THE FRANCIS E. CAMPS INTERNATIONAL SYMPOSIUM
ON SUDDEN AND UNEXPECTED DEATHS IN INFANCY
ROBERT R. ROBINSON
ALAN H. FINLEY
J. HILLSDON SMITH
THE CANADIAN FOUNDATION
FOR THE STUDY OF INFANT DEATHS
*Robert C. Reisinger, D.V.M., M.S. was associated with the National Cancer Institute, United States Public Health Service, Bethesda, Md. His present address is 3810 Dustin Road, Burtonsville, MD 20866. Phone 301-421-9377 or Fax 301-421-0854.
Absorption into the bloodstream over hours of time of small amounts of bacterial endotoxin not detoxified by a temporarily dysfunctional reticuloendothelial system results in removal of blood platelets and fibrinogen from the circulating blood.1 The result is release of relatively large amounts of serotonin from platelets into the blood plasma (in some experiments the increase of plasma serotonin is almost 100-fold). Serotonin initiates in some cases the coronary chemoreflex (Bezold-Jarisch reflex) in which there is inhibition of sympathetic outflow and increased activity of the cardiac (efferent) vagus, leading to profound brachycardia, hypotension, and cardiovascular collapse.2 The complex pathogenesis of endotoxemia, depending on time and dosage, also involves release of norepinephrine, epinephrine, corticosteriods, etc. However, if death occurs early in the course of this syndrome, it is due primarily to serotonin effect. Serotonin is associated with deep sleep and in certain circumstances strongly inhibits respiratory movements. . . Endotoxin also has a more direct effect on cellular respiration, since it interferes with oxidative metabolism of mitochondria in vitro as well as in vivo3. . . Between three and six hours, vascular capillary permeability has become more substantial, and varying amounts of edema and hemorrhage by diapedesis are apparent. After six to eight hours or more, fibrin-platelet clots have formed, and disseminated intravascular coagulation (DIC) is present in lungs, kidneys, and other organs and tissues.
There are very few Escherichia coli in the more absorptive portion of the small intestine, the duodenum, jejunum and proximal ileum. This is true of all mammalian species so far studied, including the cow and calf, the human infant and adult4. Smith and Orcutt5 demonstrated that calves suffering diarrhea have tremendously increased numbers of E. coli in the intestinal tract. They described a "great increase in the number of E. coli in the lowest third of the small intestine with a spreading of the invasion towards the duodenum as the disease gains headway. Under these conditions, a general intoxication results . . . E. coli in the digestive tract has not been in general regarded as significant. This significance appears when the quantitative factor, obtained before natural death, is determined." Their original work has since been confirmed in calves by other workers4,6 who further reported that infrequently death may occur before diarrhea is evident, with similar overgrowth of E. coli in the absorptive portion of the small intestine. Two of these studies demonstrated a triggering effect of various viral infections, and further demonstrated that the viral infections are non-lethal in the absence of E. coli, or in the presence of relatively few E. coli, in the digestive tract.
In the enterotoximic form of diarrhea and sudden death syndrome the blood and internal organs are usually free of bacteria.
In the only appropriate study thus far reported of the gastrointestinal microflora of human infant victims of SIDS, Bendig and Haenel7 reported finding an abnormal microbial flora in the intestinal tract, including greatly increased numbers of E. coli in the proximal small intestine, of 24 out of 29 cases. "Only in three infants did there exist normal, eubiotic relations." Beller and Graeff1 have demonstrated in the rabbit that continuous i.v. injection of very small amounts (50 mµg/kg/hr.) of E. coli endotoxin up to 14 hours results in quiet death which occurs at varying lengths of time after injection has begun, although some few rabbits do not die. There is a decrease in platelets and fibrinogen, so that after six hours, the blood does not clot. Animals which die early have few or no lesions, those which die somewhat later show varying amounts of edema and petechial hemorrhage of the lungs. Disseminated intravascular coagulation is observed only in rabbits which die eight hours or longer after infusion began. This extremely relevant experiment beautifully demonstrates the varying susceptibility of individual animals at various times to endotoxin effect, and the variety of lesions which may be produced from very little or nothing to moderate to severe. The pathology is consistent to that reported in SIDS in the human, the calf and other mammalian species.5,6
If one considers the relatively obscure pathogenesis and reported pathology of other infant diseases including cerebral palsy, mental retardation, hearing and speech defects, retrolental fibroplasia, respiratory distress syndrome, etc., and accepts the reality of bacterial endotoxin absorption from the gastrointestinal tract, it is not difficult to see the possibility of endotoxin involvement in at least a portion of these syndromes. It is increasingly apparent that bacteria and bacterial toxins play an important role in the pathogenosis of many "viral" diseases.4,8 Canine distemper virus (CDV), an agent closely related to measles virus, is associated in dogs with a disease syndrome often characterized by encephalitis, pneumonitis and enteritis. Brain lesions and CNS sequelae following CDV infection in the dog are similar to those characteristic of cerebral palsy in the human. However, the only signs of disease in gnotobiotic dogs infected with CDV were mild transient fever and leukopenia. Very similar findings have been reported in gnotobiotic kittens infected with feline panleukopenia virus, an agent which causes a highly lethal enteritis in conventional kittens.
Dubos et al developed and maintained a colony of mice (NCS) practically free of E. coli in which the largest percentage of intestinal flora cultivable both aerobically and anerobically consists of organisms commonly classified as Lactobacillus and Bacteroides spp. NCS mice grow faster, are more resistant to lethal effects of endotoxin and have less exacting nutritional requirements than control mice. However, administration per os of even small amounts of penicillin brings about a sudden disappearance of lactobacilli from fecal flora of NCS mice accompanied by an explosive and lasting increase in enterococci and gram negative enterobacilli. E. coli, which is not normally found in the stool cultures of the NCS mice, became abundant following treatment with penicillin. These findings are similar to those of De Somer et al in the guinea pig which also normally has a primarily gram positive intestinal microflora, and dies from E. coli overgrowth when penicillin or the tetracyclines are administered. Schaedler and Dubos found that in the mouse "the composition of the bacterial flora could be rapidly and profoundly altered by a variety of unrelated disturbances, such as sudden changes in environmental temperature, crowding in cages, handling, administration of antibacterial drugs, etc. The first effect of the change was a marked decrease in the numbers of lactobacilli and commonly an increase in the numbers of gram negative bacilli and enterococci. When tested three weeks after these disturbances some NCS animals, normally relative resistant, were found to have become susceptible to the lethal effect of endotoxin."
Dubos et al reported that the numbers of lactobacilli recovered from stools of mice fed diets of natural materials were much larger than in those mice fed a casein semi-synthetic diet.
Dubos et al and Ravin et al have shown that endotoxin is being continually absorbed into the circulatory system of animals which have appreciable numbers of E. coli in their digestive tracts. Fine and co-workers have stated that endotoxins are always at hand ready to destroy peripheral vascular integrity, and to kill when endotoxin detoxifying power is lost, and have reported that blocking the reticuloendothelial system of the rabbit with thorotrast (a sterile colloidal suspension of 25 per cent thorium dioxide in dextrins) makes this animal exquisitely susceptible to the effects of endotoxin. Rabbits so blockaded can be killed with one one-hundred-thousandth of the normally lethal dose of endotoxin.
Administration of Diphtheria-Pertussis-Tetanus Toxoid (DPT) can cause temporary liver dysfunctions in infants similar to those resulting from viral hepatitis, and inoculation of killed Bordetella pertussis organisms makes some strains of mice 200 times more sensitive to histamine and three to five times as sensitive to endotoxins for approximately 14 days.
Deaths can occur within several hours after the feeding of an apparently healthy normal calf.4,10 Pathologic lesions at autopsy are absent or minimal, as in the crib-death syndrome in the human infant. Experimentally, the diarrhea syndrome can be precipitated by exposure to various viruses, cold and wet, avitaminous A, etc. any "adverse contributing factor" that impairs optimal reticuloendothelial and/or gastrointestinal function. A feature of many cases of the calf-diarrhea syndrome is labored breathing (pneumonic signs without pneunionic lesions), but this disappears when, by appropriate antibiotic therapy, excessive numbers of E. coli are cleared from the digestive tract.
Death may occur in calves before diarrhea is evident. Writing of diarrhea in the human infant, McKay and Smith11 stated that "An occasional infant may go into shock and die from water and electrolyte loss into the intestinal lumen before a diartheal stool is passed." It is difficult to accept that lethal loss of water and electrolytes could occur without evidence of diarrhea, and bacterial endotoxin action is probably occurring in such cases.
The celiac crisis, "an acute medical emergency and a severe and immediate threat to the patient's life" is often triggered by upper respiratory infection. The child is prostrate, drowsy and dehydrated, and has the chemical and laboratory manifestations of acidosis.12
Severe pneumonias of unexplained etiology in young children have been referred to as "missed cot deaths", and the signs and course of the disease, as well as the described pathology, are compatible with known endotoxin effects.
Some factors common to many cases of SIDS, the respiratory-distress syndrome, and endotoxin shock are hyperkalemia, hyponatremia, acidosis, thrombocytopenia, noncoagulability of blood, early respiratory signs without appropriate lesions, pulmonary edema, hemorrhage by diapedesis, fast, weak pulse, and circulatory collapse. It is more illogical to consider these similarities as fortuitous than to realize the probability, or at least possibility, of a common cause or mechanism.
There is a great deal of evidence to indicate that the breast-fed baby is relatively healthier, suffers less allergy, respiratory, enteric and other disease, and less mortality than his bottle-fed counterpart.11,13-17
SIDS is preponderently a disease of the bottle-fed infant. Coombs18 stated that if SIDS were relatively as common in the breast-fed as in the bottle-fed infant he should have had 17 breast-fed cases in his series, whereas at that time he had not one. Johnstone and Lawry19 reported, "In the 46 cases where the type of feeding is definitely known all but two were fed on dried cow's milk. The two exceptions were aged four and eight days." Tonkin reported that in her series of 86 SIDS cases, only two were breast-fed. Since twenty-five percent of her control population were breast fed, she should have had 21 cases of SIDS in breast-fed infants, if the risk were the same in both breast-fed and bottle-fed. Steele20 reported SIDS to be significantly related to feeding other than at the breast.
The data of Rivera21, concerning feeding of infants in low income families attending a New York City clinic indicate that more than 30 per cent of two-month-old infants were fed evaporated milk formulas. and an additional 20 per cent or more were fed on fresh cow's milk. These data are interesting and probably significant when coupled with various reports of epidemiological studies in various countries indicating a significant increased incidence of SIDS in low income groups.
In rural Chile, infant mortality rose with income. It is reported that with higher income and better education, weaning was practiced earlier and more infants were fed by bottle only. The National Health Service of Chile attempts to provide all weanlings and preschool children with dried cow's milk.
There are important differences between bacterial flora, pH, and physical characteristics of the intestinal contents of human infants fed human milk and those fed cow's milk.13,22,23 Due to its high content of calcium and protein and its lower content of lactose, cow's milk fed to the human infant raises the E. coli count in the large intestine approximately 1,000-fold (from 106-107 /g to 109-1010 /g), raises the pH from acid (4.5-5.6) to alkaline (7.0-8.0), makes curds hard and coarse instead of soft and fine, and makes bowel movements relatively infrequent. Characteristics of a healthy human infant on human milk are a relative low coliform count, acid pH, soft, fine curds, and frequent bowel movements. Even if there were no SIDS, use of cow's milk or any other formula during the first six months of age is against all scientific reason to produce the optimally healthy child and the healthy adult he should become.
Although SIDS is primarily a disease of the bottle-fed infant, it does occur in fully breast-fed infants. Relatively long term studies on the fecal flora of breast-fed infants show that the numbers of E. coli may increase intermittently for variable periods of time, but they seldom achieve or maintain numbers characteristic of the cow's milk-fed infant. Thus the breast-fed infant is presumably at high risk for shorter periods of time.
Evidence accumulated in studies of the young of various mammalian species, including the human infant, indicates that the Sudden Infant Death Syndrome is not a disease entity, but a peracute, lethal manifestation of other disease syndromes, including gastroenteritis and respiratory disease.
The "endotoxin theory" does not detract from, but is additive to, several other theories postulated for SIDS, such as sudden vagal storms and other malfunctions of the autonomic nervous system, apnea. anoxia, allergy to cow's milk, overwhelming viral, bacterial, or mixed infection, etc.
It is not claimed that the "endotoxin theory" will explain all cases of SIDS, but it is believed that endotoxins, and other bacterial toxins, are directly involved in a high percentage of these cases.
(A more complete list of references may be obtained from the author.)