Measuring vibration sensations with graduated tuning fork. Simple and reliable means to detect diabetic patients at risk of neuropathic foot ulceration.

To find a simple and reliable means to measure vibration sensations, 189 diabetic patients and 88 control subjects were tested at different sites with a graduated tuning fork. Within-test variation at big toes reached 8.4% in diabetic patients vs. 2.2% in control subjects. Mean contralateral variation was 7.5% in diabetic patients vs. 2.5% in control subjects. Tuning-fork sensations were inversely correlated with duration of diabetes, whereas no correlation was found with HBA1c levels or the severity of retinopathy. Ninety-nine (52%) patients had vibratory sensation at big toes of less than 99th percentile of normal values for age. In addition, 51% of the patients with clinical symptoms at extremities (n = 67), 70% of the patients without tendon reflexes (n = 50), and 75% of the patients with abnormal nerve conduction velocities (n = 60) also had low vibration sensations. All patients with lower-limb injuries (n = 7) had values at big toes of less than 2. Altogether, the graduated tuning fork represents a simple and reliable alternative to quantitate vibration sensations. Long-term follow-up of asymptomatic patients will indicate whether these abnormalities reflect underlying neuropathy. Patients with abnormal values at screening will necessitate additional investigations and special foot-care education programs.

Demonstration of an antiglucocorticoid action of progesterone on the corticosterone inhibition of beta-endorphin release by rat anterior pituitary in primary culture.

The effects of progesterone or 17 alpha-hydroxyprogesterone on corticosterone regulation of beta-endorphin (beta-end) release have been studied in vitro using primary culture of rat anterior pituitaries. Incubation of pituitary cells with ovine corticotropin-releasing factor (CRF) for 2 h resulted in a dose-dependent increase in beta-end release. Maximal stimulation was obtained with 200 ng/ml CRF. Preincubation for 2 h with corticosterone resulted in a dose-dependent inhibition of CRF-induced beta-end release. When the cultures were preincubated for 2 h with 200 ng/ml corticosterone and increasing concentrations (1, 10, 100, 1,000, and 10,000 ng/ml) of progesterone, a significant decrease in the corticosterone feedback action was observed with 100 ng/ml progesterone. Complete inhibition of the action of 200 ng/ml corticosterone was achieved with 10,000 ng/ml progesterone. Moreover, when the cultures were preincubated with increasing concentrations of corticosterone in the presence of 100 ng/ml progesterone, the ED50 of corticosterone increased significantly from 212 +/- 36 to 940 +/- 42 ng/ml (mean +/- SEM; P less than 0.01). Under the same conditions, 17 alpha-hydroxyprogesterone had no effect. These data demonstrate that progesterone antagonizes the corticosterone feedback inhibition of beta-end release by rat anterior pituitary.

Serum bone Gla protein: a marker of bone turnover in hyperthyroidism.

Serum bone Gla protein (BGP) concentrations were measured in 24 hyperthyroid patients before and after treatment. Before treatment, the mean concentration was higher [11.8 +/- 3.4 ( +/- SD) ng/ml] in the patient group than in a group of 12 age-matched normal subjects (6.1 +/- 1.7 ng/ml; P less than 0.001); 16 of the 24 patients had a value above the normal range. Serum BGP concentrations in the patients correlated significantly with serum T3 (r = 0.65; P less than 0.001) and T4 concentrations (r = 0.56; P less than 0.01). Other biochemical markers of bone metabolism (serum alkaline phosphatase, serum and urinary calcium, and urinary hydroxyproline) did not correlate with circulating thyroid hormone levels. Serum BGP also was measured after the patients had become euthyroid; 23 measurements were made on 16 patients at various times after the start of treatment. All values were normal after 16 weeks; before this period, most of the values were still above the normal range despite normal plasma thyroid hormone concentrations in all patients. These results suggest that BGP is a sensitive marker of bone metabolism alterations during hyperthyroidism.

Decreased hypoglycemic effect of insulin at night in insulin-dependent diabetes mellitus and healthy subjects.

Diurnal variations in insulin-induced hypoglycemia and in plasma counterregulatory hormone concentrations were explored in eight insulin-dependent diabetic and six healthy subjects during a 100-min iv insulin infusion performed at 0300 h and 1500 h. In healthy subjects, plasma glucose concentrations (mean +/- SD) fell by 35 +/- 2% during the daytime test and by 26.5 +/- 2% during the nocturnal test (P less than 0.01). Plasma cortisol, GH, and epinephrine concentrations increased more during the daytime than during the nocturnal test. In contrast, plasma glucagon concentrations rose more during the nocturnal tests. In insulin-dependent diabetes mellitus patients, insulin infusion had to be interrupted in three subjects because plasma glucose fell below 1.9 mmol/L 80 min after the beginning of the test. In the other five patients plasma glucose fell by 34 +/- 5% during the daytime test while no significant decrease in plasma glucose was observed in any of the eight patients during the nighttime test. Counterregulatory hormone concentrations were consistent with the results of plasma glucose, with no change during the nocturnal test and significant increases in cortisol, GH, and epinephrine during the daytime test. These results show that insulin sensitivity is decreased at night in comparison to midafternoon in healthy subjects and that in insulin-dependent diabetes mellitus patients this phenomenon is exaggerated, even in patients with defective counterregulation to hypoglycemia.

Acute effect of glyburide on insulin sensitivity in type I diabetic patients.

Possible extrapancreatic effects of glyburide on insulin action were studied in six patients with insulin-dependent diabetes mellitus. Each patient was studied on two separate occasions with continuous iv infusions of either glyburide (0.3 mg/h after a 1-mg iv bolus dose) or NaCl. During the studies blood glucose concentrations were controlled by a glucose-controlled infusion system (Biostator). The study included the 12-h period after the evening meal, followed by a 4-h period during which euglycemic hyperinsulinemic clamp studies were performed at two rates of insulin infusion: 1 and 10 mU/kg.min. During the glyburide infusion, the Biostator-determined insulin delivery rate was similar to that during the NaCl infusion for the first 6 h after the meal, but it decreased by 32% between the 6th and 12th hours after the meal. During the hyperinsulinemic clamp studies, glucose was delivered at a significantly higher rate when glyburide was infused; this was true for both rates of insulin infusion [5.6 +/- 1.9 (+/- SD) vs. 3.6 +/- 1.4 mg/kg.min and 12.1 +/- 2.4 vs. 9.1 +/- 2.1 mg/kg.min; P less than 0.05, glyburide vs. NaCl, respectively]. Plasma C-peptide was undetectable in all patients during both studies. These results indicate that 1) glyburide has an acute effect on insulin action in insulin-dependent diabetic patients; and 2) this effect occurs at physiological as well as pharmacological insulin concentrations.

Sex-related difference in the growth of prolactinomas: a clinical and proliferation marker study.

Prolactinomas in women commonly present as small intrasellar tumors, but are usually much larger in men. This discrepancy has generally been attributed to differences in the delay before diagnosis. However, studies comparing clinical and pathological correlates of growth of these tumors in both sexes are lacking. We conducted a retrospective study comparing 45 men and 51 women bearing prolactinoma to determine whether the predominance of large tumors in men was due to a delay in diagnosis or, rather, to a fundamental sex-related difference in tumor growth. Basal PRL levels (mean +/- SEM, 2789 +/- 573 ng/mL) and mean tumor diameter (26 +/- 2 mm) were significantly higher in men than in women (292 +/- 74 ng/mL and 10 +/- 1 mm, respectively; P < 0.001), but were not correlated to the age at diagnosis or the duration of symptoms. Giant tumors (n = 8) occurred in males only. The frequencies of bromocriptine-resistant tumors (30 vs.5%; P < 0.01) and invasive macroadenomas (52 vs.27%; P < 0.001) were significantly greater in men than those in women. Lastly, macroprolactinomas in males exhibited higher indexes of proliferating cells by Ki-67 immunoreactivity (2.6 +/- 1.1% of positive nuclei) than did similar tumors in female patients (0.4 +/- 0.2%; P = 0.08). We conclude that the predominance of large prolactinomas in men is due to a high frequency of rapidly growing tumors, which are often invasive and frequently bromocriptine resistant.

Thyroid iodine organification defects: a case with lack of thyroglobulin iodination and a case without any peroxidase activity.

Two patients (G2, G3) with iodine organification defect were studied. The first patient (G2), a 25-year-old women with no clinical hypothyroidism, had had her goiter for 10 years; 62% of the thyroidal iodine was released by perchlorate indicating iodine organification defect. The thyroid tissue obtained at thyroidectomy contained a normal concentration of thyroid peroxidase (I2 formation from I-) when tested after solubilization of the enzyme by trypsin and digitonin treatment of the particulate material. 1. The enzymatic activity (G2-TPO) behaved on DEAE cellulose chromatography very differently from those of hog (P-TPO) or another human goiter peroxidase (G1-TPO) (Pommier, et al., J Clin Endocrinol Metab 39: 69, 1974): the molarity of elution was 2M NaCl instead of 0.15 mM. 2. Both P-TPO and G2-TPO catalyzed iodide peroxidation (I- leads to I2) but the Km (iodide) value for G2-TPO was much lower (2.3 x 10(-2) M) when compared with that of P-TPO (3.7 x 10(-3) M) or G1-TPO (3.5 x 10(-3) M). In addition, the optimum pH for this reaction differed markedly (pH 6.1 instead of 7.9). 3. G2-TPO was poorly efficient in catalyzing the oxidation of gaiacol to tetragaiacol. 4. G2-TPO was unable to perform the iodination of non-iodinated goiter thyroglobulin whatever the pH and the iodide concentration. 5. Thyroglobulin from this goiter (G2) was almost not iodinated (0.0014%), i.e., 0.07 atoms iodine/mole thyroglobulin), and its total content in the gland was very low (0.3-4 g/1000 g wet tissue instead of 25 g). A clear discrepancy was thus shown between the euthyroid state of this patient and the total lack of iodinating activity of the isolated peroxidase. The second patient (G3), a 17-year-old man with clinical hypothyroidism, had had his goiter for 5 years. 100% of the thyroidal iodine was released by perchlorate indicating a complete iodine organification defect. The thyroid tissue obtained at thyroidectomy contained no peroxidase activity when tested before and after treatment of the particulate material by trypsin and digitonin and even in the presence of hematin. Thyroglobulin from this goiter, which was almost non-iodinated (0.0014%), was present in normal amounts in the gland (congruent to 25 g/1000 g).

Somatotropic adenoma manifested by galactorrhea without acromegaly.

In two of eight premenopausal women with somatotropic adenomas, galactorrhea was the earliest clinical feature, associated in one patient with amenorrhea. These two patients did not have clinically evident acromegaly. Mean basal serum GH levels were elevated and did not decrease after glucose ingestion. Both patients had modest hyperprolactinemia. Histological and immunocytological studies of the adenomas showed numerous adenomatous somatotropic cells and some alpha-subunit- and PRL-containing cells. In these patients, the origin of the hyperprolactinemia was not clear. In one patient, elevated GH secretion was probably responsible for the galactorrhea, since it disappeared after surgical treatment despite persistence of hyperprolactinemia. In conclusion, galactorrhea, isolated or associated with amenorrhea, can be the only clinical manifestation of a somatotropic adenoma.

Isolated aldosterone synthase deficiency caused by simultaneous E198D and V386A mutations in the CYP11B2 gene.

Isolated deficiencies in aldosterone biosynthesis are caused by mutations in the CYP11B2 (aldosterone synthase) gene. Patients with this deficiency have impaired aldosterone synthesis, exhibit increased plasma renin activity, secrete increased amounts of the steroid precursors DOC, corticosterone, and 18OHDOC, and are subject to salt wasting and poor growth. Two forms are generally distinguished. The first, corticosterone methyloxidase type I (CMO I or type 1 deficiency), is characterized by no detectable aldosterone secretion, a low or normal secretion of the steroid 18OHB, and are always found to have mutations that completely inactivate the encoded CYP11B2 enzyme. The second form (CMO II or type 2 deficiency) may have low to normal levels of aldosterone, but at the expense of greatly increased secretion of its immediate precursor 18OHB. These patients usually have a CYP11B2 enzyme with some residual enzymatic activity, especially 11beta-hydroxylase activity. We have studied two twins with an isolated aldosterone synthase activity who have a clinical profile typical of the type 1 deficiency. Their CYP11B2 genes are homozygous for three sequence changes, R173K, E198D, and V386A. In transfection assays these substitutions individually have modest effects on the encoded enzyme, but when found together they result in an enzyme with a decreased 11beta-hydroxylase activity, a large decrease of 18-hydroxylase activity, and no detectable 18-oxidase activity. This residual activity is more typical of that observed in patients classified as having CMO II deficiency, rather than CMO I deficiency, where no activity is detectable. This disparity between the CYP11B2 enzyme with residual activity and a clinical phenotypic typical of the type 1 deficiency, suggests that phenotype genotype relationships are not yet fully understood.

Beta-lipotropin and cortisol responses to an intravenous infusion dexamethasone suppression test in Cushing's syndrome and obesity.

An i.v. infusion dexamethasone (Dex) test was used to investigate the ACTH feedback response in 9 normal subjects, 12 obese patients, and 11 patients with Cushing's syndrome. Dex phosphate was infused iv for 4 h, starting at 1100 h (1 mg/h). Plasma concentrations of beta-lipotropin (beta LPH) and cortisol were measured every 20 min between 0900 and 1600 h, then every 2 h until midnight and at 0900 h the next day. In normal subjects and obese patients, plasma beta LPH and cortisol concentrations fell rapidly to less than 40 ng/liter and 3 micrograms/dl, respectively, at the end of Dex infusion. Subsequent values remained low through 0900 h the next day. In 7 patients with Cushing's disease, basal plasma beta LPH and cortisol concentrations declined by greater than 50% during the Dex infusion. In these patients, rapid escape from suppression occurred between 1600 and 2400 h; by 0900 h the following day, beta LPH and cortisol levels were higher than 100 ng/liter and 10 micrograms/dl, respectively. In 3 patients with adrenal tumors, beta LPH concentrations were low, and cortisol concentrations did not decline during the Dex infusion. In 1 patient with ectopic ACTH secretion, beta LPH concentrations were high and were not suppressed by the Dex infusion. We conclude that the iv infusion Dex suppression test can distinguish patients with Cushing's syndrome from normal or obese subjects and can aid in the etiological diagnosis of Cushing's syndrome.

Mono- and plurihormonal thyrotropic pituitary adenomas: pathological, hormonal and clinical studies in 12 patients.

In a series of 12 patients (eight women and four men, aged between 20 and 62 years), operated on for a pituitary adenoma shown to be thyrotropic by immunocytochemistry, we performed a retrospective and comparative analysis of clinical and biological data, tumor studies including immunocytochemistry with double labeling, and proliferation marker (proliferative cell nuclear antigen (PCNA) and Ki-67) detection, electron microscopy and culture. Our study leads us to confirm that thyrotropic tumors are rare (12 of 1174 pituitary adenomas: 1%). The main points arising were that: (1) high or normal plasma TSH associated with an increase in plasma alpha-subunit and high thyroid hormone levels is the best criterion for diagnosis; (2) the failure of TSH to respond to TRH or Werner's test is not a reliable criterion for diagnosis; (3) thyrotropic adenomas may be 'silent', without clinical signs of hyperthyroidism and with only slight increase in TSH, tri-iodothyronine and thyroxine concentrations; (4) mitoses and nuclear atypies are frequently detected in large tumors, which are invasive in more than 50% of cases - the first analysis of two proliferation markers (PCNA and Ki-67) bears out the relative aggressiveness of thyrotropic adenomas; (5) thyrotropic adenomas are frequently plurihormonal. Immunocytochemical double labeling, complemented by in vitro study, showed that thyrotropic tumor cells sometimes can or sometimes cannot cosecrete TSH, GH or prolactin. The pathological identification of monohormonal and plurihormonal adenomas seems to be supported by clinical and biological differences.

Twenty-four hour melatonin secretory pattern in men with idiopathic hemochromatosis.

In idiopathic hemochromatosis, iron deposits in endocrine tissue can be associated with hormonal disorders including hypogonadism. We have studied the functional status of the pineal gland in this disease in relation to gonadotrophin levels and cortisol rhythm. Plasma melatonin, luteinizing hormone (LH) and cortisol concentrations were measured by radioimmunoassay every 20 min over a 24 hr period in nine men with idiopathic hemochromatosis aged 36 to 66 years. In six patients a circadian melatonin rhythm was present. The 24 hr means were in the normal range in three patients, and varied below the control values in two patients and above the control values in one patient. These variations seemed unrelated to gonadotrophin status. In the three other patients no plasma melatonin rhythm was observed; two patients with gonadotrophin insufficiency had low melatonin levels, and one with normal gonadotrophin function had high melatonin concentrations. In all cases, the plasma cortisol rhythm was normal. We concluded that the circadian melatonin rhythmicity can be disturbed in some cases of idiopathic hemochromatosis without relationship to the cortisol rhythm and associated endocrine disorders.

Pathophysiology of sex hormone binding globulin (SHBG): relation to insulin.

In humans, the plasma level of sex hormone binding globulin (SHBG) is regulated by several hormones. We have now accumulated evidence that SHBG is also intimately related to nutritional state. However, we do not yet know what specific signal, if any, may be the regulator of SHBG. There is a strong and negative correlation between fasting insulin level and SHBG in obese as in hyperandrogenic women. Under such circumstances, a high fasting insulin level, normal glycemia and a low SHBG level suggest insulin resistance in terms of glucose disposal but not in terms of SHBG inhibition. This is a rather complex situation. It is too early to judge the importance of IGF-I in the regulation of SHBG. But it may turn out that IGF-I is the main regulator of SHBG and that, by interaction with the IGF-I receptors, insulin carries on its inhibitory activity on SHBG.

Sex steroid-binding protein in nonendocrine diseases.

In humans, sex steroid-binding protein (SBP) is a protein from the liver which binds with high affinity sex steroid hormones. The plasma concentration of SBP is regulated in part by hormonal factors. It has been shown that estrogens and/or thyroid hormones increase the production of SBP by hepatoma cell lines. It is therefore assumed that the increase in SBP levels in patients given oral estrogens or thyroid hormones is the consequence of a direct stimulation of the liver production of SBP by these hormones. The effects of androgen, progestagen and glucocorticoid hormones are unclear or still a matter of controversy. Moreover, the regulation of the metabolic clearance rate of SBP and the influence of nonhormonal factors on the production of SBP are still speculative. Changes in SBP have been described in a few nonendocrine diseases. A slight hormonal dysfunction may be either the primary or the sole cause of the changes in SBP occurring in these diseases. As an example, elevated SBP levels have been reported in men with liver cirrhosis together with testicular hypofunction and increased estrogen levels. It is therefore difficult to demonstrate that the increase in SBP is due to the liver dysfunction rather than to the endocrinological side effects of cirrhosis. The aim of this review is to present some aspects of the nonhormonal regulation of SBP. There is accumulating evidence in the literature for a relation between SBP levels and body weight and fat distribution, energy balance, diet and physical activity, and lipid metabolism. Therefore, it is tempting to propose that SBP is an index which reflects the status of endocrine, metabolic and nutritional functions. Measurement of SBP may be considered of interest in the light of previous epidemiological studies and the preventive approach to diseases such as hormone dependent tumors, cardiovascular diseases and osteoporosis.

Evidence of excessive androgen secretion by both the ovary and the adrenal in patients with idiopathic hirsutism.

Fifteen patients with idiopathic hirsutism, who had no attenuated adrenal hyperplasia, obesity, enlarged ovaries, or amenorrhea, were studied. Excessive androgen secretion by adrenal tissue was suggested by the finding of increased levels of dehydroepiandrosterone sulfate, which decreased after dexamethasone administration but did not change after human chorionic gonadotropin (hCG) injection. Excessive androgen secretion by ovarian tissue was suggested by the finding that testosterone and androstenedione levels were elevated, correlated significantly with the levels of luteinizing hormone, decreased with administration of estrogen-progestagen, and increased after hCG injection. Notably, free testosterone levels, which were significantly increased, were only partially suppressed during dexamethasone or estrogen-progestagen administration. These results provide further evidence that both the adrenals and the ovaries secrete androgens excessively in patients with idiopathic hirsutism.

TRH and LH-RH distribution in discrete nuclei of the human hypothalamus: evidence for a left prominence of TRH.

Thyrotropin-releasing hormone (TRH) and luteinizing hormone-releasing hormone (LH-RH) have been measured by radioimmunoassay in individual human hypothalamic nuclei. A significant lateralization has been found for TRH in the ventromedial, dorsal and paraventricular nuclei, with higher concentration in the left side. In contrast LH-RH values did not differ between the left and the right side. This finding represents an additional example of cerebral specialization and the first report of lateralized peptide distribution in the human hypothalamus.

Treatment of hypogonadotropic hypogonadal male patients with the luteinizing hormone-relasing hormone (LH-RH) anolog D-Ser(TBU)6 EA10 LH-RH: transient disappearance of gonadotropin stimulation.

We have studied the effect of prolonged treatment with a long-acting luteinizing hormone-releasing hormone (LH-RH) analog (D-Ser-(TBU)6 EA10 LH-RH in six patients with isolated gonadotropin deficiency. Before treatment, all subjects responded to LH-RH (100 microgram intravenously [IV]); one responded immediately, and five after 5 daily infusions of LH-RH (200 microgram). Treatment by LH-RH analog (348 microgram every 2 days with a nasal spray for 90 or 120 days) is only efficient for 1 month; a consistent increase in serum LH and a slight increase in testosterone (T) were observed in all patients, but no increase of serum follicle-stimulating hormone (FSH) was detectable. Then a paradoxical effect appeared: LH and T levels returned to the basal values. Moreover, this treatment induced refractoriness of the pituitary to LH-RH for several months after the end of treatment. The appearance of antibodies to LH-RH and LH-RH analog was eliminated. A pituitary response was obtained in three patients when a new LH-RH stimulation was repeated 7 and 11 months after the end of treatment. The mechanism of this pituitary desensitization is discussed.

Magnetic resonance imaging with superparamagnetic iron oxide particles to evaluate hepatic macrophage-monocytic phagocytosis after arterial devascularization in minipigs.

RATIONALE AND OBJECTIVES: We examined the effects of arterial ischemia on the phagocytic activity of the hepatic macrophage-monocytic phagocytic system (MMPS). METHODS: Six minipigs were studied before and 24 hr after complete arterial devascularization of the liver. Magnetic resonance (MR) imaging was performed at 1.5 T using superparamagnetic iron oxide (SPIO) particles (18 mumol Fe/kg body weight) as an MMPS-specific contrast agent. Hepatobiliary scintigraphy, measurements of serum liver enzymes, and histology also were obtained. RESULTS: On MR imaging, the postcontrast-to-precontrast ratios of the arterially devascularized livers were significantly higher than the corresponding baseline values (p < .01). The greatest difference (52%) between the baseline and the postoperative values was observed on gradient-echo (GE) images. Scintigraphy, laboratory analyses, and histology results indicate that the MR imaging findings were probably predominantly attributable to a reduction in phagocytic activity of the hepatic MMPS. CONCLUSION: SPIO particles have already proved useful for improving detection of liver neoplasms on MR imaging, but they also may provide a novel way of evaluating the function of the hepatic MMPS in liver diseases.

Somatotropic adenomas without acromegaly.

Seventeen somatotropic adenomas removed from patients without acromegaly were studied. Thirteen of them presented as a prolactinoma with amenorrhea and/or galactorrhea and elevated serum PRL levels. According to basal serum GH levels, the patients were divided into two groups, namely Group I: GH slightly elevated (n = 4) and group II: GH less than or equal to 5 micrograms/l (n = 13). The tumoral GH secretion was proved by immunocytochemistry in all cases and by intratumoral RIA, in vitro study and/or in situ hybridization in five of them. Pathological, clinical and biochemical relationships suggested two anatomoclinical aspects. In group I, the tumors were small, well-differentiated somatotropic adenomas with clinically silent GH hypersecretion. It is probably an early stage of the disease. In group II, the tumors were large with normal GH serum levels. They were poorly differentiated and secreted very low amounts of GH. In nine of them, PRL and/or PRL mRNA expression were also detected. These tumors do not secrete enough GH to increase serum levels and cause acromegaly. The somatotropic adenomas without acromegaly correspond to two anatomoclinical aspects of the disease.

[Auricular fibrillation in hyperthyroidism. Incidence of associated cardiopathy and of dilatation of the left auricle]. / Fibrillation auriculaire au cours de l'hyperthyroïdie. Incidence de la cardiopathie associée et de la dilatation de l'oreillette gauche.

In this study the mechanism responsible for atrial fibrillation (AF) in hyperthyroidism was investigated by standard cardiovascular exploration and echocardiography. Fifty four patients (43 women, 11 men, mean age 44 years) were examined during, and after successful treatment of a thyrotoxicosis episode associated with Graves' disease in 43 cases, with a secondarily toxic goitre in 7 cases and with a toxic adenoma in 4 cases. Nineteen patients presented with a heart disease: mitral valve prolapse (MVP) in 11 (including 4 with AF) and another cardiopathy in 8 (including 4 with AF). Among the 34 patients without heart disease, only 2 had AF during thyrotoxicosis. In all groups the antero-posterior diameter of the left atrium was greater in patients with AF than in those with normal sinus rhythm, but it remained within normal limits in patients with MVP. It may be assumed that in these cases AF resulted from synergism between the arrhythmogenic potential of MVP and that of the thyroid hormones. In contrast, prior dilatation of the left atrium seemed to play a predominant role in patients with another cardiopathy. The 4 patients with AF in the latter group remained with AF after the thyrotoxicosis was cured, whereas the 2 patients without heart disease and 3 of the 4 MVP patients reverted to sinus rhythm without anti-arrhythmic therapy or cardioversion. It is concluded that the presence of an underlying heart disease accounts for most cases of AF developed during thyrotoxicosis, but in 1 out of 2 cases the heart disease in a minor one, consisting of MVP.(ABSTRACT TRUNCATED AT 250 WORDS)