Tutorials › Pharmacometrics › Noncompartmental Analysis: MRT and Terminal Half-Life
Pharmacokinetics · Noncompartmental Analysis

Noncompartmental Analysis: MRT and Terminal Half-Life

Learn how noncompartmental analysis turns concentration-time data into AUC, AUMC, mean residence time, terminal elimination rate constant, and terminal half-life—and understand why MRT and terminal half-life describe different aspects of pharmacokinetic behavior.

Intermediate Noncompartmental Analysis MRT Terminal Half-Life
01 · The big picture

1. What Is Noncompartmental Analysis?

Noncompartmental analysis (NCA) is a pharmacokinetic approach that summarizes concentration-time data without requiring the analyst to specify a complete compartmental structural model.

Instead of assuming that the body consists of one or more kinetically defined compartments, NCA derives quantities directly from the observed concentration-time profile. Important quantities include Cmax, Tmax, AUC, AUMC, clearance, volume measures, MRT, and the terminal elimination rate constant λz.

Concentration vs. time NCA AUC · AUMC λz · t1/2 MRT · CL other PK summaries PK parameters Observed data → numerical summaries → interpretable PK quantities

NCA works primarily from the observed concentration-time profile and numerical integration rather than from a fully specified compartmental model.

Core idea: NCA is model-light, not assumption-free. It avoids specifying a complete compartmental structure, but it still depends on assumptions about sampling, terminal-phase behavior, integration, and the relationship between observed concentrations and the quantity being estimated.
02 · What NCA asks

2. What Questions Can NCA Answer?

NCA is particularly useful when the primary goal is to summarize exposure and time-related PK characteristics rather than to build a mechanistic model of drug disposition.

Question NCA quantity Interpretation
How much systemic exposure occurred? AUC Area under the concentration-time curve
How much concentration-time exposure is weighted toward later times? AUMC Area under the first moment curve
How long does drug tend to remain in the system? MRT Mean residence time derived from AUMC/AUC
What is the slope of the terminal log-linear phase? λz Apparent terminal elimination rate constant
What is the corresponding terminal time scale? t1/2 Time associated with a 50% reduction during the terminal phase
How efficiently is drug cleared? CL Dose divided by AUC0-∞ for appropriate IV dosing conditions

These quantities are related, but they are not interchangeable. In particular, MRT and terminal half-life answer different questions. MRT is a moment-based summary of the concentration-time profile, whereas terminal half-life is determined from the slope of the terminal log-linear phase.

03 · Exposure

3. AUC: The Foundation of NCA

The area under the concentration-time curve (AUC) summarizes systemic exposure over a specified time interval.

For observations from time 0 through the last quantifiable concentration, the observed area is commonly written as:

\[ AUC_{0-t_{\mathrm{last}}} = \int_0^{t_{\mathrm{last}}} C(t)\,dt \]

Because concentration is usually observed only at discrete sampling times, the observed portion of the AUC is generally estimated numerically. A simple linear trapezoidal segment between two observations is:

\[ AUC_i = \frac{C_i+C_{i+1}}{2}(t_{i+1}-t_i) \]

Summing the individual trapezoids gives the AUC through the last measurable time point.

Important: AUC is an exposure measure. It is not itself a rate constant, half-life, or residence time. Those quantities require additional information or transformations of the concentration-time profile.
04 · First moment

4. AUMC: The Area Under the First Moment Curve

The area under the first moment curve (AUMC) weights concentration by time. Instead of integrating \(C(t)\), it integrates \(tC(t)\):

\[ AUMC_{0-t} = \int_0^t \tau C(\tau)\,d\tau \]

The extra factor of time gives later concentrations greater weight than earlier concentrations. This is the key reason AUMC contains information about time-related behavior that AUC alone does not capture.

For discrete observations, the linear trapezoidal contribution can be written:

\[ AUMC_i = \frac{t_iC_i+t_{i+1}C_{i+1}}{2} (t_{i+1}-t_i) \]

Summing these contributions provides \(AUMC_{0-t_{\mathrm{last}}}\). An extrapolated AUMC can then be calculated when the terminal phase is characterized sufficiently well.

Time Value C(t) tC(t) AUMC weights concentration according to time

The first moment curve is \(tC(t)\). Because later observations are multiplied by larger values of \(t\), AUMC emphasizes the time distribution of exposure.

05 · Mean residence time

5. What Is Mean Residence Time?

Mean residence time (MRT) is a moment-based summary describing the average residence time associated with the drug amount represented by the concentration-time profile.

For a profile that can be appropriately extrapolated to infinity:

\[ MRT_{0-\infty} = \frac{AUMC_{0-\infty}}{AUC_{0-\infty}} \]

The units make the interpretation intuitive. AUMC has concentration × time² units, whereas AUC has concentration × time units. Their ratio therefore has units of time.

MRT is not simply another name for half-life. MRT is calculated from the entire concentration-time profile through the AUMC/AUC relationship. Terminal half-life is calculated from the slope of the terminal phase. The two quantities can therefore be quite different.

For an IV bolus dose in a simple one-compartment model with first-order elimination, MRT has a particularly simple relationship with the elimination rate constant:

\[ MRT_{IV} = \frac{1}{k} \]

Because terminal half-life in that same model is \(0.693/k\), the relationship becomes:

\[ t_{1/2} = 0.693\,MRT_{IV} \]

This relationship is model-specific. It should not be used as a universal conversion between MRT and terminal half-life, especially for multi-compartment systems.

06 · Terminal phase

6. What Is the Terminal Elimination Rate Constant?

The terminal elimination rate constant, \(\lambda_z\), describes the slope of the apparent terminal log-linear portion of the concentration-time profile.

If the terminal phase follows:

\[ C(t)=C_{\mathrm{last}}e^{-\lambda_z(t-t_{\mathrm{last}})} \]

then taking logarithms gives:

\[ \ln C(t) = \ln C_{\mathrm{last}} - \lambda_z(t-t_{\mathrm{last}}) \]

Therefore, a regression of \(\ln C\) against time during the terminal phase has slope approximately equal to \(-\lambda_z\).

distribution / earlier phases terminal log-linear phase Time ln C

On a semi-log concentration-time plot, the terminal phase appears approximately linear. Its negative slope provides the estimate of \(\lambda_z\).

07 · Terminal half-life

7. How Is Terminal Half-Life Calculated?

Once \(\lambda_z\) has been estimated, the apparent terminal half-life is:

\[ t_{1/2,z} = \frac{\ln(2)}{\lambda_z} = \frac{0.693}{\lambda_z} \]

For example, if:

\[ \lambda_z=0.10\ \mathrm{h}^{-1} \]

then:

\[ t_{1/2,z} = \frac{0.693}{0.10} = 6.93\ \mathrm{h} \]

The subscript \(z\) is important. This is the terminal half-life, not necessarily the half-life associated with every phase of a multicompartment concentration-time profile.

Interpretation: terminal half-life describes the time scale of the terminal decline. It does not necessarily describe the time required for the entire drug burden in the body to fall by 50%.
08 · Extrapolation

8. Extending AUC and AUMC to Infinity

NCA frequently reports both an observed quantity through the last measurable concentration and an extrapolated quantity extending to infinite time.

For AUC:

\[ AUC_{0-\infty} = AUC_{0-t_{\mathrm{last}}} + \frac{C_{\mathrm{last}}}{\lambda_z} \]

The second term represents the estimated residual area after the last observation under the assumed terminal exponential decline.

For AUMC, the corresponding extrapolation is:

\[ AUMC_{0-\infty} = AUMC_{0-t_{\mathrm{last}}} + \frac{t_{\mathrm{last}}C_{\mathrm{last}}}{\lambda_z} + \frac{C_{\mathrm{last}}}{\lambda_z^2} \]

The final term becomes especially important because AUMC weights later time points by time. Consequently, uncertainty in the terminal phase can have a substantial influence on MRT.

Why sampling matters: if the terminal phase is poorly characterized, both the extrapolated AUC and AUMC can become highly dependent on the estimated \(\lambda_z\). This is one reason adequate late sampling is important when terminal PK parameters are of interest.
09 · Data quality

9. Percentage of AUC Extrapolated

A useful diagnostic is the fraction of the total AUC that is contributed by the extrapolated terminal portion:

\[ \%AUC_{\mathrm{extrap}} = 100 \times \frac{AUC_{0-\infty}-AUC_{0-t_{\mathrm{last}}}} {AUC_{0-\infty}} \]

A large extrapolated fraction means that a substantial part of the reported AUC depends on the terminal-phase model rather than directly observed concentrations.

Situation Potential implication
Small extrapolated fraction Most of AUC0-∞ is supported by observed concentrations.
Moderate extrapolated fraction Terminal-phase estimation contributes meaningfully to total exposure.
Large extrapolated fraction AUC0-∞, MRT, and related quantities may be strongly dependent on terminal-phase assumptions.

There is no universal percentage that automatically determines whether a particular analysis is valid. Interpretation depends on the drug, study design, sampling schedule, regulatory context, and prespecified analysis plan.

10 · Choosing λz

10. How Is the Terminal Phase Selected?

Estimating \(\lambda_z\) is one of the most consequential steps in NCA. The terminal phase should represent the portion of the profile that is reasonably consistent with the apparent terminal log-linear decline.

A typical workflow is:

  1. Plot the concentration-time data on a semi-log scale. This helps reveal whether a terminal log-linear segment is present.
  2. Identify candidate late-phase observations. The selected observations should be consistent with a plausible terminal decline.
  3. Fit a linear regression to log concentration versus time. The negative slope provides the estimate of \(\lambda_z\).
  4. Evaluate the selected points. Consider the number of points, visual linearity, residual behavior, the estimated slope, and the scientific plausibility of the resulting terminal phase.
  5. Assess sensitivity. Where appropriate, examine whether reasonable changes in terminal-point selection materially change \(\lambda_z\), half-life, AUC0-∞, or MRT.
Do not automatically choose the last three observations. Three or more observations are commonly needed for a terminal regression, but the correct terminal set depends on the actual concentration-time profile. The last observations are not necessarily the terminal phase simply because they occur last in time.
11 · The distinction

11. MRT vs. Terminal Half-Life

The most important conceptual distinction in this tutorial is that MRT and terminal half-life summarize different mathematical features of the PK profile.

Feature MRT Terminal half-life
Primary calculation AUMC / AUC 0.693 / λz
Information used Whole concentration-time profile, including extrapolation when applicable Selected terminal log-linear observations
Primary interpretation Mean residence-time summary Terminal decline time scale
Sensitive to terminal slope? Yes, especially for extrapolated AUMC Directly
Equal in general? No
Simple one-compartment IV case 1/k 0.693/k

In a simple one-compartment IV model, MRT and half-life are directly related. But in a multi-compartment system, the concentration profile can contain a distribution phase followed by a slower terminal phase. In that setting, MRT incorporates information from the broader profile, whereas terminal half-life focuses on the terminal slope.

12 · Route matters

12. MRT After IV and Extravascular Administration

The interpretation of MRT depends partly on the route of administration. After an IV dose, the observed profile begins with drug already in the systemic circulation. After an oral, subcutaneous, intramuscular, or other extravascular dose, the profile also contains information about drug input.

Under appropriate linear PK assumptions, mean absorption time can be related to the difference between extravascular and IV MRT:

\[ MAT = MRT_{\mathrm{extravascular}} - MRT_{IV} \]

This relationship illustrates an important principle: an extravascular MRT can reflect both disposition and the timing of drug absorption.

Practical implication: MRT after oral administration should not automatically be interpreted as a pure measure of elimination or systemic drug persistence. The absorption process can contribute to the observed residence-time summary.
13 · Related PK parameters

13. Clearance, Vz, and Vss From NCA

NCA can also provide useful volume and clearance measures. For an appropriate IV dose:

\[ CL = \frac{Dose}{AUC_{0-\infty}} \]

A terminal-phase volume measure can be calculated as:

\[ V_z = \frac{CL}{\lambda_z} \]

For IV administration under conditions where the usual moment relationships apply, the steady-state volume is related to MRT and clearance:

\[ V_{ss} = CL\times MRT_{IV} \]

These quantities have different interpretations. \(V_z\) is linked to the terminal phase, whereas \(V_{ss}\) is a moment-based distribution volume. Neither should automatically be interpreted as a literal anatomical volume.

14 · Worked example

14. Worked Example: Calculating MRT and Terminal Half-Life

Consider a hypothetical IV bolus study. Concentrations are measured in mg/L and time is measured in hours. The observed concentrations are:

Time (h) Concentration (mg/L)
020.000
111.264
27.122
43.963
62.868
82.272
121.507
161.010
240.454

Step 1: Calculate AUC through the last concentration

Using the linear trapezoidal rule over the observed interval gives:

\[ AUC_{0-24} \approx 66.323\ \mathrm{mg\cdot h/L} \]

Step 2: Calculate AUMC through the last concentration

Applying the same numerical integration principle to \(tC(t)\) gives:

\[ AUMC_{0-24} \approx 366.057\ \mathrm{mg\cdot h^2/L} \]

Step 3: Estimate λz

Suppose the observations from 8 through 24 hours are judged to represent the terminal log-linear phase. Regressing \(\ln C\) against time gives approximately:

\[ \lambda_z \approx 0.1006\ \mathrm{h^{-1}} \]

Step 4: Calculate terminal half-life

\[ t_{1/2,z} = \frac{0.693}{0.1006} \approx 6.89\ \mathrm{h} \]

Step 5: Calculate the extrapolated AUC

The last concentration is approximately \(0.454\) mg/L at 24 hours. Therefore:

\[ AUC_{0-\infty} = 66.323+ \frac{0.454}{0.1006} \approx 70.833\ \mathrm{mg\cdot h/L} \]

Step 6: Calculate the extrapolated AUMC

\[ AUMC_{0-\infty} = 366.057 + \frac{(24)(0.454)}{0.1006} + \frac{0.454}{(0.1006)^2} \] $$ AUMC_{0-\infty} \approx 519.16\ \mathrm{mg\cdot h^2/L} $$

Step 7: Calculate MRT

\[ MRT_{0-\infty} = \frac{519.16}{70.833} \approx 7.33\ \mathrm{h} \]

Step 8: Examine AUC extrapolation

\[ \%AUC_{\mathrm{extrap}} = 100 \times \frac{70.833-66.323}{70.833} \approx 6.37\% \]

The main numerical results are therefore:

Parameter Approximate result
AUC0-24 66.323 mg·h/L
AUC0-∞ 70.833 mg·h/L
AUMC0-24 366.057 mg·h²/L
AUMC0-∞ 519.16 mg·h²/L
λz 0.1006 h⁻¹
Terminal half-life 6.89 h
MRT0-∞ 7.33 h
AUC extrapolated 6.37%
The key observation: terminal half-life is approximately 6.89 hours, while MRT is approximately 7.33 hours. They are close in this example, but they are not the same quantity and should not be treated as interchangeable.
15 · Why they differ

15. Why Can MRT and Terminal Half-Life Be Different?

Suppose a drug follows a two-compartment concentration-time profile:

\[ C(t) = Ae^{-\alpha t} + Be^{-\beta t} \qquad \alpha>\beta \]

The early part of the profile can be dominated by the faster distribution process associated with \(\alpha\), while the later portion is dominated by the slower terminal process associated with \(\beta\).

The terminal half-life is then:

\[ t_{1/2,z} = \frac{0.693}{\beta} \]

But MRT is calculated from the integrated profile:

\[ MRT = \frac{AUMC}{AUC} \]

Therefore, MRT incorporates the temporal distribution of exposure across the profile rather than focusing exclusively on the final slope.

Conceptual shortcut: think of terminal half-life as a slope-based quantity and MRT as a moment-based quantity.
16 · Common mistakes

16. Common NCA Mistakes

1. Treating MRT as another name for half-life

MRT and terminal half-life are mathematically different quantities. They can be related under particular models, but one should not be substituted for the other.

2. Choosing the last three points automatically

The last observations may not represent the terminal elimination phase. Terminal-point selection should be based on the observed profile and a prespecified or appropriately documented analysis procedure.

3. Ignoring extrapolation

AUC0-∞ and MRT0-∞ can depend substantially on the terminal extrapolation when sampling ends too early.

4. Assuming a high R² proves the terminal phase is correct

A strong linear relationship on a semi-log plot is useful evidence, but terminal-phase selection also requires scientific and pharmacokinetic judgment. A mathematically good regression can still represent the wrong portion of the profile.

5. Treating Vz as a literal anatomical volume

NCA volume measures are apparent quantities derived from PK relationships. They do not necessarily correspond to a physical space in the body.

6. Ignoring the route of administration

For extravascular administration, absorption contributes to the observed concentration-time profile and can therefore affect MRT.

7. Extrapolating beyond what the data support

If late concentrations are sparse, noisy, or below reliable quantification, the estimated terminal slope may be unstable. In that setting, terminal half-life and extrapolated exposure can become highly uncertain.

17 · Practical workflow

17. A Practical NCA Workflow

  1. Inspect the concentration-time data. Plot the observed concentrations against time using actual sampling times.
  2. Identify quantifiable observations. Apply the study's prespecified handling rules for observations below the lower limit of quantification.
  3. Calculate observed exposure. Estimate AUC and AUMC through the last appropriate quantifiable concentration.
  4. Identify the terminal phase. Use the semi-log concentration-time profile and an appropriate documented selection procedure.
  5. Estimate λz. Fit the terminal log-linear regression and evaluate the resulting estimate.
  6. Calculate terminal half-life. Use \(t_{1/2,z}=0.693/\lambda_z\).
  7. Calculate extrapolated AUC and AUMC. Extend the observed quantities using the terminal slope where appropriate.
  8. Calculate MRT. Use \(MRT=AUMC/AUC\), with the appropriate time limits.
  9. Inspect extrapolation and sensitivity. Determine how much of AUC and AUMC depends on extrapolation and whether reasonable terminal-phase choices materially change the results.
  10. Interpret within the study design. Consider route, dose, sampling schedule, assay limitations, nonlinear PK, and whether the assumptions underlying the NCA quantities are appropriate.

NCA is often straightforward computationally, but the quality of the result depends heavily on the quality and timing of the concentration measurements. The calculations themselves are rarely the most difficult part; identifying what the data actually support is often more important.

18. Key Takeaways

  • Noncompartmental analysis summarizes pharmacokinetic behavior without requiring a complete compartmental structural model.
  • AUC summarizes concentration-time exposure, while AUMC weights that exposure by time.
  • MRT is calculated from \(AUMC/AUC\) and provides a moment-based summary of residence time.
  • λz is the apparent terminal elimination rate constant estimated from the terminal log-linear portion of the concentration-time profile.
  • Terminal half-life is calculated as \(0.693/\lambda_z\).
  • MRT and terminal half-life are not interchangeable. MRT is moment-based; terminal half-life is slope-based.
  • In a simple one-compartment IV model, \(MRT=1/k\) and \(t_{1/2}=0.693/k\), but this relationship should not be generalized to complex multicompartment profiles.
  • For extravascular administration, MRT can include the influence of drug absorption, not just systemic disposition.
  • AUC0-∞ and MRT0-∞ depend on terminal extrapolation when sampling does not continue to infinity.
  • Terminal-phase selection is therefore a critical part of NCA and should not be reduced to automatically selecting the last few observations.
  • Adequate late sampling is particularly important when terminal half-life, AUC0-∞, or MRT0-∞ are important study endpoints.
References

19. References

  1. European Medicines Agency. Pharmacokinetic studies in man. Scientific guideline. EMA/CHMP/225895/2006 Rev. 1. EMA.
  2. European Medicines Agency. Clinical pharmacology and pharmacokinetics: Questions and answers. EMA.
  3. Food and Drug Administration. Pharmacokinetic report illustrating noncompartmental PK quantities including AUC, λz, terminal half-life, AUMC, MRT, clearance, and volume measures. FDA.
  4. Karol MD. Mean residence time and the meaning of AUMC/AUC. Biopharmaceutics & Drug Disposition. 1990;11:179–181. doi:10.1002/bdd.2510110210
  5. Kasuya Y, Hirayama H, Kubota N, Pang KS. Interpretation and estimates of mean residence time with statistical moment theory. Biopharmaceutics & Drug Disposition. 1987;8:223–234. PubMed.
  6. Cheung BWY, Cartier LL, Russlie HQ, Sawchuk RJ. The application of sample pooling methods for determining AUC, AUMC and mean residence times in pharmacokinetic studies. Fundamental & Clinical Pharmacology. 2005;19:347–354. doi:10.1111/j.1472-8206.2005.00329.x
Methodology note: NCA calculations can vary according to the specified integration rule, handling of concentrations below quantification, terminal-point selection algorithm, and software implementation. For regulated analyses, the analysis method and decision rules should be prespecified and documented.
Next step

Where to Go Next

A natural next step is to study AUC and AUMC calculation in detail, including linear and log-linear trapezoidal methods, terminal extrapolation, partial AUCs, and the effect of sampling schedules on NCA estimates.

From there, the concepts can be extended to clearance and volume estimation, repeated-dose NCA, steady-state parameters, bioavailability and bioequivalence, and comparisons between NCA and compartmental PK modeling.

```

Related Tutorials

Clinical Biostats

```
← Back to Pharmacokinetics Tutorials