[Clinical, biological and developmental aspects of alcoholic ketoacidosis]. / Aspects cliniques, biologiques et évolutifs de l'acidocétose alcoolique.

OBJECTIVES: To review cases of alcoholic ketoacidosis in order to better ascertain therapeutic management. METHODS: The medical files of 32 alcoholic patients with ketoacidosis hospitalized in the Saint-Pierre general hospital of the Reunion island from January 1, 1991 through 31 August 1994 were analyzed. RESULTS: There were 18 women and 14 men, mean age 47 years. The first clinical signs were predominated by digestive (n = 22) or neurological disorders (n = 10). Acidosis was severe (mean pH = 7.12) and always associated with a wide anion gap (mean anion gap = 35). There were 3 types of glycemic status: hypoglycemia 10 cases, normal or subnormal glycemia in 19 cases (mean glycemia = 9.3 mmol/l) and hyperglycemia above 20 mmol/l in 3 cases. Hypophosphatemia, elevated serum lactate levels and cytolytic hepatitis were the main abnormalities associated. CONCLUSION: Short-term outcome was favorable in all cases after rehydration. The use of insulin may be dangerous and needs to be avoided.

Immunoradiometric analysis of circulating human glycosylated and nonglycosylated prolactin forms: spontaneous and stimulated secretions.

The monoclonal antibodies (MAbs) obtained in mice immunized with human PRL coupled to an anti-PRL MAb were screened for their ability to distinguish the glycosylated (G-) and nonglycosylated (NG-) forms of PRL. The 431-29 MAb exhibited high affinity binding for NG-PRL but little or no cross-reactivity to G-PRL. Using this antibody in conjunction with other MAbs which equally recognized both forms, we developed 2 immunoradiometric assays which were used to determine the amount of G- and NG-PRL in plasma. In 85 normal subjects, NG-PRL baseline levels averaged 6.6 +/- 3 micrograms/L, and represented 76 +/- 8% of the total PRL immunoreactivity. In 74 pregnant women, this proportion was significantly higher during the last 2 trimesters (84 +/- 4% and 85 +/- 6%), as compared to the first trimester (76 +/- 7%). In 6 healthy volunteers studied over 24 h, 79% of the NG-PRL peaks detected using the cluster algorithm occurred concomitantly to a G-PRL peak. The mean NG-PRL/PRL ratio was significantly higher during NG-PRL pulses (81 +/- 9%) than during valleys (71 +/- 12%). Similarly, this ratio was significantly increased during TRH or metoclopramide stimulated PRL secretion (to 88 +/- 7% and 86 +/- 6%, respectively). We conclude that 1) NG-PRL is the predominant immunoassayable form of PRL in plasma; 2) both G- and NG-PRL are cosecreted but NG-PRL is the main PRL form released during spontaneous or pharmacologically induced PRL secretion.