Participants

This study was approved by our institutional review board and registered on clinicaltrials.gov ( NCT01726036 ) prior to enrollment of patients. Enrollment occurred from October 2012 to March 2014. Inclusion criteria included term, male singleton infants delivered by vaginal delivery or cesarean delivery on the resident or midwife service at Good Samaritan Hospital in Cincinnati, OH, during the study period.

Exclusion criteria included infants with congenital anomalies and/or known chromosomal syndromes, preterm infants defined as less than 37/0 weeks’ gestation, delivery secondary to nonreassuring fetal heart tracing as defined by the delivering physician, infants of diabetic mothers, infants of preeclamptic mothers on magnesium sulfate within 48 hours of delivery, infants delivered by vacuum or forceps-assisted vaginal delivery, infants of mothers who took steroids, infants with a 5 minute Apgar score of less than 7, infants born to mothers with a history of drug abuse during pregnancy (with the exception of first-trimester marijuana use), and infants born to mothers currently on methadone or with hepatitis C or human immunodeficiency virus.

Extensive exclusion criteria were used to prevent inclusion of infants with conditions that might impact their hypothalamic-pituitary-adrenocorticotropic response and therefore salivary cortisol levels. Blood draws were held prior to the circumcision to prevent confounding results from other potential sources of neonatal pain and stress.

On admission to the labor and delivery suite, parents who request male neonatal circumcision sign an informed consent. The counseling is performed by the mother’s primary obstetrician-gynecologist. Infant male charts are flagged in the electronic medical record after delivery to indicate a pending circumcision. We used this circumcision identification system that was already in place in our hospital to identify potential candidates for this study. After the parents provided consent for neonatal circumcision, the infant’s chart was reviewed to determine study eligibility. If eligible, the parents were approached for possible enrollment of their infant in the study. Parents were not blinded to treatment group.

Standardized curriculum

A residency-wide curriculum for first through fourth year obstetrics and gynecology residents was implemented utilizing a pretest, didactics, hands-on training using circumcision models (Global Technologies, Davie, FL), a postwritten test, and a practical examination. This separate educational initiative had Institutional Review Board approval and supplemented the existing circumcision curriculum. Additionally, each (obstetrics-gynecology) resident was required to perform 5 each of Gomco and Mogen clamp circumcisions supervised by an attending physician prior to their involvement in the study.

Procedure standardization

All components of the circumcision were standardized between the 2 groups except for the techniques specific to the Gomco or Mogen clamp. All circumcisions were performed by obstetrics-gynecology residents on weekdays between 8:00 am and 5:00 pm when at least 1 of 2 trained research nurses were present.

Objective preoperative baseline measurements included blood pressure measurement on the lower extremity of the neonate using a pediatric cuff (GE Medical, Rochester, NY), heart rate and oxygen saturation (Masimo Rainbow Set, Irvine, CA), and salivary cortisol. Collection of salivary cortisol was obtained from infants lying in their bassinets. An absorbent swab stick (SalivaBio infant’s swab; Salimetrics, State College, PA) was used for salivary cortisol collection.

Two trained research nurses (P.M.W. and R.G.D.) collected all the salivary cortisol samples; assigned crying, requires increased oxygen administration, increased vital signs, expression, sleeplessness (CRIES) scores; and obtained preprocedure and postprocedure vital signs on all the infants. Nurses were not blinded to procedure. A sufficient amount of saliva sample was assumed when saliva oozed out of the swab end when pressed to the inner wall of the collection tube.

After collecting preoperative vital signs and salivary cortisol, a sign was placed in the infant’s crib, reminding staff to defer neonatal blood draws until after the circumcision unless medically indicated. Prior to the circumcision, the research nurses confirmed with the infant’s primary nurse that no blood draws occurred prior to the circumcision.

At the time of circumcision, all infant males were positioned on a restraint board (Olympic Circumstraint; Natus, San Carlos, CA) and given oral sucrose (TootSweet; Natus). The infants were prepped and draped. Subcutaneous ring block consisting of 0.8 mL of 1% lidocaine without epinephrine was administered using a 30-gauge needle. A subcutaneous ring block was adopted based on findings from previous randomized controlled trials that a ring block may provide the most effective form of anesthesia. A 5 minute waiting period was required between injection of anesthesia and the start of the circumcision defined as placement of the hemostat on the foreskin.

To reinforce a uniform protocol, laminated checklists were posted where the circumcisions were performed for resident physician reference if needed. After the circumcision the infant was placed back in the bassinet, postoperative vital signs were taken, and salivary cortisol was collected within 20 minutes after the procedure in the same manner they were obtained preoperatively.

Primary outcome

The primary outcome was defined a priori as the change in preoperative and postoperative salivary cortisol level and the CRIES neonatal pain score. Salivary cortisol was obtained as described in previous text and measured using high-sensitivity salivary cortisol enzyme immunoassay (Salimetrics). Minimum sample volumes required for the assay was 25 μL. Briefly, frozen samples were thawed to room temperature, mixed by vortexing, and centrifuged at 1500 × g for 15 minutes. Twenty-five microliters of cleared saliva were assayed in duplicate, according to the manufacturers’ recommended protocol. Cortisol levels were fit against a standard curve.

Samples outside the limits of the standard curve were diluted as needed and assayed again. Levels were reported as micrograms of cortisol per deciliter of saliva. The intra- and interassay coefficients of variation were 3.65% and 6.41%, respectively. Salivary cortisol changes were calculated as absolute and percentage increase or decrease comparing pre- and postoperative samples. The percentage cortisol change was defined as follows: percentage cortisol change = (postoperative-preoperative)/preoperative * 100. For the salivary cortisol analysis, outliers greater than 3.417 micrograms of cortisol per deciliter of saliva were excluded based on the salivary cortisol expected ranges published by the manufacturer that gave an overall range of a nondetectable range of 3.417 micrograms in their study of 275 neonates.

The neonatal pain scale was scored based on CRIES, a validated neonatal pain score ranging from 0 to 10. Two trained research nurses assigned the CRIES scores for all 251 infants as described. The CRIES score assigns points for crying characteristics, oxygen requirements, change in vital signs, facial expressions, and the infant’s sleep state. The minimum and maximum scores of CRIES is 0 to 10, respectively. The CRIES tool was chosen because it includes both behavioral and physiological scores and was developed to be used in term postoperative infants.

Secondary outcomes

Secondary outcomes included the amount of bleeding, length of procedure, parental satisfaction survey, and pediatrician assessment at the first office visit. Bleeding was assessed by subtracting the average clean gauze weight from the weight of the gauzes after the procedure. Prior to randomization we weighed the 2 × 2 gauzes (n = 20) with a balance (H & C Weighing Systems, Columbia, MD). Procedure time was recorded (in minutes and seconds) by the research nurse and was defined as the time when the hemostat was placed on the foreskin to the moment the surgeon removed the clamp.

Parental surveys were conducted by the 2 research nurses via telephone beginning 6 weeks after the circumcision. Three attempts were made to reach each family through 3 postoperative months. Pediatrician surveys were completed by the pediatrician or a designated office staff member via chart review. Three attempts were made to reach the pediatrician office between 6 weeks and 3 months after the circumcision.

Sample size

The sample size was calculated using R statistical software (version 2.15.3; R Core Team 2013) and the extension package samplesize . The original sample size calculation was based on the results of a previous clinical trial of Gomco and Mogen clamps. We considered a 60% reduction (from 0.57 [0.8] μg/dL) in salivary cortisol in the Mogen group compared with the Gomco group to be clinically significant. Based on this assumption, it was determined that 86 patients per arm would be required to detect this difference with 80% power and a significance level of P = .05. To account for up to 20% loss to follow-up and insufficient samples, the sample size was set at 103 per arm for a total sample of 206 infants.

After 6 months of enrollment, an interim analysis was performed and no adverse events were reported. Because of the fact that there is no standard curve for neonatal cortisol level during these procedures, we recalculated the sample size based on our own population’s cortisol levels. Although a mean difference of greater than 60% was seen in this preliminary data, the amount of variance in our study population was larger than that of the previous study. Using the preliminary data (mean difference of salivary cortisol 1.22 μg/dL; Gomco SD, 3.34; Mogen SD, 0.81), it was determined that 238 participants would be required. To account for 20% loss to follow-up, the final sample size was set at 274 infants.

Randomization

Blocked randomization was used to assign participants to the Gomco or Mogen clamp. A randomization schedule matching study identification numbers to clamp type was created by the statistician using a random number generator in Microsoft Excel. Study identification numbers were sequentially assigned to patients upon consenting to the study, which determined the type of clamp used. For the first 106 patients, there were 10 blocks of 10 participants followed by a final block of 6 participants. After the sample size was increased to 274, the final 68 participants were randomized using 6 blocks of 10 participants followed by a final block of 8 participants. The random number generator sorted an equal number of Gomco and Mogen assignments in each block.

Statistical analysis

Normality of all continuous variables was tested using Shapiro-Wilk tests, Q-Q plots, and box plots. All variables were normally distributed. The precortisol levels was the only variable that showed evidence of having borderline nonnormal distributions according to the Shapiro-Wilk test ( P = .03). Because the Shapiro-Wilk test is more apt to give a significant indication of nonnormality with larger sample sizes, it is best used in conjunction with a visual analysis of the Q-Q plots and box plots, which confirmed a close to normal distribution.

A nonparametric analysis was conducted if appropriate. Student t tests or Mann-Whitney U tests were used to compare continuous variables; χ ² tests or a Fisher exact test was used for categorical variables. Linear regression was used to compare the effect of type of clamp while controlling for the duration of the procedure. A Pearson’s correlation was used in the analysis of pain and cortisol levels. All statistics were performed using SPSS, version 22 (SPSS Inc, Armonk, NY).