Factors Associated with Early Deaths Following
Neonatal Male Circumcision in the United States, 2001-
Authors: Brian D. Earp1, Veerajalandhar Allareddy2, Veerasathpurush Allareddy3, Alexandre
Affiliations: 1 Yale-Hastings Program in Ethics and Health Policy, Yale University and The Hastings Center
2Division of Pediatric Critical Care, Stead Family Children’s Hospital, University of Iowa, Iowa City, IA;
3Department of Orthodontics, University of Illinois at Chicago College of Dentistry; 4Division of Pediatric
Critical Care Medicine, UH Rainbow Babies & Children’s Hospital, Case Western
Reserve University School of Medicine, Cleveland, OH.
Key words: Circumcision, Predictors of Mortality, Neonates.
We sought to quantify early deaths following neonatal circumcision (same hospital
admission) and to identify factors associated with such mortality. We performed a
retrospective analysis of all infants who underwent circumcision in an inpatient hospital
setting during the first 30 days of life from 2001-2010 using the National Inpatient Sample
(NIS). Over 10 years, 200 early deaths were recorded among 9,833,110 subjects (1 death
per 49,166 circumcisions). Note: this figure should not be interpreted as causal but
correlational as it may include both under-counting and over-counting of deaths attributable
to circumcision. Compared to survivors, subjects who died following newborn circumcision
were more likely to have associated co-morbid conditions, such as cardiac disease (OR:
697.8 [378.5-1286.6] p<0.001), coagulopathy (OR: 159.6 [95.6-266.2] p<0.001), fluid and
electrolyte disorders (OR: 68.2 [49.1-94.6] p<0.001), or pulmonary circulatory disorders
(OR: 169.5 [69.7-412.5] p<0.001). Recognizing these factors could inform clinical and
parental decisions, potentially reducing associated risks.
Key words: circumcision, risks, death, complications, co-morbid conditions
This is the authors’ copy of an accepted manuscript, now published. Citation for the published version:
Earp, B. D., Allareddy, V., Allareddy, V., & Rotta, A. T. (in press). Factors associated with early deaths
following neonatal male circumcision in the United States, 2001-2010. Clinical Pediatrics, Vol. 57, No.
13, 1532-1540. https://doi.org/10.1177/0009922818790060
Version of record: http://journals.sagepub.com/eprint/zkEeqjcANkZQm2xq8gK9/full
Factors Associated with Early Deaths Following
Neonatal Male Circumcision in the United States, 2001-
Male circumcision is a surgery that removes part or all of the penile prepuce or foreskin.1,2 In
rare instances, the procedure is medically indicated, as in patients with recurrent, pathologic
phimosis due to balanitis xerotica obliterans.3,4 Most often, however, it is performed
nontherapeutically on healthy newborns and other male minors.5 In the United States,
circumcision is typically performed by a physician shortly after birth. With an out-of-hospital
rate of approximately 55% as of 2010,6,7 such newborn male circumcision is the most common
pediatric surgery performed in the USA.8 With the exception of Israel, it is much less common in
other Western developed nations.9,10 It is the only pediatric surgery regularly performed on
children who have not been diagnosed with a disease.
Like any surgery, circumcision carries risks. These include infection, bleeding, inflammation,
development of fistulae or skin bridges, acquired torsion, buried penis, meatal stenosis, partial
or complete penile amputation or other injury to the penis, and death.11–13 Death is typically
characterized as a rare complication of circumcision, insofar as the surgery is performed in a
modern clinical setting by a well-trained provider.14 However, precise estimates of death rates
following such medicalized newborn circumcision have been notoriously difficult to establish. In
this study, we attempt to provide a large-sample estimate of the frequency of early deaths (that
is, deaths occurring during the same hospital admission) associated with newborn circumcision,
along with an account of the major factors that are correlated with such mortality.
MATERIALS AND METHODS
Study Design and Description of Database
We performed a retrospective descriptive analysis of the Nationwide Inpatient Sample (NIS) for
the years 2001 to 2010. The NIS is the largest all-payer inpatient database in the United States
that provides information on close to 100 patient and hospital level variables during the index
hospitalization.15 NIS is a component of the Healthcare Cost and Utilization Project (HCUP)
family of databases that include the State Inpatient Databases, Nationwide Emergency
Department Sample, State Emergency Department Databases, Ambulatory Databases, and
Hospital Re-admission Databases. The HCUP is sponsored by the Agency for Healthcare
Research and Quality (AHRQ).
Up to the year 2010, the NIS database was designed such that approximately 20% of all acute
care hospitals provided data. Each of the included hospitals provided information on 100% of
hospitalizations that occurred in the selected years. Each hospitalization in the NIS database is
assigned a discharge weight variable that can be used to project to nationally representative
estimates of approximately all hospitalizations in the United States. Thus, the NIS can be used
to examine close to 20 million hospitalizations occurring in the United States annually.16 The NIS
databases have been extensively validated and to date over 1000 publications have resulted
from its analysis.17 The NIS provides information on variables such as age, sex, race, insurance
status, type of admission, co-morbid conditions, primary reason for hospitalization, procedures
performed during hospitalization, disposition status, hospital charges, length of stay in hospital,
type of hospital, geographic region of hospital, etc. The NIS has been designed to answer
questions pertaining to health care utilization, health costs, and health outcomes, including
those associated with uncommon conditions, procedures that are performed infrequently, or
outcomes that are designated clinically rare events (such as death). Using single-center studies
for examining these issues would be unviable owing to the lower statistical power with limited
samples of data and a lack of generalizability or external validity.
Institutional Review Board (IRB) Approval
The present study used de-identified secondary datasets already in existence. The Office of
Human Subjects Oversight Board at the University of Iowa deemed this project not to meet the
regulatory definition of human subjects research and hence institutional review board (IRB)
approval was not needed.
Males with ages between 0 and 30 days who underwent a circumcision procedure were
selected for inclusion in the present study. Circumcision procedures were identified by using
ICD-9-CM procedure codes in the 15 procedure fields available in the database.18 All subjects
meeting these criteria were selected regardless of race, disposition status, or other related
The following variables were examined in the present study.
Demographics: The demographic variables that were examined included race and insurance
status (any governmental insurance programs, private insurance, or uninsured).
Co-morbid Burden: For each patient in the study cohort, the co-morbid burden severity was
assessed using the NIS co-morbid severity files.15 NIS severity files provide data on the presence
of multiple comorbid conditions (deficiency anemias, rheumatoid arthritis/collagen vascular
diseases, chronic blood loss anemia, congestive heart failure, chronic pulmonary disease,
coagulopathy, diabetes [uncomplicated], diabetes [with complications], hypertension, liver
disease, lymphoma, fluid and electrolyte disorders, metastatic cancer, neurological disorders,
obesity, paralysis, peripheral vascular disorders, pulmonary circulation disorders, renal failure,
solid tumor without metastasis, peptic ulcer disease excluding bleeding, valvular disease, and
weight loss). For the purpose of this study, congestive heart failure, hypertension, and valvular
disease were clustered as cardiovascular disease. The presence of different co-morbid
conditions was coded as a binomial variable: “1” for presence of a specific co-morbid condition
and “0” for absence of a specific co-morbid condition. In accordance with the data user
agreement with HCUP-AHRQ, any individual cell count with ≤ 10 events was not reported.
When such low cell counts were encountered, the term “DS” was used. “DS” refers to
“discharge information suppressed.”
Hospital Characteristics: These included bed size of hospital (small, medium, or large bed-size
hospitals as defined by HCUP-AHRQ), teaching status of hospital (teaching or non-teaching
status as defined by the American Hospital Association), location of hospital (rural or urban
location), and geographic region of hospital.15
Outcome: The outcome of interest in the present study was “early” or in-hospital death. Thus,
for the purpose of this study, early circumcision deaths were deaths that occurred following the
circumcision procedure but prior to discharge, that is, within the same hospital admission as
Simple descriptive statistics were used to summarize the data, including patient and hospital
related variables. The descriptives were summarized by in-hospital mortality status. For
computing nationally representative estimates, the discharge weight variable assigned to each
hospitalization was used. The stratification unit was the stratum to which each hospital
belonged. The stratum to which a hospital belonged was based on its geographic region
(Northeast, Midwest, South, or West), locality (rural or urban), hospital ownership, teaching
status of hospital, and bed size (number of beds in hospital based on geographic region and
location). Descriptive statistics were computed using SAS Version 9.4, SAS Callable SUDAAN
Version 11.0.1, and SigmaPlot Version 13.0 software.
During the study period, we identified 9,833,110 subjects who underwent a circumcision
procedure within the first 30 days of life in an inpatient hospital setting. Among these, there
were 200 deaths within the same hospital admission as the procedure, or 1 death occurring
shortly after every 49,166 circumcisions (10.2 deaths per 500,000 circumcisions). The
distribution of early circumcision deaths per study year is shown in Table 1.
Subjects who died soon after circumcision were predominantly white (63.9%), and had medical
care funded by private insurance (52.1%) or Medicaid (39.7%) (Table 2). Most deaths occurred
at teaching hospitals (80.1%), large bed-size hospitals (59.6%), and those predominantly located
in the South (33.8%) or Midwest (33%) (Table 3).
The most prevalent co-morbid conditions for this sample are shown in Table 4. Compared to
survivors, subjects who died soon after being circumcised were significantly more likely to have
associated co-morbid conditions. The presence of cardiovascular comorbidities was associated
with significantly greater odds of death (OR: 102.9 [61.7 – 171.5] p<0.001). Likewise, the
presence of coagulopathy (OR: 159.6 [95.6-266.2] p<0.001), fluid and electrolyte disorders (OR:
68.2 [49.1-94.6] p<0.001), pulmonary circulatory disorders (OR: 169.5 [69.7-412.5] p<0.001),
and neurological disorders (OR: 111.7 [45.9 – 271.6] p<0.001), were significantly associated
with greater odds of death following circumcision.
Data from this large national cohort of infants who underwent circumcision in an inpatient
hospital setting in the first 30 days of life can help to inform estimates of the risk of early death
following this procedure, especially in patients with co-morbid conditions. In the present
analysis, we identified 1 early death for every 49,166 newborn inpatient circumcisions, or 10.2
early deaths per 500,000 newborn inpatient circumcisions. We stress that this figure refers to
deaths that occurred subsequent to the circumcision but prior to discharge within the same
hospital admission; it should not be taken to refer to the frequency of deaths caused by
newborn circumcision, as the present data cannot directly support such an inference. To see
this, consider the following hypothetical cases:
Case A. Infant Adam is a seemingly healthy newborn infant who has a ritual circumcision
performed as an outpatient. He develops severe bleeding immediately following the
procedure, and is transferred to the local hospital where he requires multiple blood
transfusions. Unfortunately, he does not respond to treatment and dies two days later.
Tests later confirm that he had hemophilia.
Case B. Infant Ben is thought to be a healthy newborn delivered by cesarean section and
has a circumcision on the first day of life. On day 3, he develops refractory shock and
dies. An autopsy reveals an undiagnosed ductal-dependent cardiac lesion (interrupted
Case C. Infant Charlie is circumcised in the first days of his life in a hospital inpatient
setting, but suffers a severe hypoxic-ischemic brain injury due to cardiorespiratory
arrest that occurred during the procedure. Six years later, still with severe neurologic
sequelae, Charlie dies from acute respiratory failure following aspiration pneumonia.
Case D. Infant David is circumcised in the first days of his life in a hospital inpatient
setting, but suffers a complication: too much skin is removed, leaving a tight scar around
his penis, ultimately preventing urination. A month later, David is re-admitted to the
hospital for a correction surgery to relieve his urethral obstruction, but dies due to
complications from anesthesia.
In the case of Adam, the death would be appropriately causally attributed to the newborn
circumcision in conjunction with undetected hemophilia; however, Adam’s death would not be
included in our data set since the circumcision itself was performed in an outpatient setting
whereas we looked only at data from inpatient circumcisions (according to a recent study,
between 1970 and 1996, an average of 45.4% of all circumcisions were non-hospital based or
post-discharge, i.e., “unrecorded”).19,20 Adam, then, would be an instance of undercounting a
death due to newborn circumcision for our purposes.
By contrast, Ben would be an instance of overcounting such a death: although he died during
the same hospital admission as his circumcision, and would thus be included in our data set, it
would not be appropriate to causally attribute the death to the circumcision, because of the
significant confound of the undiagnosed cardiac defect.
Finally, Charlie and David, like Adam, would also represent cases of undercounting: both deaths
are plausibly attributable to circumcision, in that, had the circumcision not occurred, neither
would the deaths from the more immediate causes21 (although this is less certain in the case of
Charlie). However, the deaths occurred sometime after the original hospital discharge and so
would not be included in our data set.
To summarize, our figure of 10.2 early deaths per 500,000 newborn inpatient circumcisions
may (1) include some deaths that were not caused by such circumcision, as well as (2) exclude
some deaths that were caused by such circumcision. It is not possible to determine how many
instances of each of these is represented in our data set. However, it must be emphasized that
our data do not cover outpatient, post-discharge, or ritual circumcisions, which, as noted, may
make up nearly half of all US circumcisions. Thus, any deaths caused by such circumcisions are
invisible to our study. This makes it likely that we have underestimated the total number of
deaths associated with newborn circumcision in the United States during the study period. With
these important caveats in mind, we ask: How does our figure compare with the previous
Prior death estimates
In 1953, Speert reported 1 death from newborn circumcision over the period 1939-1951, based
on records from the New York City Health Department.22 The total number of newborn
circumcisions during this period was unknown, but based on the assumption of a 61.27%
circumcision rate (the rate observed at a single hospital, Sloane Hospital, generalized to all
hospitals in New York City) and the further assumption that this rate remained constant over
the 12 year period, Speert estimated 1 operative death per 566,483 newborn circumcisions
(i.e., 61.27% of the 924,569 known live male births in New York City). This corresponds to <1
death per 500 000 newborn circumcisions; the death was attributed to fatal hemorrhage
following a ritual circumcision performed at-home by a Jewish mohel. The limitations of this
estimate include: (1) it is more than six decades old, and (2) its denominator is based on an
extrapolation from 1 hospital combined with data from a single city. Therefore, generalizability
may be limited.
In 1982, King reported in a commentary that 500,000 consecutive neonatal circumcisions were
performed over an unidentified period of time in New York City without any fatalities, which
entails less than 1 death per 500,000 circumcisions.23 As with the Speert estimate, this figure
concerns a single city, and therefore cannot speak to death rates elsewhere, much less
nationally. In addition, King did not provide any references for his claim, including who
performed the circumcisions, in what setting, with what instruments, etc., rendering it difficult
to interpret. In fact, it is possible that King was simply alluding to the earlier study by Speert
from nearly 30 years earlier.
In 1989, Wiswell and Geschke reported no deaths from circumcision during the first month of
life over a series of 100,157 circumcisions performed at US Army hospitals between 1980 and
1985.24 The authors acknowledge that an unknown number of circumcised boys may have later
been admitted to a civilian hospital, in which case any complications or deaths would not have
been included in their records review. There is also uncertainty about the extent to which such
findings can be generalized to non-Army hospitals during the period in question. Finally, the
estimate is more than three decades old.
In 2010, based on an analysis of US hospital discharge records, Bollinger estimated that 117
deaths related to newborn circumcision occurred in the United States in 2007 out of a total of
1,299,000 circumcisions estimated to be performed that year. This amounts to 9.01 deaths per
100,000 circumcisions or 45.03 deaths per 500,000 circumcisions—far higher than the earlier
estimates as well as our own.25 However, Bollinger’s figure included deaths that occurred
during the entire neonatal period, up to 28 days of life. When only “early” deaths are
included—that is, deaths occurring prior to hospital release, as in our study—Bollinger’s
estimate drops to 14.5 deaths per 1,299,000 circumcision procedures, or 5.6 deaths per
500,000 circumcisions, a figure that is more closely aligned with ours.
Morris et al. have critiqued the estimate by Bollinger, arguing that he must have mistakenly
assumed that sex differences in infant mortality in the United States are entirely due to
newborn circumcision, when in fact similar sex differences occur in countries with low rates of
such circumcision.26 However, Morris et al. appear to have misinterpreted the report by
Bollinger. While Bollinger does refer to a 40.4% relative higher death rate for infant males
compared to infant females in the United States, this is not reported as an overall relative death
rate, but rather one “from causes that are associated with male circumcision complications,
such as infection and hemorrhage,” specifically during the period of “one hour after birth to
hospital release (day 2.4), the time frame in which circumcisions are typically performed.”
Constraints on estimation
One of the major challenges in estimating the number of deaths in which newborn circumcision
is likely to have played a contributory role is that death certificates typically fail to mention the
underlying condition or procedure that led to the subject’s death, and instead list only the
immediate cause of death such as exsanguination or overwhelming sepsis.27 For example, a
study of children with heritable disorders who died while in a pediatric intensive care unit
showed that that the underlying disorder was left off the death certificate 41% of the time.28
Adding to the challenge is a lack of adequate record-keeping in countries where newborn
circumcision remains customary. For example, if complications are treated in a hospital or clinic
other than the one in which the original circumcision was performed, the general inability to
cross-link register data in countries such as the United States that do not have a unique medical
ID for every citizen can make it difficult or impossible to confirm that the complications were in
Yet another problem is “the natural tendency of physicians not to ascribe a poor outcome to an
elective procedure,” which may lead to “gross underestimates” of the frequency of life
threatening infectious complications from circumcision.30 In support of this view, Cleary and
Kohl draw attention to a significant gap in the relevant data (internal references omitted):
In one retrospective series of more than 900,000 male births, there were no deaths
attributed to infection. The incidence of neonatal sepsis is estimated at 1:300 to 1:1,600
live births. With the usual male predominance (approximately 2:1), one might have
expected between 700 and 3,900 in this large study. Since the rate of circumcision
during the years of study was 61%, there must have been many males who were
circumcised and who became septic but in whom the two events were not judged to
have been related.30
According to the American Academy of Pediatrics (AAP), the “true incidence” of surgical
complications attributable to newborn circumcision is unknown.6 This lack of certainty is due to
disagreements about appropriate diagnostic criteria for certain complications as well as other
limitations with existing data, such as those described above.31 Other risks, including risks to the
developing nervous system and long-term risks to neuroendocrine and immune system stress
responses are even less well-studied.32 These issues should be a priority for future research.33
Our study has several limitations inherent to its retrospective design, including the lack of
granularity involving patient-level observations that is typical of a large de-identified dataset.
Nevertheless, we believe that the use of a large multi-year national dataset was the only way to
capture this relatively rare outcome that may follow neonatal male circumcision along with its
associated factors. We chose to include a more remote (2001 to 2010) sample to decrease the
likelihood that one could attempt to re-identify low prevalence events from this de-identified
sample. In addition, the NIS data sampling strategy was constant between 2001 and 2010,
changing in 2012 and once again in 2015 with the implementation of ICD 10 codes.
Another obvious limitation of our study is that it lacks a control group of intact (non-
circumcised) newborns, which would not have been possible given the design of the study.
Including intact neonates as a control group would have captured deaths that occurred in
children with severe co-morbidities that were deemed too ill to undergo circumcision, but who
otherwise would have been circumcised, or children with non-survivable malformations and
those for whom life support measures had been limited. In this context, it should be noted that
circumcision is typically performed only when the physician or other operator believes that the
newborn is stable and healthy enough to undergo surgery. Such judgments are of course
sometimes mistaken, for example when proper screening is not undertaken. However, an
infant that has been deemed to be in sufficiently good health to be operated on, and then dies
shortly after the operation, is relatively unlikely to be one with sufficiently severe co-
morbidities that these would independently explain his death without the operation. Thus,
although we emphasize once again that our findings should be taken as correlational rather
than causal, it is clear that baby boys can and do die as a result of their circumcision
procedures, including in hospital settings.34–44
In keeping with the correlational nature of our results, we observed a significant increase in the
odds of death following circumcision in children with certain co-morbidities, including cardiac
disease, coagulopathy, fluid and electrolyte disorders, and pulmonary circulatory disorders.
Many, if not all of these co-morbidities can be identified a priori in neonates and their presence
should inspire caution as to whether these more vulnerable neonates are indeed able to
tolerate this particular surgical procedure and its attendant consequences (for example, pain
and related sequelae).45–48 Pending further research, we suggest that the specific co-morbidities
we have identified in this study should be considered as potential contra-indications for
In this study, using a validated multi-year national dataset, we identified 1 early death for every
49,166 newborn inpatient circumcisions, or 10.2 deaths per 500,000 newborn inpatient
circumcisions; however, this figure cannot be simply interpreted as a causal relationship, since
our study design allows for both undercounting and overcounting of deaths due to such
circumcisions and we do not know the true proportion of over-counted versus under-counted
In addition, we identified a number of co-morbid conditions that were associated with a
significant increase in the odds of death following newborn circumcision, and we have
suggested that these may need to be considered as potential contra-indications for the surgery
pending further research.
According to the AAP, repeated exposure to painful stimuli soon after birth can cause
immediate and long- term adverse outcomes, including physiologic instability and altered brain
development, as well as abnormal somatosensory and stress response systems, which can
persist into later childhood.48 The AAP therefore recommends that “every health care facility
caring for neonates should implement a pain-prevention program that includes strategies for
minimizing the number of painful procedures performed.”48 Since newborn circumcision is a
surgical procedure that inevitably causes pain as well as carries other risks, up to and including
death, and since it is performed on a healthy subject who does not require surgery on medical
grounds, it is a clear candidate for a procedure whose incidence should be minimized, all things
In this context, it should be remembered that the disvalue of any given risk is a function of its
likelihood as well as its magnitude, which is in turn affected by which alternative courses of
action are available.49 While the most common surgical risks associated with circumcision, such
as pain, bleeding, or infection may be considered tolerable when a surgery is medically
necessary, their disvalue to an individual is much greater when he is healthy and does not
require surgery.50 In addition, while other risks associated with circumcision may be less
common, they may also be extremely serious—as in the case of death—and must therefore be
weighted more heavily in assessments of the value or disvalue of performing the surgery.51
An abstract based on these data was presented at the 2017 National Congress of the American
Academy of Pediatrics in Chicago, IL.52 We thank Professor Dominic Wilkinson of the University
of Oxford for valuable feedback on an earlier draft of this manuscript.
BDE drafted the Introduction and Discussion sections as well as revised the entire manuscript.
VA and VA carried out and reported all data analyses and revised the entire manuscript. ATR
drafted the Methods and Results sections and revised the entire manuscript.
1. Cold CJ, Taylor JR. The prepuce. BJU Int. 1999;83(S1):34–44.
2. Taylor JR, Lockwood AP, Taylor AJ. The prepuce: specialized mucosa of the penis and its loss to circumcision.
Br J Urol. 1996;77(2):291–295.
3. Rickwood AMK. Medical indications for circumcision. BJU Int. 1999;83(S1):45-51. doi:10.1046/j.1464-
4. Sneppen I, Thorup J. Foreskin morbidity in uncircumcised males. Pediatrics. 2016;137(5):e20154340.
5. UNAIDS. Neonatal and child male circumcision: a global review. Jt U N Program HIVAIDS. 2010:1-103.
6. AAP. Male circumcision (technical report). Pediatrics. 2012;130(3):e756-e785. doi:10.1542/peds.2012-1990
7. Maeda JL, Chari R, Elixhauser A. Circumcisions Performed in U.S. Community Hospitals, 2009. Agency for
Healthcare Research and Quality; 2012. https://www.hcup-us.ahrq.gov/reports/statbriefs/sb126.pdf.
8. Witt WP, Weiss AJ, Elixhauser A. Overview of hospital stays for children in the United States, 2012. Healthc
Cost Util Proj. 2006;statistical brief #187:1-17.
9. Morris BJ, Wamai RG, Henebeng EB, et al. Estimation of country-specific and global prevalence of male
circumcision. Popul Health Metr. 2016;14(4):1-13. doi:10.1186/s12963-016-0073-5
10. Wallerstein EJ. Circumcision: the uniquely American medical enigma. Urol Clin North Am. 1985;12(1):123-
11. Frisch M, Simonsen J. Cultural background, non-therapeutic circumcision and the risk of meatal stenosis and
other urethral stricture disease: Two nationwide register-based cohort studies in Denmark 1977–2013. The
Surgeon. 2016:in press. doi:10.1016/j.surge.2016.11.002
12. Krill AJ, Palmer LS, Palmer JS. Complications of circumcision. Sci World J. 2011;11:2458-2468.
13. Lau G, Kim J, Schaeffer A. Identification of circumcision complications using a regional claims database. In:
The Societies for Pediatric Urology. San Francisco; 2018. http://spuonline.org/abstracts/2018/P21.cgi.
14. Hutcheson JC. Male neonatal circumcision: indications, controversies and complications. Urol Clin.
15. NIS. Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project (HCUP). 2001 - 2010. Agency
Healthc Res Qual Rockv MD. 2018. http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed March 6, 2018.
16. HHS. HCUPnet - national statistics on all hospital stays. Agency Healthc Res Qual - US Dep Health Hum Serv.
2018. http://hcupnet.ahrq.gov/. Accessed March 6, 2018.
17. NIS. Publication Search Engine: Nationwide Inpatient Sample (NIS). Healthcare Cost and Utilization Project
(HCUP). Agency Healthc Res Qual Rockv MD. 2018. https://www.hcup-
us.ahrq.gov/reports/pubsearch/pubsearch.jsp. Accessed March 4, 2018.
18. CDC. Classification of diseases, functioning, and disability. International Classification of Diseases, Ninth
Revision, Clinical Modification (ICD-9-CM). US Cent Dis Control Prev. 2015.
www.cdc.gov/nchs/icd/icd9cm.htm#ftp. Accessed March 4, 2018.
19. Morris BJ, Bailis SA, Wiswell TE. Circumcision rates in the United States: rising or falling? What effect might
the new affirmative pediatric policy statement have? Mayo Clin Proc. 2014;89(5):677-686.
20. Jenkins I. Bias and male circumcision. Mayo Clin Proc. 2014;89(11):1588. doi:10.1016/j.mayocp.2014.09.001
21. Darby R. To avoid circumcision complications, avoid circumcision. Can Urol Assoc J. 2014;8(7-8):231.
22. Speert H. Circumcision of the newborn: an appraisal of its present status. Obstet Gynecol. 1953;2(2):164-172.
23. King LR. Neonatal circumcision in the United States in 1982. J Urol. 1982;128(1):1135-1136.
24. Wiswell TE, Geschke DW. Risks from circumcision during the first month of life compared with those for
uncircumcised boys. Pediatrics. 1989;83(6):1011-1015.
25. Bollinger D. Lost boys: an estimate of U.S. circumcision-related infant deaths. Thymos J Boyhood Stud.
26. Morris BJ, Tobian AAR, Hankins CA, et al. Veracity and rhetoric in paediatric medicine: a critique of Svoboda
and Van Howe’s response to the AAP policy on infant male circumcision. J Med Ethics. 2014;40(7):463-470.
27. Svoboda JS, Adler PW, Van Howe RS. Circumcision is unethical and unlawful. J Law Med Ethics.
28. Cunniff C, Carmack JL, Kirby RS, Fiser DH. Contribution of heritable disorders to mortality in the pediatric
intensive care unit. Pediatrics. 1995;95(5):678-681.
29. Frisch M, Earp BD. Circumcision of male infants and children as a public health measure in developed
countries: a critical assessment of recent evidence. Glob Public Health. in press.
30. Cleary TG, Kohl S. Overwhelming infection with group B beta-hemolytic streptococcus associated with
circumcision. Pediatrics. 1979;64(3):301-303.
31. Earp BD, Shaw DM. Cultural bias in American medicine: the case of infant male circumcision. J Pediatr Ethics.
32. Fitzgerald M, Walker SM. Infant pain management: a developmental neurobiological approach. Nat Clin
Pract Neurol. 2009;5(1):35-50. doi:10.1038/ncpneuro0984
33. Earp BD, Darby R. Circumcision, sexual experience, and harm. Univ Pa J Int Law. 2017;37(2-online):1-57.
34. Fagen CR. Tot’s shock hosp death. New York Post. https://nypost.com/2011/05/05/tots-shock-hosp-death/.
Published May 5, 2011.
35. Newell TEC (Coroner). Judgment of Inquiry into the Death of McWillis, Ryleigh Roman Bryan. Province of
British Columbia - Office of the Chief Coroner; 2002. http://www.circumstitions.com/death-exsang.html.
36. Rachter P. Grand jury to probe death of baby after circumcision. The Des Moines Register.
http://www.cirp.org/news/desmoinesregister11-20-82/. Published November 20, 1982.
37. Scurlock JM, Pemberton PJ. Neonatal meningitis and circumcision. Med J Aust. 1977;1(10):332-334.
38. Vogel-Hockley S. Death of local baby subject of investigation. The Penticton Western.
http://www.cirp.org/news/thewestern08-31-02/. Published August 31, 2002.
39. Curtis P. Police probe death of baby after circumcision. The Guardian.
http://www.theguardian.com/uk/2007/feb/17/religion.world1. Published February 17, 2007. Accessed
March 6, 2018.
40. Hiss J, Horowitz A, Kahana T. Fatal haemorrhage following male ritual circumcision. J Clin Forensic Med.
41. Offit PA. Ban the ritual that can kill Jewish newborns. The Daily Beast.
April 22, 2017. Accessed July 22, 2017.
42. Robbins L. Infant’s death renews debate over ritual circumcision. The New York Times.
ritual.html. Published March 7, 2012. Accessed March 6, 2018.
43. Edler G, Axelsson I, Barker GM, Lie S, Naumburg E. Serious complications in male infant circumcisions in
Scandinavia indicate that this always be performed as a hospital-based procedure. Acta Paediatr.
44. Sauer LW. Fatal staphylococcus bronchopneumonia following ritual circumcision. Am J Obstet Gynecol.
45. Frisch M, Simonsen J. Ritual circumcision and risk of autism spectrum disorder in 0- to 9-year-old boys:
national cohort study in Denmark. J R Soc Med. 2015;108(7):266-279. doi:10.1177/0141076814565942
46. Macke JK. Analgesia for circumcision: effects on newborn behavior and mother/infant interaction. J Obstet
Gynecol Neonatal Nurs. 2001;30(5):507-514. doi:10.1111/j.1552-6909.2001.tb01570.x
47. Taddio A, Katz J. The effects of early pain experience in neonates on pain responses in infancy and childhood.
Pediatr Drugs. 2005;7(4):245-257. doi:10.2165/00148581-200507040-00004
48. AAP. Prevention and management of procedural pain in the neonate: an update. Pediatrics.
49. Allhoff F. Risk, precaution, and emerging technologies. Stud Ethics Law Technol. 2009;3(1):1-27.
50. Hutson JM. Circumcision: a surgeon’s perspective. J Med Ethics. 2004;30(3):238-240.
51. Fox M, Thomson M. Short changed? The law and ethics of male circumcision. Int J Child Rights.
52. Earp BD, Allareddy V, Allareddy V, Rotta AT. Factors associated with early deaths following neonatal
circumcision. Pediatrics. 2018;42(1-meeting abstracts):224-224.
FIGURES AND TABLES
Table 1. Annual distribution of early deaths following neonatal circumcision
DS = Discharge information suppressed due to low cell count (as per data user agreement with AHRQ)
Table 2. Insurance status and demographic information
DS = Discharge information suppressed due to low cell count (as per data user agreement with AHRQ)
Missing race information
Missing insurance information
Table 3. Hospital characteristics
Table 4. Co-morbid condition and odds of death following circumcision
DS = Discharge information suppressed due to low cell count (as per data user agreement with AHRQ)
Odds Ratio (95% CI)
102.9 (61.7 – 171.5)
159.6 (95.6 – 266.2)
Fluid and electrolyte
68.2 (49.1 – 94.6)
169.5 (69.7 – 412.5)
111.7 (45.9 – 271.6)