REVIEWS OF INFECTIOUS DISEASES. VOL. 1,NO. 4 • JULY-AUGUST 1979 © 1979 by The University of Chicago. 0162-0886/79/0104-0005-$00.75
The Sudden Infant Death Syndrome and Infant Botulism From the Department of Epidemiology, School of Public Health and Community Medicine, University of Washington; and the National Oceanic and Atmospheric Administration, Seattle, Washington
Donald R. Peterson, Melvin W. Eklund, and Nina M. Chinn
and, for comparison, from infants of similar age who died from other causes.
In this paper, we shall describe our efforts to determine whether infant botulism, which was first recognized as a unique disease entity in 1976 [1, 2] is responsible for some cases of sudden infant death syndrome (SIDS). This hypothesis is predicated on the striking similarity in age distribution of these two ostensibly disparate disorders. Our investigations began in King County, Wash., on November I, 1976, and concluded on October 10, 1977. King County is a unique resource for research on SIDS because all suspected cases of sudden infant death are referred to a single facility-the Children's Orthopedic Hospital and Medical Center in Seattle-for definitive pathologic diagnosis (pathologic examinations performed under the supervision of Dr. J. Bruce Beckwith). This situation has existed since 1963, when a special state statute [3] was enacted to stimulate research into SIDS. We planned to obtain specimens of stool and of heart blood from each case diagnosed as a sudden infant death
Materials and Methods
Forty-one cases of SIDS occurred in King County during the period of investigation; specimens of stool and heart blood were obtained from all but II of these infants. In addition to the 30 specimens from SIDS infants, we obtained specimens from eight infants who were diagnosed as having died from causes other than SIDS. Our original plan called for the study of a larger group of non-SIDS infants than was accomplished. This departure from our original plan resulted from the press of other professional responsibilities and commitments, and the investigation was terminated with only a single positive finding. Nevertheless, we feel that our findings should be reported. Despite the ostensibly optimal circumstances in King County for this sort of investigation, we encountered a variety of situations that delayed laboratory examination of specimens (Botulism Laboratory, National Fisheries Service, Seattle, Wash.). Table I describes this aspect of our experience. Most of the delays involved the interval between the time of the infants' deaths and the time the autopsies were performed. Most specimens
This investigation was supported in part by grant HD 09297 from the National Institute of Child Health and Human Development. Please address requests for reprints to Dr. Donald R. Peterson, Department of Epidemiology SC-36, School of Public Health, University of Washington, Seattle, Washington 98195.
630
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
Fecal and serum specimens taken from 30 cases of sudden infant death and from eight cases of nonsudden infant death that were diagnosed at a single facility in King County, Wash., were examined for the presence of Clostridium botulinum organisms and toxin. Organisms, but not toxin, were recovered from a fecal specimen in one case of sudden intant death, results that parallel those from studies previously reported by investigators in California. Studies made in our laboratory of a nonfatal case of infant botulism revealed that an estimated 366,000 mouse minimal lethal doses of toxin were excreted in feces collected by purging the infant. Organisms and toxin were excreted for at least 15 days after the infant was hospitalized. Observations made in our laboratory of atypical responses in mice to both fecal and serum extracts, coupled with recently described experiments in which mice were used as an animal model for infant botulism in humans, provide a biologically plausible foundation for the hypothesis that C. botulinum may be implicated etiologically in some sudden infant deaths. Additional microbiologic, physiologic, and toxicologic data are needed to adequately test this hypothesis.
631
Sudden Infant Death and Infant Botulism
Table 1. Interval between death of infant and initiation of laboratory examination of fecal and blood specimens for the presence of Clostr£dium botul£num. King County, Washington, November 1, 1976-0ctober 10, 1977.
Interval (hr)
Total
Cases of non-sudden infant death
5 5 10 7 3
2 1 3 1 1
30
8
were delivered to the laboratory within 1-2 hr after the postmortem examination. All of the specimens were maintained at refrigerator temperature (2 C). The age and sex distribution of the SIDS group and the non-SIDS group are shown in table 2. Twenty-two infants in the SIDS group were white, five were black, and three were of mixed racial extraction. All eight of the infants in the nonSIDS group were white. Stool specimens were obtained from the infants by milking the contents of the resected colon into sterile jars or large test tubes (Vacutainer® anaerobic specimen collector, Becton-Dickson, Rutherford, N .J.) that were specially prepared to maintain anaerobic conditions during transport to the laboratory. A swab of a portion of the lumen of the small intestine of nine infants who died of SIDS during the latter phase of the investigation was taken for subsequent culture and was transported in the special test tubes. Specimens of he~rt blood were obtained by aspiration with a stenle needle and syringe. Upon arrival at the laboratory, the swab specimen and an aliquot of the stool specimen were inoculated into cooked meat medium, incubated, and assayed for the presence of botulinal toxin by injecting the culture supernatant ip into mice [4J. The remainder of the stool was extracted with gelatin-phosphate buffer and the supernatant fluid was assayed. Assays for the presence of toxin in specimens of blood were performed with undiluted serum; in addition, six serum specimens and four stool extracts were concentrated four-
Results
Gelatin-phosphate extracts of 15 of the 30 stool specimens from SIDS infants produced an atypical lethal reaction in mice; these extracts, although lethal, were not neutralized by botulinal antisera. Two of the 30 serum specimens evoked a similar response. The lethality to mice of a few of these samples was eliminated by neutralizing the sample with antisera to Clostridium perfringens (welchii), type A. No Clostridium botulinum organisms were isolated from swab specimens taken from the small intestine. Extracts from stool specimens from five of the eight non-SIDS infants induced nonpathognomonic symptoms and signs of toxicity when injected ip into mice. The interval between the death of the infant from whom samples exhibiting nonspecific toxicity were taken and the time the samples were examined in the laboratory ranged from 6-80 hr with a fairly even distribution over this range. A stool specimen from a black baby girl who died on October 9, 1977 at 206 days of age, and whose death was attributed to SIDS, was examined in the laboratory --10 hr after her death. Type A botulinal toxin was identified in the supernatant of a subculture of the stool specimen grown in
Table 2. Age and sex distribution of cases of sudden infant death and of non-sudden infant death, King County, Washington, November 1, 1976-0ctober 10, 1977. Sudden infant death Age (days)
Male
Non-sudden infant death
Female
Male
Female
0-30 31-60 61-90 91-120 121-150 151-180 181-210 211-270
0 3 6 8 1 0 0 0
3
1 3 1 2 1 1 0
0 0 1 0 0 1 0 0
0 0 3 0 1 0 0 2
Total
18
12
2
6
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
0-12 13-24 25-48 49-72 73-96'
Cases of sudden infant death
fold by dialysis against polyethylene glycol before being injected into mice. Laboratory personnel identified all specimens examined in this study only by code.
Peterson et
632
Discussion
The isolation of C. botulinum organisms, but not botulinal toxin, from the stool specimen from one of the 30 (3%) SIDS infants examined parallels the results of the more extensive studies made in California [6], where nine of 211 (4%) of the stool specimens from SIDS infants showed evidence of C. botulinum organisms. In two of these nine infants, both C. botulinum organisms and toxin were identified in a swab taken from the small intestine but were not identified in the feces. We recognize that these results are not impressive statistically, but some credibility is lent to the association of SIDS and the presence of C. botulinum by the similarity in the two studies of the proportions of samples in which these organisms were found. In fact, these percentages probably underestimate the actual relative frequency of the involvement of C. botulinum because of logistic and technological problems encountered in both the California and the King County studies. In King County there were few well-identified controls. In California, however, as of September 29, 1978, 76 infants who died of causes not attributed to SIDS had been examined in the same manner as were the 211 SIDS infants; there wasn't a single positive identification of either C. botulinum organisms or toxin (Stephen Arnon, personal communication). The facts that an infant with manifest clinical botulism can excrete prodigious quantities of toxin and still survive and that toxin can occasionally be identified in specimens from SIDS infants, poses a curious paradox regarding the hypothesis suggested by the investigators in California-that infant botulism may account for some cases of SIDS. If some cases of SIDS constitute one extreme of the spectrum of severity of infant botulism, one would have thought that previous investigations of SIDS would have revealed antemortem signs of cranial nerve paresis and of constipation in affected infants. However, a relatively low frequency of these signs in SIDS infants could account for the lack of their mention in published reports of SIDS. Nevertheless, the demonstration that botulinal toxin activity persists in feces kept under laboratory conditions does not discount the possibility
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
cooked meat medium. A pure culture of C. botulinum was later isolated. The swab specimen from the small intestine of this infant was negative for C. botulinum organisms. To assess the stability of toxin in stool specimens, feces from two infants in the SIDS group were each combined with type A botulinal toxin to achieve a final concentration of 1,000 mouse minimal lethal doses (MLDs)jg of feces (wet weight). The two mixtures were refrigerated at 2 C for 18 hr, extracted with gelatin-phosphate buffer and assayed for the presence of toxin. The assay was positive for only one of the samples. In November, 1977, approximately one month after terminating our surveillance of SIDS infants for evidence of the presence of C. botulinum, a 1O~-week-old white baby girl with suspected infant botulism was admitted to the Children's Orthopedic Hospital. Her symptoms included loss of head control, weak suck, floppiness, and constipation that had begun 10 days before her admission to the hospital (clinical details provided by Dr. A. Thomas Collins). The initial stool specimen was extracted with gelatin-phosphate buffer, and the extract yielded 500 MLDs of type A botulinal toxinjg of feces. A dose of 0.5 ml of her serum, collected on the same day as the stool specimen, was found to be nontoxic in mice injected ip. However, the one mouse injected with serum concentrated four-fold by dialysis against polyethylene glycol developed symptoms suggestive of botulism after 16 hr; these symptoms disappeared after 24 hr. The mouse died six days later. Insufficient amounts of serum precluded further study of these effects. On the day after the initial stool specimen was collected, feces obtained from the infant after administration of a laxative yielded 4,000 MLDs j g of feces. This purge resulted in the excretion of an estimated 366,000 MLDs of toxin, an amount far greater than that which is lethal to humans [5]. The infant continued to excrete type A botulinal toxin and type A C. botulinum organisms for at least 15 days after she w.as admitted to the hospital; nevertheless, she survived. A portion of the feces was refrigerated at 2 C for 40 days and then was extracted and assayed for the presence of toxin. More than 32 mouse MLDs of toxinjg of feces were still present in the specimen.
at.
633
Sudden Infant Death and Infant Botulism
J
tion of in vivo toxin production provide a biologically plausible explanation for the relatively benign course of clinical infant botulism. It is conceivable that a temporarily incompetent ileocolic valve could allow retrograde passage of a spurt of toxin from the cecal end of the colon into the ileum, from which toxin could readily be absorbed, producing concentrations of toxin in the circulation sufficient to cause SIDS. Likewise, clinically manifest but nonfatal botulism could be explained either by reflux of nonlethal doses of toxin from the large intestine to the small intestine or by absorption directly into the circulation of only a small fraction of the toxin concentrated in the large intestine. An adequate test of this hypothesis will depend upon further research into the ecologic development of microflora in the intestine of the human infant, the physiologic variation in function of the ileocecal valve in infants, and the development of more sensitive methods for the detection of circulating and/or synaptically fixed toxin. Conclusion
Our surveillance of SIDS for evidence of infection with C. botulinum confirms the results of more extensive research conducted in California. Based on observations made in our laboratory and in others, we have formulated a biologically plausible hypothesis for the pathogenesis of SIDS. References 1. Pickett, J., Berg, B., Chaplin, E., Brunstetter-Shafer, M. Syndrome of botulism in infancy: clinical and electrophysiologic study. N. Eng!. J. Med. 295 :770772,1976. 2. Midura, T. F., Arnon, S. S. Infant botulism: identification of Clostridium botulinum and its toxin in faeces. Lancet 2:934-936, 1976. 3. State of Washington Laws of 1963, Chapter 178, RCW 68.08.100 (1973 Suppl.). 4. Eklund, M. W., Poysky, F. T., Wieler, D. I. Characteristics of Clostridium botulinum type F isolated from the Pacific Coast of the United States. Appl. Microbiol. 15:1316-1323,1967. 5. Morton, H. E. The toxicity of Clostridium botulinum, type A toxin for various species of animals, including man. Project Summit, Contract no. DAI8-064-Cml2757, Institute for Cooperative Research, University of Pennsylvania, 1961.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
that, in vivo, toxin is inactivated after death during the appreciable interval between death and the diagnosis of SIDS at autopsy. The toxicity that was found in both stool and serum specimens that was not attributable to the presence of botulinal toxin may have resulted from toxins or metabolic products of other organisms, such as C. perfrigens that thrive in a dead host and that may suppress the growth of C. botulinum and inactivate its toxin, as Smith has demonstrated for samples of soil [7]. Because sudden infant deaths are relatively rare events (average, two in every 1,000 live births), are only occasionally witnessed, and usually occur away from research centers, it seems unlikely that surveillance alone will provide definitive data concerning this syndrome. On the basis of investigations in California and King County, infant botulism accounts for only a small fraction of cases diagnosed as SIDS. Epidemiological studies of SIDS that are directed at such small subset populations pose great difficulties and little promise of success. For this reason, a recent report by Sugiyama and Mills [8] that describes experiments involving intragastric injection of mice with type A C. botulinum spores is noteworthy. They demonstrated that in vivo production of toxin occurred only after the first week of life and could not be induced after the age of two weeks; this characteristic is analogous to the circumscribed pattern of age distribution that is common to both infant botulism and SIDS. Sugiyama and Mills postulated that this sharply demarcated period of susceptibility to small inocula of C. botulinum spores is probably related to the kinds of microorganisms that successively colonize the intestinal tract during the first weeks of life. Unfortunately, relatively little is known about the microflora of the intestinal tract of human infants [9]. In the mice studied, Sugiyama and Mills found toxin only in the large intestine, from which absorption into the systemic circulation would be negligible. None of the mice showed signs of botulism. Miyazaki and Sakaguchi [10] experimented with two-week-old chicks and demonstrated that production and absorption of lethal quantities of botulinal toxin is localized in the cecum of these animals. These observations concerning the loca-
634
6. Arnon, S. S., Midura, T. F., Damus, K., Wood, R. M., Chin, J. Intestinal infection and toxin production by Clostridium botulinum as one cause of sudden infant death syndrome. Lancet 1:1273-1277, 1978. 7. Smith, L. DS. Inhibition of Clostridium botulinum by strains of Clostridium perfringens isolated from soil. Appl. Microbiol. 30:319-323, 1975. 8. Sugiyama, H., Mills, D. C. Intraintestinal toxin in infant mice challenged intragastrically with Clostridium botulinum spores. Infec. Immun. 21:59-63, 1978.
Peterson et al.
9. Mata, L. J., Urrutia, J. J. Intestinal colonization of breast-fed children in a rural area of low socioeconomic level. Ann. N.Y. Acad. Sci. 176:93-109, 1971. 10. Miyazaki, E., Sakaguchi, G. Experimental botulism in chickens: the cecum as the site of production and absorption of botulinum toxin. Jpn. J. Med. Sci. BioI. 31:1-15,1978.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
Discussion
of C. botulinum positivity in stools of SIDS infants of a particuular age group? DR. CHIN. The highest isolation rate was in the children under one month of age: three positive stools of 19. In all the other age groups, the isolation rate was approximately 4%-6%, except in the group of 40 infants aged two months, among whom there were no isolates. DR. RONALD ALTMAN. Has anybody looked for botulinal or other toxins or organisms in children with other illnesses? DR. CHIN. Not yet. DR. V. R. DOWELL, JR. I just want to make a comment in regard to the stability of the toxin. Holding the specimen in the refrigerator is not getting at the question of whether the toxin deteriorates in the cadaver. Postmortem changes in the microbial popUlation can occur, and the toxin can be destroyed before the specimen is collected. It is well known that Clostridium perfringens, for example, is commonly found in heart blood of cadavers if they are not studied early. I think this may account for some of the problems of toxin testing of the intestinal contents-the changes that occur postmortem before the specimen is collected. DR. ALTMAN. You reported nonspecific deaths in mice. This means that toxin was not heat-labile and/or not neutralized by botulinal antitoxin. When we lack a known etiology and something in stools kills mice, it makes you wonder whether there isn't something else present. DR. DOWELL. We have encountered Clostridium ditficile toxin in the stools of one of the babies. DR. LOUIS SMITH. In relation to the stability of the toxin, Dr. Peterson stated that he and his associates added 100 LD50 /g to two different samples of feces. In one of those samples, the toxin all disappeared overnight. So toxin can be destroyed very rapidly at room temperature. In two other cases, the toxin disappeared more slowly. There is a great deal of individual variation from one specimen to another. I think it would be foolish to generalize about toxin stability in stool.
635
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
DR. ROGER FELDMAN. One of the characteristics of infant botulism is a peak onset at two months of age. Is that peak reproduced with the sudden infant death syndrome (SIDS)? Are SIDS cases seasonal throughout the country? DR. DONALD PETERSON. The age distribution curve of SIDS is almost superimposable in every study that's been done. The distribution by age is one of the few constants that we have found in this research. The seasonal distribution is similarly reproducible. In the Southern Hemisphere, it is reversed but is still striking. From the family study data that I described earlier, we have just sorted all 1,200 cases by latitudes and then looked at the seasonal distribution in the northern, middle, and lower tiers of states, and we get the same pattern. I'm convinced that the seasonal association for SIDS is reproducible. There is a clear relation between birth order and sudden death, with a relative sparing of the firstborn. There is an inverse relation with maternal age. The highest risk accrues to children of young multiparas. DR. JAMES CHIN. We had 211 SIDS and supposedly 69 non-SIDS deaths, but the one nonSIDS death in which we found Clostridium botulinum organisms was a 14-day-old infant who died suddenly and unexpectedly at home. A local professor of pathology, also a SIDS Foundation member, considered this a typical SIDS case and recommended that the County Health Department provide SIDS counseling to the family. The coroner's pathologist attributed the death to "cardiopulmonary failure secondary to acute tracheolaryngiobronchitis." DR. JOHN NELSON. There were 192 infants under six months of age in the SIDS group and 49 in the non-SIDS group. There were nine infants positive for C. botulinum of the 192 in the SIDS group and one infant positive of the 49 in the non-SIDS group. DR. CHIN. We prefer to say that we found C. botulinum organisms in 10 infants (and also toxin in two of them) who died suddenly and unexpectedly. DR. FELDMAN. Was there a higher frequency
636
State Health Department has done to define the frequency of C. botulinum isolation. DR. CHIN. When we did our infant death study, some of our methodology evolved during the year. Once we started collecting stool specimens from infants with SIDS, we postulated where the toxin would be produced and asked for specimens from specific parts of the gastrointestinal tract. Our method finally evolved to the point at which we tied off the piece and had a sort of sausage package. But even then, with this procedure, we couldn't be absolutely certain that the contents were from that specific area. DR. DOWELL. Do you think it's possible that other clostridia could be involved in SIDS? DR. PETERSON. It's a question I wanted to ask myself. Do we know of other clostridia that produce neurotoxins besides C. botulinum? One of the problems with a new disease of low frequency is funding. The CDC should have a funding arrangement for pilot studies comparable to those of infant botulism, so that investigators have a chance to move when their interest is high and when the opportunity is best and not have to piece together things on a day-by-day basis. The work done in Seattle [Wash.] has been piggybacked on other grants and contracts. DR. FELDMAN. There are funds at the CDC for the study of a new disease; witness our efforts on Legionnaires' disease. The question is the priority given to this clinical problem and its association with SIDS. DR. ALTMAN. That's not a fair analogy. The money spent on Legionnaires' disease was spent to give services. It was spent on people running to New York, Philadelphia, Dallas, Memphis, and other cities. What is being asked for here is money for a non-CDC group that is competent to study a problem of concern to the CDC. Funds should be available for that purpose.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
DR. THADDEUS MIDURA. Some of the infant botulism cases that were diagnosed as myasthenia gravis were tested with pyridostigrnine, and when their stools were studied the test mice died. Fortunately, the work that Dr. Hatheway did with dialyzing the stool to get rid of the drug was then reported. This was one situation where what was seen in adults was used in studying infants. DR. CHIN. In Utah, an infant was brought to the llOspital outpatient clinic for possible failure to thrive and was examined by the pediatric neurologist, who found nothing. The child was sent home. Two weeks later, he was admitted to the hospital after a respiratory arrest at home and expired after several days in the hospital's intensive-care unit. The pediatric neurologists asked for testing for botulism, and the test was positive. That was a case that resembled SIDS but was infant botulism, as determined by the state and the Center for Disease Control (CDC). On the death certificate, SIDS was the stated cause of death. DR. PETERSON. My reaction to the relation between SIDS and infant botulism is that I feel frustrated at this point. It is obviously low-yield research. It's not very exciting if the proportion of SIDS that we can attribute to infant botulism is only 3%-5~o' But the other cause of frustration is the technology. If we're going to study SIDS and screen many babies' stools with the present techniques, we have a very expensive proposition. The techniques are not necessarily all that specific, so we are not sure about our negative results. And there are all of the other factors that make interpretation of what we think we know difficult. I think we are in a situation in which we have to let basic science do whatever it can do and take off again when we have better methodology. I guess I'm thinking that the emphasis should be on infant botulism and not on further case control studies of the kind that the California
Discussion
The Sudden Infant Death Syndrome and Infant Botulism From the Department of Epidemiology, School of Public Health and Community Medicine, University of Washington; and the National Oceanic and Atmospheric Administration, Seattle, Washington
Donald R. Peterson, Melvin W. Eklund, and Nina M. Chinn
and, for comparison, from infants of similar age who died from other causes.
In this paper, we shall describe our efforts to determine whether infant botulism, which was first recognized as a unique disease entity in 1976 [1, 2] is responsible for some cases of sudden infant death syndrome (SIDS). This hypothesis is predicated on the striking similarity in age distribution of these two ostensibly disparate disorders. Our investigations began in King County, Wash., on November I, 1976, and concluded on October 10, 1977. King County is a unique resource for research on SIDS because all suspected cases of sudden infant death are referred to a single facility-the Children's Orthopedic Hospital and Medical Center in Seattle-for definitive pathologic diagnosis (pathologic examinations performed under the supervision of Dr. J. Bruce Beckwith). This situation has existed since 1963, when a special state statute [3] was enacted to stimulate research into SIDS. We planned to obtain specimens of stool and of heart blood from each case diagnosed as a sudden infant death
Materials and Methods
Forty-one cases of SIDS occurred in King County during the period of investigation; specimens of stool and heart blood were obtained from all but II of these infants. In addition to the 30 specimens from SIDS infants, we obtained specimens from eight infants who were diagnosed as having died from causes other than SIDS. Our original plan called for the study of a larger group of non-SIDS infants than was accomplished. This departure from our original plan resulted from the press of other professional responsibilities and commitments, and the investigation was terminated with only a single positive finding. Nevertheless, we feel that our findings should be reported. Despite the ostensibly optimal circumstances in King County for this sort of investigation, we encountered a variety of situations that delayed laboratory examination of specimens (Botulism Laboratory, National Fisheries Service, Seattle, Wash.). Table I describes this aspect of our experience. Most of the delays involved the interval between the time of the infants' deaths and the time the autopsies were performed. Most specimens
This investigation was supported in part by grant HD 09297 from the National Institute of Child Health and Human Development. Please address requests for reprints to Dr. Donald R. Peterson, Department of Epidemiology SC-36, School of Public Health, University of Washington, Seattle, Washington 98195.
630
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
Fecal and serum specimens taken from 30 cases of sudden infant death and from eight cases of nonsudden infant death that were diagnosed at a single facility in King County, Wash., were examined for the presence of Clostridium botulinum organisms and toxin. Organisms, but not toxin, were recovered from a fecal specimen in one case of sudden intant death, results that parallel those from studies previously reported by investigators in California. Studies made in our laboratory of a nonfatal case of infant botulism revealed that an estimated 366,000 mouse minimal lethal doses of toxin were excreted in feces collected by purging the infant. Organisms and toxin were excreted for at least 15 days after the infant was hospitalized. Observations made in our laboratory of atypical responses in mice to both fecal and serum extracts, coupled with recently described experiments in which mice were used as an animal model for infant botulism in humans, provide a biologically plausible foundation for the hypothesis that C. botulinum may be implicated etiologically in some sudden infant deaths. Additional microbiologic, physiologic, and toxicologic data are needed to adequately test this hypothesis.
631
Sudden Infant Death and Infant Botulism
Table 1. Interval between death of infant and initiation of laboratory examination of fecal and blood specimens for the presence of Clostr£dium botul£num. King County, Washington, November 1, 1976-0ctober 10, 1977.
Interval (hr)
Total
Cases of non-sudden infant death
5 5 10 7 3
2 1 3 1 1
30
8
were delivered to the laboratory within 1-2 hr after the postmortem examination. All of the specimens were maintained at refrigerator temperature (2 C). The age and sex distribution of the SIDS group and the non-SIDS group are shown in table 2. Twenty-two infants in the SIDS group were white, five were black, and three were of mixed racial extraction. All eight of the infants in the nonSIDS group were white. Stool specimens were obtained from the infants by milking the contents of the resected colon into sterile jars or large test tubes (Vacutainer® anaerobic specimen collector, Becton-Dickson, Rutherford, N .J.) that were specially prepared to maintain anaerobic conditions during transport to the laboratory. A swab of a portion of the lumen of the small intestine of nine infants who died of SIDS during the latter phase of the investigation was taken for subsequent culture and was transported in the special test tubes. Specimens of he~rt blood were obtained by aspiration with a stenle needle and syringe. Upon arrival at the laboratory, the swab specimen and an aliquot of the stool specimen were inoculated into cooked meat medium, incubated, and assayed for the presence of botulinal toxin by injecting the culture supernatant ip into mice [4J. The remainder of the stool was extracted with gelatin-phosphate buffer and the supernatant fluid was assayed. Assays for the presence of toxin in specimens of blood were performed with undiluted serum; in addition, six serum specimens and four stool extracts were concentrated four-
Results
Gelatin-phosphate extracts of 15 of the 30 stool specimens from SIDS infants produced an atypical lethal reaction in mice; these extracts, although lethal, were not neutralized by botulinal antisera. Two of the 30 serum specimens evoked a similar response. The lethality to mice of a few of these samples was eliminated by neutralizing the sample with antisera to Clostridium perfringens (welchii), type A. No Clostridium botulinum organisms were isolated from swab specimens taken from the small intestine. Extracts from stool specimens from five of the eight non-SIDS infants induced nonpathognomonic symptoms and signs of toxicity when injected ip into mice. The interval between the death of the infant from whom samples exhibiting nonspecific toxicity were taken and the time the samples were examined in the laboratory ranged from 6-80 hr with a fairly even distribution over this range. A stool specimen from a black baby girl who died on October 9, 1977 at 206 days of age, and whose death was attributed to SIDS, was examined in the laboratory --10 hr after her death. Type A botulinal toxin was identified in the supernatant of a subculture of the stool specimen grown in
Table 2. Age and sex distribution of cases of sudden infant death and of non-sudden infant death, King County, Washington, November 1, 1976-0ctober 10, 1977. Sudden infant death Age (days)
Male
Non-sudden infant death
Female
Male
Female
0-30 31-60 61-90 91-120 121-150 151-180 181-210 211-270
0 3 6 8 1 0 0 0
3
1 3 1 2 1 1 0
0 0 1 0 0 1 0 0
0 0 3 0 1 0 0 2
Total
18
12
2
6
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
0-12 13-24 25-48 49-72 73-96'
Cases of sudden infant death
fold by dialysis against polyethylene glycol before being injected into mice. Laboratory personnel identified all specimens examined in this study only by code.
Peterson et
632
Discussion
The isolation of C. botulinum organisms, but not botulinal toxin, from the stool specimen from one of the 30 (3%) SIDS infants examined parallels the results of the more extensive studies made in California [6], where nine of 211 (4%) of the stool specimens from SIDS infants showed evidence of C. botulinum organisms. In two of these nine infants, both C. botulinum organisms and toxin were identified in a swab taken from the small intestine but were not identified in the feces. We recognize that these results are not impressive statistically, but some credibility is lent to the association of SIDS and the presence of C. botulinum by the similarity in the two studies of the proportions of samples in which these organisms were found. In fact, these percentages probably underestimate the actual relative frequency of the involvement of C. botulinum because of logistic and technological problems encountered in both the California and the King County studies. In King County there were few well-identified controls. In California, however, as of September 29, 1978, 76 infants who died of causes not attributed to SIDS had been examined in the same manner as were the 211 SIDS infants; there wasn't a single positive identification of either C. botulinum organisms or toxin (Stephen Arnon, personal communication). The facts that an infant with manifest clinical botulism can excrete prodigious quantities of toxin and still survive and that toxin can occasionally be identified in specimens from SIDS infants, poses a curious paradox regarding the hypothesis suggested by the investigators in California-that infant botulism may account for some cases of SIDS. If some cases of SIDS constitute one extreme of the spectrum of severity of infant botulism, one would have thought that previous investigations of SIDS would have revealed antemortem signs of cranial nerve paresis and of constipation in affected infants. However, a relatively low frequency of these signs in SIDS infants could account for the lack of their mention in published reports of SIDS. Nevertheless, the demonstration that botulinal toxin activity persists in feces kept under laboratory conditions does not discount the possibility
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
cooked meat medium. A pure culture of C. botulinum was later isolated. The swab specimen from the small intestine of this infant was negative for C. botulinum organisms. To assess the stability of toxin in stool specimens, feces from two infants in the SIDS group were each combined with type A botulinal toxin to achieve a final concentration of 1,000 mouse minimal lethal doses (MLDs)jg of feces (wet weight). The two mixtures were refrigerated at 2 C for 18 hr, extracted with gelatin-phosphate buffer and assayed for the presence of toxin. The assay was positive for only one of the samples. In November, 1977, approximately one month after terminating our surveillance of SIDS infants for evidence of the presence of C. botulinum, a 1O~-week-old white baby girl with suspected infant botulism was admitted to the Children's Orthopedic Hospital. Her symptoms included loss of head control, weak suck, floppiness, and constipation that had begun 10 days before her admission to the hospital (clinical details provided by Dr. A. Thomas Collins). The initial stool specimen was extracted with gelatin-phosphate buffer, and the extract yielded 500 MLDs of type A botulinal toxinjg of feces. A dose of 0.5 ml of her serum, collected on the same day as the stool specimen, was found to be nontoxic in mice injected ip. However, the one mouse injected with serum concentrated four-fold by dialysis against polyethylene glycol developed symptoms suggestive of botulism after 16 hr; these symptoms disappeared after 24 hr. The mouse died six days later. Insufficient amounts of serum precluded further study of these effects. On the day after the initial stool specimen was collected, feces obtained from the infant after administration of a laxative yielded 4,000 MLDs j g of feces. This purge resulted in the excretion of an estimated 366,000 MLDs of toxin, an amount far greater than that which is lethal to humans [5]. The infant continued to excrete type A botulinal toxin and type A C. botulinum organisms for at least 15 days after she w.as admitted to the hospital; nevertheless, she survived. A portion of the feces was refrigerated at 2 C for 40 days and then was extracted and assayed for the presence of toxin. More than 32 mouse MLDs of toxinjg of feces were still present in the specimen.
at.
633
Sudden Infant Death and Infant Botulism
J
tion of in vivo toxin production provide a biologically plausible explanation for the relatively benign course of clinical infant botulism. It is conceivable that a temporarily incompetent ileocolic valve could allow retrograde passage of a spurt of toxin from the cecal end of the colon into the ileum, from which toxin could readily be absorbed, producing concentrations of toxin in the circulation sufficient to cause SIDS. Likewise, clinically manifest but nonfatal botulism could be explained either by reflux of nonlethal doses of toxin from the large intestine to the small intestine or by absorption directly into the circulation of only a small fraction of the toxin concentrated in the large intestine. An adequate test of this hypothesis will depend upon further research into the ecologic development of microflora in the intestine of the human infant, the physiologic variation in function of the ileocecal valve in infants, and the development of more sensitive methods for the detection of circulating and/or synaptically fixed toxin. Conclusion
Our surveillance of SIDS for evidence of infection with C. botulinum confirms the results of more extensive research conducted in California. Based on observations made in our laboratory and in others, we have formulated a biologically plausible hypothesis for the pathogenesis of SIDS. References 1. Pickett, J., Berg, B., Chaplin, E., Brunstetter-Shafer, M. Syndrome of botulism in infancy: clinical and electrophysiologic study. N. Eng!. J. Med. 295 :770772,1976. 2. Midura, T. F., Arnon, S. S. Infant botulism: identification of Clostridium botulinum and its toxin in faeces. Lancet 2:934-936, 1976. 3. State of Washington Laws of 1963, Chapter 178, RCW 68.08.100 (1973 Suppl.). 4. Eklund, M. W., Poysky, F. T., Wieler, D. I. Characteristics of Clostridium botulinum type F isolated from the Pacific Coast of the United States. Appl. Microbiol. 15:1316-1323,1967. 5. Morton, H. E. The toxicity of Clostridium botulinum, type A toxin for various species of animals, including man. Project Summit, Contract no. DAI8-064-Cml2757, Institute for Cooperative Research, University of Pennsylvania, 1961.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
that, in vivo, toxin is inactivated after death during the appreciable interval between death and the diagnosis of SIDS at autopsy. The toxicity that was found in both stool and serum specimens that was not attributable to the presence of botulinal toxin may have resulted from toxins or metabolic products of other organisms, such as C. perfrigens that thrive in a dead host and that may suppress the growth of C. botulinum and inactivate its toxin, as Smith has demonstrated for samples of soil [7]. Because sudden infant deaths are relatively rare events (average, two in every 1,000 live births), are only occasionally witnessed, and usually occur away from research centers, it seems unlikely that surveillance alone will provide definitive data concerning this syndrome. On the basis of investigations in California and King County, infant botulism accounts for only a small fraction of cases diagnosed as SIDS. Epidemiological studies of SIDS that are directed at such small subset populations pose great difficulties and little promise of success. For this reason, a recent report by Sugiyama and Mills [8] that describes experiments involving intragastric injection of mice with type A C. botulinum spores is noteworthy. They demonstrated that in vivo production of toxin occurred only after the first week of life and could not be induced after the age of two weeks; this characteristic is analogous to the circumscribed pattern of age distribution that is common to both infant botulism and SIDS. Sugiyama and Mills postulated that this sharply demarcated period of susceptibility to small inocula of C. botulinum spores is probably related to the kinds of microorganisms that successively colonize the intestinal tract during the first weeks of life. Unfortunately, relatively little is known about the microflora of the intestinal tract of human infants [9]. In the mice studied, Sugiyama and Mills found toxin only in the large intestine, from which absorption into the systemic circulation would be negligible. None of the mice showed signs of botulism. Miyazaki and Sakaguchi [10] experimented with two-week-old chicks and demonstrated that production and absorption of lethal quantities of botulinal toxin is localized in the cecum of these animals. These observations concerning the loca-
634
6. Arnon, S. S., Midura, T. F., Damus, K., Wood, R. M., Chin, J. Intestinal infection and toxin production by Clostridium botulinum as one cause of sudden infant death syndrome. Lancet 1:1273-1277, 1978. 7. Smith, L. DS. Inhibition of Clostridium botulinum by strains of Clostridium perfringens isolated from soil. Appl. Microbiol. 30:319-323, 1975. 8. Sugiyama, H., Mills, D. C. Intraintestinal toxin in infant mice challenged intragastrically with Clostridium botulinum spores. Infec. Immun. 21:59-63, 1978.
Peterson et al.
9. Mata, L. J., Urrutia, J. J. Intestinal colonization of breast-fed children in a rural area of low socioeconomic level. Ann. N.Y. Acad. Sci. 176:93-109, 1971. 10. Miyazaki, E., Sakaguchi, G. Experimental botulism in chickens: the cecum as the site of production and absorption of botulinum toxin. Jpn. J. Med. Sci. BioI. 31:1-15,1978.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
Discussion
of C. botulinum positivity in stools of SIDS infants of a particuular age group? DR. CHIN. The highest isolation rate was in the children under one month of age: three positive stools of 19. In all the other age groups, the isolation rate was approximately 4%-6%, except in the group of 40 infants aged two months, among whom there were no isolates. DR. RONALD ALTMAN. Has anybody looked for botulinal or other toxins or organisms in children with other illnesses? DR. CHIN. Not yet. DR. V. R. DOWELL, JR. I just want to make a comment in regard to the stability of the toxin. Holding the specimen in the refrigerator is not getting at the question of whether the toxin deteriorates in the cadaver. Postmortem changes in the microbial popUlation can occur, and the toxin can be destroyed before the specimen is collected. It is well known that Clostridium perfringens, for example, is commonly found in heart blood of cadavers if they are not studied early. I think this may account for some of the problems of toxin testing of the intestinal contents-the changes that occur postmortem before the specimen is collected. DR. ALTMAN. You reported nonspecific deaths in mice. This means that toxin was not heat-labile and/or not neutralized by botulinal antitoxin. When we lack a known etiology and something in stools kills mice, it makes you wonder whether there isn't something else present. DR. DOWELL. We have encountered Clostridium ditficile toxin in the stools of one of the babies. DR. LOUIS SMITH. In relation to the stability of the toxin, Dr. Peterson stated that he and his associates added 100 LD50 /g to two different samples of feces. In one of those samples, the toxin all disappeared overnight. So toxin can be destroyed very rapidly at room temperature. In two other cases, the toxin disappeared more slowly. There is a great deal of individual variation from one specimen to another. I think it would be foolish to generalize about toxin stability in stool.
635
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
DR. ROGER FELDMAN. One of the characteristics of infant botulism is a peak onset at two months of age. Is that peak reproduced with the sudden infant death syndrome (SIDS)? Are SIDS cases seasonal throughout the country? DR. DONALD PETERSON. The age distribution curve of SIDS is almost superimposable in every study that's been done. The distribution by age is one of the few constants that we have found in this research. The seasonal distribution is similarly reproducible. In the Southern Hemisphere, it is reversed but is still striking. From the family study data that I described earlier, we have just sorted all 1,200 cases by latitudes and then looked at the seasonal distribution in the northern, middle, and lower tiers of states, and we get the same pattern. I'm convinced that the seasonal association for SIDS is reproducible. There is a clear relation between birth order and sudden death, with a relative sparing of the firstborn. There is an inverse relation with maternal age. The highest risk accrues to children of young multiparas. DR. JAMES CHIN. We had 211 SIDS and supposedly 69 non-SIDS deaths, but the one nonSIDS death in which we found Clostridium botulinum organisms was a 14-day-old infant who died suddenly and unexpectedly at home. A local professor of pathology, also a SIDS Foundation member, considered this a typical SIDS case and recommended that the County Health Department provide SIDS counseling to the family. The coroner's pathologist attributed the death to "cardiopulmonary failure secondary to acute tracheolaryngiobronchitis." DR. JOHN NELSON. There were 192 infants under six months of age in the SIDS group and 49 in the non-SIDS group. There were nine infants positive for C. botulinum of the 192 in the SIDS group and one infant positive of the 49 in the non-SIDS group. DR. CHIN. We prefer to say that we found C. botulinum organisms in 10 infants (and also toxin in two of them) who died suddenly and unexpectedly. DR. FELDMAN. Was there a higher frequency
636
State Health Department has done to define the frequency of C. botulinum isolation. DR. CHIN. When we did our infant death study, some of our methodology evolved during the year. Once we started collecting stool specimens from infants with SIDS, we postulated where the toxin would be produced and asked for specimens from specific parts of the gastrointestinal tract. Our method finally evolved to the point at which we tied off the piece and had a sort of sausage package. But even then, with this procedure, we couldn't be absolutely certain that the contents were from that specific area. DR. DOWELL. Do you think it's possible that other clostridia could be involved in SIDS? DR. PETERSON. It's a question I wanted to ask myself. Do we know of other clostridia that produce neurotoxins besides C. botulinum? One of the problems with a new disease of low frequency is funding. The CDC should have a funding arrangement for pilot studies comparable to those of infant botulism, so that investigators have a chance to move when their interest is high and when the opportunity is best and not have to piece together things on a day-by-day basis. The work done in Seattle [Wash.] has been piggybacked on other grants and contracts. DR. FELDMAN. There are funds at the CDC for the study of a new disease; witness our efforts on Legionnaires' disease. The question is the priority given to this clinical problem and its association with SIDS. DR. ALTMAN. That's not a fair analogy. The money spent on Legionnaires' disease was spent to give services. It was spent on people running to New York, Philadelphia, Dallas, Memphis, and other cities. What is being asked for here is money for a non-CDC group that is competent to study a problem of concern to the CDC. Funds should be available for that purpose.
Downloaded from http://cid.oxfordjournals.org/ at New York University on September 5, 2015
DR. THADDEUS MIDURA. Some of the infant botulism cases that were diagnosed as myasthenia gravis were tested with pyridostigrnine, and when their stools were studied the test mice died. Fortunately, the work that Dr. Hatheway did with dialyzing the stool to get rid of the drug was then reported. This was one situation where what was seen in adults was used in studying infants. DR. CHIN. In Utah, an infant was brought to the llOspital outpatient clinic for possible failure to thrive and was examined by the pediatric neurologist, who found nothing. The child was sent home. Two weeks later, he was admitted to the hospital after a respiratory arrest at home and expired after several days in the hospital's intensive-care unit. The pediatric neurologists asked for testing for botulism, and the test was positive. That was a case that resembled SIDS but was infant botulism, as determined by the state and the Center for Disease Control (CDC). On the death certificate, SIDS was the stated cause of death. DR. PETERSON. My reaction to the relation between SIDS and infant botulism is that I feel frustrated at this point. It is obviously low-yield research. It's not very exciting if the proportion of SIDS that we can attribute to infant botulism is only 3%-5~o' But the other cause of frustration is the technology. If we're going to study SIDS and screen many babies' stools with the present techniques, we have a very expensive proposition. The techniques are not necessarily all that specific, so we are not sure about our negative results. And there are all of the other factors that make interpretation of what we think we know difficult. I think we are in a situation in which we have to let basic science do whatever it can do and take off again when we have better methodology. I guess I'm thinking that the emphasis should be on infant botulism and not on further case control studies of the kind that the California
Discussion
