Absolute and Baseline-Relative Emax Parameterizations

Why Two Emax Equations Can Describe the Same Curve

In a direct concentration-effect model, the response at zero drug concentration may already be above or below zero. A physiological measurement, for example, can have a nonzero baseline before the drug contributes any effect. This creates two natural ways to describe the response ceiling: report how much the drug raises the response above baseline, or report the absolute response at the ceiling. Both descriptions can represent the same curve, but the symbol \(E_{\max}\) is used differently in them. A parameter value cannot be interpreted correctly until its convention is stated.

The equations below describe a direct, stimulatory, saturating effect with concentration \(C\). They assume that response is measured on a fixed scale, \(E_0\) is the response at zero concentration, and the response approaches a plateau as concentration increases. The concentration-effect relationship is assumed to be adequately represented by the Emax form; the equations themselves do not establish that the model is appropriate for every drug or data set.

The subscripts “inc” and “abs” are included here to make the two meanings explicit. In applications, authors may write \(E_{\max}\) for either quantity. Therefore, an equation or a clear verbal definition is needed before interpreting an \(E_{\max}\) estimate. The baseline-relative form calls the maximum change \(E_{\max}\); the absolute-plateau form calls the maximum response \(E_{\max}\).

Converting Between the Parameterizations

The conversion follows by matching the coefficients of the concentration-dependent term. In the baseline-relative form, the response can be written as \(E(C)=E_0+\left(E_{\max,\mathrm{inc}}\right)C/(EC_{50}+C)\). In the absolute form, that coefficient is \(E_{\max,\mathrm{abs}}-E_0\). Since the curves are to be identical at every concentration, these coefficients must be equal.

The conversion leaves \(E_0\) and \(EC_{50}\) unchanged. It changes only the meaning and value assigned to the maximum-response parameter. If \(E_0\) is not zero, the increment and absolute plateau are not numerically equal. If \(E_0=0\), they happen to have the same numerical value, which can conceal the distinction.

What the Parameters Mean at Key Concentrations

At zero concentration, the fraction \(C/(EC_{50}+C)\) is zero, so \(E(0)=E_0\). At \(C=EC_{50}\), the fraction is \(1/2\). Consequently, the response is halfway from baseline to plateau. As \(C\) grows without bound, the fraction approaches 1 and the response approaches the upper plateau. These statements apply to both equations; only the label attached to the plateau-related parameter differs.

The middle expression is the arithmetic midpoint between baseline and the upper plateau. It equals half the absolute plateau only when the baseline is zero. For example, if baseline is 10 response units and the absolute plateau is 50 units, the response at \(EC_{50}\) is 30 units—not 25. This is why the phrase “half-maximal concentration” should be interpreted as half of the maximum drug-related change, with the baseline specified.

Parameterization and Response Scale

A parameter’s meaning also depends on the response scale and units. Both the baseline and the absolute plateau are measured in response units; the baseline-relative maximum is measured in the same units because it is a difference between responses. By contrast, \(EC_{50}\) has concentration units. The concentration fraction \(C/(EC_{50}+C)\) is dimensionless, so the term added to \(E_0\) has response units.

These parameterizations are mathematically equivalent when converted consistently. The baseline-relative form often makes the size of the drug-related change easy to communicate. The absolute-plateau form can be convenient when the scientific question concerns a response ceiling or when a reference range is expressed in absolute response units. Neither form changes the underlying curve, and neither automatically makes the parameters easier to estimate from limited data.

Assumptions, Diagnostics, and Common Errors

The reparameterization is an algebraic identity, but interpreting either model still requires assumptions about the response relationship. The model treats \(E_0\) as the response at zero relevant concentration and assumes a monotone approach to a finite upper plateau for a positive drug-related increase. It also treats \(EC_{50}\) as a positive concentration on the same scale as \(C\). A concentration value and its \(EC_{50}\) must use compatible units. The model does not account for a delayed effect, concentration-history dependence, or a changing baseline; such features require additional modeling rather than a different label for \(E_{\max}\).

When checking a fitted or reported model, verify that the zero-concentration prediction equals \(E_0\), that the limiting response agrees with the stated plateau, and that the response at \(EC_{50}\) is halfway between baseline and plateau. Check the reported units and whether \(E_{\max}\) denotes an increment or an absolute response. A negative inferred increment in a model presented as stimulatory, a plateau below baseline, or predictions that contradict the stated direction of effect should prompt review of the model specification and parameter labels.

A frequent mistake is to call \(E_{\max}\) the absolute maximum response in one sentence, then substitute it as an increment in the equation \(E_0+E_{\max}C/(EC_{50}+C)\). For nonzero \(E_0\), that mixes conventions and yields a plateau of \(E_0+E_{\max}\), not \(E_{\max}\). The correction is to state the equation and convert first: if \(E_{\max}\) is an absolute plateau, use \(E_{\max}-E_0\) as the coefficient of the concentration fraction.

Another mistake is to interpret \(EC_{50}\) as the concentration giving an absolute response of \(E_{\max}/2\). That only follows when baseline is zero. In the general case, the absolute response at \(EC_{50}\) is \((E_0+E_{\max,\mathrm{abs}})/2\). Finally, a good numerical fit does not by itself prove that the baseline, plateau, or \(EC_{50}\) is well determined: if observations do not represent the baseline or approach the plateau, the data may provide limited information about those quantities. Distinguishing parameter meanings prevents a labeling error, but it cannot resolve inadequate information in the concentration range.