomized clinical trial of EMLA.2 Infants were fed I to 3 hours before the circumcision. All of the infants were restrained on a circumcision board and were circumcised by another single operator (P.R.).2Since the results of our randomized controlled trial were published, EMLA has been considered the minimum standard for analgesia during routine circumcision in our institution. We chose not to perform a randomized controlled trial because these data were available for comparison and it was considered unethical to prospectively enroll infants into a group that was expected to experience more pain during a circumcision.We included healthy full-term infants without jaundice and methemoglobinemia and not receiving analgesic or sedative drugs outside of the study protocol. Infants in group 2 were circumcised in the first week of life. However, infants in group 1 also included older infants, to assess the effect of postnatal age on pain response.Infants in both groups were videotaped during the procedure. Pain was scored from the videotape by a research assistant using the same techniques as our previous study (ie, Neonatal Facial Coding System [NFCS] and infant cry duration).2 The circumcision procedure was divided into phases. The analyses included the baseline phase and the circumcision phase (ie, forceps application, lysis of adhesions, application of clamp, cutting foreskin, and removal of clamp).The primary outcome was facial activity score. The facial activity score was composed of the sum of the percentage of time that 3 discrete facial actions from the NFCS (ie, brow bulge, eyes squeezed shut, and nasolabial furrow) were observed for each phase of the procedure. These facial actions are considered the most sensitive and specific to pain. As in our clinical trial, each facial action was coded as present or absent every 2 seconds for a maximum of 20 seconds per phase.2 Then, the percentage of time that each of the 3 facial actions was observed was calculated. The 3 percentage scores were weighted equally using a ratio of 1:3 and added together for an overall facial activity score that could range from 0 to 1. The percentage of time spent crying per phase of the procedure was calculated by dividing the duration of time spent crying by the duration of time of the phase.

Adverse events (skin reactions, bleeding, and infection) were noted during the circumcision and during follow-up telephone interviews with the parents at 24 hours and 1 week after the circumcision.

SAMPLE SIZE CALCULATION
A sample size of 60 infants (30 per group) was considered sufficient based on the ability to detect a difference in pain scores between infants that was 0.8 SD, with 80% power and 95% confidence (ie, large effect size). We recruited an additional 30 infants in group 1 to account for the effects of age on pain response.

STATISTICAL ANALYSES
Infant responses during circumcision were compared between groups using repeated-measures analysis of variance, with the baseline value as the covariate. Regression analysis were used to compare responses of infants of different ages. Demographic data and adverse effects were analyzed using X2 test and t test as appropriate.