Although underappreciated by practitioners, there exists good evidence that thyroid dysfunction is common in patients with chronic kidney disease (CKD). Documentation of this association has come from large, well-conducted epidemiological studies that have demonstrated an inverse relationship between thyroid function (generally best assessed by measurement of serum thyrotropin [TSH] levels) and estimated glomerular filtration rates.
- Lo J.C.
- Chertow G.M.
- Go A.S.
- Hsu C.Y.
Increased prevalence of subclinical and clinical hypothyroidism in persons with chronic kidney disease.
- Chonchol M.
- Lippi G.
- Salvagno G.
- Zoppini G.
- Muggeo M.
- Targher G.
Prevalence of subclinical hypothyroidism in patients with chronic kidney disease.
Notably, although it is well recognized in the general population that more severe levels of both hypothyroidism and hyperthyroidism are associated with alterations in cardiovascular function, these potential relationships have been less well characterized in patients with CKD with the exception of studies in patients with end-stage renal disease. End-stage renal disease is characterized by exceedingly high rates of cardiovascular disease and attendant morbidity and mortality, and the majority of studies in end-stage renal disease show that both hypothyroidism and hyperthyroidism are associated with further increases in mortality in this patient population.
- Rhee C.M.
- Kim S.
- Gillen D.L.
- et al.
Association of thyroid functional disease with mortality in a national cohort of incident hemodialysis patients.
- Drechsler C.
- Schneider A.
- Gutjahr-Lengsfeld L.
- et al.
Thyroid function, cardiovascular events, and mortality in diabetic hemodialysis patients.
However, whether thyroid dysfunction is associated with the risk of death in patients with less severe renal impairment is unclear.
It is in this context that the article by Rhee et al
- Rhee C.M.
- Kalantar-Zadeh K.
- Ravel V.
- et al.
Thyroid status and death risk in US veterans with chronic kidney disease.
in the current issue of Mayo Clinic Proceedings
is of particular interest. As described in their study, the authors performed a retrospective analysis of a large (N=227,422) population of US military veterans with non–dialysis-dependent CKD in order to analyze the impact of thyroid status (as defined by TSH measurement) on all-cause mortality. Rather than studying patients with near dialysis dependence, the authors chose to examine mortality in patients with moderate renal dysfunction (CKD stage 3, defined as an estimated glomerular filtration rate of 30-60 mL/min per 1.73 m2
), a patient population commonly found in clinical practice.
Importantly, the authors astutely acknowledge that in a population with chronic disease, mortality might reflect concomitant comorbidity burden, with thyroid function test abnormalities a manifestation of what has been frequently referred to as “sick euthyroid syndrome” and now better known as acquired transient central hypothyroidism.
Clinical review 86; euthyroid sick syndrome: is it a misnomer?.
To address this issue, Rhee et al performed 2 sensitivity analyses to examine thyroid status and mortality. The first involved the use of a 30-day “lag period” from TSH measurement to follow-up date to minimize the risk for transient changes in TSH levels that were readily reversible; the second used incident cases of thyroid dysfunction.
Overall, assessments of both thyroid status and the potential impact of therapies used to modulate thyroid function were carefully evaluated. As described in the article, the statistical analysis plan was sufficiently detailed and utilized multiple layers of analytical assessment to determine the impact of a large number of variables and potential confounders (eg, laboratory values, demographic variables, comorbidities, nutritional status, thyroid therapies) on the outcome of interest. As an example of the authors' impressive efforts at comprehensiveness is their inclusion of variables that could be construed to reflect physician and patient attentiveness to care such as information regarding whether patients had received influenza vaccination or had prescriptions for angiotensin-converting enzyme inhibitor, angiotensin receptor blocker, or statin therapies within the preceding year.
As clearly demonstrated, the authors unequivocally identify a significant association between thyroid dysfunction (both hypothyroidism and hyperthyroidism) and increased mortality in patients with stage 3 CKD, with an increased risk of death 17% to 27% above the risk identified in euthyroid patients with CKD, over a median follow-up period of 5.5 years. This finding remained fairly consistent both when sensitivity analyses were carried out with the 30-day lag period from time of TSH testing to outcome and when only incident cases of thyroid dysfunction were included. Furthermore, when TSH values were subclassified into 4 broad categories (0.5 to <3.0, 3.0-5.0, >5.0-10.0, and >10 mIU/L), an increased risk of mortality with increasing TSH values was observed. Importantly, similar results were noted for TSH values within the hyperthyroid range with 2 caveats: first, only patients in the subclinical hyperthyroidism group (defined as a TSH level of 0.1 to <0.5 mIU/L) had those results confirmed in sensitivity analyses; and, second, the impact of subclinical hyperthyroidism on mortality was stronger than for overt hyperthyroidism. Not surprisingly, these results led to a U-shaped association of thyroid function/dysfunction with mortality, with TSH values in the 1.7 to 2.1 mIU/L range identified as optimal. Notably, adjustments related to quality of care did not impact the identified associations between thyroid dysfunction and mortality, both with respect to TSH values, which were suppressed (<0.5 mIU/L), and those that started from the upper-normal range (TSH >3.0 mIU/L). Body mass index was identified as a covariate that impacted thyroid function–associated mortality for both hypothyroidism and hyperthyroidism, with a higher risk seen in patients who were classified as overweight (body mass index >25 kg/m2) or above. Intriguingly, in female patients the impact of hyperthyroidism on mortality was also evinced, whereas an association of hypothyroidism and mortality was not observed. Other notable findings of the study were that the impact of both hypothyroidism and hyperthyroidism on mortality risk was actually stronger in younger (ie, <60 years) as compared with older individuals, and that pharmacological treatment of subclinical hypothyroidism with thyroid hormone replacement may be beneficial, as levothyroxine treatment was associated with reduced mortality in those who achieved a TSH value within the normal range. Finally, multiple sensitivity analyses revealed results consistent with the main analysis, thereby endorsing the internal validity of the authors' findings.
Collectively, these findings provide reasons for care providers to more carefully consider the relationship between thyroid status and renal function in patients with moderate renal impairment (CKD stage 3), a common finding in clinical practice, particularly in aged patients. Given that the observed increased mortality risk in patients with renal dysfunction contrasts with normal population data in which age-associated subclinical hypothyroidism was associated with longevity,
- Atzmon G.
- Barzilai N.
- Surks M.I.
- Gabriely I.
Genetic predisposition to elevated serum thyrotropin is associated with exceptional longevity.
it seems likely that the concomitant occurrence of CKD and thyroid dysfunction must be exerting a negative impact on overall health. The basis for this negative effect, however, remains unknown. Accordingly, this observation warrants closer investigation into potential mechanisms underlying this relationship in order to determine if this association plays a causal role in the increased mortality seen in patients with moderate renal dysfunction. Potential guidance for such investigations may come from studies of the relationship between dialysis-dependent CKD and thyroid function, in which progressive hypothyroidism associates with increased mortality.
- Rhee C.M.
- Alexander E.K.
- Bhan I.
- Brunelli S.M.
Hypothyroidism and mortality among dialysis patients.
Whether well-recognized pathophysiologic alterations that occur with both hypothyroidism (such as hyperlipidemia, accelerated atherosclerosis, and endothelial dysfunction) and hyperthyroidism (such as increased cardiac demand with resultant ischemia, atrial fibrillation, and cardiac remodeling) are exacerbated even in moderate (CKD stage 3) renal impairment remains an open question.
Once these questions are satisfactorily answered, it will be important to discuss the potential need for a public health policy in which thyroid monitoring is routinely considered in patients with non–dialysis-dependent CKD and also to identify therapeutic approaches that might mitigate this increased mortality risk. In this context, it is germane to recognize that in the population of hypothyroid individuals included in this study who were treated to target TSH levels, the mortality risk was reduced to below that of euthyroid individuals (95% hazard ratio, 0.85-0.91). This finding suggests that the likely low risk of thyroid hormone replacement therapy may be justifiable in patients with non–dialysis-dependent CKD, although this would require formal prospective study. Additional points of consideration include the potential target TSH level, as well as the risk for the development of iatrogenic hyperthyroidism, a risk that would be expected to increase as the TSH target goal was lowered.
Overall, this study provides considerable food for thought. Strengths of the study are myriad and include the use of a very large database with extensive statistical assessment and sensitivity analyses that verified the validity of the data. Additionally, the duration of follow-up was adequate to assess for the primary end point, and access to care was similar for all the individuals. The study is not without limitations, and these are clearly outlined by the authors and are implicit in this type of retrospective study, particularly the finding of association as opposed to causation.
The interesting results reported in this study will require careful consideration if we are to improve the health of the target population. Subsequent studies should pursue TSH screening in patients with non–dialysis-dependent CKD, both to understand the magnitude of the problem and to allow for prospective evaluation. Future research may also benefit from the evaluation of additional parameters such as thyroxine and triiodothyronine values so as to understand whether this additional data provide nuance to the mortality risk prediction models. As the article by Rhee et al clearly demonstrates, there is much to do if we are to understand and harness this new insight and improve the lives of our patients with CKD.