Portal de Periódicos da CAPES

Fatal Late Vitamin K-Deficiency Bleeding After Oral Vitamin K Prophylaxis Secondary to Unrecognized Bile Duct Paucity

1999; Lippincott Williams & Wilkins; Volume: 29; Issue: 5 Linguagem: Inglês

10.1097/00005176-199911000-00023

1536-4801

Tilman Humpl, K. Brühl, Ryta Brzezinska, Gerd Häfner, Wiltrud Coerdt, Martin J. Shearer,

Alcoholism and Thiamine Deficiency

Bleeding due to vitamin K deficiency may occur spontaneously in the first few months of life. Traditionally, this syndrome has been called hemorrhagic disease of the newborn, a term coined by the Boston physician Townsend who first identified it (1). More recently, the term vitamin K-deficiency bleeding (VKDB) has been introduced so that the cause of the bleeding is unambiguous. Three forms of VKDB have been differentiated according to the age of the infant at occurrence and the different causes (2,3). They are: early, 0-24 hours; classic, 1-7 days; and late, typically, 2-12 weeks. The introduction in 1961 of parenteral vitamin K prophylaxis in the United States (4) resulted in a dramatic decrease in the incidence of classic VKDB (2) and is also of proven efficacy in preventing late VKDB which became a major concern in the 1980s (2,5). However, in 1992, an association was found in an epidemiological study in the United Kingdom between vitamin K when administered intramuscularly, but not orally, and later childhood cancer, especially leukemia (6). Although this risk remains unconfirmed (7), the concern generated has resulted in a shift toward oral prophylaxis (3,5). Since 1992, it has been the practice in several countries, including Germany, to provide healthy newborns three oral doses of vitamin K. Originally, German practice was to administer an oral dose of 1 mg vitamin K (Konakion; Hoffmann-La Roche AG, Germany; with Cremophor EL as a solubilizing agent) on day 1, with doses repeated between days 4 and 10 and days 28 and 42. Later, the dose was increased to 2 mg (5). We report here an infant with late VKDB who had a fatal cerebral hemorrhage despite having received the standard three-dose oral regimen of vitamin K (2 mg Konakion). An autopsy revealed intrahepatic bile duct deficiency. CASE REPORT The patient was the first-born male child of a 28-year old mother and a 29-year old father. After a normal term pregnancy, he was delivered by vacuum extraction because of a prolonged second stage of labor. He was exclusively breast-fed, the neonatal period was uneventful, and his mother was not taking any medication. Three oral doses of vitamin K (2 mg Cremophor-based Konakion) had been administered to the boy on the day of birth, on the fourth day of life, and at the age of 1 month. Neither parents nor physicians noted jaundice or abnormal color of the stool and urine. Weight gain and infant development were normal. At 10 weeks he was admitted to our hospital with clinical signs of meningitis and shock. The parents reported the sudden onset of irritability on the same day with vomiting and paleness. The house pediatrician noted neck stiffness and bradycardia with episodes of apnea. After stabilization of the respiratory and circulatory systems by mechanical ventilation and volume support, the ophthalmologist ruled out papilledema but detected streaky retinal bleeding on the left side. The computed tomographic scan (Fig. 1) detected intracerebral, subdural, subarachnoid, and intraparenchymal hemorrhages, which were of different ages. These findings raised the suspicion of child abuse. The follow-up ultrasound studies confirmed the increased cerebral pressure caused by general edema (Fig. 2). Radiologic procedures (full skeletal survey) revealed no evidence of trauma.FIG. 1: Cranial computed tomography: Subdural hemorrhage on the left hemisphere. Bleeding of different durations intraparenchymal in the right frontal and left temporoparietal area. New bleeding along the falx cerebri.FIG. 2: Sonographic imaging of the head, coronal scan: Diffusely increased gyrate echogenicity with obliteration of the normal sulcal interfaces as a sign of general edema and bleeding.Hematologic and Coagulation Tests Blood samples obtained before treatment indicated hemoglobin 73 g/l, hematocrit 21.2%, thrombocyte count 669 × 109/l, activated partial thromboplastin time more than 120 seconds, antithrombin III 93%, thrombin time 15.5 seconds, fibrinogen 3.44 g/l, Quick value (prothrombin time) less than 10%, international normalized ratio (INR), 6.4. Coagulation factor assays showed: factor II, less than 10%; factor V, 75%; factor VII, less than 3%; factor VIIIc, 178%; and factor IX, less than 3%. In samples drawn after treatment protein C was 94% (normal, >60%) and protein S, 91% (normal, >55%). Reassessed 60 hours after intravenous vitamin K (1 mg) and coagulation factor supplementation (total of 200 ml fresh-frozen plasma), protein C was 72% and protein S, 75%. Liver Function Tests Liver function findings were (upper limits of normal in parentheses): γ-glutamyl transpeptidase, 103 U/l (135 U/l); aspartate aminotransferase, 42 U/l (110 U/l); and alanine aminotransferase, 40 U/l (60 U/l). Total bilirubin was 62 μmol/l (20 μmol/l), and conjugated bilirubin was 29 μmol/l (2 μmol/l). Further Tests of Vitamin K Status In a sample drawn 1 hour after 1 mg vitamin K and 60 ml fresh-frozen plasma were administered intravenously, undercarboxylated prothrombin (PIVKA-II), assayed by an in-house enzyme immunoassay (Haemophilia Centre, St. Thomas' Hospital) using a specific monoclonal antibody C4B6 (8), was 67.9 arbitrary units (AU)/ml. Plasma vitamin K1 in a posttreatment sample, measured by high-performance liquid chromatography with electrochemical detection (9) was 12.9 ng/ml (normal adult range, 0.17-0.68 ng/ml). Other Laboratory Tests Hepatitis (A, B, and C), human immunodeficiency virus infection, toxoplasmosis, rubella, cytomegalovirus infection, and herpes simplex virus infection 1 and 2 were ruled out by appropriate tests. Serum amino acids, serum triglycerides, cholesterol, high-density lipoproteins, low-density lipoproteins, serum electrolytes, and creatinine were normal. Lactate was 13.5 mmol/l (normal, <2.5 mmol/l), serum urea nitrogen was 66 μmol/l (normal, 80μmol/l), and serum α1-antitrypsin was 2.28 g/l (range, 0.9-2.4 g/l). Vitamin A (retinol) was 2.3 μmol/l (range, 1.0-2.3 μmol/l), 1-25-di-OH-vitamin D was 54 nmol/l (range, 60-110 nmol/l), 25-(OH)-vitamin D was 7.8 nmol/l (range, 35-105 nmol/l), and vitamin E was 10.7 μmol/l (range, 11.6-46.4 μmol/l). Chromosomal study showed 46,XY. Sweat sodium concentration was 42 mmol/l. Blood, urine, and tracheal bacterial cultures were negative. Management and Outcome After repeated intravenous injections of vitamin K and coagulation factor supplementation, the coagulation status normalized within 6 hours. Seventy-two hours after admission, brain death was declared. The parents did not allow a full autopsy but consented to postmortem examination of the liver. Histology confirmed cholestasis, edema, and focal degeneration of hepatocytes, as well as a paucity of bile ducts (Fig. 3).FIG. 3: Histology of the liver: Developmental defect resulting in paucity of bile ducts. No functional bile ducts in central position and only epithelial cords in ductal plate position in the periphery of periportal fields. Immunohistochemical stain for cytoceratin AE1/AE3; magnification, ×15.6.DISCUSSION The laboratory diagnosis of severe vitamin K deficiency in this infant was secure, based on the specific lowering of the four vitamin K-dependent coagulation factors (II, VII, IX, and X), the presence of undercarboxylated species of prothrombin (PIVKA-II) at levels found in adults receiving anticoagulation therapy (mean, 40 AU/ml; upper limit in normal adults, 0.15 AU/ml), and the correction of the coagulopathy by vitamin K and supplementation of coagulation factors. The relatively high plasma vitamin K reflected the fact that only a posttreatment sample was available for assay. This did not invalidate the assay for PIVKA-II because of its long plasma half-life (approximately 50 hours), thus enabling vitamin K deficiency to be diagnosed retrospectively. Despite the improved understanding of vitamin K metabolism and nutrition in newborns, late VKDB is still a significant and worldwide cause of infant morbidity and mortality (3). Without neonatal vitamin K prophylaxis, the incidence of late VKDB in Europe has been reported to be 5 to 7 per 100,000 live births (5,10). Therefore, there is a broad consensus that vitamin K prophylaxis is necessary. The debate sparked by the publication of the 1992 epidemiological study by Golding et al. (6) led many countries to change from intramuscular to oral vitamin K prophylaxis. This has led to a problem of switching from the known reliability of a single intramuscular injection at birth to oral regimens of uncertain effectiveness. In those countries with active surveillance of VKDB it has been possible to monitor the change to an oral prophylactic policy (5). In two countries, Germany and Australia, the failure rates after three oral doses of 1 mg vitamin K1 (Cremophor-based Konakion; Roche) administered at similar times (day 1, days 3-10, and 4-6 weeks) were similar (approximately 2.5 per 100,000 births) compared with the almost complete protection afforded by parenteral vitamin K prophylaxis (5). This led both countries to change their policies: Australia reverted to intramuscular vitamin K and Germany increased the oral dose from 1 to 2 mg at each of the three time points. In Australia, subsequent surveillance data since the policy reversal showed no cases of VKDB in a birth population of 325,000, compared with the previous eight cases in the same birth population when the oral regimen was in place (5). The presence of chronic liver disease significantly increases the risk of late VKDB, by decreasing the intestinal absorption or hepatic use of vitamin K. Moreover, the liver disease can easily remain undiagnosed, and VKDB may be the first manifestation of an underlying hepatic insufficiency. In our patient, the mild icterus was not detected by either health practitioners or the parents, and the diagnosis of bile duct paucity was determined only after death. Other reasons for impaired vitamin K metabolism (α1-antitrypsin deficiency, cystic fibrosis, abetalipoproteinemia, viral hepatitis, and cephalosporin exposure) (2,10-12) and congenital deficiencies of vitamin K-dependent coagulation factors were ruled out. Protein C and protein S were measured twice after supplementation of vitamin K and fresh-frozen plasma. Even after the initial peak, the levels did not drop to the pathologic range. Although doses administered orally may sometimes be unreliable because infants occasionally spit it out, the vitamin K preparation was not vomited by our patient. Our patient was given the earlier preparation of Konakion, which contains Cremophor EL as a solubilizing agent. A newer mixed-micellar preparation (Konakion MM), in which the vitamin is solubilized in the bile acid glycocholic acid and the phospholipid lecithin, seems to be better absorbed in healthy infants (13). However, although there is one report that Konakion MM is also much better absorbed than the earlier preparation in children with cholestasis, the numbers were very limited and the doses large (14). In a recent report of three infants in Switzerland with late VKDB, all had received two doses of 2 mg Konakion MM (on days 1 and 4), and two were subsequently found to have biliary atresia (15). This calls into question the effectiveness of absorption of Konakion MM in infants with cholestasis, particularly when only two doses are administered in the first week of life. The prophylaxis for "well infants" in the Netherlands was changed to 1 mg oral vitamin K on the day of birth and a daily dose of 25 μg vitamin K for breast-fed infants up to week 13 of life (5). This low, multidose daily regimen, in which the vitamin preparation is marketed as a food supplement and administered by the parents, has been very effective (5). This dramatic case is important for another reason. The ophthalmologic examination revealed retinal bleeding which in most cases has been associated with child abuse (16). In our child, retinal bleeding was most likely a result of the coagulopathy, which is seen in patients with congenital bleeding disorders (17). It is therefore important to exclude a pathophysiologic cause, such as VKDB, when retinal bleeding is detected. In conclusion, further investigations on the safety and efficacy of oral vitamin K prophylaxis as well as the bioavailability and metabolism of different preparations is needed to prevent the late form of vitamin K deficiency bleeding.