Parenteral Nutrition-Associated Anaphylaxis in a 4-Year-Old Child

Abstract

Hypersensitivity reactions occur infrequently in patients receiving parenteral nutrition (PN), but they have been reported in children(1-5). The length of exposure to PN before the reaction, the severity of the response, and the component responsible are variable. We report a case of child who had an anaphylactic reaction when PN was resumed after a 5-day interruption in therapy. CASE REPORT A 4-year-old boy came to Le Bonheur Children's Medical Center in Memphis, Tennessee with a 24-hour history of worsening abdominal pain and bilious emesis. His medical history was significant for the diagnosis of Wilms' tumor at 10 months of age, for which he underwent chemotherapy and radiation followed by total resection of the right kidney and partial resection of the left. He had had no abdominal symptoms before this time, and his weight and height were in the 75th and 60th percentiles for his age, respectively. The diagnosis was presumed small bowel obstruction. There was no history of allergy in the patient or in his family. Oral feeding was prohibited, and the patient had a nasogastric tube placed for gastric decompression and a peripheral intravenous catheter placed for intravenous fluids. Laboratory values were within normal limits, white blood cell (WBC) count was 10,400/mm3 with 0% eosinophils. The symptoms continued, and on the fifth hospital day, the patient underwent exploratory laparotomy with lysis of multiple areas of dense adhesions throughout the bowel. A central venous catheter was placed the next day, and age-appropriate PN was started. On day 8 the nasogastric tube was removed, and a clear liquid diet was ordered. On day 11, the patient developed a fever, and vancomycin was prescribed for presumed central-line infection. The following day, the patient complained of pruritic palms and soles, and diphenhydramine was prescribed for symptomatic relief. On day 15 the patient's mother noted a faint, urticarial rash on the lower abdomen and neck that did not appear to be related to vancomycin infusion. On day 17, vancomycin was discontinued and the central line was changed. At this time the patient's oral intake was approximately half of the estimated requirements, and PN was cycled off for 8 hours. The rash did not change during the cycle-off or -on period. Because the etiology of the rash had not been determined and a contact allergy had not been ruled out, hypoallergenic sheets were used, and the patient's clothes were washed in a fragrance-free soap. Although no temporal relationship between PN infusion and the appearance of the rash or requests for diphenhydramine existed, the lipid emulsion and multivitamins were withheld from the PN formula for one infusion period. Symptoms did not change, and lipid emulsion and multivitamins were reintroduced the following evening. On day 21, central access was lost. For the next few days, oral intake did not improve, and the rash persisted. After five days without PN, a Broviac catheter was placed, and PN was resumed with the same solution formulation and cycling regimen as that used previously. Within 40 minutes of beginning the infusion, the patient had 3+ edema of the lips, tachycardia (140 beats/minute), and tachypnea (44 respirations/minute) with mild stridor. Parenteral nutrition was stopped, and intravenous diphenhydramine and subcutaneous epinephrine were given. The tachycardia and stridor resolved quickly; the edema resolved more gradually. By the next morning the patient had recovered completely. The dextrose-amino acid solution and lipid emulsion were saved and refrigerated. The only medications that the child had received during the preceding week were diphenhydramine and central-line flushes containing heparin. Five days after anaphylaxis and 4 days after the last dose of diphenhydramine, dermal allergy testing and metabisulfite oral challenge were performed. The dermal testing strategy consisted of an initial skin-prick test that, if results were negative, was followed by an intradermal injection of the same substance. The original dextroseamino acid solution; a freshly compounded solution; the original lipid emulsion (Liposyn III 20%, Abbott Laboratories; Chicago, IL); and lipid emulsion from a newly opened bottle produced negative skin-prick results with a positive histamine control. Intradermal injections of both 1:10 and 1:100 dilutions of these solutions were positive only for the freshly compounded solution(Table 1). Metabisulfite oral challenge with doses from 5 to 100 mg given in lemonade every 30 minutes produced negative results. With appropriate precautions, lipids were reintroduced that night without incident. Two days later, skin-prick testing with reconstituted undiluted multivitamin(MVI-pediatric, Armour Pharmaceutical; Kankakee, IL, U.S.A.) resulted in a slight response and undiluted Aminosyn 10% wt/vol (Abbott Laboratories) produced no response. However, intradermal injection of Aminosyn 0.1% (1:100 dilution of 10% solution) resulted in a moderate response(Table 1). The same skin testing strategy with MVI-pediatric and Aminosyn in a control subject was negative. Wheal-and-flare response to skin-prick testing for latex allergy was negative, and latex-specific immunoglobulin E was 0.13 IU/ml (no sensitivity = class 0,<0.35 IU/ml, Henry Ford Healthsystem Allergy Research Laboratory, Detroit, MI, U.S.A.). By the time the dermal testing was completed, the patient was consuming≈70% of nutritional needs orally and was maintaining his weight. Allergy test results were inconclusive, and the risks of rechallenge with vitamin-free PN was thought to outweigh the benefits. On the day of discharge, total immunoglobulin E was 38 IU/ml (range, 0-70 IU/ml) and circulating eosinophil count was 3,093/mm3 (range, 0-450/mm3). The patient was discharged with anaphylaxis kit and the central line in place. Two months after anaphylaxis, WBCs were 10,800/mm3 with 14% eosinophils (estimated circulating eosinophil count of 1,512/mm3) and by 6 months, WBCs were 7,200/mm3 with 10% eosinophils (estimated circulating eosinophil count of 720/mm3). One year later, the circulating eosinophil count was 270/mm3. DISCUSSION Until now, PN-associated hypersensitivity reaction had been described in five children (1-5), and only one had anaphylaxis (5). Two children were sensitive to fat emulsion (1,2). The first was a child with a history of allergies who developed wheezing, erythema, wheals, and urticaria after 11 days of receiving Intralipid 10% (Kabi Vitrum, Clayton, NC, U.S.A.)(1). The second was a 9-year-old boy with a pancreatic pseudocyst who developed generalized pruritic urticaria on the twentieth day of therapy with intralipid 20% (2). In both of these cases, symptoms resolved promptly when lipid emulsion was discontinued and rechallenge was positive. Two children receiving treated for cancer, a 34-month-old girl with neuroblastoma (3) and a 4-year-old boy with Burkitt's lymphoma (4), had an immediate urticarial rash when PN with multivitamins was infused. Symptoms resolved quickly when the infusion was stopped, and both received vitamin-free PN without difficulty (3,4). Before this report, the only report of PN-associated anaphylaxis in a child occurred in a 4-year-old boy in whom PN was initiated after colostomy closure (5). Within 15 minutes of his first PN infusion, he developed tachycardia, hypotension, stridor, and dyspnea that resolved quickly with discontinuation of PN and drug treatment. Dermal allergy tests 10 weeks after the episode demonstrated sensitivity to intradermal injections of undiluted amino acid solution (Travasol, Travenol Laboratories; Deerfield, IL, U.S.A.), vial 2(biotin, folic acid, vitamin B12, propylene glycol, citric acid, and sodium citrate) of the multivitamin preparation (MVI-12, Armour Pharmaceutical), and the magnesium sulfate solution. Sulfites are added to crystalline amino acid products to decrease the oxidation of amino acids. Both acute and delayed sensitivity reactions have been reported with oral ingestion of sulfite-containing foods and other products (6,7). Parenteral administration of local anesthetics, gentamicin, metoclopramide, and a parenteral vitamin B preparation resulted in adverse reactions that were related to sulfite hypersensitivity (6). Oral metabisulfite testing resulted in a positive response; most were to low doses (6). MVI-pediatric contains the preservatives butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) and polysorbate emulsifiers. Oral ingestion of BHA and BHT has been associated with allergy symptoms in sensitive individuals(8,9), but allergy to parenterally delivered BHA or BHT has not been reported. Polysorbates have been implicated as the most likely cause of morbidity and mortality in infants given the vitamin E parenteral product, E-Ferol, intravenously (10). These neonates and infants experienced a toxic reaction to very high doses of the emulsifiers. Unlike the experience with other emulsifiers, hypersensitivity to polysorbates has not been reported. Vitamin K1 is contained in MVI-pediatric, and anaphylactic reactions have been reported with parenteral vitamin K1 in adults(11,12). In one sensitive patient, the individual components in the product failed to evoke a response; however, the intact product resulted in a positive response, suggesting that the allergic potential of individual components had been altered by the combination(11). In an animal study, polyoxyethylated castor oil, the emulsifier in the parenteral vitamin K1 product, caused increases in renal vascular resistance and decreases in renal blood flow, whereas polysorbate 80, an emulsifier in MVI-pediatric, produced no changes in vascular resistance or in renal blood flow in a rat model(13). Perhaps some of the adults who had a hypersensitivity reaction to parenteral vitamin K were responding to Cremophor EL and not to vitamin K. The case we present is different from the previous reports of PN-related hypersensitivity in children and was confounded by an urticarial rash that may have been unrelated to PN. This patient had been receiving PN for several days before the onset of urticaria, which did not change when the patient was cycled off PN for 8 to 12 hours, nor did it change when when the vitamins and lipids, the components most often associated with PN hypersensitivity, were withheld. Frequently, the causative agent of urticaria is not identified and, in this case, may have been unrelated to the anaphylaxis. Interestingly, anaphylaxis occurred in this child when PN was reinstituted after a 5-day interruption in PN infusion. Allergy testing ruled out fat emulsion as the cause for the hypersensitivity response, metabisulfite challenge ruled out sulfite sensitivity. Skin-prick response to the multivitamin was slight and negative to the amino acid solution. Thus, the allergic potential of these two PN components individually may not have been significant enough to cause anaphylaxis; but taken together, the allergic potential may have been amplified. The patient in this case received PN without incident for more than 2 weeks, including at least 5 days of intermittent exposure in a cycling regimen, followed by a 5-day interruption in therapy. Based on the presentation, we believe this to be an immunoglobulin E-mediated type I allergic response; however, the causative agent was not identified with certainty. It is not likely that individual amino acids stimulated an allergic response, although we cannot rule out the possibility of aggregated amino acids as a potential sensitizing agent. It is possible that a component of MVI-pediatric was the cause by itself or in combination with aggregated amino acids. Least likely but possible, the response may have resulted from an unknown contaminant of the solution. It is intriguing that this case and two others (3,4), 50% of those reported, occurred in children with a history of malignancy. It is interesting to speculate that cancer or anti-cancer therapies that modulate immune function may place children at greater risk for hypersensitivity reactions.