1. The review consolidates findings from nearly 100 studies and identifies numerous metabolites (e.g., phenylalanine, serine, valine, hippurate) that are reported as potential biomarkers for multiple, distinct diseases (mastitis, lameness, ketosis, acidosis) as well as for physiological states (pregnancy, lactation, diet, parity). For example, phenylalanine and serine are elevated in both mastitis and lameness; hippurate levels fluctuate in mastitis but are also strong biomarkers for pasture-based diets.
The review fails to address the lack of disease-specificity of these metabolites. If the same metabolite (e.g., phenylalanine) is altered in inflammatory disease (mastitis), pain-related stress (lameness), and even in response to diet or heat stress, how can it be considered a reliable diagnostic biomarker for any single condition? This raises a fundamental question about the confounding effects of farm management, nutrition, and physiological stage on the metabolome. Given that many studies did not control for these variables, how can the authors justify the claim that metabolomics can lead to “early disease detection” without establishing a disease-specific metabolic fingerprint that is orthogonal to background metabolic noise?
2. The review presents a wealth of correlative data linking metabolite alterations to disease states. For instance, in mastitis, increased amino acids are interpreted as “immune activation” and “protein catabolism.” In lameness, increased amino acids are attributed to “negative energy balance” or “pain-induced stress.” However, the review does not critically assess whether these metabolic changes are causally linked to the disease etiology or are merely secondary effects of reduced feed intake, pain, stress, or inflammation common to all illnesses.
Many of the diseases discussed (mastitis, lameness, metritis, ketosis) share common pathophysiological features: inflammation, oxidative stress, and negative energy balance. How can the authors differentiate between disease-specific metabolic drivers and generic stress/inflammation responses? Without mechanistic studies (e.g., isotope tracing, intervention trials) or studies that control for pain and anorexia, the metabolic changes reported may simply reflect a common sickness behavior phenotype rather than unique disease pathways. This severely limits the translational utility of these “biomarkers” for precision diagnosis or targeted therapy. How do the authors propose to disentangle these confounding systemic effects in future study designs?