We conclude that the early-night peak of ketone concentrations is related to growth hormone release, whereas the fasting levels are largely determined by insulin concentration. Inadequate insulin delivery in the presence of the high growth hormone concentrations characteristic of diabetic adolescents could lead to rapid decompensation and ketoacidosis.

We report a case of neonatal symptoms of irritability, increased tonus and convulsions after in-utero exposure to paroxetine 30 mg/day. The infant's symptoms commenced on the first day after birth and persisted for 10 days. Paroxetine levels were undetectable on day 6. Extensive investigations excluded infective and metabolic causes. Serotonin toxicity due to paroxetine seems the most likely mechanism, though an important differential diagnosis is a paroxetine discontinuation (withdrawal) syndrome. Differentiating between these two syndromes in the neonate presents a dilemma for clinicians. Irrespective of the mechanism, we recommend that all neonates exposed to antidepressants, particularly serotonin reuptake inhibitors (SSRIs), during the last trimester should be followed-up closely for adverse symptoms commencing in the first 10 days after birth. The possibility of such symptoms needs to be discussed with women who are considering starting or continuing antidepressant treatment in pregnancy. All neonatal adverse drug events should be reported to a pharmacovigilance centre. Further research is warranted.

In patients with insulin-dependent diabetes mellitus there is an increase in both the amplitude and frequency of pulsatile GH secretion compared to normal subjects, which is not affected by maintenance of overnight normoglycaemia. The anticholinergic drug pirenzepine appears to suppress the amplitude of GH pulse secretion but has no effect on frequency.

Plasma growth hormone profiles in adolescents with Type 1 (insulin-dependent) diabetes mellitus are characterized by both increases in pulse amplitude and higher baseline concentrations. To determine which of these abnormalities adversely affect metabolic control, we studied six young adults overnight on three occasions. On each night somatostatin (50-100 micrograms.m2-1.h-1) and glucagon (1 ng.kg-1.min-1) were infused continuously and 18 mU/kg of growth hormone was given as either: three discrete pulses of 6 mU.kg-1.h-1 at 180-min intervals or a 12-h infusion (1.5 mU.kg-1.h-1) or buffer solution only on a control night. Euglycaemia was maintained by an insulin-varying clamp. Blood samples were taken every 15 min for glucose and growth hormone and every hour for intermediate metabolites and non-esterified fatty acids. Comparable normoglycaemic conditions were achieved on all three nights. Growth hormone levels achieved (mean +/- SEM) on study nights were: 32.8 +/- 2.2 mU/l (peak level during growth hormone pulses); 9.8 +/- 0.8 mU/l (continuous growth hormone) and 1.1 +/- 0.3 mU/l (control level). Pulsatile growth hormone administration led to an increase in insulin requirements (mean +/- SEM: 0.17 +/- 0.03 vs control 0.09 +/- 0.01 mU.kg-1.min-1, p less than 0.05) whereas insulin requirements following continuous growth hormone administration were unchanged. Cross-correlation confirmed an increase in insulin requirements occurring 135 min after a growth hormone pulse (r = 0.21, p less than 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)