Waist circumference, metabolic syndrome and coronary artery disease in a Pakistani cohort.

BACKGROUND: Metabolic syndrome (M-IRS) is common in Asians. This study investigated the relationship of two definitions of M-IRS to atherosclerosis in Indian Asians with suspected coronary arterial disease (CAD). METHODS: 400 patients with chest pain selected for the presence or absence of angiographic disease were recruited from a tertiary referral centre in Pakistan into a prospective case-control study. Patients were categorized by the National Cholesterol Education Program adult treatment panel 3 (NCEP) and International Diabetes Federation (IDF) definitions of the metabolic syndrome and the relationship of these to the presence of CAD and extent of atheroma burden was investigated. RESULTS: M-IRS was present in 53% by IDF criteria and in 44% using the Asian criteria for NCEP. The 2 populations identified were only 69% concordant. No relationship existed between the presence of NCEP M-IRS and atheroma burden. In contrast, the presence of IDF M-IRS was associated with CAD (65 vs. 34%; RR=1.88; p<0.001) and angiographic disease burden (28 [0-224] vs. 0 (0-198); RR=1.83; p<0.001). This association persisted (beta=18.4; p<0.001) after correction for C-reactive protein (beta=8.67; p<0.001), lipoprotein (a) (beta=8.14; p=0.002), and estimated glomerular filtration rate (beta=-0.22; p=0.01). Differences in presumed underlying factors were found in the 2 populations identified by the definitions though both agreed on the separate weightings given to blood pressure and HDL-C/apolipoprotein A1. CONCLUSIONS: The specific Asian IDF and NCEP definitions of M-IRS show limited concordance in Pakistanis. The IDF criteria in contrast to the NCEP criteria are associated with the presence of CAD even after allowing for other risk factors identified in this population.

Differential effect of pioglitazone (PGZ) and rosiglitazone (RGZ) on postprandial glucose and lipid metabolism in patients with type 2 diabetes mellitus: a prospective, randomized crossover study.

BACKGROUND: Postprandial metabolism is impaired in patients with type 2 diabetes (T2Dm). Two thiazolidinediones pioglitazone (PGZ) and rosiglitazone (RGZ) have similar effects on glycaemic control but differ in their effects on fasting lipids. This study investigated the effects of RGZ and PGZ on postprandial metabolism in a prospective, randomized crossover trial. METHODS: Seventeen patients with T2Dm were randomized to RGZ or PGZ for 12 weeks, with an 8-week wash-out period. Fasting blood samples were taken for glucose (FPG), insulin, HbA(1c), lipids, apolipoproteins (apo), lipoprotein (LPL) and hepatic lipase (HL), and cholesterol ester transfer protein (CETP) activity. A standardized breakfast was served and postprandial glucose, insulin, and lipid subfraction profiles were determined. RESULTS: RGZ and PGZ treatment resulted in a similar improvement in FPG, HbA(1c) and homeostasis model assessment. Fasting and postprandial triglyceride (TG) levels were significantly lower following PGZ therapy (fasting: -0.35 vs 0.44 mmol/L; p < 0.04; postprandial AUC-TG: -195.6 vs 127.9 mmol/L/min; p < 0.02) associated with changes in VLDL-2-TG (-0.10 vs 0.21 mmol/L; p = 0.23) and VLDL-3-TG (0.0 vs 0.34 mmol/L; p < 0.04). Fasting cholesterol increased with RGZ compared to PGZ (0.06 vs 0.59 mmol/L; p < 0.04), particularly in VLDL-2-C (-0.30 vs 0.59 mmol/L; p < 0.03) and VLDL-3-C (-0.85 vs 2.11 mmol/L; p < 0.02). Postprandial VLDL lipid and protein content increased after RGZ and decreased after PGZ. Fasting apoB, apoA-I, apoC-II/C-III-ratio, and LPL activity did not differ. CETP activity decreased after RGZ and increased after PGZ (-6.2 vs 4.2 p/mol/mL/min; p < 0.002). CONCLUSIONS: Both the glitazones had similar effects on glucose metabolism. The additional beneficial effect of PGZ on lipid metabolism may be related to its effects on insulin-independent VLDL production and CETP activity.

Cardiovascular risk factors determine erectile and arterial function response to sildenafil.

BACKGROUND: Erectile dysfunction is related to endothelial function. Cardiovascular risk factors determine endothelial function. Sildenafil is effective in treatment of erectile dysfunction but shows variable results. This study investigated the relationship of cardiovascular risk factors to acute and chronic responses to sildenafil. METHODS: Cardiovascular risk factors and acute and chronic pulse wave responses to a single 50-mg dose of sildenafil were assessed in 45 patients with erectile dysfunction confirmed by low international index of erectile function (IIEF) score before and after chronic therapy with sildenafil. RESULTS: On recruitment all patients showed evidence of erectile dysfunction with an IIEF score of 5 points (1 to 20 points). Chronic sildenafil therapy resulted in an increase of IIEF score of 13 points (range -1 to +24 points) and 24 patients (53%) achieved an IIEF score>or=21 points. Improvement in erectile function in response to sildenafil (rn=0.79; P<.001) was dependent on initial erectile function (P=.002) and baseline apolipoprotein B (P=.01). Vascular responses to acute treatment with sildenafil were assessed using pulse wave analysis. Acute changes in stiffness index induced by sildenafil (rn=0.65; P<.001) were related to apolipoprotein A-1 (P=.006), B (P=.02), and lipoprotein(a) (P=.008) concentrations, whereas reflection index (rn=0.69; P<.001) was related to pulse pressure (P<.001), albumin-to-creatinine ratio (P=.007), and lipoprotein(a) (P=.02). CONCLUSIONS: The extent of acute and chronic effects of sildenafil on erectile function and pulse wave profiles is determined by metabolic cardiovascular risk factors. Improved cardiovascular risk factor control is likely to increase the efficacy of phosphodiesterase-5 inhibitor therapy in the treatment of erectile dysfunction.

Percentage of body fat and plasma glucose predict plasma sialic acid concentration in type 2 diabetes mellitus.

Circulating sialic acid is an independent risk factor for cardiovascular disease and is higher in people with type 2 diabetes mellitus. Sialic acid is associated with body mass index, but it is uncertain whether body fat contributes to the higher levels of sialic acid in type 2 diabetes mellitus. Therefore, we have investigated whether the higher levels of sialic acid observed in type 2 diabetes mellitus persist when controlling for fatness. Fasting plasma samples were collected from 24 individuals with type 2 diabetes mellitus and 24 controls. Percentage of body fat was measured by bioelectrical impedance. Plasma sialic acid was quantified by an enzymatic method. Plasma sialic acid was higher in the group with type 2 diabetes mellitus than controls (602 +/- 14 vs 545 +/- 14 mg/L, P = .007). Percentage of body fat was associated with plasma sialic acid concentration in both the control group (r = 0.481, P = .020) and the group with type 2 diabetes mellitus (r = 0.527, P = .007). Fasting glucose was also associated with plasma sialic acid in the group with type 2 diabetes mellitus (r = 0.700, P < .001). Adjustment for percentage of body fat accounted for the higher levels of sialic acid in type 2 diabetes mellitus. Using linear regression, 54.3% of the variation of plasma sialic acid was explained by percentage of body fat and glucose concentrations in the whole group. Seventy-four percent of sialic acid variation was explained by the same model in type 2 diabetes mellitus. In conclusion, this is the first study to show that percentage of body fat predicts plasma sialic acid concentration and contributes toward higher levels of sialic acid in type 2 diabetes mellitus.

Low-density lipoprotein apolipoprotein B100 turnover in hypopituitary patients with GH deficiency: a stable isotope study.

BACKGROUND: Epidemiological studies suggest that hypopituitary patients have an increased risk for cardiovascular mortality. The dyslipidaemia associated with this condition is often characterised by an increase in total cholesterol (TC) and low-density lipoprotein (LDL) cholesterol (LDL-C) and may contribute to these findings. The underlying mechanisms are not fully elucidated. MATERIALS AND METHODS: LDL apolipoprotein B (apoB) production rate and metabolic clearance rate were measured in seven patients with hypopituitarism (including GH deficiency) under stable conventional replacement therapy (three males and four females; age 40-16.1 years; body mass index 29.0-6.1 kg/m(2) (means +/- s.d.)) and seven age-, gender- and body mass index-matched control subjects with an infusion of 1-(13)C-leucine. Fasting lipid profile and lipid composition of LDL were also measured. RESULTS: Fasting TC, triglycerides (TG), high-density lipoprotein-C, LDL-C and free fatty acid concentrations were not different between hypopituitary patients and control subjects. LDL-TG (P < 0.006) and LDL-TG/LDL apoB ratio (P < 0.02) were significantly increased in hypopituitary patients. LDL apoB pool size was not statistically different between patients and control subjects. In the hypopituitary patients, LDL apoB metabolic clearance rate (P < 0.05) and LDL apoB production rate (P < 0.02) were lower than in the control subjects. CONCLUSIONS: The present results suggest that LDL apoB turnover and LDL composition is altered in hypopituitary patients. Whether these findings explain the increased risk for cardiovascular disease in hypopituitary patients remains to be established.

Asymmetric dimethyl arginine levels correlate with cardiovascular risk factors in patients with erectile dysfunction.

BACKGROUND: Erectile dysfunction is related to penile arterial endothelial nitric oxide production. Asymmetric dimethylarginine (ADMA) and E-selectin are often considered plasma markers of endothelial function. OBJECTIVE: This study investigated the relationship between these plasma markers and cardiovascular risk factors in patients with erectile dysfunction. METHODS AND RESULTS: Cardiovascular risk factors, ADMA and E-selectin were assessed in 45 patients with erectile dysfunction. Plasma markers showed associations with baseline risk factors. E-selectin levels showed an inverse relationship with age (p = 0.005) and statin therapy (p = 0.03) and a weak association with concomitant beta-blocker therapy (p = 0.05). Compared to these relatively weak associations with cardiovascular risk factors, ADMA levels showed strong associations with pulse pressure (p < 0.001), lack of smoking (p = 0.002) and lipoprotein (a) (p = 0.004) concentrations and weak associations with LDL-cholesterol (p = 0.02), and C-reactive protein levels (p = 0.04). ADMA levels correlated with E-selectin (partial r = 0.76; p < 0.001) after adjustment for lipoprotein (a), pulse pressure and smoking. No change in E-selectin or ADMA levels was seen after 70 days therapy with sildenafil and no relationship was found between either plasma marker and the acute pulse wave response to a single challenge dose of sildenafil. CONCLUSION: ADMA levels correlate at baseline with some cardiovascular risk factors including inflammatory markers and lipoprotein (a) in patients with erectile dysfunction.

Impaired renal function and atherosclerosis in a Pakistani cohort.

OBJECTIVE: To investigate the relationship of creatinine and calculated glomerular filtration rate (GFR) with coronary arterial disease (CAD) in Pakistani patients. SUBJECTS: Four hundred individuals with chest pain; 200 with angiographic disease matched with 200 without occlusive disease. DESIGN: A prospective case-control study. SETTING: A tertiary referral cardiology unit in Pakistan. RESULTS: Impaired renal function as estimated by calculated GFR was common in this population. Creatinine and glomerular filtration rate, as calculated by the Cockcroft-Gault (CG) and Modification of Diet in Renal Disease (MDRD) formulae, were associated with CAD and atherosclerotic burden in Pakistani patients. Calculation of creatinine clearance, correcting for age, sex and body mass index, showed that clearance was 81 (17-257) mL/min/1.73 m2 in patients with CAD compared with 88 (23-167) mL/min/1.73 m2 in controls with a significant number of patients (18.5 vs. 6.5%; RR = 2.85; p < 0.001) showing significant renal impairment (< 60 mL/min/1.73 m2) by CG and more by the MDRD equation (26 vs. 9%; RR = 2.88; p < 0.001). The unadjusted odds ratios for CAD for a GFR < 60 mL/min/1.73 m2 were 3.66 (1.87-7.16) and 3.29 (1.81-6.01), respectively and, after adjustment for diabetes, smoking, insulin resistance, inflammation and apolipoprotein A1, 1.04 (1.02-1.09) and 1.04 (1.02-1.09), respectively. CONCLUSIONS: Impaired renal function is common in Pakistani patients with coronary arterial disease and is strongly associated with a risk of atherosclerosis independent of insulin resistance.

Efficacy of ezetimibe in patients with statin-resistant and statin-intolerant familial hyperlipidaemias.

OBJECTIVE: To investigate the efficacy of the cholesterol absorption inhibitor ezetimibe in patients with refractory familial hyperlipidaemia or intolerant to statin therapy. METHODS: This prospective study assessed the safety and efficacy of ezetimibe in 200 patients with refractory familial hyperlipidaemias not achieving a low-density-lipoprotein (LDL) cholesterol < 3 cholesterol < 3 mmol/L (116 mg/dL) including 22% intolerant to all statin therapy, many consuming intolerant to all statin therapy, many consuming sterol-containing products. RESULTS: Ezetimibe monotherapy resulted in 7% and 11% reductions in LDL-cholesterol and apolipoprotein B respectively. Ezetimibe-statin combination therapy reduced LDL-cholesterol by an additional 11 +/- 27% and apolipoprotein B by 11 (+79 to -18)%. There was a similar response between various sub-groups but a wide variation within groups with the greatest effect seen in patients groups with the greatest effect seen in patients under-responding to statins. The number of patients achieving the LDL-C target of 3 mmol/L rose from 5.5% to 18%. Non-significant effects included a 5 (+78 to -470)% reduction in triglycerides, 8 +/- 36% increment in HDL-cholesterol, 21 (+35 to -82)% reduction in C-reactive protein and a 1 (+20 to -50)% increase in alanine transaminase. No effects were seen on creatinine, creatine kinase, or insulin resistance. Fourteen patients (7%) discontinued ezetimibe: seven due to gastrointestinal side-effects, one patient developed an ezetimibe-induced hypercholesterolaemia (x 1.5), one developed ezetimibe-induced hypertriglyceridaemia (x 7) and five discontinued for other reasons. CONCLUSION: Ezetimibe is a useful addition to statins in patients with familial hyperlipidaemias but shows a highly variable response profile.

Waist-hip ratio and low HDL predict the risk of coronary artery disease in Pakistanis.

OBJECTIVE: To establish risk factor causal associations for coronary artery disease (CAD) in the native Pakistani population. METHODS: We conducted a hospital-based, case-control study of 200 cases with angiographically documented CAD and 200 age- and sex-matched controls without angiographic evidence of CAD. Patients on lipid lowering therapy were excluded. Lifestyle, anthropometric and biochemical risk factors were assessed in both groups. RESULTS: The presence of CAD was associated with current, past or passive smoking, a history of diabetes and high blood pressure, a positive family risk factors in this study; levels were below history of CAD, body fat percentage, waist-hip ratio (WHR), low apolipoprotein A1 or low HDL, lipoprotein (a), glucose, insulin, insulin resistance, C-reactive protein (CRP), total cholesterol to HDL ratio (TC/HDL) and creatinine on univariate conditional logistic regression analysis. In multiple regression analysis, significant independent associations were found with low HDL (OR 0.11; 95% CI 0.04-0.34; p < 0.001) positive family history (OR 1.79; 95% CI 1.09-2.93; p = 0.02), CRP (OR 1.45; 95% CI 1.19-1.75; p < 0.001) and WHR (OR 1.04; 95% CI 1.01-1.08; p = 0.01). Angiograms were also quantified for the extent and severity of CAD by the Gensini scoring system. Quantitative angiographic data showed associations with age (p = 0.01), the duration of diabetes (p = 0.04), WHR (p = 0.06), low HDL (p < 0.001), lipoprotein (a) (p = 0.001), creatinine (p < 0.001) and CRP (p = 0.007). Results indicate that total and LDL cholesterol were not significant currently accepted thresholds for treatment. CONCLUSIONS: The cardiovascular risk profile in this population is consistent with metabolic syndrome where low HDL and WHR can be used to predict the risk of CAD. Results suggest the need to redefine the currently practised approach to CAD management in this population to fit local needs.

Raised plasma total sialic acid levels are markers of cardiovascular disease in renal dialysis patients.

Cardiovascular disease (CVD) rates in dialysis patients are very high. One of the many associated risk factors is chronic inflammation. The relationship of baseline markers of chronic inflammation with the presence of CVD was assessed in a large cohort of stable dialysis patients. Median time (IQR) on dialysis treatment was 20(9-52) months. Forty-one patients had CVD (as defined by the history / clinical presence of ischemic heart disease, peripheral vascular disease, or cerebrovascular disease). Patients with CVD were significantly older than patients without (67 + /- 11 vs. 54 + /- 10 yrs, p < 0.03). Time from dialysis, urea reduction ratio (hemodialysis only) and smoking history were similar between the two groups. Patients with CVD had significantly higher levels of sialic acid (SA) (91.2 +/- 24.2 vs. 82.0 + /- 18.2 mg/dL, p = 0.03). Body weight, plasma fibrinogen, C-reactive protein (CRP), homocysteine, creatinine, total-, LDL (low density lipoprotein)-, or HDL (high density lipoprotein)-cholesterol, systolic, diastolic and pulse pressures did not differ between the CVD and CVD(-) groups. Patients on chronic ambulatory peritoneal dialysis (CAPD) had more elevated lipid fractions, inflammatory markers, and SA levels than did patients on hemodialysis (HD). The presence of diabetes, the use of lipid-lowering therapy, and smoking history was not associated with any difference in SA levels. In contrast to C-reactive protein (CRP) concentrations, SA levels were unaffected by the hemodialysis session. SA was strongly correlated with CRP (r = 0.59, p < 0.0001), but not with patient age, any measure of blood pressure (BP), urea reduction ratio, plasma creatinine, lipid fractions or homocysteine. Levels of the chronic inflammation marker sialic acid correlate strongly with CRP and are increased in patients with cardiovascular disease, but show no relationship to hemodialysis session. Thus sialic acid may be a superior marker to CRP for assessment of chronic inflammation in patients undergoing dialysis.

The apolipoprotein E2 allele modulates activity and maximal velocity of the sodium-lithium countertransporter.

BACKGROUND: Alterations in erythrocyte sodium-lithium countertransport (SLC) activity and its maximal velocity (Vmax) are associated with hypertension and hypertriglyceridemia. The presence of apolipoprotein (apo) E variants is associated with hypertriglyceridemia. This study investigated the relationship between apoE phenotype and SLC kinetics. METHODS: Cardiovascular risk factors and SLC kinetics were measured in 171 subjects and 69 controls. Apolipoprotein E phenotypes were determined by Western blotting. RESULTS: Patients were 51% male, aged 56+/-13 years, with a blood pressure (BP) of 134+/-22/81+/-11 mm Hg, total cholesterol of 6.71+/-1.57 (256+/-61 mg/dL); median triglycerides 1.65 mmol/L (146 mg/dL) (range, 0.31 to 9.85 mmol/L; 27 to 872 mg/dL) and high-density lipoprotein (HDL) 1.39+/-0.43 mmol/L (54+/-16.6 mg/dL); fasting glucose 4.91+/-0.61 mmol/L (88.5+/-11.0 mg/dL); median insulin 11.7 IU/L (range, 3.7 to 39.8 IU/L). Phenotype frequencies were E3/E3 56%, E2/E3 14%, E2/E2 1%, E3/E4 27%, and E4/E4 2%. The SLC activity, Vmax, and sodium affinity (Km) were not significantly different with respect to apoE phenotype in simple analysis by Kruskal Wallis test. However, in multiple regression analysis after exclusion of BP, a strong co-correlate of SLC activity, the presence of an apoE2 allele was associated reduced activity (beta = -0.061; P = .01) along with HDL:apoA1 ratio (beta = -0.170; P < .001), whereas for the kinetic parameter Vmax, associations were found with triglyceride (beta = 0.029; P = .04), HDL:apoA1 ratio (beta = -0.186; P = .03) and the presence of an apoE2 allele (beta = -0.089; P = .04). CONCLUSIONS: These findings suggest that the apoE phenotype may modulate SLC activity and that the presence of an apoE2 allele phenotype is associated with lower SLC activity and Vmax.

Effect of IGF-I therapy on VLDL apolipoprotein B100 metabolism in type 1 diabetes mellitus.

Abnormal lipid metabolism may be related to the increased cardiovascular risk in type 1 diabetes. Secretion and clearance rates of very low density lipoprotein (VLDL) apolipoprotein B100 (apoB) determine plasma lipid concentrations. Type 1 diabetes is characterized by increased growth hormone (GH) secretion and decreased insulin-like growth factor (IGF) I concentrations. High-dose IGF-I therapy improves the lipid profile in type 1 diabetes. This study examined the effect of low-dose (40 microg.kg(-1).day(-1)) IGF-I therapy on VLDL apoB metabolism, VLDL composition, and the GH-IGF-I axis during euglycemia in type 1 diabetes. Using a stable isotope technique, VLDL apoB kinetics were estimated before and after 1 wk of IGF-I therapy in 12 patients with type 1 diabetes in a double-blind, placebo-controlled trial. Fasting plasma triglyceride (P < 0.03), VLDL-triglyceride concentrations (P < 0.05), and the VLDL-triglyceride-to-VLDL apoB ratio (P < 0.002) significantly decreased after IGF-I therapy, whereas VLDL apoB kinetics were not significantly affected by IGF-I therapy. IGF-I therapy resulted in a significant increase in IGF-I and a significant reduction in GH concentrations. The mean overnight insulin concentrations during euglycemia decreased by 25% after IGF-I therapy. These results indicate that low-dose IGF-I therapy restores the GH-IGF-I axis in type 1 diabetes. IGF-I therapy changes fasting triglyceride concentrations and VLDL composition probably because of an increase in insulin sensitivity.

The uptake of lipoprotein-borne phylloquinone (vitamin K1) by osteoblasts and osteoblast-like cells: role of heparan sulfate proteoglycans and apolipoprotein E.

Vitamin K is essential for the gamma-carboxylation of Gla-containing bone proteins such as osteocalcin and a suboptimal vitamin K status has been linked to osteoporosis but nothing is known of how the lipoprotein-borne vitamin accesses the bone matrix. We have studied the mechanism of transport of lipoproteins labeled with [3H]-phylloquinone (vitamin K1 [K1]) into osteoblasts using both tumor-derived cell lines and normal osteoblast-rich cell populations. We also investigated the effect of heparin in this model since long-term heparin treatment causes osteopenia and the anticoagulant is known to impair normal lipoprotein metabolism. Heparinase treatment, which removes heparan sulfate proteoglycans (HSPG), reduced uptake of [3H]-K1 from triglyceride-rich lipoproteins (TRL) and low-density lipoproteins (LDL). The effect of heparin in this model was complex depending on cell type, concentration, and time but, overall, the results were consistent with an inhibition of vitamin K uptake by osteoblasts. Anti-apolipoprotein E (apoE) antiserum reduced uptake of TRL-[3H]-K1 by 55 +/- 4% and LDL-[3H]-K1 uptake by 35 +/- 2%. Exogenous apoE4 increased uptake of TRL-[3H]-K1 by 90 +/- 1% compared with 53 +/- 11% for apoE3 and 52 +/- 5% for apoE2. Our findings show that HSPG on the cell surface and apoE in the lipoprotein particles contribute to lipoprotein-K1 uptake by osteoblasts as is known for lipoprotein uptake by hepatocytes. This mechanism is significant in view of the epidemiological association of both undercarboxylation of osteocalcin and the presence of an apo epsilon4 allele with increased fracture risk and reduced bone mineral density (BMD). The inhibition by heparin of lipoprotein-mediated carriage of vitamin K and possibly other lipids to bone may provide a basis for the future understanding of heparin-induced osteoporosis.

A 6-month randomized trial of thyroxine treatment in women with mild subclinical hypothyroidism.

PURPOSE: The role of thyroxine replacement in subclinical hypothyroidism remains unclear. We performed a 6-month randomized, double-blind, placebo-controlled trial to evaluate the effects of thyroxine treatment for mild subclinical hypothyroidism, defined as a serum thyroid-stimulating hormone level between 5 to 10 microU/mL with a normal serum free thyroxine level (0.8-16 ng/dL). SUBJECTS AND METHODS: We randomly assigned 40 women with mild subclinical hypothyroidism who had presented to their family practitioners to either thyroxine treatment (n = 23; 50 to 100 microg daily) or placebo (n = 17). Health-related quality of life (Hospital Anxiety and Depression scale, 30-item General Health Questionnaire), fasting lipid profiles, body weight, and resting energy expenditure were measured at baseline and 6 months. RESULTS: The most common presenting symptoms were fatigue (n = 33 [83%]) and weight gain (n = 32 [80%]). At presentation, 20 women (50%) had elevated anxiety scores and 22 (56%) had elevated scores on the General Health Questionnaire. Thirty-five women completed the study. There were no significant differences in the changes from baseline to 6 months between women in the thyroxine group and the placebo group for any of the metabolic, lipid, or anthropometric variables measured, expressed as the mean change in the thyroxine group minus the mean change in the placebo group: body mass index, -0.3 kg/m(2) (95% confidence interval [CI]: -0.9 to 0.4 kg/m(2)); resting energy expenditure, -0.2 kcal/kg/24 h (95% CI: -1.3 to 1.0 kcal/kg/24 h); and low-density lipoprotein cholesterol, -4 mg/dL (95% CI: -23 to 15 mg/dL). There was a significant worsening in anxiety scores in the thyroxine group (scores increased in 8 of 20 women and were unchanged in 2 of 20) compared with the placebo group (scores increased in 1 of 14 women and were unchanged in 6 of 14; P = 0.03). CONCLUSIONS; We observed no clinically relevant benefits from 6 months of thyroxine treatment in women with mild subclinical hypothyroidism.