After two serious complications had alerted physicians to the potential risks of routine neonatal circumcision, the circumcision rate in a regional general hospital decreased significantly (p < 0.001), from approximately 40% to 20%, settling at the level claimed to prevail when physicians oppose circumcision. Following this decrease 219 consecutive pregnancies resulting in male infants were prospectively studied in order to identify factors associated with insistence on circumcision. Factors significantly associated with circumcision were the existence of an older brother (p < 0.001), especially if circumcised (p < 0.001), and delivery of prenatal care by an obstetrician rather than a general practitioner (p < 0.05). Factors significantly associated with no circumcision were the fact that this infant was the first male born in the family (p = 0.001), delivery of prenatal and infant care by the same general practitioner (p < 0.05) and a maternal age of 20 years or less (p < 0.02). The circumcision status of the father, the marital status of the mother, the mother's intention to breast-feed, attendance of a primigravida at prenatal classes, delivery of infant care by a pediatrician and socioeconomic status did not appear to influence whether circumcision was performed. These data may assist physicians in understanding the potential effectiveness and limitations of counselling against circumcision.

Après que deux complications sérieuses eurent alerté les médecins sur les risques potentiels de la circoncision néonatale de routine, le taux de circoncision dans un hôpital général régional s'est abaissé significativement (p < 0.001), passant de 40% à 20% environ, et se stabilisant au niveau qu'on considère exister quand les médecins s'opposent à la circoncision. Suite à cette diminution 219 grossesses consécutives d'enfants mâles ont été etudiées de manière prospective dans le but d'identifier les facteurs qui contribuent à ce qu'on insiste pour obtenir une circoncision. Les facteurs qui étaient associés de façon significative avec la circoncision étaient l'existence d'un frère aîné (p < 0.001), surtout s'il avait été circoncis (p < 0.001), et la distribution des soins prénatals par un obstetricien plutôt que par un praticien général (p < 0.05). Les facteurs associés de façon significative avec l'absence de circoncision étaient le fait que l'enfant soit un premier né de sexe masculin (p = 0.001), la distribution des soins prénatals et néonatals par le même praticien général (p < 0.05) et un âge maternel de 20 ans ou moins (p < 0.02). La circoncision du père, la situation de famille de la mère, l'intention de la mère d'allaiter, la participation d'une primigeste à des cours prénatals, le recours à un pédiatre pour les soins du nourrisson et la situation socio-économique ne semblaient pas influencer la décision de circoncire. Ces résultats sont susceptibles d'aider le médecin à comprendre le succès et les limites qu'il peut envisager quand il déconseille la circoncision,

Routine neonatal circumcision remains a controversial procedure in North America. Circumcision rates are reported to vary between 69% and 97% in the United States[1] and between 2% and 70% in Canada.[2] Despite the statements of the American Academy of Pediatrics that routine circumcision of the newborn is not medically indicated,[3,4] and the recommendations of the Canadian Paediatric Society against routine neonatal circumcision,[5,6] these rates appear to have remained relatively constant over the past decade.[7,8] The strongest determinant of whether a male infant is circumcised appears to be the attitude of the attending physician toward circumcision.[9] In one study the circumcision rate was 100% when the physician favoured circumcision, compared with 20% when the physician opposed the procedure.[9] The following study was undertaken after two surgical mishaps increased physician opposition to circumcision, resulting in a sudden decline in circumcision rates from approximately 40% to 20%. The study attempted to identify characteristics of families that correlated with either insistence on circumcision or acceptance of no circumcision.

Methods

The study was performed in the Dr. Everett Chalmers Hospital, Fredericton, an acute care regional genera! hospital with approximately 1500 births per year. The population served by the hospital is predominantly Caucasian and Protestant. During the period of the study pregnant mothers had unrestricted access to 5 active obstetricians and more than 20 general practitioners for maternity care. Prenatal classes were available to all expecting couples or women. Most newborn care was delivered by general practitioners. The four pediatricians in the hospital delivered newborn care to infants born to women without a general practitioner on staff, but usually they did not accept new primary care cases from the local community. Neonatal circumcisions were performed by general practitioners and pediatricians, but only when requested by parents. Obstetricians did not perform circumcisions.

A survey of parental attitudes performed at the start of the study showed that the main reasons given by parents who wanted their infants circumcised were: to avoid later circumcision (38%), to facilitate penile hygiene (29%), because of a perceived cosmetic or sexual advantage (14%), because they thought the procedure was routine (14%) and because other members of the family were circumcised (7%).

The monthly circumcision rates for the 28-month period commencing January 1979 are shown in Fig. 1. The period is divided into four 6-month epochs and one 4-month epoch, and the average rate for each epoch is given.

At the point depicted by the arrow two male infants required plastic surgery following unintentional denuding of the shaft of the penis during circumcision. Although these mishaps were not publicized outside the hospital, they focused the attention of the hospital administration and medical staff on the potential hazards of circumcision and resulted in a formal proposal to suspend neonatal circumcisions. After deliberation the proposal was rejected on the grounds that suspension of circumcisions might result in an adverse reaction among the segment of the community served by the hospital who wanted their infants circumcised. Simultaneously the circumcision rate declined, so that the average rates for the last two epochs (18.9% and 18.6%) were significantly lower (p < 0.001) than the average rates during the first two epochs (42.9% and 39.9%). The decline in rates is assumed to have occurred because physicians who were previously ambivalent about circumcision had now been motivated to oppose the procedure.

The study was undertaken when the resistance to further decline in the circumcision rate became apparent. Data were prospectively collected on 220 consecutively born male infants to yield a study group of 50 circumcised infants. Of the 170 uncircumcised infants, 1 born to Jewish parents and ritually circumcised after discharge from hospital was excluded from the control group. The data were reviewed at discharge, and if necessary the mother was immediately contacted by telephone to obtain missing information.

Socioeconomic status was determined from the occupation of the father, or of the mother if she was single, and scored according to the 1976 revision of the Blishen Scale.[10] This scale, based on Canadian census data, links occupation with income and education levels. The socioeconomic score thus obtained is divided into six intervals. Unemployed individuals were assigned to group 6.

Chi-square analysis and analysis of variance were used to determine whether significant differences in the distribution of study variables existed between the circumcised and control groups, with a p value of 0.05 considered significant. Yates's correction was used in all chi-square calculations when the number in any contingency table cell was less than 10.

Results

The study variables that were significantly more often associated with circumcision are listed in Table I. The data show that all of the older brothers of the circumcised infants had been circumcised, compared with only 15 (34%) of the 44 older brothers of the uncircumcised infants (p < 0.001).

Table II lists the variables that were significantly less often associated with circumcision. Although there was no significant difference in mean maternal age between the two groups (25.0 years for the uncircumcised group v. 25.8 years for the circumcised group), the age distribution of the mothers of the uncircumcised group was skewed toward youth.

Table III lists the variables that were independent of circumcision. Of particular interest are the identical rates of circumcision (64%) of the fathers of the two groups of infants.

As Fig. 2 shows, the distribution of socioeconomic status was similar for the circumcised and uncircumcised groups.

Discussion

Very little is known about factors associated with nonritual neonatal circumcision in North America. Apart from religious belief the only determinants of circumcision identified to date are the cultural background[7,11] and socioeconomic status[9] of the parents and, most powerful of all, the attitude of the attending physician.[8,9] The factors that predicate physician attitude to circumcision are variable and often associated with an imprecise knowledge of the medical rationale for circumcision.[8] The attitudes of the physicians who were asked by parents to perform circumcision in this study ranged from total opposition to full cooperation. However, circumcision was never performed unless the parents requested it. The study was undertaken when it became apparent that a significant change in the neonatal circumcision rate had occurred in temporal relation to a series of unusual events that may have increased opposition to circumcision in physicians who had not previously held a strong opinion about the procedure. The study did not attempt to measure initial parental pressure for circumcision or the extent to which physician opposition was effective. However, the circumcision rate of 23% during the study period is similar to the 20% rate claimed by Patel[9] to prevail when physicians are opposed to circumcision, suggesting that circumcision only occurred as a result of parental insistence.

This study provides new data that identify characteristics of families whose sons are likely to be circumcised compared with families from the same population whose sons are not. The factor most strongly associated with circumcision was the existence of a circumcised older brother. In every family in which a subsequent male infant was circumcised his older male sibling had been circumcised. Conversely, there was a significant association between the infant's being the first male child of the family and no circumcision.

Significantly more parents of the circumcised infants chose an obstetrician rather than a general practitioner for prenatal care. The factors underlying this association remain speculative. The obstetricians in this study did not perform circumcision, nor did they include advice on circumcision during routine prenatal care. However, when asked about circumcision by prospective parents they indicated that it would be available on request. This may have created an expectation of circumcision in some parents, thus reducing the effectiveness of any postnatal counselling against circumcision by the infant's physician. This is consistent with the finding that circumcision was independent of delivery of infant care by a pediatrician, whose relationship with the parents usually started only after the birth of the infant.

Surprisingly, perhaps, circumcision was independent of the reported circumcision status of the father. Despite reports that many mothers are unaware of the circumcision status of their partner,[12] none of the mothers in this study were unable to provide this information. However, without physical examination of the infant's father it is impossible to be certain that the mother reported the status correctly.

Intention to breast-feed and attendance at prenatal classes were chosen as variables in an attempt to assess whether attitude to pregnancy was related to circumcision. The absence of an association between attendance at prenatal classes and circumcision may reflect, in part, the fact that circumcision was deliberately excluded from the prenatal curriculum in an effort to avoid confrontation between prenatal educators and physicians holding different opinions about circumcision.

Although Patel[9] affirmed that socioeconomic status influences the likelihood of circumcision, the findings of this study indicate that this is not the case in this population. The data on which Patel based his statement were not presented in his paper. In this study socioeconomic status was derived from a scale based on census data directly relevant to the study population.[10] The effects of religious and cultural factors on circumcision[7,11] may have been erroneously attributed to socioeconomic differences in Patel's study. The absence of religious and cultural heterogeneity in this study population may allow more precise appreciation of the absence of an association between socioeconomic status and circumcision.

In summary, these data indicate that when circumcision is requested by parents and subsequently negotiated with physicians who are generally not in favour of the operation the circumcision status of the infant's brother is highly predictive of the outcome. Moreover, when viewed against the relatively stable circumcision rates in other North American populations,[7] the significant local change in rate that preceded this study suggests that both the fact that the child is the first son and low maternal age, which were strongly associated with the decision not to circumcise, are specific indicators of susceptibility to counselling against circumcision rather than indicators of a fall in the demand for circumcision. Thus, it seems that a selective approach to counselling against neonatal circumcision aimed at the young mother or the mother with a first male infant will result in a progressive fall in circumcision rates while at the same time safeguarding freedom of choice in families who already have one or more circumcised sons. It remains to be shown whether obstetricians and prenatal educators might influence circumcision rates further by specific prenatal counselling.

I thank Dr. Ada Steele, who helped with the attitudinal survey, and Kathrine Peters, RN, BNSc, who helped with the data analysis.

References
 

1.	Kaplan GW. Circumcision—an overview. Curr Probl Pediatr 1977; 7: 3-33.
2.	Wirth JL. Current circumcision practices—Canada. Pediatrics 1980; 66: 705-708.
3.	American Academy of Pediatrics. Committee on fetus and newborn. Lucey JF (chmn). Hospital Care of Newborn Infants, 5th ed, Am Acad Pediatr. Evanston, Ill. 1971; 110.
4.	Thompson HC, King LR, Knox E, Korones SB. American Academy of Pediatrics. Committee on fetus and newborn: Report of the ad hoc task force on circumcision. Pediatrics 1975; 56: 610-611.
5.	Canadian Paediatric Society, fetus and newborn committee. Circumcision in the newborn period. CPS News Bull Suppl 1975; 8 (2).
6.	Fetus and newborn committee, Canadian Paediatric Society. Benefits and risks of circumcision: another view. Can Med Assoc J 1982; 126: 1399.
7.	Warner E, Strashin E. Benefits and risks of circumcision. Can Med Assoc J 1981; 125: 967-976,992.
8.	Patel DA, Flaherty EG, Dunn J. Factors affecting the practice of circumcision. Am J Dis Child 1982; 136: 634-636.
9.	Patel H. The problem of routine circumcision. Can Med Assoc J 1966; 95: 576-581.
10.	Bushen BR, McRoberts HA. A revised socioeconomic index for occupations in Canada. Can Rev Soc Anrhropol 1976; 13: 71-79.
11.	Grossman E, Pooner NA. Surgical circumcision of neonates: a history of its development. Obstet Gynecol 1981; 58: 241-246.
12.	Grimes DA. Routine circumcision of the newborn infant: a reappraisal. Am J Obstet Gynecol 1978; 130: 125-129.