Are Thyroid Hormone Values Obtained in Hospitalized Elderly Patients Reproducible? - A Cohort Study

It is unclear whether TSH should be obtained in elderly hospitalized patients as several factors associated with hospitalization influence these levels, which may not truly represent the thyroid status of the individual. All patients≥65 years old hospitalized in a geriatric ward in the years 2012-2014 had TSH measured irrespective of the cause for hospitalization. All patients in whom TSH levels were performed in an ambulatory setting 2-12 months following hospitalization were identified and these TSH levels were correlated with levels recorded during hospitalization. Factors influencing TSH reproducibility were identified through review of patients' medical records. Of 562 patients hospitalized during the study period, 198 had repeat ambulatory TSH measurements during follow-up. The Katz Index of Independence was higher (9.43±2.98 vs. 8.43±3.67 p=0.002) and cerebrovascular disease was less prevalent (15.6 vs. 25.2% p=0.014) in those who had a repeat TSH measurement compared with those who did not, but other baseline characteristics and TSH levels on admission were similar. Ambulatory TSH values were significantly correlated with those obtained during hospitalization (correlation coefficient=0.677), irrespective of baseline systolic blood pressure, cause of hospitalization, or admission albumin levels. Mean TSH difference between in-hospital and ambulatory was 0.65±2.36 mIU/l and in more than 94% of cases TSH values differed by less than 2 mIU/l between the 2 measurements. In hospitalized patients≥65 years old, TSH levels are highly concordant with ambulatory values. TSH obtained during hospitalization may be used for making treatment decisions and has clinical utility in this population.