1 1 4 Clinical notes The Journal of Pediatrics January 1978 thin and occasionally hypoxic intestinal wall of the low-birth- weight infant to be responsible for perforation. In view of the above possibilities it seems reasonable that silicone, as well as potyvinyl, tubes should not be repositioned and that any abdominal palpation should be performed with extreme care and gentleness in low-birth-weight infants who are being fed through nasojejunal tubes. J. P~rez-Rodr~gues J. Quero E. G. F/ms k: Omehaca Climca lnfantil "'La Paz'" Servicio de Neonatologia A. Mart~nez Servicio de Radiologia C~ General Yague ll-lO~ Madrid 20, Spain REFERENCES 1. Rhea JW, Ahmad MS, and Mange MS: Nasojejunal (trans- pyloric) feeding: A commentary, J PEDIATR 86:45I, 1975. 2. Siegle RL, Rabiniwitz JG, and Sarashon C: Intestinal perloration secondary to nasojejunal feeding tubes, Am J Roentgenol Radium Ther Nucl Med 126:1229, 1976. 3. Heird WC, and Driscoll JM: Newer methods for feeding low birth weight infants, Clin Perinatol 2:309, 1975. 4. Tyson JE, deSa DJ, and Moore S: Thromboatheromatous complications of umbilical arterial catheterization in the newborn period: Ctinicopathological study, Arch Dis Child 51:744, 1976. Yellow pulmonary hyaline membranes This communication is prompted by the recent interest shown in yellow hyaline membranes in the lungs of infants with respiratory distress syndrome and artificial ventilation? 4 We have recently evaluated 299 infants as part of a clinicopa- thologic study of br0nchopulmonary dysplasia ~ seen at this institution from 1962 through 1973 with the clinical diagnosis of respiratory distress syndrome who were ventilated for at least 24 hours. Our initial findings were presented at the Carmel, Califor- nia, Meeting of the Western Society for Pediatric Research on February 6, 19767 Of these 299 patients, 143 died and had autopsies which have been reviewed. The amount of yellow hyaline membranes was graded from 1 + to 2 + when present in at least one of the multiple lung sections of each case. Multiple clinical, radiologic, and other pathologic features were also examined and intercorrelated. RESULTS Of the 143 pulmonary sections pathologically examined, bright yellow material ("yellow hyaline membranes") was found in 51. The material varied from homogeneous ~yellow hyaline membranes to variably sized, "slightly refractile intrac3)toplasmic Table I Year Total no. of infants No. with yellow material (autopsied infants) Percentage with yellow hyaline membranes 1962 2 0/1 0 1963 0 0/0 0 1964 14 1/8 12 1965 22 0/14 0 1966 20 3/17 18 1967 22 1/12 8 1968 16 1/6 17 1969 21 9/11 82 1970 28 10/18 55 1971 41 8/15 57 1972 42 10/20 50 1973 71 8/21 38 299 51/143 yellow granules. A continuous spectrum of changes was seen between these extremes but the former predominated. Of the 51 sections with yellow hyaline membranes, 28 were judged 2 + ; 23, 1+. Ten specimens with 2+ involvement were selected at random and sections stained for iron and bile (Prussian blue and Fourchet methods, respectively). In all ten specimens the yellow hyaline membranes stained positively for bile and negatively for iron. The membranes were always more prominent and impres- sive on the hematoxylin and eosin sections than on those sections stained to demonstrate bile. iron was usually found near lymphatic vessels but not in proximity to the yellow membranes, and it was present in much smaller amounts. Yellow hyaline membranes were seen only in patients who lived from 45 to 547 hours; the average survival of infants with 1+ yellow material was 207 hours; it was 163 hours lbr those with 2+ involvement. Statistically, this difference is not signifi- cant. Hyperbilirubinemia, defined according to a weight-related scale (Stanford Hospital Printing No. 15-29), was seen in 26 of the 51 infants with yellow membranes and in only 13 of the remaining 79 for whom data were available. Interestingly, the incidence of yellow hyaline membranes increased in later years as shown in Table I. No significant correlation (by Pearson correlation coefficient) was found between the amount of yellow membranes (as graded 0, 1+, or 2+) and the following features: amount and percentage of oxygen delivered, survival time, radiographic severity of respiratory distress at presentation, birth weight, duration of ventilation, severity of hyaline membranes histologi- cally, and maximal ventilation pressures. A significant negative correlation (P = --0.38 for n of 51) was found with the estimated gestational age. DISCUSSION Our observations support those of others that yellow hyaline membranes are probably related to serum hyperbilirubinemia and not to meconium aspiration or to intrapulmonary hemor- rh'age. 1-~ A significant association with hyperbilirubinemia is