Correlates of Protection

A correlate of protection is an immune measurement, most often an antibody titre or neutralizing activity, that is associated with protection against a disease and can therefore be used to predict whether a vaccine will protect. Identifying such a marker allows immunogenicity data to stand in, at least partly, for large clinical-endpoint trials.

Definition

A correlate of protection is an immune response, such as a specific antibody titre, that is statistically associated with protection from disease and, when validated, can serve as a surrogate endpoint for vaccine protection.

Scope

The topic covers what qualifies a measurement as a correlate of protection, the distinction between mechanistic and non-mechanistic correlates, how thresholds of protection are derived, and the limits of using a single marker. It is a reference treatment of an evaluation concept, not clinical guidance.

Core questions

• What evidence is needed to establish an immune marker as a correlate of protection?
• How do mechanistic and non-mechanistic correlates differ in their inferential weight?
• How are protective thresholds estimated, and how far can they be generalized?

Key concepts

• Mechanistic correlate of protection
• Non-mechanistic correlate (surrogate)
• Protective threshold or titre
• Neutralizing antibody
• Surrogate endpoint validation
• Bridging immunogenicity to efficacy

Key theories

Mechanistic versus non-mechanistic correlates

Plotkin and Gilbert distinguish a mechanistic correlate of protection, an immune response that is causally responsible for protection, from a non-mechanistic correlate, a marker that predicts protection without directly causing it; the distinction governs how confidently a marker can be used as a surrogate.

Mechanisms

When an immune response causally mediates protection, the level of that response predicts the probability of protection, allowing a threshold to be estimated above which most vaccinees are protected. Validation requires showing that the marker tracks protection across trials or populations; immune-correlates analyses nested in efficacy trials, such as the analysis of the mRNA-1273 COVID-19 trial, estimate how protection rises with binding and neutralizing antibody levels.

Clinical relevance

Correlates of protection shape how vaccines are licensed, updated, and compared, and how immunogenicity results are interpreted. The concept describes how vaccine protection is inferred from immune measurements and is reference material rather than a basis for individual clinical decisions.

Epidemiology

A validated correlate can reduce the need for repeated large efficacy trials, for example when adapting a vaccine to a new variant or extending it to a new age group, by allowing immunogenicity to predict protection. However, the strength of a correlate can vary by pathogen, vaccine platform, and outcome, so generalization must be checked.

Evidence & guidelines

The conceptual framework comes from Plotkin's reviews, and quantitative examples come from immune-correlates analyses of efficacy trials such as the COVID-19 vaccine studies. These are reference syntheses and research analyses, not prescriptive guidelines.

History

The notion that a measurable immune level predicts protection has long-standing roots in serology, with protective antibody thresholds described for diseases such as diphtheria and tetanus. Plotkin's reviews from 2010 onward formalized the modern terminology, and the COVID-19 pandemic produced detailed quantitative correlates analyses linking neutralizing antibody levels to protection.

Debates

Can a single marker fully capture protection?

Protection often depends on multiple, interacting immune mechanisms, so a single correlate such as antibody titre may be incomplete and can underestimate contributions from cellular immunity or non-neutralizing antibody functions.

How transferable is a protective threshold?

A correlate established for one population, variant, or outcome may not hold for another, so thresholds derived in one setting require validation before being applied elsewhere.

Key figures

• Stanley A. Plotkin
• Peter B. Gilbert
• David S. Khoury
• Miles P. Davenport

Seminal works

• plotkin-2010
• plotkin-gilbert-2012
• plotkin-2013
• khoury-2021
• gilbert-2022

Frequently asked questions

What is the difference between a mechanistic and a non-mechanistic correlate of protection?

A mechanistic correlate is an immune response that directly causes protection, while a non-mechanistic correlate is a marker that reliably predicts protection without itself being the cause; both can be useful, but the mechanistic correlate supports stronger inference.

Why are correlates of protection useful?

Once validated, a correlate lets immunogenicity measurements predict whether a vaccine will protect, which can reduce the need for large, lengthy clinical-endpoint trials when updating or extending a vaccine.