1 figure omitted
In February 1999, pertussis was diagnosed in six neonates born at hospital A in Knoxville, Tennessee. Because a health-care worker at hospital A was most likely the source of exposure, the local health department recommended on February 25, 1999, that erythromycin be prescribed as postexposure prophylaxis for the approximately 200 infants born at hospital A during February 1-24, 1999. In March 1999, local pediatric surgeons noticed an increased number of cases of infantile hypertrophic pyloric stenosis (IHPS) in the area, with seven cases occurring during a 2-week period. All seven IHPS cases were in infants born in hospital A during February who were given erythromycin orally for prophylaxis following possible exposure to pertussis, although none had pertussis diagnosed. The Tennessee Department of Health and CDC investigated the cluster of IHPS cases and its possible association with use of erythromycin. This report summarizes the results of the investigation, which suggest a causal role of erythromycin in this cluster of IHPS cases.1
IHPS cases occurring during 1997-1999 were ascertained by reviewing medical records in the two area hospitals that provide IHPS treatment. IHPS was defined as a hospital diagnosis of pyloric stenosis (International Classification of Diseases, Ninth Revision, Clinical Modification, code 750.5) that required pyloromyotomy in an infant born in one of the six birthing facilities in the region during 1997-1999. The rate of IHPS cases per 1000 live-born infants for each month was calculated using the number of live-born infants at the six birthing facilities as the denominator. The incidence of IHPS among infants born at hospital A peaked during February 1999 with seven IHPS cases among 217 live-born infants (rate: 32.3 cases per 1000 live-born infants), a rate that was nearly seven times higher than during 1997-1998 (relative risk = 6.8; 95% confidence interval [CI] = 3.0-15.7). No additional IHPS cases were reported among infants born during March-May 1999 at hospital A, and the risk for IHPS in the region returned to the background rates following the peak in February 1999.
To compare the clinical characteristics of the seven index IHPS cases with those of historical IHPS cases, a detailed chart review of IHPS cases from January 1998 through March 1999 was conducted at the two hospitals in the region that had pediatric surgery services. The diagnostic features of the seven index cases were similar to 40 historical cases. Compared with historical cases, index case-patients were younger at the time of admission for IHPS (mean age = 25.6 days versus 35.4 days) and were less likely to have a family history of IHPS (0% versus 17.5%). The mean pyloric thickness and length as measured on ultrasound were similar in the two groups. All index case-patients had received oral erythromycin, compared with none of the historical case-patients.
To validate the IHPS diagnoses, a pediatric radiologist, who was blinded to the original readings, reviewed ultrasound films for the seven index case-patients and seven infants without IHPS. The ultrasound review showed perfect agreement with the original readings (Kappa = 1.0; 95% CI = 0.48-1.0).
A retrospective cohort study of 282 infants born during January-February 1999 at hospital A was conducted to assess a possible association between erythromycin use, gastrointestinal symptoms, and IHPS. In the cohort, 157 infants (55.7%) had a history of oral erythromycin use. The prevalence of erythromycin use was 8.6% among 116 infants born during January 1999 and 88.6% among 166 infants born during February 1999. The erythromycin preparations administered to the infants included ethyl succinate (n = 83), estolate (n = 59), both ethyl succinate and estolate (n = one), and unknown (n = 14). No differences were observed in gastrointestinal symptoms or risk for IHPS in relation to the type of erythromycin preparation.
The infants who were given erythromycin but who did not develop IHPS were aged 1-53 days when they began erythromycin (median age = 13 days; mean = 14.1 days), and the duration of erythromycin exposure ranged from 1 to 21 days (median duration = 14 days; mean = 12.2 days). The seven index IHPS case-patients were aged 2-17 days when they began erythromycin (median = 5 days; mean = 9.3 days), and the duration of their erythromycin exposure ranged from 10-18 days (median duration = 14 days; mean = 13.3 days). Seven IHPS cases occurred among infants who were exposed to erythromycin and none among infants not exposed to erythromycin (relative risk = infinity, lower bound of exact 95% CI = 1.7).
L Patterson, MD, J Peeden, MD, S Sirlin, MD, East Tennessee Children's Hospital; S Hall, MD, IM Himelright, MD, Knox County Health Dept, Knoxville; AS Craig, MD, WL Moore, MD, State Epidemiologist, Tennessee Dept of Health. B Lee, MD, Johns Hopkins School of Medicine, Baltimore, Maryland. Child Vaccine Preventable Diseases Br, Epidemiology and Surveillance Div, National Immunization Program; Birth Defects and Pediatric Genetics Br, Div of Birth Defects, Child Development, and Disability and Health (proposed), National Center for Environmental Health, CDC.
IHPS is a hypertrophy of the pyloric muscle that usually results in nonbilious, projectile vomiting that begins at about 3.5 weeks of age.2 IHPS affects approximately one to three infants per 1000 live-born infants and affects about four to five times as many male as female infants.3,4 Evidence suggests that the pyloric muscle hypertrophy of IHPS develops postnatally.5 The first case reports of a possible association between IHPS and erythromycin in five neonates were published in 1976,6 but the association was considered improbable and had remained unconfirmed. The only subsequent report of this association was a single case report of IHPS in a breastfed infant whose mother had taken erythromycin.7 The findings in this report provide further evidence that erythromycin has a causal role in the etiology of IHPS and raise concerns about the use of erythromycin in neonates.
The peak in IHPS incidence in this region corresponded temporally with the use of erythromycin following the county health department recommendation. All index IHPS case-patients began having symptoms of either vomiting or excessive irritability while taking erythromycin.
The study described in this report is not population-based but includes all live-born infants at facilities in the Knoxville metropolitan area. Local clinicians and public health workers considered it unlikely that an infant born at one of these facilities would be referred outside the region for pediatric surgery, but this possibility cannot be completely eliminated. No evidence indicated a change in case definition, in referral patterns, or in pediatric surgeons or pediatric radiologists that could account for this increase in IHPS incidence. It is unlikely that children with severely hypertrophied pylori would not exhibit symptoms, and evaluation of the pyloric muscle of normal children versus those with IHPS has not demonstrated the existence of severe hypertrophy among asymptomatic children.8 Therefore, it is unlikely that IHPS cases were missed.
Previous epidemiologic studies of IHPS have not identified erythromycin as a risk factor, possibly because few neonates included in such studies were exposed to erythromycin. In most mass prophylaxis situations, the number of neonates treated may be small, possibly explaining why an increased risk for IHPS with erythromycin had not been established.
The prevention of pertussis in infants is important; most hospitalizations for and deaths from pertussis occur in children aged <1 year.9 Although no data exist to confirm a safe and effective alternative to erythromycin for prophylaxis of neonates exposed to pertussis, these findings indicate a need for further examination of recommendations for erythromycin prophylaxis.10 The high case-fatality ratio of pertussis in neonates demonstrates the need to prevent pertussis in this age group, as was done successfully in Tennessee. However, public health officials should continue to use caution in defining risk groups to minimize unnecessary prophylaxis. Physicians who prescribe erythromycin to newborns should inform parents about the possible risks for IHPS and counsel them about signs of developing IHPS.
Cases of pyloric stenosis following use of oral erythromycin should be reported to the Food and Drug Administration (FDA) MedWatch, telephone (800) 332-1088, or through the World-Wide Web, http://www.fda.gov/medwatch. * Additional information on use of erythromycin for treatment of ophthalmia neonatorum and infant pneumonia caused by Chlamydia trachomatis in newborns is available at http://www.cdc.gov/nchstp/dstd/eryth.htm or by fax, (800) 332-0178.
Hypertrophic Pyloric Stenosis in Infants Following Pertussis Prophylaxis With Erythromycin—Knoxville, Tennessee, 1999. JAMA. 2000;283(4):471-472. doi:10.1001/jama.283.4.471