TABLE 6.

criteria to select biomarkers of aging in clinical trials.

Cohort Size:	Large cohorts require biomarkers that are easy to extract and process at low cost (e.g., serum markers). Studies with smaller cohorts and more specific aging-associated questions may require (and can afford) difficult-to-use and/or more expensive markers (e.g., fibroblast cultures, measurement of telomere length or the methylation level of CpG islands).
Cohort type:	Depending of the aims of the trial, usually reflected by inclusion criteria, it is often not appropriate to consider biomarkers which are usually used as markers for specific tissue damage or organ failure (e.g., creatinine, cystatin C, Pro-BNP) or markers that reflect a general activation of immunological processes such as CRP and IL6 or markers that reflect a higher risk for typical age-related diseases such as lipids, HbA1c or other cardiovascular risk factors. Additionally, organ specific markers should be controlled because these can be strong confounders in a study. A possible strategy to increase the informative value for all aging aspects could be the combination of organ-specific and more general markers.
Compartment of disease:	If the disease (or dysfunction) that is specifically considered in a trial features strong effects not in general but in distinct compartments (organs, tissues, combinations of these, or parts thereof), e.g., the brain, the overall question is which compartment to sample for biomarker analysis, e.g., peripheral blood vs. cerebrospinal fluid. Markers in the blood can often but not necessarily be attributed to more general aging processes.
Assessment of potential pitfalls:	Even if easy-to-handle biomarkers have a high sensitivity for aging-related processes, they often lack clinical specificity. This is true for many inflammatory markers (e.g., CRP, interleukins), which are more valuable markers of aging in populations without an overrepresentation of infections. For most questions acute infection must be ruled out by standard criteria (fever, feeling unwell, B-symptoms, etc.). Specific tissue/organ checks (e.g., physical examination, echocardiography etc.) can be added to rule out acute diseases. Strictly speaking, the biomarkers excluded on this basis may also reflect some acceleration of aging-related processes. However, they are less relevant than biomarkers reflecting more general aspects of aging, and, more importantly, they would lead to misinterpretations in individual patients. Furthermore, in addition to standard preanalytics precautions such as control of patient’s position, application of the tourniquet, fasting vs. non-fasting and diurnal fluctuations, special aspects must be taken into account. For example, measurements that may be influenced by habits such as exercise should not be done on Mondays; exercise on weekends may influence cytokine levels, etc. In general, the same days should be used for all participants and all longitudinal time points.
Future directions:	There is a strong need to investigate biomarkers of aging more systematically. This should include promising markers such as the methylation of CpG islands and the standardization for specific sampling procedures (e.g., of peripheral blood cells for specific measurements) and the clarification as to whether and in what context acute disease markers, which at the same time can also reflect chronic processes of aging, are useful biomarkers of aging. Furthermore, biomarkers might be put together into composite markers, also known as “aging panels.” Finally, the assessment of very sophisticated but highly informative measures with high potential validity to monitor aging such as MRI (“Brain age“) or PET-Scans (e.g., TAU-PET, detecting the continuous increase of TAU deposition in temporo-parietal-occipital lobes) should be considered (Sowell et al., 2003; Cole et al., 2015; Schöll et al., 2016).