Evidence for altered intestinal permeability in human research

What observational evidence can establish

Human observational research can determine whether measured permeability-related features differ between defined groups or correlate with defined outcomes. This is scientifically valuable because it identifies patterns in the populations of interest and tests whether further investigation is warranted.

A cross-sectional difference does not establish which feature came first. It cannot distinguish whether altered measurement contributed to the clinical state, resulted from it, shared a cause with it, or was influenced by treatment, diet, age, inflammation or another factor.

The systematic review as a worked example

Birna’s verified systematic review and meta-analysis examined zonulin-dependent intestinal permeability in children diagnosed with mental disorders. Its inferential population and outcomes should remain exactly that narrow. The publication is not evidence about every age group, every mental-health condition or every permeability method.

Systematic review makes search, selection and synthesis explicit. Meta-analysis can estimate a combined association where data are sufficiently comparable. Neither method repairs limitations in the included studies; the synthesis inherits variation in design, measurement and population.

Heterogeneity is biological and methodological

Heterogeneity can arise from differences in diagnostic groups, age, sample selection, specimen, marker, assay, timing and analysis. It can also reflect real biological diversity. A pooled estimate should therefore be read alongside the comparability and quality of its components.

Biomarker heterogeneity is especially important when different measurements are treated under one conceptual label. Shared terminology does not guarantee a shared molecular target or functional construct.

Confounding, reverse causation and time

Confounding occurs when another factor is associated with both the measured exposure and outcome. Reverse causation is possible when the outcome or its context influences the marker rather than the marker preceding the outcome. These alternatives cannot be removed by stronger wording.

Longitudinal measurement can improve temporal interpretation because it observes change through time, but it still requires valid measurements and adequate control of competing explanations. Intervention evidence addresses a further question: whether deliberately changing an exposure changes an outcome.

What the evidence supports

The verified review supports the existence of an observational evidence base in the specified paediatric populations and supports quantitative synthesis where study data permitted it.

It supports identifying heterogeneity and methodological limitation as central findings rather than treating them as secondary caveats.

What remains uncertain

The evidence does not establish temporal direction, causation, diagnostic thresholds or treatment effects.

Conclusions should not be generalized beyond the populations, markers and study designs included in the verified publication.