 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
PEDIATRICS Vol. 105 No. 1 Supplement January 2000, pp. 246-249
,,, 
From the * Department of Pediatrics, University of Washington;
Child Health Institute, University of Washington; § Robert Wood
Johnson Clinical Scholars Program, University of Washington;
Department of Health Services, University of Washington, Seattle,
Washington.
 |
ABSTRACT |
|---|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Background. The risks associated with newborn circumcision have not been as extensively evaluated as the benefits.
Objectives. The goals of this study were threefold: 1) to derive a population-based complication rate for newborn circumcision; 2) to calculate the number needed to harm for newborn circumcision based on this rate; and 3) to establish trade-offs based on our complication rates and published estimates of the benefits of circumcision including the prevention of urinary tract infections and penile cancer.
Methods. Using the Comprehensive Hospital Abstract Reporting System for Washington State, we retrospectively examined routine newborn circumcisions performed over 9 years (1987-1996). We used International Classification of Diseases, Ninth Revision codes to identify both circumcisions and complications and limited our analyses to children without other surgical procedures performed during their initial birth hospitalization.
Results. Of 354 297 male infants born during the study period, 130 475 (37%) were circumcised during their newborn stay. Overall 287 (.2%) of circumcised children and 33 (.01%) of uncircumcised children had complications potentially associated with circumcision coded as a discharge diagnosis. Based on our findings, a complication can be expected in 1 out every 476 circumcisions. Six urinary tract infections can be prevented for every complication endured and almost 2 complications can be expected for every case of penile cancer prevented.
Conclusions. Circumcision remains a relatively safe procedure. However, for some parents, the risks we report may outweigh the potential benefits. This information may help parents seeking guidance to make an informed decision. Key words: newborn circumcision, complications, urinary tract infections.
Routine circumcision of newborn infants remains
controversial.1-3 Although a recent policy statement by
the American Academy of Pediatrics recommended that parents be
given accurate and unbiased information regarding the risks and
benefits of the procedure,3 accomplishing this task of
well informed consent is hindered both by the lack of precise information about potential harm and by the lack of clear ways to
present this information.
The benefits of circumcision have been described in numerous previous
studies using a variety of methodologies.4-8 Reported
benefits include reduction in the risk of penile cancer,9 urinary tract infections (UTIs),4,5 and sexually transmitted diseases.10,11 Although the extent to which
circumcision decreases the risk of each of these outcomes has been
debated,612-15 a consensus appears to be emerging that
there are some small protective effects.3
These protective effects of circumcision (or any other therapy) can be
meaningfully conveyed in terms of a number needed to treat (NNT). The
NNT is calculated from the reciprocal of the absolute risk reduction
associated with a given treatment.16 In the case of
circumcision, this number represents the number of children who would
need to be circumcised to prevent 1 undesired outcome such as
UTI.
Although NNTs inform both providers and patients of the benefits of a
given therapy or procedure, by themselves the NNTs present only half of
the pertinent equation. What is also needed to make an informed
decision is the number needed to harm (NNH). By analogy to NNTs, NNHs
are based on the absolute difference in complication rates between
treatment and control groups and tell one in effect how many patients
would need to be circumcised before an adverse event can be expected to
occur. Taken together, NNTs and NNHs can be used to construct a
framework of "trade-offs"; in essence, one can compute how many
undesirable outcomes will be prevented per complication incurred.
Trade-offs can be useful to patients and providers seeking to weigh the
pros and cons of a given procedure.
Although existing data are sufficient to enable us to estimate the NNT
with circumcision for given outcomes, the risks of the procedure have
received considerably less attention, hindering our ability to
calculate NNHs. Previous studies of newborn circumcision complications
have reported rates between .19 and .60%, but because these data were
either derived from single institutions or from the children of army
recruits they are not population-based.6,17,18 Therefore
we undertook a large, retrospective, population-based cohort study of
circumcision of newborn infants. Our goals were threefold: 1) to obtain
population-based estimates of complications associated with
circumcision; 2) to convert these complication rates into NNHs; and 3)
to use these NNHs in conjunction with published data of the benefits of
the procedure to develop a framework for presenting NNT versus NNH
trade-offs to better inform both practitioners and parents about
routine newborn circumcision.
Data Source
We used the Comprehensive Hospital Abstract Reporting System
(CHARS) to conduct a retrospective analysis of 9 years (1987-1996) of
routine newborn circumcisions performed in hospitals throughout Washington State. The CHARS contains International Classification of Diseases, Ninth Revision diagnosis and procedure codes for all
discharges from short-stay hospitals in Washington State including birth hospitalizations. Each patient is assigned a unique yet anonymous
identifier, enabling the analysis of patient level data without
compromising confidentiality.
Patients
A retrospective cohort was assembled by using the CHARS to
identify all newborn male infants who had circumcision coded as a
procedure during their birth hospitalization. We excluded children who
had any other surgical procedure documented in the database to ensure
that all surgical complications we discovered were appropriately attributable to the circumcision procedure. Using the unique patient identifiers, we ensured that our analysis counted individual patients who had 1 or more complications, and not simply complications themselves.
NNTs to Prevent Undesired Outcomes
Based on published reports in the literature, we calculated NNTs
for several outcomes. In the case of UTI in the first year of life, the
NNT is 100 based on data from previous cohort
studies19-21; that is, 100 children need to be
circumcised to prevent 1 UTI. For penile cancer, although the
association with circumcision status remains
controversial,1222-24 conservative estimates (based on
life table analyses and assuming circumcision to be 100% effective at
preventing penile cancer), suggest a NNT of 909.24,25 Data
for the reduction of the risk of human immunodeficiency virus
exist, but have been derived from very different populations than US
adults and have led to conflicting conclusions.11,26,27 Data from a sexually transmitted disease clinic suggest that
circumcision reduces the risk of contracting syphilis,10
although results from a population-based study appear to refute this
association.28 Moreover, sexual behavioral practices
remain the most important modifiable risk factor for sexually
transmitted diseases.3 Therefore, we focused on the
potential role of circumcision in the prevention of UTIs and penile
cancer. These 2 outcomes also usefully frame all others: UTI is
the most prevalent serious outcome deemed modifiable through
circumcision and penile cancer is the least prevalent one.
NNH Based on Adverse Events Related to Circumcision Procedure
We limited our analysis to complications coded only during the
postpartum hospital stay, because the overwhelming majority of
complications related to circumcision occur shortly after the procedure,10,11,17,18 and because we could not ascertain
whether children who were not circumcised during their initial newborn
stay were not subsequently circumcised as outpatients. However, we did
include children who were transferred from one hospital to another and
thus were continuously hospitalized from birth. Based on previous
reports of complications attributable to
circumcision,6,1829-31 we searched for specific
International Classification of Diseases, Ninth Revision
codes during postpartum hospitalizations. These codes are presented in
Table 1.
TABLE 1
METHODS
Top
Abstract
Methods
Results
Discussion
References
Complications Attributable to Newborn Circumcision Among Circumcised
and Noncircumcised Males During the Birth Hospitalization
Analysis
2 was used to compare proportions of
circumcised and uncircumcised children with complications. Exact tests
were used where indicated based on expected cell frequencies.
Independent sample t tests were used to compare mean lengths
of stay. NNH were calculated based on the reciprocal of the
absolute risk difference between complications in the circumcised
versus the uncircumcised cohorts. Trade-offs were calculated based on
the number of complications that could be expected conditioned on
circumcising 100 children to prevent 1 UTI and 909 children to prevent
1 case of penile cancer.
| |
RESULTS |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
There were 354 297 male births during the study period. Of these, 130 475 (37%) had circumcision coded during the birth hospitalization and had no other procedures coded. Various complications occurred during the birth hospitalization, as summarized in Table 1. Overall, there were 287 complications (.2%) in the circumcised children and 33 (.01%) in the uncircumcised group. Circumcised children with any complication had significantly longer newborn stays than circumcised children without complications (mean: 2.81 days vs 2.26 days; P = .003) although the only single complication associated with an increased newborn stay was "suture penile laceration" (644.1; 7.0 days vs 2.26 days; P < .001).
Based on these findings, the aggregated NNH is 476: that is, some form of complication occurs in 1 out of every 476 children who are circumcised. The NNH for each of the separate complications is also displayed in Table 1.
The NNT versus NNH trade-offs, each representing a discrete comparison of a particular benefit and a particular harm, are summarized in Table 2. Overall, 1.14 immediate circumcision-related complications can be expected for every 6 UTIs prevented, and 1.90 complications can be expected for every case of penile cancer prevented.
|
| |
DISCUSSION |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
Because circumcision is not a lifesaving procedure,6 and its modest medical benefits may be offset by its complications, its indications are discretionary. Parents who opt to circumcise their child for religious or cultural reasons do so appropriately within this discretionary boundary. For those parents seeking objective medical advice regarding the procedure, our trade-off analysis is intended to provide parents and providers with a sense of the relative medical risks and benefits of the procedure.
Our study found an overall complication rate associated with newborn circumcision of approximately .20%, or 1 adversely affected child for every 476 boys circumcised. We believe this estimate to be an accurate reflection of the true complication rate, as it is based on a large and diverse patient sample, and it is consistent with 2 previous studies that reported rates of .19% and .20%6,18 although it is significantly less than that of another which reported a rate of .60%.17 In part, this inconsistency may be accounted for by the method of data collection as some of these studies involved chart reviews that may be more apt to discern minor complications not coded on discharge.
Beyond this summary of the overall complication rate, our analysis also provides separate NNH estimates for different kinds of complications, which is important because the complications clearly are not all of equal consequence, medically or emotionally. Similarly, the value placed on preventing a UTI in an infant or a case of penile cancer in late adult life also are best reported as 2 distinct NNT figures, for they too are not necessarily of equal consequence.
Parents' subjective weighing of the benefits of prevention and the
harm of complications is fundamental to this decision-making process.
Accordingly, we did not assign relative weights to the outcomes.
Further, we believe that attempts to designate utilities for these
outcomes
whether they be based on expert panels or community surveyswould be misguided. The weighing process in this decision should remain individualized and subjective, taking fully into account
the parents' general degree of aversion to risk, and in particular
whether the risk arises from either omitted or committed actions.
Parents might well have greater feelings of guilt associated with
adverse events arising from circumcision, such as needing a penile
wound repaired, than from a different adverse and somewhat preventable
event, such as a UTI, occurring in uncircumcised boys. This aversion to
committed action risk and its associated feelings may counterbalance or
outweigh any potential benefits.
Two specific limitations of this study warrant consideration. First, as with any large administrative database study, the records likely contain some incorrect or missing codes. Because it is both common and reimbursable, circumcisions were likely to be accurately coded in the database when they were performed. However, at first glance, our reported circumcision rate of 37% may seem low. Unpublished estimates suggest circumcision rates of approximately 50% for the state of Washington. However, our data collection method would only capture children who were circumcised during their birth hospitalization and not those subsequently circumcised as outpatients. Nevertheless it is possible that circumcisions were performed but not coded as was possibly the case with the 1 uncircumcised boy for whom "penis wound" was coded. Misclassifications of this kind would have biased our findings toward the null. We attempted to control for other misclassifications, either other procedures performed but not coded as may have been the case for children with interoperative bleeding in the absence of other documented procedures or miscoding of complications, by using a comparison group of "uncircumcised" children. For misclassifications then to have meaningfully impacted our results, they would have to have occurred a greater proportion of the time in one group versus the other. There is no a priori reason to believe this to be the case.
Second, because our sample was limited to the immediate postbirth observation period, we did not capture the very rare but serious and delayed complications associated with circumcisions (eg, necrotizing fasciitis, cellulitis). Our analysis also did not capture the less serious but far more common "complications," such as the healthy but unsightly healing of a circumcision that often unnerves parents, or a less than ideal cosmetic result, either of which may lead parents to seek medical attention. Our results therefore represent conservative estimates of complications attributable to circumcision.
Overall, routine newborn circumcision appears to be a relatively safe procedure. It is not without some risks, however, and these risks do not seem to be mitigated by the hands of more experienced physicians.32 Given that these risks then are likely to remain, we hope that the methods we used to present both the risks and the benefits as trade-offs are both meaningful and helpful.
| |
ACKNOWLEDGMENTS |
|---|
We thank Frederick P. Rivara, MD, MPH, for his thoughtful review and to Alta Lyn Bassett for her assistance with manuscript preparation.
| |
FOOTNOTES |
|---|
Dr Christakis is a Robert Wood Johnson Generalist Physician Faculty Scholar.
Received for publication Jun 9, 1999; accepted Sep 1, 1999.
Address correspondence to Dimitri A. Christakis, MD, MPH, Child Health Institute 146 N Canal St, Suite 300, Seattle, WA 98103. E-mail: dachris{at}u.washington.edu
| |
ABBREVIATIONS |
|---|
UTI, urinary tract infection; NNT, number needed to treat; NNH, number needed to harm; CHARS, Comprehensive Hospital Abstract Reporting System.
| |
REFERENCES |
|---|
|
Top
Abstract
Methods
Results
Discussion
References
|
|---|
a rare complication.
Br J Urol.
1996;
77:924-925 [Medline]This article has been cited by other articles:
 |
|
 |
|
  D. Andres Should newborns be circumcised?: NO Can Fam Physician, December 1, 2007; 53(12): 2097 - 2099. [Full Text] [PDF]
|
|
|
|
|
|
D. Andres Faut-il circoncire les nouveau-nes?: NON Can Fam Physician, December 1, 2007; 53(12): 2101 - 2103. [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Fergusson, J. M. Boden, and L. J. Horwood Circumcision Status and Risk of Sexually Transmitted Infection in Young Adult Males: An Analysis of a Longitudinal Birth Cohort Pediatrics, November 1, 2006; 118(5): 1971 - 1977. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S Van Howe, J S. Svoboda, and F. M Hodges HIV infection and circumcision: cutting through the hyperbole The Journal of the Royal Society for the Promotion of Health, November 1, 2005; 125(6): 259 - 265. [Abstract] [PDF]
|
|
|
|
 |
|
 R. S. Van Howe A Cost-Utility Analysis of Neonatal Circumcision Med Decis Making, November 1, 2004; 24(6): 584 - 601. [Abstract] [PDF]
|
|
|
|
|
|
R. Frank;, D. Harrison;, D. A. Christakis, E. Harvey, D. M. Zerr, C. Feudtner, J. A. Wright, and F. A. Connell A Trade-off Analysis of Routine Newborn Circumcision Pediatrics, October 1, 2000; 106(4): 954 - 954. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2000 American Academy of Pediatrics. All rights reserved. |
||
|
||
