Low levels of triiodothyronine (the active form of thyroid hormone) are commonly observed in patients with end-stage renal disease (ESRD), and have been linked to inflammation and adverse cardiovascular effects in this setting.OBJECTIVE
To establish whether triiodothyronine is associated with survival in ESRD.DESIGN AND INTERVENTION
All patients with ESRD who were receiving thrice-weekly standard bicarbonate hemodialysis at one of two centers with the same catchment area during a calendar year, and all patients who began dialysis at these centers within the subsequent 15 months, were eligible to enrol in this prospective Italian study. Exclusion criteria were other illnesses (e.g. infection or malignancy) and use of drugs that could influence thyroid hormone levels. Blood was collected during the midweek interdialytic interval, between 0800 h and 1000 h, for measurement of plasma free triiodothyronine (by radioimmunoassay) and interleukin (IL-)6. To determine the relationship between triiodothyronine and mortality, two control groups were recruited. The first group comprised healthy individuals who were sex-matched and age-matched to within 5 years of the patients; the second group was made up of consecutive individuals with clinically normal thyroid function and normal renal function (serum creatinine level <106 μmol/l [<1.2 mg/dl]), who had been referred for specialist treatment of osteoporosis, hypertension or gastrointestinal disease.OUTCOME MEASURE
Death was the primary endpoint.RESULTS
The 200 patients (48% female; mean age 61 years) were followed up for a mean of 42 months (range 0.2-70 months). All patients had clinically normal thyroid function. The 31 healthy individuals and the 262 individuals with clinically normal thyroid and renal function both had higher plasma levels of free triiodothyronine than the study cohort (370 pg/dl [5.7 pmol/l] and 360 pg/dl [5.5 pmol/l] vs 330 pg/dl [5.1 pmol/l]; P<0.001). When the cohort was divided into tertiles according to triiodothyronine level, patients in the first tertile were significantly older, were more likely to have a history of cardiovascular events and had higher IL-6 levels than those in the other tertiles (P<0.001 for all). There were 102 deaths in total, of which 68 (66%) were cardiovascular. Mean plasma free triiodothyronine levels were significantly lower in patients who died than in those who did not (310 pg/dl [4.8 pmol/l] vs 370 pg/dl [5.7 pmol/l]; P<0.001). Kaplan-Meier analysis revealed that the risk of death increased significantly across tertiles of decreasing triiodothyronine level (log-rank statistic 25.3; P<0.001). This association remained significant when triiodothyronine level was analyzed as a continuous variable by multivariate Cox regression analysis, taking into account the clinical and demographic differences between tertiles, and univariate risk factors for death (e.g. age, smoking status and presence of diabetes). In this model, a 100 pg/dl [1.5 pmol/l] increase in plasma free triiodothyronine level corresponded to a relative all-cause mortality risk of 0.50 (95% CI 0.35-0.72; P<0.001).CONCLUSION
Low plasma level of free triiodothyronine independently predicts death in patients on hemodialysis.