As Acute Ischemic Stroke (AIS) becomes a leading cause of mortality and long-term disability, many researchers have focused on preventative measures as well as cutting-edge treatment options for this disease. These treatment options could be dependent on accurate assessment of stroke severity as well as the correct prediction of functional outcomes. The optimal therapeutic practice could then be provided to stroke patients and health care resources could be sufficiently allocated.
Accurate assessment and prognostication could not be performed without the evaluation of biomarkers. Various blood serum biomarkers have been evaluated for their prognostic values, including thyroid axis hormones, copeptin, CRP, glutamate, glucose, and vitamin D [17–20]. Among these values, the thyroid-related hormones could be easily tested with stroke patients’ peripheral blood and they are known to play a neural protective role in brain development, morphogenesis, neuron maturation process, and inflammation reduction [21].
A stroke event would trigger the stress response in the body resulting in blood serum level changes. The body responds to the stress by accumulating pro-inflammatory cytokines (IL-1, IL-6, IL-10, TNF-2, etc.) that would interfere with normal thyroid hormone metabolism leading to changes in hormone levels [22].
Researches have focused on the physiological relevancy of thyroid hormone profiles. T3 is known to be negatively correlated to stroke severity (expressed as NIHSS scores) and some studies have concluded that fT3 and T3/T4 ratios are negatively associated with the NIHSS scores, but results are not confirmed in other researches [9–14]. The relationship between other thyroid hormones (TSH, T4, and fT4) and NIHSS score have inconclusive results. Most of the previous analyses have been focusing on the whole population analysis to explore the predictive and prognostic values. Those studies did not take the effect of aging into account, or they simply adjusted age as a confounding factor in the multivariate analysis.
Aging is also a significant prognostic factor concerning severity and outcome in stroke patients [23]. The elderly patients have on average increased or unchanged levels of T4 and decreased serum T3, due to declined conversion from T4 to T3 by deiodinase [24]. The special physiologic status of aged patients sometimes leads to subclinical hypothyroidism, which complicates the assessment of thyroid status with precision. A low T3 value of elderly patients may be the result from stroke stress response, or it could be related to subclinical hypothyroidism caused by age.
Li et al have conducted a set of statistical analyses on stroke victims in two age groups, the below-65 group and the over-65 group [15]. It was concluded that the prognostic value of T3 for a poor functional outcome is clinically more meaningful in the over-65 group. But the study did not mention the degree of association between different factors and stroke severity in different age groups. Also, there was few studies focusing on the effect of aging in patients over 80 years old. It was clinically meaningful to explore the prognostic value of thyroid-related hormones in stroke patients over 80 years old, taking into account the complication of the special physiologic status.
Aging can significantly affect the results and complicate the analysis.Besides conducting a whole population analysis, more delicate age-dependent grouping methods could be used to reveal the true and accurate association between biomarkers and outcomes.
In our analysis, the correlation results vary in different age groups. For patients above 80 years old, only T3 and fT3 were negatively correlated to NIHSS scores (both P-values = 0.005). In the 65-80 Group, four out of the five thyroid hormones analyzed (except for T4) were found to have a significantly negative association with the NIHSS scores. In the Below 65 group, the TSH and fT3 were found to have a significant relation to stroke severity assessment. Looking at the correlation coefficients across different age groups, the absolute values of rho are lower in the younger age group (the 65-80 Years Group) than that in the Above 80 group, suggesting smaller magnitudes of correlation in younger age groups. For fT3 values across the three age groups, the absolute values of rho were lower in younger age groups, suggesting an increasing magnitude of correlation as age went up.
The age-dependent grouping method could be potentially applied to other analyses for disease severity assessment and prognostication. A non-significant association between any two factors in the entire population might present significant results in sub-group analysis.
However, the association between endocrine alternation does not necessarily imply causation. It is still debatable whether thyroid hormones disorder should be treated as part of the therapeutic process after stroke [25].
Although no data on the measurements of functional outcomes such as mRS or NIHSS at discharge were available in the study, there are a variety of studies suggesting that initial NIHSS scores are directly associated with functional outcomes upon discharge and in the follow-up visits [9–14].
There are a few limitations to this study. Firstly, this was a single-center study that could be inevitably biased towards a certain set of populations. Secondly, the nature of a retrospective analysis limited the ability to obtain more sufficient data, such as NIHSS score and mRS upon hospital discharge and in follow-up visits. More than one thyroid hormone testing after the initial admission test could help reflect more dynamic changes because blood serum levels had large variations during a day. Thirdly, although the NIHSS was evaluated by a neurologist, there might still be observer variability and human error in the assessment thus affecting the accuracy. Lastly, a larger cohort and multiple center population analysis would potentially more statistical insight into the association and the magnitude of the correlation between thyroid hormones and disease severity, functional outcomes, and long-term survival.