Why Regional Luminal pH Matters
A drug moving through the gastrointestinal tract encounters a changing chemical environment. Luminal pH generally rises from the strongly acidic stomach into the small intestine, then may be lower again in parts of the colon. This broad pattern can help organize physiological inputs for a PBPK model, but it is not a fixed pH profile shared by every person or every measurement.
The regional compartments and anatomical boundaries established in the earlier material on regional GI compartments provide the locations for describing this pattern. Here the focus is the pH assigned to those locations: what ranges are reasonable at a broad physiological level, what a regional summary can and cannot represent, and how to preserve important variation when converting measurements into model inputs.
A Broad Stomach-to-Colon Pattern
At a broad level, the stomach is acidic, commonly with luminal pH in the approximate range of 1 to 3. After gastric contents enter the duodenum, pH is generally higher; broad representative ranges are about 5 to 6 in the duodenum, 6 to 7 in the jejunum, and 7 to 8 in the ileum. Colonic values are often described as mildly acidic to near neutral, approximately 5.5 to 7. These are orientation ranges, not strict boundaries or universal reference intervals.
The broad pattern is therefore a rise from stomach to small intestine, followed by a possible reduction in the colon. A single uninterrupted increase from stomach to colon would be an oversimplification. The ranges overlap, and a value near a boundary does not by itself establish that a measurement came from one particular region. Anatomical location, measurement method, and sampling conditions matter alongside the number.
A pH range describes the spread of values considered plausible in a broad context; it is not interchangeable with a mean. Nor does a range imply that every value within it is equally frequent. Regional measurements can differ between people, within a person over time, and within the same broad anatomical region. A report of a mean without the observed spread can hide that variation; a range without the number and distribution of observations can be equally difficult to interpret.
The ranges above support physiological orientation, not a universal set of fixed segment values. In particular, the stomach and colon each merit more detailed treatment of their own physiological conditions. For this tutorial, use the broad pattern to frame regional inputs without treating any listed range as a prescription for a particular subject or study condition.
Worked Example: Summarizing Hypothetical Regional Measurements
Suppose a hypothetical teaching data set contains three pH readings for each of five GI regions. These values are invented for illustration and are not presented as measurements from a real study.
| Region | Readings | Mean pH | Observed range |
|---|---|---|---|
| Stomach | 2.0, 2.5, 3.0 | 2.50 | 2.0–3.0 |
| Duodenum | 5.4, 5.8, 6.2 | 5.80 | 5.4–6.2 |
| Jejunum | 6.1, 6.5, 6.9 | 6.50 | 6.1–6.9 |
| Ileum | 6.7, 7.1, 7.5 | 7.10 | 6.7–7.5 |
| Colon | 5.6, 6.0, 6.4 | 6.00 | 5.6–6.4 |
For the duodenum, the arithmetic mean is \( (5.4+5.8+6.2)/3=17.4/3=5.80 \). The three readings span 0.8 pH units, from 5.4 to 6.2. The mean gives a concise central summary; the range shows that the three observations were not identical.
Across these hypothetical regional summaries, the mean rises from 2.50 in the stomach to 7.10 in the ileum and then is 6.00 in the colon. That is consistent with the broad pattern, but three readings per region cannot establish the full physiological distribution or the frequency of values. The table is an illustration of summarization, not a population reference profile.
Regional Means, Variation, and the Meaning of an Average
A reported mean is meaningful only when its observations and weighting are clear. A regional mean could summarize measurements across participants, repeated measurements in one participant, or measurements taken at different positions within a region. Those are different data structures and answer different descriptive questions. A useful summary therefore states the anatomical region, measurement context, number of observations, central estimate, and a measure of spread when available.
Because pH is logarithmic, averaging pH values is not the same as averaging hydrogen-ion activities on their original scale. For two equally weighted values, the arithmetic mean of pH corresponds to the negative logarithm of the geometric mean of the corresponding activities. It does not correspond to the negative logarithm of their arithmetic mean activity.
This distinction does not make an arithmetic mean of measured pH values invalid. It means the mean should be described as a mean on the pH scale, and the analyst should not silently reinterpret it as an average hydrogen-ion activity. For a model that uses pH to determine a nonlinear process, such as pH-dependent drug ionization, a single mean may also fail to reproduce the average process across a varying pH profile.
Worked Example: Comparing Two Ways to Average pH
Consider two hypothetical, equally weighted pH readings: 5.0 and 7.0. Their arithmetic mean on the pH scale is \( (5.0+7.0)/2=6.0 \). The corresponding relative hydrogen-ion activities are \(10^{-5}\) and \(10^{-7}\).
The arithmetic mean activity is \( (10^{-5}+10^{-7})/2=5.05\times10^{-6} \). Converting this arithmetic mean activity back to pH gives \( -\log_{10}(5.05\times10^{-6})\approx5.297 \), not 6.0.
By contrast, the geometric mean activity is \( \sqrt{10^{-5}\times10^{-7}}=\sqrt{10^{-12}}=10^{-6} \), whose pH is 6.0. Thus, the arithmetic mean of the two pH readings corresponds to the geometric mean activity in this equal-weight example. The two summaries differ because the pH transformation is logarithmic.
For a descriptive report of pH readings, reporting a mean of 6.0 can be appropriate if it is identified as the arithmetic mean on the pH scale. If the intended quantity is mean hydrogen-ion activity, calculate that on the activity scale instead. The appropriate summary depends on the question and on how the model uses the input.
Turning Measurements into Physiological Inputs
The purpose of a summary should determine how measurements are combined. If the model requires one representative pH per region, a regional mean or median may be a practical input, provided the chosen statistic matches the intended interpretation. The observed range or another spread measure should be retained as context rather than discarded. If the purpose is to represent variation, use multiple plausible values or a region-specific profile instead of presenting one central value as universal.
Weighting matters when observations represent unequal durations or unequal contributions to the quantity being summarized. For example, a simple mean of readings gives each reading equal weight, even if one reading represents a longer interval than another. A time-weighted descriptive mean should give each interval a weight proportional to its duration. That still summarizes pH on the pH scale; it does not automatically yield the average response of a pH-dependent drug process.
Worked Example: Choosing a Time-Weighted Regional Summary
Suppose a hypothetical proximal small-intestinal record assigns pH 5.8 to a one-hour interval, pH 6.2 to the next one-hour interval, and pH 7.2 to a two-hour interval. The question is the time-weighted arithmetic mean pH across the four hours, not a mean hydrogen-ion activity or a drug-response average.
The simple mean of the three recorded values would be \( (5.8+6.2+7.2)/3=19.2/3=6.40 \), giving each recorded value equal weight. But the 7.2 reading represents twice as long as either other reading. Using interval durations as weights gives
The time-weighted arithmetic mean is 6.60 pH units. The observed values still span 5.8 to 7.2, so the summary does not imply that pH remained at 6.60. If a model’s response changes nonlinearly with pH, representing the intervals separately may be more appropriate than using only 6.60.
A concise physiological input record can make this distinction explicit. It should identify the region and the value’s role (for example, representative central input or scenario value), state the statistic and units, and preserve the measurement context and uncertainty. Earlier material on parameter sourcing and provenance covers the broader importance of recording how an input was obtained and transformed.
| Input feature | What to record | Why it matters |
|---|---|---|
| Location | Named GI region and, if relevant, sampling position | Prevents different anatomical locations from being treated as interchangeable |
| Summary | Mean, median, individual values, or specified scenario | Clarifies what the model input represents |
| Spread | Observed range or another reported variability measure | Shows whether one value masks meaningful variation |
| Context | Measurement conditions and source-specific qualifiers | Helps assess whether the value applies to the intended model setting |
Diagnostics, Alternatives, and Limitations
Several checks help identify when a regional summary is being asked to do too much. First, check that the anatomical location is known and consistently defined. Second, inspect whether the reported statistic is a mean, median, range, or individual measurement. Third, compare the reported spread with the precision implied by the proposed input: specifying many decimal places does not remove biological or measurement variability. Finally, review whether a broad profile is physiologically plausible, including the general possibility of a reduction from small-intestinal to colonic pH.
If measurements within one nominal region differ substantially, do not automatically average them and move on. Differences may reflect real temporal or spatial variability, different sampling procedures, or incomparable measurement contexts. A summary is defensible only after considering what the observations represent. Likewise, a pH measurement should not be assumed to describe every local environment relevant to drug absorption; bulk luminal pH and conditions close to the mucosal surface need not be identical.
There are three common modeling choices. A single regional value is simple and can be useful when the model requires a compact reference input, but it suppresses variation. A regional range or set of scenarios preserves uncertainty for sensitivity or scenario analysis, but does not by itself specify how likely each value is. A time- or region-specific profile can preserve more structure, but requires suitable data and a model that can use that detail. None is universally best: choose based on the scientific question, the available evidence, and the role of pH in the model.
Practical Summary
Assign it to a clearly named GI region and retain relevant sampling context.
State the number of observations, chosen summary, and observed spread where available.
Use a central value for a representative input only when that role is justified; preserve variation when it could matter.
Clarify that pH is logarithmic and that a mean pH need not represent the response of a nonlinear pH-dependent process.
Check Your Understanding
Answer each question, then compare with the solutions below.
- What is the broad pattern of luminal pH from stomach through small intestine to colon, and why is it not accurately described as a continuous rise?
- Three hypothetical jejunal readings are 6.0, 6.4, and 6.8. Calculate their arithmetic mean and observed range. What does the mean not establish?
- Two equally weighted readings are pH 4 and pH 6. What is their arithmetic mean pH, and what activity does that mean correspond to geometrically?
- Why might a time-weighted mean differ from the simple mean when measurements represent unequal durations?
- Name two details, beyond a central pH value, that should be preserved when documenting a regional physiological input.
Solutions. (1) pH is generally low in the stomach, higher through the small intestine, and may be lower again in the colon; regional ranges overlap and values vary, so a continuous monotonic rise is an oversimplification. (2) The mean is \( (6.0+6.4+6.8)/3=19.2/3=6.40 \); the observed range is 6.0–6.8. The mean does not establish that pH was constant at 6.40 or describe the full population distribution. (3) The arithmetic mean pH is \( (4+6)/2=5.0 \). It corresponds to the geometric mean of activities \( \sqrt{10^{-4}\times10^{-6}}=10^{-5} \), not the arithmetic mean activity. (4) A simple mean weights each recorded value equally, whereas a time-weighted mean gives longer intervals more influence; the summaries answer different questions when durations differ. (5) Suitable details include the anatomical region, sampling context, number of observations, summary statistic, observed spread, and source-specific qualifiers.