Distribution and Terminal Phases in Two-Compartment Profiles

Why a Two-Compartment Curve Has More Than One Apparent Phase

After an intravenous bolus, a two-compartment model can produce a concentration curve that falls steeply at first and then declines more slowly. The early steep segment is commonly called the distribution phase; the later, shallower segment is called the terminal phase. These labels describe features of the observed concentration-time profile. They do not mean that one phase is caused only by drug “moving into tissues” while the other is caused only by drug “being eliminated.”

The central and peripheral compartments are mathematical representations of drug amounts in connected spaces. Concentration measured in the central compartment reflects the combined effects of drug leaving and returning to that compartment, as well as elimination from it. The earlier material on two-compartment eigenvalues and macroparameters expresses this behavior as a sum of exponential terms. Understanding the phases means interpreting that sum over time, rather than assigning each exponential to one isolated anatomical event.

Definition: For a two-compartment IV bolus profile written as \(C(t)=A e^{-\alpha t}+B e^{-\beta t}\), with \(A,B>0\) and \(\alpha>\beta>0\), the distribution phase is the early period in which the faster-decaying term, \(A e^{-\alpha t}\), makes an important contribution to the observed concentration decline. The terminal phase is the late period in which the slower-decaying term, \(B e^{-\beta t}\), dominates the observed concentration. The terms are descriptions of the curve’s behavior, not names for separate anatomical mechanisms.

Here, \(t\) is time, \(C(t)\) is central concentration, and \(A\) and \(B\) are the amplitudes of the fast and slow exponential terms. The rate constants \(\alpha\) and \(\beta\), with units of inverse time, are macro-rate constants: they describe the decay rates of the curve’s exponential components. They are not, in general, identical to individual transfer or elimination microconstants.

Why the Dominant Contribution Changes

The fast component decreases more quickly because \(\alpha>\beta\). The ratio of fast to slow contributions at time \(t\) is

\frac{A e^{-\alpha t}}{B e^{-\beta t}} =\frac{A}{B}e^{-(\alpha-\beta)t}.

Since \(\alpha-\beta>0\), this ratio decreases as time increases. Thus, even if the fast term is substantial initially, it eventually becomes small relative to the slow term. When the two contributions are equal, their crossover time is

t_{\mathrm{cross}}=\frac{\ln(A/B)}{\alpha-\beta},

provided \(A>B\), so that the fast contribution is initially larger and the crossover occurs at a positive time. This is a useful description of the relative component sizes, not a universal boundary between two sharply separated biological events. In real data, the transition may be gradual, and the sampling schedule determines whether it is visible.

Worked Example: Finding When the Slow Term Takes Over

Suppose a hypothetical IV bolus profile is described by \(C(t)=12e^{-1.2t}+3e^{-0.12t}\) mg/L, where \(t\) is in hours. The amplitudes are in mg/L and the rate constants are in h\(^{-1}\). Compare the contributions at selected times and find their crossover.

At \(t=0\), the fast contribution is \(12e^0=12\) mg/L and the slow contribution is \(3e^0=3\) mg/L, so \(C(0)=15\) mg/L.

At \(t=1\) h, the fast contribution is \(12e^{-1.2}=3.614\) mg/L and the slow contribution is \(3e^{-0.12}=2.661\) mg/L. Therefore, \(C(1)=3.614+2.661=6.275\) mg/L, or 6.28 mg/L to three significant figures. Both terms still contribute appreciably.

At \(t=8\) h, the fast contribution is \(12e^{-9.6}=0.0008128\) mg/L and the slow contribution is \(3e^{-0.96}=1.1487\) mg/L. Their sum is approximately 1.1495 mg/L, which is 1.15 mg/L to three significant figures. The fast term is now negligible in comparison.

For the crossover, set the contributions equal: \(12e^{-1.2t}=3e^{-0.12t}\). Rearranging gives \(4=e^{1.08t}\), so \(t=\ln(4)/1.08=1.283\) h. The fast contribution is larger before this time and the slow contribution is larger afterward. Neither contribution should be interpreted as the amount in one particular anatomical compartment.

The Observed Log-Slope Is a Changing Combination

A straight line on a plot of log concentration against time has a constant negative slope and corresponds to a single exponential. A biexponential curve is generally not a straight line on that scale across its full time course. Its local log-slope can be found by differentiating the concentration and dividing by the concentration:

-\frac{d\ln C(t)}{dt} =-\frac{C'(t)}{C(t)} =\frac{\alpha A e^{-\alpha t}+\beta B e^{-\beta t}} {A e^{-\alpha t}+B e^{-\beta t}}.

The numerator weights each decay rate by that component’s current concentration contribution. Because the contributions are positive, the local rate lies between \(\beta\) and \(\alpha\). As the fast contribution fades, the local rate approaches \(\beta\); the curve therefore becomes shallower on a log-concentration scale. The terminal slope is not an average of the two rates taken once for all time: its value emerges from the slow component dominating at late times.

Result: For \(A,B>0\) and \(\alpha>\beta>0\), the instantaneous decline rate satisfies \(\beta< -d\ln C(t)/dt<\alpha\) at finite times. As \(t\) becomes large, \(A e^{-\alpha t}\) becomes negligible relative to \(B e^{-\beta t}\), and \(-d\ln C(t)/dt\) approaches \(\beta\). The terminal half-life calculated from the terminal slope is \(\ln(2)/\beta\); it is a property of the terminal concentration decline, not necessarily the half-life of a single physical compartment.

This result follows directly by taking the late-time limit. Divide the numerator and denominator of the slope expression by \(B e^{-\beta t}\). The ratio \(A e^{-\alpha t}/(B e^{-\beta t})=(A/B)e^{-(\alpha-\beta)t}\) tends to zero because \(\alpha-\beta>0\). The remaining numerator and denominator are both \(\beta\) and 1, respectively, so the slope tends to \(\beta\). The same argument gives \(C(t)\approx B e^{-\beta t}\) at sufficiently late times.

Worked Example: Interpreting a Terminal Slope

Consider the hypothetical profile \(C(t)=10e^{-1.0t}+2e^{-0.10t}\) mg/L, with time in hours. Calculate its initial local decline rate, its local rate at 2 h, and its terminal half-life.

At \(t=0\), the component contributions are 10 and 2 mg/L. The local rate is \((1.0\times10+0.10\times2)/(10+2)=10.2/12=0.850\) h\(^{-1}\). The initial observed decline is therefore not simply the fast component’s rate of 1.0 h\(^{-1}\).

At \(t=2\) h, the fast contribution is \(10e^{-2}=1.353\) mg/L and the slow contribution is \(2e^{-0.2}=1.637\) mg/L. The concentration is \(1.353+1.637=2.991\) mg/L. The local rate is \((1.0\times1.353+0.10\times1.637)/2.991=0.5073\) h\(^{-1}\), to four significant figures. The rate has decreased because the slower component has become relatively more important.

At late times the slope approaches \(\beta=0.10\) h\(^{-1}\). The corresponding terminal half-life is \(\ln(2)/0.10=6.931\) h. This half-life summarizes the late concentration decline under the model; it does not show that drug is undergoing a single process with a 6.931-hour half-life.

Connection to the Two-Compartment Structure

For the standard linear two-compartment model with elimination from the central compartment, the microconstants are \(k_{10}=CL/V_c\), \(k_{12}=Q/V_c\), and \(k_{21}=Q/V_p\). Here \(CL\) is clearance, \(Q\) is intercompartmental clearance, and \(V_c\) and \(V_p\) are the central and peripheral volumes. The two positive macro-rate constants are eigenvalues of the amount-balance system. In the standard notation their sum and product are

\alpha+\beta=k_{10}+k_{12}+k_{21}, \qquad \alpha\beta=k_{10}k_{21}.

These relationships connect the curve’s decay rates to the complete model structure. Both eigenvalues depend on the system’s elimination and exchange parameters. Consequently, it is generally incorrect to equate \(\alpha\) with “distribution only” or \(\beta\) with “elimination only.” Exchange continues to matter for the eigenvalues, and elimination from the central compartment can affect the overall distribution of drug between compartments.

Worked Example: Reading the Eigenvalues Without Assigning One Process to Each

Suppose a hypothetical model has \(CL=6\) L/h, \(Q=4\) L/h, \(V_c=20\) L, and \(V_p=40\) L. Calculate the microconstants and macro-rate constants, then interpret the faster and slower rates.

The microconstants are \(k_{10}=6/20=0.300\) h\(^{-1}\), \(k_{12}=4/20=0.200\) h\(^{-1}\), and \(k_{21}=4/40=0.100\) h\(^{-1}\). Thus \(\alpha+\beta=0.300+0.200+0.100=0.600\) h\(^{-1}\), while \(\alpha\beta=(0.300)(0.100)=0.0300\) h\(^{-2}\).

The two rates are the roots of \(r^2-0.600r+0.0300=0\). Using the quadratic formula, \(r=[0.600\pm\sqrt{0.600^2-4(0.0300)}]/2=[0.600\pm\sqrt{0.240}]/2\). Therefore, \(\alpha=0.5449\) h\(^{-1}\) and \(\beta=0.05505\) h\(^{-1}\), to four significant figures.

The larger rate is associated with the faster-decaying exponential contribution, and the smaller rate with the slower terminal contribution. Both were calculated from the combined values of elimination and exchange. The calculation does not identify either eigenvalue as a rate for one anatomical movement or one compartment alone.

Assumptions, Diagnostics, and Alternatives

The phase interpretation assumes that the concentration profile is adequately represented by a linear, time-invariant two-compartment model over the interval being interpreted. In the examples, the dose is an IV bolus, elimination is first order from the central compartment, transfer between compartments is linear, and the parameters are constant over time. The positive-amplitude biexponential form also assumes two decaying terms with \(A,B>0\) and \(\alpha>\beta>0\). If the route includes absorption, if kinetics are nonlinear, or if the data contain substantial model misspecification, the same phase labels may not have the same interpretation.

A practical diagnostic is to examine the concentration-time profile on both ordinary and log scales, together with the sampling times and measurement variability. Curvature in a log-concentration plot over the early period is consistent with more than one exponential contribution, but it does not by itself prove that the two-compartment model is correct. A few late observations may appear nearly linear because the fast component has faded; that apparent line cannot establish that early samples also follow a single exponential. Conversely, sparse early samples may miss the fast change and make the profile look almost monoexponential.

A one-compartment exponential is a simpler alternative when one exponential adequately describes the observed time course for the intended purpose. It has one constant log-slope and does not represent a distinct fast and slow phase. A two-compartment model can describe curvature and the transition in slopes, but adds parameters and may be difficult to estimate if samples do not cover both phases. The observed phases are useful summaries of profile shape; the underlying compartment model is a mechanistic representation with its own assumptions, not a direct anatomical map.

Common pitfall: “The distribution phase is tissue uptake, and the terminal phase is elimination from the body.” This is too literal. The correction is that the phases describe the relative dominance of exponential components in central concentration. Both eigenvalues reflect the coupled exchange-and-elimination system, and the terminal decline can be influenced by drug returning from the peripheral compartment as well as by elimination.

A related mistake is to treat the terminal half-life as the time required for the total amount of drug in the body to halve. The terminal half-life is calculated from the late concentration slope, \(\ln(2)/\beta\). It describes the late concentration profile under the model; the amount distribution across compartments can change while that concentration declines. Interpretation should therefore name the quantity being described—central concentration, a curve component, or a model parameter—and retain its units.

Check Your Understanding

Answer each question, then compare with the solutions below.

For \(C(t)=6e^{-0.8t}+2e^{-0.08t}\) mg/L, which term dominates at sufficiently late times, and what is the terminal half-life?
For \(C(t)=4e^{-0.5t}+1e^{-0.1t}\), write the fast-to-slow contribution ratio as a function of time. Does it increase or decrease over time?
Why is it generally incorrect to call \(\alpha\) the distribution rate constant and \(\beta\) the elimination rate constant?
A two-compartment profile looks linear on a log scale in late samples. Does that establish that the entire profile is monoexponential? Explain.

Solutions. (1) The slower term \(2e^{-0.08t}\) dominates at late times because \(0.8>0.08\). Its terminal half-life is \(\ln(2)/0.08=8.664\) time units; if \(t\) is measured in hours, this is 8.664 h. (2) The ratio is \((4/1)e^{-(0.5-0.1)t}=4e^{-0.4t}\). It decreases because the exponent becomes more negative as \(t\) increases. (3) Each eigenvalue depends on the coupled exchange and elimination parameters, so neither is generally a rate for one isolated anatomical process. (4) No. The fast component may be negligible in late samples, leaving an approximately straight terminal segment even though the earlier profile is biexponential.