Volume 103 Editorial correspondence 5 0 1 Number 3 muramidase was 11.3 #g/ml (normal 7 to 14 /zg). Response to exchange transfusion and prednisone and vineristine was dramat- ic. The WBC decreased to 3400/ram 3 within four days of the initiation of treatment, all peripheral blasts were gone by 14 days, and a bone marrow study 2 89 weeks after presentation revealed no evidence of leukemia. However, four weeks after diagnosis, bone marrow aspirate was hypercellular, with 74% myeloblasts and monocytoid blasts. Cytochemical analysis of the marrow showed the blasts to be positive for sudan black, peroxidase, and a- napthyl-esterase, with moderate PAS positivity. The serum muramidase had increased to 30 ~g/mL Although subsequent treatment with doxorubicin, cytosine arabinoside, and 6-thiogua- nine decreased the percentage of blasts, complete remission was never achieved and the child died 4 89 months after diagnosis. Chromosome study of peripheral blood at a time when the WBC was 140,000/mm 3 with 89% blasts revealed the predominant karyotype to be 47,XY,11 q+,-4,+mar,+mar. The problems of morphologic classification of neonatal leuke- mia have been pointed out in the past, 2 although routine morphol- ogy and cytochemical analysis were concordant in our patient. Our interpretation of the patient's course was similar to that proposed for the patient of Perentesis eta[.; namely, the malignant cell line was multipotent with differential sensitivity to chemother- apeutic agents. The initial cytoreductive chemotherapy allowed the expansion of a resistant clone with myeloid characteristics. This case of a neonate with biphenotypic leukemia further supports the concept of a common lymphoid-myeloid stem cell. A. Kim Ritchey, M.D. Department of Pediatrics Yale University School of Medicine New Haven, CT 06510 REFERENCES I. Perentesis J, Ramsay NKC, Bruning R, Kersey, JH, Filipo- vich AH: Biphenotypic leukemia: Immunologic and morpho- Iogic evidence of a common lymphoid-myeloid progenitor in humans. J PEDIATR 102:63, 1983. 2. Wolk JA, Stuart M J, Davey FR, Nelson DA: Congenital and neonatal leukemia: Lymphocytic or myelocytic? Am J Dis Child 128:864~ 1974. Hypothalamic effects of sodium valproate To the Editor: The observations of Kritzler et al. 1 focus on the ACTH suppression occurring in children receiving sodium valproate and suggest a hypothalamic mechanism for suppression of corticotro- pin releasing factor. Some time ago we were faced with the problem of increasing gynecomastia and galactorrhea in an infant girl who was receiving sodium valproate. This child had developed generalized tonic- clonic seizures shortly after birth, which were not controlled with phenobarbital and phenytoin (Dilantin) (serum drug levels in the therapeutic range). Sodium valproate (12.5 mg/kg) was started at 6 weeks of life, and increased to 25 mg/kg, which controlled the seizure frequency. Increasing gynecomastia and galaetorrhea were noted three days after start of sodium valproate therapy. After five days of treatment, the serum prolactin concentration was 29 ng/ml, and increased further to 43 ng/ml. Because of developing hepatic toxicity, the sodium valproate therapy was discontinued. The galactorrhea stopped three days later and the gynecomastia subsequently resolved. Unfortunately, we were unable to obtain any serum prolactin measurements after sodium valproate was stopped. If we agree that sodium valproate contributed to the prolactin increase and consequent galactorrhea, then perhaps this is another bypothaiamic effect. But then the observation on ACTH and prolactin are totally opposite. Although dopaminergic interruption results in prolactin secretion, tbis should also enhance, rather than suppress, the release of ACTH. We have not read of any effects of sodium valproate on prolactin secretion, but the report of Kritzler et al. called to mind such a possibility. Sobha Kollipara, M.D. Matthew H. Connors, M.D. University of California, Davis School of Medicine Department of Pediatrics Davis, CA 95616 REFERENCE 1. Kritzler RK, Vining EPG, Plotnick LP: Sodium valproate and corticotropin suppression in the child treated for sei- zures. J PEDIATR 102:142, 1983. Low oral dose of clonidine: An effective test of growth hormone reserve To the Editor" The oral administration of elonidine, a selective alpha-adrener- gic agonist, has been proposed as a reliable test of growth hormone reserve in children?'2 We report our experience showing that an orally administered dose of clonidine lower than that (0.150 mg/m 2) administered by others 1,2 is effective for evaluating the hGH secretion in children, with no or insignificant side effects. We compared hGH responses in 12 prepubertal children, referred because of short stature (nine boys, age 7 to 14 years, linear height 3 SD below the median for age), after a single dose of clonidine orally (0.035 mg/m2/body surface area) and a standard L-arginine monohydrochloride infusion (0.5 gm/kg/body weight). The tests were performed randomly, within one week, after an overnight fast and 30 minutes of recumbent rest; samples for hGH, cortisol, and blood glucose determinations were drawn at 0, 30, 60, 90, and 120 minutes; blood pressure was monitored every half hour during the clonidine test. All patients were found to be endocrinologically normal. Informed consent was obtained prior to the study. The hGH responses are shown in the Table. In two subjects the hGH increments (normal >7 mg/mP) were found blunted after clonidine ( 1.4 and 4.7 ng/ml, respectively) but normal after arginine (27.5 and 10.6 ng/ml, respectively); on the contrary, three children showed insufficient responses after argi-