Original Articles
May 1970
Vol 119 No 5

Visceral Manifestation of Shock in Congenital Heart Disease
Ronald Coen, MD, and
A. James McAdams, MD, Cincinnati

Visceral lesions in 49 infants with ductus dependent malformations of the heart were surveyed from postmortem material collected over a seven-year period. These lesions, all of which have previously been ascribed to ischemia, were found to be related to the duration of clinical illness dating from the occurrence of closure of the ductus arteriosus. The liver and adrenal glands were found to be the most consistently affected organs in peripheral circulatory collapse, whereas the kidneys and brain seem to be significantly spared. The frequent occurrence of hypoglycemia in infants with the hypoplastic left heart syndrome apparently has its origin in liver ischemia and its consequent glycogen depletion.
he infant with congenital heart disease who appears normal early in life and then suddenly de-
Received for publication Dec 5, 1969.
From the departments of pediatrics (Drs. Coen and McAdams) and pathology (Dr. McAdams), University of Cincinnati Medical Center, and Children’s Hospital, Cincinnati.
teriorates is most often found to have one of the three following cardiac anomalies: (1) hypoplastic left heart syndrome, (2) preductal coarctation of the aorta, or (3) transposition of the great vessels without a ventricular septal defect. Functional patency of the ductus arteriosus is known to be essential in maintaining the circulation in the infant with the hypoplastic left heart syndrome 1 and appears to be equally as important to the infant with coarctation of the aorta or transposition. Because blood flow through the ductus arteriosus is critical in these three conditions, they can be referred to as “ductus- dependent” cardiac malformations. The purpose of this investigation was to study the visceral consequences of perfusional failure which results from functional closure of the ductus arteriosus in three groups of infants with ductus-dependent malformations. It is believed that this information is necessary if temporizing
procedures are to be devised for ductus-dependent malformations.
Seventeen infants with the hypoplastic left heart syndrome, 15 with preductal coarctation of the aorta, and 17 with transposition of the great vessels, selected from the pathology files of the Cincinnati Children’s Hospital during the period 1960 through 1966, were studied. Their clinical records were surveyed for signs and symptoms which heralded the onset of the terminal illness, and, by our definition, the onset of functional closure of the ductus arteriosus. The duration of the terminal illness was determined in order to tabulate the visceral lesions according to the period of circulatory failure rather than the patient’s chronological age.
Autopsy material, including the anomalous hearts and microscopic sections from the liver, adrenal glands, kidneys, brain, pancreas, myocardium, and testes were available for examination. In each of the formaldehyde-solution fixed hearts, the ductus arteriosus was inspected grossly for thickening of the vessel wall and for wrinkling of the intimal layer. According to previously documented evidence,2 these findings

Amer J Dis Child/Vol 119, May 1970

indicate that the ductus arteriosus is functionally closed»
Sections from the body viscera and central nervous system (CNS) were reviewed microscopically for evidence of ischemic injury. The changes considered to be significant were assessed only from the standpoint of their presence and not their severity. It is not possible to determine in any one case the degree to which perfusion across a closing ductus is reduced. Some infants, for example, have no detectable femoral pulsations while others may have thready pulsations. This observation and the absence of any obvious correlation. of severity of organ change with duration of clinical symptomatology made quantitation unprofitable and possibly misleading.
Those data tabulated were hepatic sinusoidal thrombi (Fig 1), centrilobular necrosis of the liver (Fig 2), focal necrosis in the adrenal fetal cortex (Fig 3), necrobiosis of pancreatic islet cells (Fig 4), necrobiosis of testicular spermatogonia (Fig 5), focal myocardial necrosis (usually located in a papillary muscle), and selective neuronal necrosis in the CNS. Those areas of the CNS examined included the outer cortex, hippocampus, central nuclei, pons, cerebellum, and medulla. Apparent ovag- onial necrobiosis is a more or less general phenomenon in newborns and not subject to analysis. Appropriate sections were available for examination in all but one of the patients in whom only the liver was represented.
To compare the severity of the circulatory disturbances in these infants with ductus dependent malformations of the heart, 14 infants with hyaline membrane disease were surveyed in a similar fashion. This group was selected because hyaline membrane disease is the only sizable, single entity of the early neonatal period with relatively uniform symptom complex and pathologic findings. Furthermore, although hypotension has been noted in infants with hyaline membrane disease,3 septicemic shock is a relatively infrequent complication. Finally, an estimation of the duration of clinical illness in this group of infants is made with ease, since the onset of the disease occurs at birth.
The results of this study are included in Table 1 to 3 for the infants
with ductus dependent cardiac malformations and Table 4 for the infants with hyaline membrane disease.
Clinically, in the infants with cardiac anomalies, symptoms elicited at the onset of the terminal illness were feeding difficulties, respiratory distress, and ashen cyanosis. Direct evidence of circulatory collapse was noted in the form of absent femoral pulses in 31 of the 32 infants with a hypoplastic left heart or coarctation, but this was present in only two of the 17 infants with transposition. Admission blood pressures were recorded in only 18 of the infants with cardiac disease. Differences in pressure of 30 to 90 mm Hg between the upper and lower extremities were recorded in two infants with coarctation (No. 2 and 15) and two with transposition (No. 13 and 16). The remaining 14 infants had normal admission blood pressures. One infant (transposition No. 9) had been studied by cardiac catheterization at age 36 days. At that time the systolic blood pressure in the right ventricle exceeded that in the aorta by 28 mm Hg. Ballooning of the atrial septum (Rashkind procedure) was not performed.
As a group, patients with the hypoplastic left heart syndrome were younger and had a shorter clinical illness than patients with coarctation of the aorta or transposition of the great vessels. It is notable that 65% of the patients, whatever the cardiac defect, had a total clinical illness of not more than four days, and in 49%, the interval was less than two days. Each of the 49 malformed hearts showed gross evidence that the ductus arteriosus was functionally closed; ie, the vessel wall was thickened and its intimal layer wrinkled. The hearts from the infants with hyaline membrane disease were not available for study. In the infants with cardiac anomalies, microscopic alterations resulting from peripheral circulatory collapse occurred most commonly in the liver and adrenal

Table 1.—Infants With Hypoplastic Left Heart Syndrome
Case No. Sex Age at Admission (Days) Cyanosis Femoral Pulses * Age at Death (Days) Terminal Illness (Days) Liver
一 Adrenal Fetal Cortex Necrosis Other Findings
Thrombi Necrosis’
1 M 1.5 + — 1.5 0.5 + + Neuronal necrosis
2 M 4.0 + 4.5 0.5 + + Neuronal necrosis
3 M 1.5 + 2.0 0.5 + + Pancreatic islet cell necrosis
4 M 1.0 + 1.5 0.5 + + Pancreatic islet cell necrosis
5 F 2.0 + 2.0 1.0 +
6 M 5.0 + 6 1.0 + Pancreatic islet cell necrosis; spermatogonial necrosis
7 F 1.0 + 1.5 1.0 + +
8 M 4.0 + 4 2.0 + + + Myocardial focal necrosis
9 M 3.0 + 3 2.75 + +
10 F 21.0 + 26 2-7 + + Pancreatic islet cell necrosis
11 F 3.0 + 3 3
12 F 3.0 + 4 3.5 + + Neuronal necrosis; renal tubular necrosis
13 M 11 + 18 3-7
14 M 3.0 + 7 5 + + + Neuronal necrosis
15 M 15.0 + 17 5 + + Spermatogonial necrosis
16 M 10.0 + + 13 6 + Renal tubular necrosis; spermatogonial necrosis
17 M 3.0 + — 10 7 + + + Neuronal necrosis
* + indicates present; —, absent.

Table 2.—Coarctation of the Aorta (Juvenile Type)
Case No. Sex Age at Admission (Days) Cyanosis Femoral Pulses * Age at Death (Days) Terminal Illness (Days) Liver
人 Adrenal Fetal Cortex Necrosis Other Findings
Thrombi Necrosis’
1 F 14 + 14 0,5 +
2 M 2.5 + 3 03 + + +
3 M 4.0 + 4.5 0.5 + + Spermatogonial necrosis
4 F 11 + 11 1.0 十 + Not examined
5 M 90 + 90 1.0
6 M 9 + 9 1.0 + +
7 F 5 + 6 2.0 + +
8 F 9 + 11 2.0 + + Pancreatic islet cell necrosis
9 M 3 + 4 3.5 Spermatogonial necrosis
10 M 35 + 36 3-15 十 Spermatogonial necrosis; pancreatic islet cell necrosis
11 F 5 + 6 4 + + + Pancreatic islet cell necrosis
12 F 42 + 45 4-36 +
13 F 13 + 17 7
14 M 9.0 + 11 9 + + Neuronal necrosis
15 F 2.0 — 10 10 + Renal tubular necrosis
* + indicates present; —, absent.

glands, whereas the brain and kidneys were the least affected. Surprisingly, CNS and renal involvement was no more frequent than pancreatic islet cell necrosis. Myocardial changes were rare, but this may be more attributable to sampling error. Spermatogonial degenerative lesions were present in 26% of the boys in this study.
More specifically, a pattern of change seems evident in the livers from the infants with the hypoplastic left heart syndrome and transposition of the great vessels. Sinusoidal thrombi were present alone for the greater portion of the first 24 hours of the terminal illness, whence centrilobular necrosis became evident, with or without sinusoidal thrombi. By contrast, centrilobular necrosis was present very early after the onset of clinical symptoms in the infants with coarctation of the aorta. Passive congestion of the liver by either microscopic or weight criteria was not a prominent finding.

386 Amer J Dis Child/Vol 119, May 1970
Table 3.—Transposition of the Great Vessels
Case No. Sex Age at Admission (Days) Cyanosis Femoral Pulses * Age at Death (Days) Terminal
Illness (Days) Liver
人 Adrenal Fetal Cortex Necrosis Other Findings
Thrombi Necrosis’
1 F 18 + + 19 1.0
2 F 0.75 + + 1 1,0
3 M 1.0 + 十 1.5 1.5 +
4 F 0.75 + + 2 1.75 + +
5 M 1.0 十 — 2 2.0 + Neuronal necrosis; pancreatic islet cell necrosis; spermatogonial necrosis
6 F 3.0 + 4 3-4 4- +
7 M 3.5 + + 4 4 +
8 M 4.0 十 + 66 4.0 + Renal tubular necrosis
9 M 1.5 + + 4.5 4.0 +
10 F 60 + 60 5.0
11 M 4.0 + + 7.5 7.0 + + Neuronal necrosis
12 M 45 + — 49 7.0
13 F 16 + + 18 8.0
14 M 14.0 + + 21 16.0 Neuronal necrosis; myocardial focal necrosis
15 F 32 + + 43 17.0 + Renal tubular necrosis; myocardial focal necrosis
16 F 2.75 + + 19 18.0 Pancreatic islet celt necrosis
17 F 35 + + 84 84.0 +
* + indicates present; —, absent.

Focal necrosis of the adrenal fetal cortex occurred frequently within the first 48 hours of terminal illness in infants with hypoplastic left heart or transposition. The fetal cortex was involved less frequently and later in the terminal illness of the infants with coarctation of the aorta. There were few microscopic lesions noted in the organs from the infants with
hyaline membrane disease. In three instances hepatic sinusoidal thrombi were demonstrated after 24 hours of clinical illness. Necrosis of the adrenal fetal cortex was present after 48 hours. Central nervous system ischemic lesions were noted in only one infant although nine of the 24 infants lived longer than 24 hours.

Although the liver, one of the principally affected organs, is perfused by both venous and arterial blood, it is apparent that a diminution in arterial flow accounts for the thrombotic and necrotic lesions observed. In 34 instances of central liver necrosis reported by Ellenberg and Osserman,4 94% 4(were clearly associated with shock.” These same authors provide evidence that congestive heart failure does not play an important role in the pathogenesis of centrilobular necrosis. Similarly, Clarke 5 reported nine cases of centrilobular necrosis associated with myocardial infarction, in five instances of which hypotension was a complicating factor. In 1961, Wade- Evans 6 reported thrombi in the hepatic sinusoids of infants dying with hyaline membrane disease and suggested that the thrombi were formed as a result of % disturbance in circulation.” Gruenwald7 also noted degenerative liver changes, including focal necrosis, in infants who

Amer J Dis Child/Vol 119, May 1970

died shortly after birth of shock and asphyxia.
Variability in the severity and extent of hepatic lesions is noted by each of these authors. Ellenberg and Osserman 4 concluded that the mere presence of shock was not necessarily sufficient to result in central necrosis, but a direct relationship was found between the duration of shock and the appearance of centrilobular necrosis. Our data corroborate their findings. Centrilobular necrosis was found only after 24 hours of circulatory failure in two of the three groups with congenital heart disease. The occurrence of liver necrosis in patients with coarctation of the aorta after a short terminal illness would suggest that the onset of ischemia had occurred even before clinical symptoms became manifest.
The alterations found in the adrenal gland must be interpreted with greater reservation because of the normal changes that occur in the fetal cortex following birth. Under normal conditions the fetal cortex undergoes involution from the time of birth, developing advanced changes by two to three weeks post- natally.8 At the same time the definitive cortex begins to differentiate into the glomerulosa and fasciculata zones. Gruenwald7 related necrosis and hemorrhage in the fetal cortex to asphyxia and shock, comparing these changes to those of the alarm reaction of Selye. Variations in reported changes in the fetal adrenal cortex of infants with congenital heart disease is probably due to the selection of cases studied. Gold- zieher9 reported delayed involution of the fetal cortex in infants with cyanotic congenital heart disease, whereas in the series reported by Tahka,10 26% (4 of 15) had advanced involutional changes. Of the 17 cases of cyanotic congenital heart disease
reported by Lanman,8 six were 11 days of age or less at the time of death, and four of these had accelerated involutional changes. The 33% incidence of adrenal glands with accelerated involutional changes in the fetal cortex in the present series would appear to be representative.
Pancreatic lesions were first associated with asphyxia by Selye.11 Gruenwald7 also described hemorrhagic necrosis and inflammatory changes in the pancreas of infants dying of asphyxia and shock soon after birth and Bernstein12 related islet cell necrosis to asphyxia. In our series islet cell necrosis was found in 18% of the heart cases but was absent in the 14 cases of hyaline membrane disease.
Acute renal tubular necrosis is rarely encountered in the newborn period.12 Jonsson 13 first related the findings of lower nephron nephrosis in a 6-day-old infant to asphyxia and shock which were assumed to be the common denominators. Only one of our five cases with tubular necrosis was under 10 days of age at the time of death. Bernstein 12 found that the renal tubular epithelium was affected earlier and more severely than the islet cells of the pancreas, but our observations conflict with this. Islet cell necrosis preceded the renal abnormalities if one relates the changes to the duration of the terminal illness or the period of circulatory failure. Furthermore, the incidence of renal tubular necrosis (8%) in our series was less tKan that of islet cell necrosis (18%), and in none of the cases were both lesions present.
Neuronal degenerative changes in the brain were relatively few (18%). While it would appear that the brain may receive preferential perfusion, an anatomical basis for this was not demonstrable in these patients.
The additional findings of sper- matogonial necrosis in the testes and of myocardial necrosis lend support to the hypothesis that decreased peripheral perfusion is the basis of the lesions described. The severity of the disturbances in circulation initiated by functional closure of the ductus arteriosus is evident if one compares the nature of the pathological alterations in the livers from infants with cardiac disease to those with hyaline membrane disease. Wade-Evans 6 using phosphotungstic acid hematoxylin PT AH staining techniques reported an incidence of hepatic sinusoidal thrombi in hyaline membrane disease of approximately 20%, similar to ours. In none of his cases or those of hyaline membrane disease presented in this study, however, was there liver cell necrosis. Furthermore, in the infants of this study with hyaline membrane disease, major lesions were not encountered in the kidneys, pancreas, or testes, and neuronal necrosis was present in only one infant. This is to be compared with the incidence of major lesions in these organs in the infants with cardiac anomalies where the brain was involved in 9 of 49, the kidneys in 5 of 49, the pancreas in 9 of 49, and the testes in 7 of the 27 boys.
The fact that the liver was more commonly involved than the brain or kidney in the infants with cardiac disease is of interest. One inference could be that with failure of peripheral blood flow, perfusion of some organs, viz, the kidney and brain, may be selectively maintained over that of the liver. Another inference could be that the liver is simply more susceptible to the adverse effects of underperfusion. Although other organ involvement was seldom observed in the absence of liver lesions (Table 5), the differences in. this small population are not statistically significant. A more positive argument against this possibility of some regulatory mechanism of blood flow in states of shock is the obvious discrepancy between the occurrence of liver and pancreatic lesions. Since these organs have a common blood supply, one would tentatively conclude that inherent organs5 susceptibility is the significant factor. A similar argument can be applied to explain the greater frequency of lesions in the adrenal gland than in the kidney.
At present there is interest in the finding that hypoglycemia is a frequent complication of the hypoplastic left heart syndrome. In this series of patients with ductus dependent cardiac malformations, blood glucose determinations were performed in 26 of the 49 cases. Hypoglycemia was documented in 13 cases, all of which had hepatic lesions. (Hypoplastic left heart syndrome, 7 of 12; coarctation of the aorta, 3 of 7; transposition of the great vessels, 3 of 7.) The etiology of the hypoglycemia has been linked to hyperinsulinemia.14 However, the association of hyperglycemia, shock, and hepatic centrilobular necrosis, as reported by Ellenberg and Osserman,[ Sinha SN, et al: Hypoplastic left heart syndrome: Analysis of thirty autopsy cases in infants with surgical considerations. Amer J Cardiol 21: 166-173, 1968.] [ Hudson REB: Cardiovascular Pathology. Baltimore, Williams & Wilkins Co, 1965, vol 2, p 2026.] [ Rudolph AM, et al: Studies on the circulation in the neonatal period: The circulation in the respiratory distress syndrome. Pediatrics 27:551-566, 1961.] [ Ellenberg M, Osserman KE: The role of shock in the production of central liver cell necrosis. Amer J Med 11:170-178, 1951.] [ Clarke WTW: Centrilobular hepatic necrosis following cardiac infarction. Amer J Path 26:249253, 1950.] [ Wade-Evans T: Thrombi in the hepatic sinusoids of the newborn and their relation to pulmonary hyaline membrane formation. Arch Dis Child 36:286-292, 1961.] [ Gruenwald P: Asphyxia, trauma and shock at birth. Arch Pediat 67:103-115, 1950.] [ Lanman JT: The fetal zone of the adrenal gland: Its developmental course, comparative anat] would suggest that hyperinsulinemia is a secondary response to the hyperglycemia. Recently, Benzing et al15 have found that the liver glycogen content was significantly reduced in four infants with the hypoplastic left heart syndrome. Therefore it would appear that hyperglycemia is an early phenomenon secondary to glycogen release from injured liver cells. The more frequently observed hypoglycemia may well be explained by the ensuing depletion of glycogen stores in the liver, with or without hype— insulinemia.
From the pathological findings in this series, it would seem that once symptoms appear, degenerative changes in the body viscera have already begun. Any therapeutic attack, then, will be dependent upon methods of detection better than those generally available at the present time.


omy, and possible physiologic functions. Medicine 32:389-430, 1953.
Goldzieher MA: Effects of interrenal function on fat metabolism and tissue respiration. Endocrinology 18:179-187, 1934.
Tahka H: On the weight and struction of the adrenal glands and the factors affecting them, in children of 0-2 years. Acta Paediat 40 (suppl) :81, 1951.
Selye H: The general adaptation syndrome and the diseases of adaptation. J Clin Endocrinol 6:117-230, 1946.
Bernstein J: Renal tubular and pancreatic islet necrosis in. newly born infant, Amer J Dis Child 96:705-710, 1958.
Jonsson B: Lower nephron nephrosis in asphyxia neonatorum. Acta Paediat 40:401-408, 1951.
Hait G, et al: Abnormal insulin release in cyanotic heart disease, abstracted. Soc Ped Res, p 11, 1968.
Benzing G III, et al: Simultaneous hypoglycemia and acute congestive heart failure. Circvla- tion 40:209-216, 1969.