Secondary Logo

Journal Logo

The Continuing Enigma of Pyloric Stenosis of Infancy: A Review

MacMahon, Brian

doi: 10.1097/01.ede.0000192032.83843.c9
Review Article
Free

Striking features of the descriptive epidemiology of pyloric stenosis of infancy have been identified but until recently have not suggested any useful etiologic lead. The disease first received serious attention approximately 100 years ago and has since occurred throughout the Western world at a rate between 2 and 5 per thousand live births; it appears to be uncommon elsewhere. Its age distribution is essentially limited to the period between the third and eighth weeks after birth. It is 4 to 5 times more common in boys than girls. It is less common in blacks than whites in the United States and less common among Asians than whites in the United States and elsewhere. Its incidence is highest in first-born infants. Evidence on a role for maternal age is not consistent. The disease reoccurs in families with sufficient frequency to incite the interest of geneticists, although no genetic model yet proposed offers a better basis for counseling than do the empiric observations on which it is based. Monozygous twins are concordant for the disease not much more frequently than are dizygous twins, and indeed not a great deal more often than nontwin siblings, which should prompt a search for environmental explanations of the disease's familial nature. Sharp declines in the incidence of the disease in Denmark and Sweden during the 1990s led to the hypothesis that infants sleeping in the prone position—a practice discouraged with some success by Scandinavian campaigns to reduce the frequency of sudden infant death syndrome (SIDS)—may also be at increased risk of pyloric stenosis. If supported, this hypothesis may offer the first-ever possibility of reducing the frequency of this disease, as well as SIDS. If the hypothesis is not supported, the recent declines in the disease in Denmark and Sweden add another facet to its enigmatic nature.

From the Department of Epidemiology, School of Public Health, Harvard University, Boston, MA.

Editors’ note:A commentary on this article appears on page 136.

Correspondence: Brian MacMahon, 89 Warren Street, Needham, MA 02192. E-mail: BrianMacMahon@aol.com.

Pyloric stenosis of infancy—often described as hypertrophic—has a dramatic onset, usually in a healthy infant between the third and eighth weeks of life, and characterized by forceful (projectile) vomiting and a palpable mass (tumor or olive) in the upper abdomen. At laparotomy, the circular muscle of the pylorus is found hypertrophied, and after surgical incision of this muscle in the line of the gut (an operation described by Ramstedt in 19121), recovery is rapid. Before Ramstedt's operation became routine, fatality rates around 30% were common, but with improved anesthetic, surgical, and biochemical techniques, the fatality rate is now usually less than 1%.

The frequency of pyloric stenosis is best described by the rate of its occurrence among live-born infants. The temporal boundaries of this rate are in terms of age rather than calendar time, but because of the limited period of risk, it is most accurately called a cumulative incidence rate (over a specified age range). However, the term is cumbersome, and in this disease, the age range cannot be limited precisely. In this article, we use rate and risk to refer to a number of cases per 1000 live-born infants. However, the terms incidence and incidence rate are common currency and easily understood. In recent decades, in Western Europe and North America, the disease is reported with remarkably uniform rates between 2 and 5.

Back to Top | Article Outline

MALE EXCESS

The higher risk of pyloric stenosis in boys than girls is the earliest noted and most consistent epidemiologic feature of the disease. In a combination of data from 13 series published before 1951,2 among 2506 cases, 82.8% were male—a male to female (M:F) ratio of 4.8:1. In the 5 largest series published since 1995,3–7 there were 8160 cases of which 81.5% were male. Four of these series were from the United States, but very similar findings have been reported from Canada,8,9 Australia,10 Denmark,11 Sweden,5 and England.12 It appears that a 4- to 5-fold higher risk of this disease in boys has been characteristic of the disease for many years and in all countries in which it is common.

Back to Top | Article Outline

BIRTH ORDER

Still,13 in 1927, was the first to report that pyloric stenosis is more common in first-born than later-born children. This opinion has generally been supported, but the focus has been on comparing first-born with subsequent infants as a group. There are only 6 published series of cases of pyloric stenosis providing data on birth order in which cases could be related to a defined population of births, or to a sample or reasonable facsimile thereof, and with more than 200 cases.11,14–18 Combined estimates from these of odds ratios in birth orders 1, 2, 3, and 4+ are 1.9, 1.5, 1.3, and 1.0, respectively. A large series from California4 that included almost as many cases as these 6 series combined could not be added to this analysis because raw numbers are not given but gives rates of 2.2, 1.7, and 1.3 for birth orders 1, 2 to 4, and 5+, respectively. These trends do not suggest a unique position for first births, but rather a general decline in risk with increasing birth order over all levels, at least up to the fourth birth.

Back to Top | Article Outline

MATERNAL AGE

Maternal age and birth order are closely correlated. In 2 of the studies14,18 referred to here, maternal age was not examined. In 3 series, the association with birth order remained after adjustment for maternal age but not vice versa.11,15,16 However, 2 studies showed independent associations with maternal age. In California, with 1925 cases, an interesting feature is that for mothers aged 35 or more, there is essentially no variation in risk with birth order.4 Numbers in Oxford were much smaller, but in the larger cells, the crosstabulation suggested that the decline of risk of pyloric stenosis with maternal age was somewhat more consistent than that with birth order, a finding unique to this study.17

Back to Top | Article Outline

MARITAL STATUS

Limited data from Oxford,17 Birmingham,15 and Denmark19 give no reason to suspect that marital status of the mother is related to risk of pyloric stenosis.

Back to Top | Article Outline

RECURRENCE RISK IN FAMILIES

Siblings

Many case reports of pyloric stenosis20 have been stimulated by the occurrence of more than one affected member in a family; they have included sibships with 2, 3, and even 4 affected. Combining data from 4 series17,20–22 in which the number of unaffected as well as affected siblings are given, 51 affected infants were found among 885 siblings, or 5.8%—approximately 30 times the rate in the general population. Excluding affected infants, the sibships in these 4 series include 417 males and 419 females, indicating that affected families are not at unusual risk of producing male children.

Back to Top | Article Outline

Half-Siblings

Sibships including an infant with pyloric stenosis and with only one parent in common may be informative, but their numbers are few and reports even fewer. Combining the only 2 published series of half-siblings,20,23 there was one infant affected among 61 paternal half-siblings and 2 among 81 maternal half-siblings. There are other reports of affected half-siblings but, without better information on frequency, they provide little useful information.

Back to Top | Article Outline

Offspring

A direct estimate of the risk of pyloric stenosis among children of parents who had been affected as infants comes from the follow up of children born to 569 cases treated at the London Hospital for Sick Children between 1920 and 1930.24 Of 882 infants traced, 54 (6.1%) had developed the disease, confirming an earlier impression that the risk to the offspring of an affected individual is of the same order as risk to siblings.21 The risk to the children of affected mothers (27 of 199, or 14%) was 3 times greater than that to those of affected fathers (27 of 683, or 4%), despite the fact that there was essentially no difference between affected fathers and mothers in the ratio of males to females among their offspring. The increase in risk with an affected parent was independent of the sex of the offspring, being a little over 10 times the risk to an infant of the same sex in the general population. Notable is the one-in-5 risk for male offspring of affected mothers (20 of 103).

Seven reports since 195020,22,25–29 have taken a retrospective approach, that is, inquiring as to the frequency of affected parents among affected infants. These give a total of 1643 affected infants of whom 27 (1.6%) had a history of an affected parent; 15 were fathers and 12 were mothers. The overall figure of 1.6% seems to be approximately 5 times the expected, although there may be biases of ascertainment in parents and more ready recognition of the disease in infants by affected parents. Still, the similarity of the numbers of fathers and mothers affected is consistent with the conclusion from the prospective data that the risk to children of an affected mother is substantially greater than that to offspring of affected fathers, because 4 times as many affected fathers as affected mothers would otherwise have been expected.

Back to Top | Article Outline

Grandchildren

The London Hospital group also provides a direct study23 of the risk of pyloric stenosis among the grandchildren of affected persons. In a follow up of the families of patients treated before 1930, 163 affected men produced 120 grandchildren of whom 2 were affected. Among 34 affected women, there were 18 grandchildren of whom 2 were affected. The ratio of the risk to the offspring of affected grandmothers to that of affected grandfathers (6:7) was even greater than the ratio of offspring of affected mothers to those of affected fathers. One pedigree in this report is particularly interesting, if only anecdotally; it includes an affected woman who had 2 daughters and one son, all 3 affected; the son himself had an affected daughter and an unaffected son.

Back to Top | Article Outline

Other Relatives

Carter et al23 enumerated the nephews and nieces of 66 index cases, finding 11 affected among 720 enumerated—approximately 5 times the number expected on the basis of population rates. There were 7 affected among 542 who were related to male index cases and 4 among 178 related to female index cases. An early speculation (based on case reports) that risk of pyloric stenosis may be increased among cousins of cases30 was not confirmed in the only study in which the number of cousins at risk was determined.20 There are numerous case reports of affected relatives, but without enumeration of the numbers at risk, no conclusion can be drawn.

Back to Top | Article Outline

Frequency of Pyloric Stenosis in Twin

The frequency of twin births in most white populations is approximately 1%.31 Because a twin birth results in 2 infants, the expected frequency of twins among all infants is approximately 2% or lower if account is taken of the high early mortality of infants born as twins. The frequency of twin infants has been reported in 7 series of consecutive cases of pyloric stenosis: they include 12,149 cases among whom 257 (2.1%) were twins.3,4,10,17,29,32,33 Thus, it appears that twins are no more susceptible to pyloric stenosis than are single-born infants.

Back to Top | Article Outline

Concordance Within Twin Pairs

The 257 affected twins just referred to came from a total of 228 pairs, of which both members were affected in 29 (8.3%). The risk to the twin sibling of an affected infant, while almost 30 times higher than that in the general population (3 of 1000), appears to be not much higher than that to a nontwin sibling (5.8%).

Back to Top | Article Outline

Concordance Within Monozygous and Dizygous Twin Pairs

Many published “series” of twins consist of accumulations of isolated reports and are subject to publication bias. The first of such reports, in 1938, asserted that in monozygous pairs, “in general,” both members are affected, whereas in dizygous sets, “usually” only one is affected.34 Subsequent authors have increasingly noted “exceptions,”35 and in 1961, Carter concluded that, although the concordance rate for monozygous twins is higher than that for dizygous sets, “as many as 50%” of monozygous sets are discordant.21

An alternative to determining the zygosity of each twin pair in a series is to estimate the frequency of dizygous sets in the series from the sex distribution of the set members and from this to derive the number of monozygous sets by subtracting the dizygous number from the whole. The method is dependent on the belief that in dizygous sets, unlike-sex sets represent half the total. This fraction is independent of the frequency of dizygous sets among all twins and of the sex affinity of the trait that leads to inclusion of the set in a series. Only one published series of twins gives the data necessary for this calculation.32 In that series, the sex distribution in the 65 pairs in which both members survived to 3 months of age were M:M 31, F:F 8, and M:F 26. In 4 M:M and 2 M:F pairs, both infants were affected. The estimated number of monozygous sets is 13 (65 − [2 × 26]), of which no more than 4 were concordant. Two of the 26 clearly dizygous (M:F) sets were also concordant, and it is likely that one or 2 of the like-gender, concordant sets were also dizygous. The author of one other small series of twins,32 although not giving detail on sex for all infants, states that of the 11 pairs of twins, one was concordant (said to be monozygous) and 3 were of unlike sex. If this series is added to that just described, there are 18 monozygous pairs of which at most 5 (28%) and more likely 3 (17%) were concordant. These numbers are too small to determine with confidence that the frequency of concordance of monozygous twins is higher than that of dizygous sets.

Back to Top | Article Outline

Genetic Models

In 1943, the suggestion was made that a single recessive gene, with penetrance dependent on sex and birth order, was responsible for pyloric stenosis.30 The model did not fit the observations available even at the time.20 Subsequent authors36–38 have concluded that no single-gene model—regardless of penetrance—is consistent with the observed risks for various categories of relatives, and they have focused on differentiating between models involving multiple genes, all of which require some environmental component for their expression. Even multifactor models offer no better prospects for accurate counseling than do the empirically observed risks on which they are based. Furthermore, it is at least possible that even the familial patterns have explanations in terms of common environments; this situation is particularly true of the environments shared by infants of the same mother, but even crossgeneration patterns might also be explained by, for example, the occasional participation of grandparents in infant care. The upper bound of any genetic contribution to the etiology of pyloric stenosis is suggested by the limited data on twins.

Back to Top | Article Outline

ETHNICITY

There are many small series of cases and clinical opinions suggesting infrequent or rare occurrence of pyloric stenosis outside the Western world; most are from regions where medical facilities are inadequate or difficult of access, and where pyloric stenosis competes for attention with more common causes of infant morbidity.17,39 These impressions may be the result of poor ascertainment and are weak bases for speculation about causes. A few apparent ethnic differences do, however, deserve attention.

In the United States, white-to-black rate ratios of approximately 2:1 have been reported from Pittsburgh,40 Atlanta,18 Tennessee,6 California,4 and New York state3 with no change in the usual sex ratio. The consistency of these findings suggests that they are not due entirely to underascertainment in black communities and should be given serious attention.

We have found no reports of pyloric stenosis from China or Japan, although our literature search was limited to publications in English. In Hawaii, infants born to parents of Japanese ancestry experienced a risk of 0.5 per 1000—approximately one third of the rate for whites in the same population.41 There were no cases of infants with pyloric stenosis whose parents were classified as Chinese; 21 cases would have been expected if they had experienced the rate of white infants. In Singapore, a risk of 0.2 was reported for infants born to Chinese parents, the same as in the remaining population.42 In California, the rate of pyloric stenosis among infants classified as “Asian” was one fourth of that in the white population.4

Back to Top | Article Outline

INFANT CHARACTERISTICS

Birth Weight

In 2 large studies,4,17 no substantial differences in mean birth weights were found between cases of pyloric stenosis and births in the same population, although neither report gives distributions by birth weight. In another study,18 a trend toward higher birth weight in cases of pyloric stenosis disappeared on adjustment for sex and race, although there were still 3 times as many cases among infants weighing 4500 g or more as expected.18 In a smaller study, the percentage of male cases weighing 4000 g or more was 3 times that in the general population.22 Thus, although there is little overall relationship between birth weight and pyloric stenosis, very large infants may be at increased risk.

Back to Top | Article Outline

Preterm Delivery

There is evidence that preterm delivery, although unrelated to overall risk of pyloric stenosis, is associated with timing of onset of the disease. In Denmark, the mean age at onset of symptoms for 60 cases born before the 38th week of pregnancy was 4.9 weeks compared with 3.7 weeks for 594 born in the 38th week or later.10 In Oxford, cases with mean gestations of <38, 38 to 40, and 41+ weeks had mean ages at admission of 52, 40, and 37 days, respectively (modes were 7, 6, and 5 weeks).17 From California, infants born at <34, 34 to 36, and 37+ weeks had mean ages at diagnosis of 54, 40, and 38 days, respectively.4 Linear regression, modeling change in time of diagnosis against week of gestational age, suggested a regression coefficient for change in age at diagnosis of −0.81 or approximately 1 week later diagnosis for each earlier week of pregnancy. Apparently, onset of symptoms is more closely related to maturity as dated from conception than from birth or to something experienced later by prematurely-born than full-term infants. Age at onset of symptoms would be a useful addition to incidence as an outcome variable in epidemiologic studies.

Back to Top | Article Outline

ECONOMIC STATUS

Rates from Belfast16 and Oxford17 were estimated according to 5 socioeconomic classes defined by the Registrar General for England and Wales. These classes are based on occupation of father categorized by assumed economic status with some adjustments for British perceptions of social status. In these data, numbers are small in class I (the highest); if classes are combined as I plus II, III, and IV plus V, rates in Belfast were 7.2, 2.0, and 2.7, respectively, and in Oxford 3.1, 2.3, and 2.7. They suggest that the higher classes had a higher frequency of the disease but with little difference between the middle and lower economic groups. In Birmingham, rates were estimated for 3 residential districts characterized by several measures of economic status.43 There was no consistent difference in rates in the 3 groups; it is likely, however, that location of residence is a less direct measure of economic status than is occupation of father. Overall, these data leave the impression that risk of diagnosed pyloric stenosis is high among infants of parents in the most favorable economic circumstances, but because of uncertainties relating to ascertainment, this does not offer a strong basis for etiologic inference.

Back to Top | Article Outline

MEDICAL HISTORIES

Several authors10,29,44 have considered pregnancy and delivery histories of series of cases of pyloric stenosis with respect to other diseases and abnormalities in mother or infant with no notable differences being observed. The evidence relating pyloric stenosis to the ABO and Rhesus blood groups does not suggest any important relationship.11,17,29

Back to Top | Article Outline

SIZE OF PYLORIC TUMOR

It seems clear that the muscle hypertrophy of pyloric stenosis is a postnatal phenomenon. Evidence of this observation includes published cases of typical pyloric stenosis after the observation of a normal pylorus documented by radiography45 or even surgery29 and evidence that pyloric tumors increase in size with age at operation, whether the latter is based on surgeons’ subjective opinions46 or physical measurements47 and with duration of symptoms.46 This does not exclude the possibility that circumstances present at or before birth initiate the muscle stimulation, but suggestions as to what these circumstances might be are either untested or, in most instances, untestable.

Back to Top | Article Outline

SUSPECTED CAUSES

Erythromycin

Erythromycin is the most common member of a group of antibiotics known as the macrolides. The most suggestive evidence for a role of erythromycin in pyloric stenosis comes from 2 apparent clusters of the disease—one cluster of 5 cases after erythromycin exposure of infants born in a naval hospital in 197348 and the other of 7 cases in a community hospital in Tennessee in 1999 in which erythromycin was given prophylactically after possible exposure to whooping cough.49 In both these clusters, all the cases were younger than 18 days old when exposed. In another study50 among 496 infants for whom erythromycin was prescribed in a single hospital, 6 developed pyloric stenosis; on the basis of the rate in unexposed infants in the same institution (2.6 per 1000), only 1.2 would have been expected. Five of the 6 cases were exposed in the first week of life.

In a Tennessee Medicaid program, prescriptions for erythromycin were filled for 7138 infants of whom 9 developed pyloric stenosis.51 Using “child-years” of risk after exposure as denominators, the authors estimated adjusted rates of 7.9 for infants exposed in the interval 3 to 13 days after birth and 1.1 for those receiving no prescription. However, the rate in exposed infants was based on only 2 cases. Moreover, among infants exposed at any time in the first 3 months of life, 18.4 cases of pyloric stenosis would have been expected, but only 9 were observed. The apparent discrepancy is puzzling. Overall, the evidence on direct exposure of infants suggests that, if there is any increased risk of pyloric stenosis associated with direct exposure to erythromycin, it is very small.

Erythromycin prescribed for the mother may cross the placenta and increase risk of pyloric stenosis in the offspring. This possibility was evaluated in a surveillance program of congenital defects in which mothers were interviewed within 6 months after the birth.7 Data on erythromycin use were obtained for 1044 infants with pyloric stenosis, 15,356 infants with “other malformations,” and 1704 normal infants. Odds ratios for pyloric stenosis were estimated separately for exposures before and after the 25th week of pregnancy; all were close to 1.0. In another study based on patients of a single hospital,50 rates of pyloric stenosis among infants whose mothers were not exposed during pregnancy were compared with those of mothers who were exposed at any time in pregnancy as well as with those whose mothers were exposed in the 10 weeks before delivery; differences were small. In a study of Tennessee Medicaid patients,6 rates of pyloric stenosis among the infants of 247,032 mothers who had had no prescriptions for macrolides during pregnancy and 638 whose mothers had had prescriptions for erythromycin after the 31st week of pregnancy were 2.6 and 2.9, respectively; the odds ratio for erythromycin exposure was 1.2 (95% confidence interval [CI] = 0.8–1.6).6 Evidence suggests that erythromycin exposure in the pregnant woman is not associated with increased risk of pyloric stenosis in the infant.

The question of whether exposure of the mother to erythromycin while breast-feeding increases the risk of pyloric stenosis in the offspring was raised by a report of a single case in 1986.52 Sorensen et al.53 used computerized pharmacy and National Health Service records for the county of North Jutland, Denmark, to address the issue.53 The comparison births were to mothers who received no prescription of any type during the relevant period. Among 41,778 mothers who had no prescriptions in the first 42 days after delivery, 50 of their infants developed pyloric stenosis, a rate of 1.2 per 1000. Applying this rate to the 624 mothers receiving prescriptions for erythromycin in this period, the expected number of cases in the group would be 0.8; 2 were observed. In the whole period (birth to 90 days), 1.4 cases of pyloric stenosis would have been expected among 1166 possibly exposed infants; 3 were observed. No data on breast-feeding were collected in this study; the relevance of the findings to exposure from breast-feeding depends on results of a referenced national study indicating that 80% of Danish neonates received breast milk for at least 3 months.

Back to Top | Article Outline

Infant Feeding

It is surprising that food and its contaminants have not been more intensively investigated in pyloric stenosis—perhaps because of the unpromising results of early studies. The role of breast-feeding has been examined primarily by correlating changes in risk of pyloric stenosis over time with what was known about changes in rates of breast-feeding—often in populations only marginally similar to those providing disease rates and without reference to comparable ages. The results are not consistent; some authors report no correlation,17 others a positive correlation54,55 (that is, high rates of pyloric stenosis during periods of higher frequency of breast-feeding), others negative,18 and yet others are equivocal.12 Case–control data are few. Dodge16 found 22% of cases breast-fed at age 1 week compared with 16% of controls. Webb et al56 found no meaningful difference between cases breast-fed at 1 week and infants in the source population. Pisacane and colleagues57 found only a slightly higher rate of formula feeding in the first week of life among cases of pyloric stenosis than in comparison infants. Clearly, the data on infant feeding are both inadequate and inconsistent.

Back to Top | Article Outline

Maternal Smoking

In a study of 578 cases of pyloric stenosis from the county of Jutland, Denmark,19 information on maternal smoking was obtained from all mothers in the county. The adjusted odds ratio of pyloric stenosis for infants of mothers who were smokers (not further defined) relative to those of nonsmokers was 2.1 with 95% CI = 1.3–3.1. This study was stimulated by coincidental declines in rates of pyloric stenosis and maternal smoking in Denmark during 1991 through 2000, but maternal smoking could not have accounted for more than a small fraction of the decline in rates of the disease, because pyloric stenosis rates declined from 2.2 in 1991 to 0.5 in 1997, whereas the decline in maternal smoking was only from 33% to 28% over the same period. There are no other important data in the literature on maternal smoking and risk of pyloric stenosis.

Back to Top | Article Outline

TIME TRENDS

Did pyloric stenosis originate—or at least flower—around the beginning of the 20th century? Sporadic reports of infants with symptoms resembling what we now call pyloric stenosis appeared during the 18th and 19th centuries (see Spicer58) but Hirschsprung's report59 of 2 cases—often said to open the modern investigation of the disease—did not appear until 1880. The first series (of 7 cases) appeared in 1902,60 the authors noting that 30 cases of pyloric stenosis had been reported in the medical literature in the 5 years before their report but only 20 before that. In 1940, a series of 38 cases was reported from a single pediatric department in Oslo,61 the author noting that the number of cases seen in the previous decade at that institution was 3 times the number seen in the preceding decade.

Epidemiologic study of pyloric stenosis dates to Wallgren,62 who in 1941 reported a rate of 4.0 per 1000 live births in Gothenburg, Sweden. Thereafter, studies of the frequency of pyloric stenosis and its secular trends flourished. Until around 1990, most local and national rates showed reasonable stability over time within a range of 2 to 5.9,11,16,18,43,63,64 Nearly all time trends have been modestly upward within the range between 2 and 5, but there have also been a few suggestions of larger increases—usually in studies based on small numbers or geographic areas.65,66 It is difficult to attribute any of these increases to a change in frequency of the disease. Parental sensitivity to infant illness, access to competent medical care, and medical familiarity with the disease have all increased over time. Quicker admission to hospital care, which has been almost universal in developed countries, may also increase the likelihood of diagnosis before spontaneous recovery.67

Recent data from Denmark and Sweden are of a different order. In Denmark as a whole,63 and in Funen County, representing approximately 9% of the country,11 rates were reasonably steady between 1955 and 1986; however, in Denmark between 1986 and 1997, the rates fell from 3.2 in 1986 to 1.1 in l996 and 1997. A similar pattern was reported from Sweden; in the Stockholm Health Care Region, after increases between 1970 and 1976, rates remained between 2.3 and 3.2 from 1976 to 1990, after which there was a sharp decline to 0.4 in 1999.68 Data for all of Sweden confirmed the downward trend showing a decline from 2.8 to 0.8 between 1989 and 1996, and suggested that the decline occurred in both sexes and in both the north and south of the country.5 In their steepness and rapidity, these declines are without precedent in any country, although a less striking decline in Gothenburg had been reported 40 years earlier.69

In discussing the Danish data, Nielsen and colleagues63 observed that the beginning of the decline coincided with reports of the American Academy of Pediatrics and the Danish National Board of Health in 1992 recommending a change in infant sleeping position from prone to supine. They considered it “quite hypothetical” that this situation might have relevance to the decline in pyloric stenosis but, in light of radiographic evidence that the rate of passage of food through the stomach depends on the position of the infant, thepossibility was worth investigating. Persson and colleagues68 noted that secular trends in SIDS and pyloric stenosis in Sweden were parallel before 1990 as well as afterward, and that the Nordic campaigns against SIDS included a recommendation for the avoidance of the prone sleeping position—the effectiveness of which had been documented by Wennergen.70 (In Sweden, the percentage of infants placed in the prone position for sleeping declined from 42% to15% between 1991–1992 and 1994–1995.70 In Denmark, this position was used in only 4% of infants at the beginning of the period, but the percentage placed supine—from which position it is more difficult for an infant to wriggle toprone—increased from 11% to 17%.70) Persson et al68 also invoked radiographic evidence that the stomach contents accumulate in the proximal part of the stomach if the infant is supine, whereas if the infant is prone, they accumulate toward the pylorus. They concluded that the combination of ecologic data and a possible biologic mechanism “suggest a causal association” between sleeping position and risk of pyloric stenosis.

Although the hypothesis that pyloric stenosis and SIDS share a common risk factor remains to be evaluated, it is challenging. It is new—never having even been considered previously—simple in concept and, if supported, offers a simple and inexpensive means of prevention. It relates to the most striking decline in rates of pyloric stenosis ever observed, and it is most unlikely to be due to artifacts of ascertainment. If this hypothesis is not confirmed, the recent decline in rates of pyloric stenosis in Denmark and Sweden itself becomes an enigma.

Back to Top | Article Outline

ACKNOWLEDGMENTS

For assistance with the literature search, I thank Audrey Smith and Mark Gehret.

Back to Top | Article Outline

REFERENCES

1. Ramstedt C. Zur operation der angeborenen pylorus-stenose. Med Klinik. 1912;8:1702–1705.
2. MacMahon B. Hypertrophic Pyloric Stenosis of Infancy [Dissertation]. Birmingham, UK: School of Medicine, Birmingham, The University of Birmingham; 1952.
3. Appelgate MS, Druschel CM. The epidemiology of infantile pyloric stenosis in New York State: 1983 to 1990. Arch Pediatr Adolesc Med. 1995;149:1123–1129.
4. Schechter R, Torfs CP, Bateson TF. The epidemiology of infantile hypertrophic pyloric stenosis. Pediatr Perinat Epidemiol. 1997;11:407–427.
5. Hedbach G, Abrahamsson K, Husberg B, et al. The epidemiology of infantile hypertrophic pyloric stenosis in Sweden 1987–1996. Arch Dis Child. 2001;85:379–381.
6. Cooper WO, Ray WA, Griffin MR. Prenatal prescription of macrolide antibiotics and infantile pyloric stenosis. Obstet Gynecol. 2002;100:101–106.
7. Louik C, Werler MM, Mitchell AA. Erythromycin use during pregnancy in relation to pyloric stenosis. Am J Obstet Gynecol. 2002;186:288–290.
8. Habbick BF, To T. Incidence of infantile hypertrophic pyloric stenosis in Saskatchewan, 1970–85. CMAJ. 1989;140:395–398.
9. Walpole IR. Some epidemiological aspects of pyloric stenosis in British Columbia. Am J Med Genet. 1981;10:237–244.
10. Hitchcock NE, Gilmour AI, Gracey M, et al. Pyloric stenosis in Western Australia. Arch Dis Child. 1987;62:512–513.
11. Rasmussen L, Green A, Hansen LP. The epidemiology of infantile hypertrophic pyloric stenosis in a Danish population, 1950–1984. Int J Epidemiol. 1989;18:413–417.
12. Walsworth-Bell JP. Infantile pyloric stenosis in Greater Manchester. J Epidemiol Community Health. 1983;37:149–152.
13. Still GF. Place in family as a factor in disease. Lancet. 1927;ii:795–853.
14. Ford N, Ross MA, Brown A. Primogeniture as an etiologic factor in pyloric stenosis. Am J Dis Child. 1941;61:745–751.
15. McKeown T, MacMahon B, Record RG. The incidence of congenital pyloric stenosis related to birth rank and maternal age. Ann Eugen. 1951;16:249–259.
16. Dodge JA. Infantile hypertrophic pyloric stenosis in Belfast, 1957–1969. Arch Dis Child. 1975;50:171–178.
17. Adelstein P. Fedrick J. Pyloric stenosis in the Oxford Record Linkage Study area. J Med Genet. 1976;13:439–448.
18. Lammer EJ, Edmonds LD. Trends in pyloric stenosis incidence, Atlanta, 1968 to 1982. J Med Genet. 1987;24:482–487.
19. Sorensen HT, Norgard L, Pedersen NL, et al. Maternal smoking and risk of hypertrophic infantile pyloric stenosis: 10 year population based cohort study. BMJ. 2002;325:1011–1012.
20. McKeown T, MacMahon B, Record RG. The familial incidence of congenital pyloric stenosis. Ann Eugen. 1951;16:260–281.
21. Carter CO. The inheritance of congenital pyloric stenosis. Br Med Bull. 1961;17:251–254.
22. Czeizel A. Birth weight distribution in congenital pyloric stenosis. Arch Dis Child. 1947;47:978–980.
23. Carter CO, Evans K, Warren J. The grandchildren of patients with pyloric stenosis. J Med Genet. 1980;17:411–415.
24. Carter CO. Genetics of common disorders. Br Med Bull. 1969;25:52–57.
25. Carter CO, Savage TH. Pyloric stenosis in four first cousins. Arch Dis Child. 1951;26:50–51.
26. Dougall AJ. Infantile pyloric stenosis. A review of 200 cases. Scot J Med. 1969;14:156–161.
27. Stang H. Pyloric stenosis revisited in St Paul. Minn Med. 1985;68:661–662.
28. Breaux CW Jr, Georgeson KE. Hypertrophic pyloric stenosis: a review of 216 cases from 1980 to 1986. Alabama Med. 1986;55:34–37.
29. Dodge JA. Infantile pyloric stenosis inheritance, psyche and soma. Irish J Med Sci. 1973;142:6–18.
30. Cockayne EA, Penrose LS. The genetics of congenital pyloric stenosis. Ohio J Sci. 1943;43:1–16.
31. Jeanneret O, MacMahon B. Secular changes in rates of multiple births in the United States. Am J Hum Genet. 1962;14:410–425.
32. MacMahon B, McKeown T. Infantile hypertrophic pyloric stenosis: data on 81 pairs of twins. Acta Gen Med Gemel. 1955;IV:320–329.
33. Scharli A, Sieber WK, Kiesewetter WB. Hypertrophic pyloric stenosis at the Children's’ Hospital of Pittsburgh from 1912 to 1967. J Pediatr Surg. 1969;4:108–1147.
34. Sheldon W. Hypertrophic pyloric stenosis in one of uniovular twins. Lancet 1938;I:1048–1049.
35. Powell BW, Carter CO. Pyloric stenosis in twins. Arch Dis Child. 1951;26:45–49.
36. Kidd KK, Spence MA. Genetic analysis of pyloric stenosis suggesting a specific maternal effect. J Med Genet. 1976;13:290–294.
37. Lalouel JM, Morton NE, MacLean CJ, et al. Recurrence risks in complex inheritance with special regard to pyloric stenosis. J Med Genet. 1977;14:408–414.
38. Mitchell LE, Risch N. The genetics of infantile hypertrophic pyloric stenosis. A reanalysis. Am J Dis Child. 1993;147:1203–1211.
39. Jain SP. Congenital hypertrophic pyloric stenosis in African infants in Somalia and Ethiopia. East Afr Med J. 1977;54:332–337.
40. Laron Z, Horne LM. The incidence of infantile pyloric stenosis. Am J Dis Child. 957;94:151–154.
41. Shim WK, Campbell A, Wright SW. Congenital hypertrophic pyloric stenosis in Hawaii. II. Racial aspects. Hawaii Med J. 1970;29:292–295.
42. Chong AYH, Lee HP. Pyloric stenosis in the ethnic groups of Singapore. Singapore Med J. 1976;17:181–183.
43. MacMahon B, Record RG, McKeown T. Congenital pyloric stenosis. An investigation of 578 cases. Br J Soc Med. 1951;5:185–192.
44. Hannon VR. Prenatal and Early Postnatal Factors in Pyloric Stenosis of Infancy [Thesis]. Boston: Department of Epidemiology, School of Public Health, Harvard University; 1967.
45. Wallgren A. Preclinical stage of infantile hypertrophic pyloric stenosis. Am J Dis Child. 1946;72:371–376.
46. McKeown T, MacMahon B, Record RG. Size of tumor in infantile pyloric stenosis related to age at operation. Lancet. 1951;ii:556–558.
47. Gerard JW, Waterhouse JAH, Maurice DG. Infantile pyloric stenosis. Arch Dis Child. 1955;30:493–496.
48. SanFillippo JA. Infantile hypertrophic pyloric stenosis related to ingestion of erythromycin estrolate. J Pediatr Surg. 1976;11:177–180.
49. Honein MA, Paulozzi LJ, Himelright IM, et al. Infantile hypertrophic pyloric stenosis after pertussis prophylaxis with erythromycin: a case review and cohort study. Lancet. 1999;354:2101–2105.
50. Mahon BE, Rosenman MB, Kleinman MB. Maternal and infant use of erythromycin and other macrolide antibiotics as risk factors of infantile hypertrophic pyloric stenosis. J Pediatr. 2001;139:380–384.
51. Cooper WO, Griffen MR, Arbogast P, et al. Very early exposure to erythromycin and infantile pyloric stenosis. Arch Pediatr Adolesc Med. 2002;156:647–650.
52. Stang H. Pyloric stenosis associated with erythromycin ingested through breast milk. Minn Med. 1986;69:669–670, 682.
53. Sorensen HT, Skriver MV, Pedersen L, et al. Risk of infantile hypertrophic pyloric stenosis after maternal use of macrolides. Scand J Infect Dis. 2003;35:104–106.
54. O’Donoghue JM, Connolly KD, Gallagher MM, et al. The increasing incidence of infantile pyloric stenosis. Irish J Med Sci. 1993;162:175–176.
55. Knox EG, Armstrong E, Haynes R. Changing incidence of infantile pyloric stenosis. Arch Dis Child. 1983;58:582–585.
56. Webb AR, Lari J, Dodge JA. Infantile pyloric stenosis in South Glamorgan 1970–9. Effects of change. Arch Dis Child. 1983;58:586–590.
57. Pisacane A, de Luca U, Criscuolo L, et al. Breast-feeding and hypertrophic pyloric stenosis: population based case–control study. BMJ. 1966;312:745–746.
58. Spicer RD. Infantile hypertrophic pyloric stenosis; a review. Br J Surg. 1982;69:128–135.
59. Hirschsprung NP. Falle von angeborenen Pylorusstenose beobacht bei Sauglingen. Jahrb f Kinderheilk. 1888;28:61–68.
60. Cautley E, Dent CT. Congenital hypertrophic stenosis of the pylorus and its treatment by pyloroplasty. Lancet. 1902;ii:1679–1685.
61. Rinvik R. Investigations of congenital stenosis of the pylorus, its treatment and prognosis, with reports of 107 cases examined 1 to 22 years after treatment. Acta Paediatr. 1940;27:296–298.
62. Wallgren A. Incidence of hypertrophic pyloric stenosis. Am J Dis Child. 1941;62:751–756.
63. Nielsen JP, Haahr P, Haahr J. Infantile hypertrophic pyloric stenosis. Dan Med Bull. 2000;47:223–225.
64. Sule ST, Stone DH, Gilmour H. The epidemiology of infantile hypertrophic pyloric stenosis in Greater Glasgow area, 1980–1996. Paediat Perinat Epidemiol. 2001;15:379–380.
65. Kinane B, Cosgrove JF, Counahan R. Increasing incidence of infantile hypertrophic pyloric stenosis (IHPS). Irish Med J. 1986;79:298.
66. Kerr AM. Unprecedented rise in incidence of infantile hypertrophic pyloric stenosis. BMJ. 1980;281:714–715.
67. Hansen LP, Rasmussen L. Does earlier admission for infantile hypertrophic pyloric stenosis (IHPS) apparently increase incidence rates? J Pediatr Gasteroenterol Nutr. 1988;7:148.
68. Persson S, Ekbom A, Granath F, et al. Parallel incidences of sudden infant death syndrome and infantile pyloric stenosis: a common cause? Pediatrics. 2001;108:E70.
69. Wallgren A. Is the rate of hypertrophic pyloric stenosis declining? Acta Paediatr. 1960;49:530–535.
70. Wennergren G, Alm B, Oyen N, et al. The decline in the incidence of SIDS in Scandinavia and its relation to risk-intervention campaigns. Acta Paediatr. 1997;86:963–968.
© 2006 Lippincott Williams & Wilkins, Inc.