No abstract

complication it might be difficult to conclude which was the most significant risk factor for death. None the less, overall case fatality rate was highest among children with hyponatraemia.The multiple linear regression model showed that deaths associated with hyponatraemia, hyperkalaemia, and hypoglycaemia were significant at the 010 confidence level, but the effect of confounding factors could not be measured.In the logit regression model, which measures the effect of confounding factors, all the variables except serum sodium concentration and coma lost their effects. This observation that death was inversely related to serum sodium concentration supports our earlier findings that hyponatraemia is the most significant risk factor among children with complicated diarrhoea in Bangladesh.4 We reported earlier that the incidence of hyponatraemia was directly related to the degree of malnutrition, but the results of the logit regression analysis did not show that malnutrition was a significant predictor of death. The reason for the direct relation of the incidence of hyponatraemia to the degree of malnutrition might be that serum albumin concentration carries a negative charge and is largely responsible for the normal anion gap by holding serum sodium in the intravascular space.'0 11 Although nutritional state was not found to be a predictor of death, the serum albumin concentration in these children was not run as a variable in the logit regression model owing to lack of data. None the less, hyponatraemia remained a significant predictor of death, though we do not have any satisfactory explanation for this. The direct relation between the incidence of hyponatraemia and the degree of malnutrition necessitates further, prospective studies to assess the effect of serum albumin concentration on the physiopathology of hyponatraemia.The finding of coma as a predictor of death was not expected. As postmortem examinations were not performed we are unable to discuss the cause of the coma and its relation to death. Progesterone and the premenstrual syndrome: a double blind crossover trial L DENNERSTEIN, C SPENCER-GARDNER, G GOTTS, J B BROWN, M A SMITH, G D BURROWS Abstract A double blind, randomised, crossover trial of oral micronised progesterone (two months) and placebo (two months) was conducted to determine whether progesterone alleviated premenstrual complaints. Twenty three women were interviewed premenstrually before treatment and in each month of treatment.

No abstract

Past research into the phenomenon of reduced latency to repeated events in serial reaction time tasks resulted in attempts to partition reaction time into components attributable to separate cognitive processes. These attempts focused on many-to-one condensing paradigms, which permitted direct assessment of the effect of response-only repetitions but not stimulus-only repetitions. The three experiments reported in the present study used a technique designed to overcome this limitation. It was found that repetition of only the symbolic content of the total stimulus-response event was a sufficient condition for a reduced response latency and that response-only repetitions tended to be facilitative only when stimulus-response relations were compatible. These and related findings were discussed with respect to their bearing on contemporary issues concerning the analysis of cognitive operations into separable and additive components.

The two issues of premenstrual syndrome (PMS) symptom heterogeneity and symptom severity are addressed in an examination of the idiosyncratic nature of premenstrual complaints. Analyses of PMS symptoms reported during the first assessment month by a group of treatment-seeking women revealed that when presented with a prepared list of symptoms (Menstrual Distress Questionnaire), they select many more symptoms than when self-reporting their complaints in ranked order of severity. In this study of a sample of 98 women, most of the women reported only four rank-ordered symptoms. Symptom profiles from five symptom categories were constructed for each of the women and from a possible 86 different profiles 84 were discovered. Of these, psychological profiles dominated the total reports. These findings, together with results of previous research, lead to two conclusions. First, that PMS, though multidimensional, may be defined for each woman by a limited number of symptoms; and second, that PMS is highly idiosyncratic in nature.