Clinical Trial of Protein and Blood Pressure: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Clinical Trial of Protein and Blood Pressure (NCT00107744), a randomized na study. The primary outcome is Change From Baseline in Average Systolic Blood Pressure at 8 Weeks.

2. Clinical Question

3. Trial Design

The registry identifies this as a randomized crossover trial of dietary supplementation. Its title is “Clinical Trial of Protein and Blood Pressure,” and the listed sponsor is Tulane University. The trial is marked completed. The record gives a start of 2003-09 and a primary completion of 2008-04. The phase is listed as NA.

The intervention descriptions identify three treatment sequences: soy protein followed by milk protein and carbohydrate; milk protein followed by carbohydrate and soy protein; and carbohydrate followed by soy protein and milk protein. In each sequence, participants received 40 grams daily of each supplement for 8 weeks. These are sequence descriptions, not separate parallel-group assignments in which each participant received only one supplement. The record gives 352 as enrollment and reports treatment-specific endpoint summaries for each supplement.

The registry lists masking as triple, while its brief summary calls the trial double-blind. Both descriptions indicate masking, but the difference in wording is not explained in the registry record. The primary endpoint description says trained and certified observers were masked to group assignment. Other details about who was masked, how sequence allocation was concealed, or how treatment periods were separated are not reported in the registry record.

The primary outcome was change from baseline in average systolic blood pressure at 8 weeks. Secondary outcomes listed are change from baseline in serum LDL-cholesterol at 8 weeks and body weight at 8 weeks. The registry lists three outcome measures and two posted secondary outcomes. Details of the planned statistical model for the crossover structure, period effects, sequence effects, or carryover are not reported in the registry record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Intervention 1 (8 Weeks)

Intervention 2 (8 Weeks)

Intervention 3 (8 Weeks)

Analysis population for the primary outcome, as defined in the registry: We conducted analysis according to intention-to-treat principle. All participants with data were included in analysis.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Participants randomized to this three-phase cross-over trial.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Average Systolic Blood Pressure at 8 Weeks, measured over Baseline and 8 Weeks.

The primary endpoint is change from baseline in average systolic blood pressure at 8 weeks, expressed in mmHg. It is a continuous measure, summarized as a mean. The endpoint description defines the change using the mean of 6 blood pressure values from two 8-week visits minus the mean of 6 values from 2 baseline visits within each intervention phase. At each visit, 3 blood pressure readings were taken. Measurements used a Hawksley random-zero sphygmomanometer, with participants seated quietly for 5 minutes before measurement. Observers were trained and certified, and masked to group assignment. The instructions also asked participants to avoid eating, smoking, drinking alcohol, or exercising for at least 30 minutes before measurement.

Repeated measurements can reduce the influence of short-term fluctuations in a single reading; averaging also defines a more stable endpoint for analysis. That is general statistical context, not a claim about the amount of measurement error in this trial. Blood pressure is the direct outcome of interest in the registry’s stated purpose: examining the effect of soybean protein supplementation on blood pressure in persons with prehypertension or stage-1 hypertension.

The registry reports treatment-specific mean changes and 95% confidence intervals. Those summaries describe the average observed change and its estimated precision for each treatment condition. A treatment contrast, however, would directly estimate how changes differ between conditions. Such a comparison is not reported as estimable in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

A mean is a measure of the central value of a continuous outcome. Here, the endpoint is change in systolic blood pressure from baseline to the 8-week assessment within each intervention phase, and the unit is mmHg. A negative mean change indicates that the reported average at the endpoint was lower than the baseline average under the endpoint’s stated subtraction definition; a positive mean indicates the reverse. This is a description of the sign convention and the posted summaries, not a claim about an individual participant’s response.

What does a confidence interval express?

A confidence interval communicates uncertainty in an estimated mean. The interval for soy protein is -2.4 to -0.6, centered on the reported mean of -1.5. The interval for milk protein is -2.7 to -1.0 around the mean of -1.8. The interval for carbohydrate is -0.4 to 1.3 around the mean of 0.5. In general statistical terms, a confidence interval is generated by a method that would cover the target parameter at its stated confidence level over repeated use under the method’s assumptions. It is not a range of individual outcomes, nor a guarantee that the true value lies inside a particular interval.

The interval widths reflect estimated precision, but they do not by themselves describe all uncertainty. Their interpretation depends on the analysis and its assumptions. The registry record gives the 95% confidence intervals but does not report the formula, model, or standard error used to calculate them. The intervals for treatment-specific means should not be confused with intervals for comparisons between treatments.

Why is a crossover comparison different?

In a crossover design, participants receive more than one treatment condition in an assigned sequence. The trial record describes three sequences, each including soy protein, milk protein, and carbohydrate periods. A direct comparison can use each participant’s outcomes under different conditions, accounting for the fact that measurements from the same person are related. In general, this pairing can affect the estimate’s precision: the variability of a within-participant contrast depends on the relationship between that participant’s outcomes across periods.

Separate means and confidence intervals for each treatment do not provide that relationship. Consequently, a contrast cannot be recovered simply by visually comparing interval positions or treating each treatment summary as if it came from an unrelated group. The registry record identifies the continuous two-arm comparison as missing or not estimable. It does not report a paired contrast, a model for repeated treatment periods, or an explanation of why the comparison could not be estimated.

What assumptions can affect crossover analyses?

As general statistical context, crossover analyses may need to account for treatment, period, sequence, and within-participant correlation. Depending on the design and data, carryover from an earlier treatment period may also matter. These are general considerations, not claims about the analysis performed in this trial. The registry record does not report whether these terms were included, how period effects were assessed, or whether any carryover assessment was planned or conducted.

Measurement conditions are described in some detail: the endpoint averages multiple readings at baseline and follow-up visits, uses a specified instrument, and standardizes participant preparation and measurement posture. Repeated readings can reduce the influence of moment-to-moment variation in general. The record does not quantify measurement reliability or say how measurement variability was modeled.

What do planning illustrations mean?

General sample-size calculations estimate the number of participants that would be needed under specified assumptions to detect a chosen effect with a chosen power. The planning assumptions described for illustrations associated with this trial use a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and an effect defined as a stated fraction of the observed effect. Such calculations are planning illustrations, not observed power and not a report of the trial’s original design calculation. The registry record does not report the trial’s own sample-size justification. No planning output values are reported in the registry record discussed by these results, so no required sample size can be stated.

What cannot be inferred from the posted summaries?

The reported means and confidence intervals do not establish a causal treatment contrast on their own, and they do not give individual-level distributions. They also do not show how many participants contributed data to each period or how missing measurements were handled beyond the stated intention-to-treat description. A complete interpretation of a crossover analysis would require the contrast estimate and its uncertainty, along with the analysis details and relevant assumptions. Those details are not reported in the registry record. No recalculated treatment contrast should replace a sponsor analysis; the record does not report a sponsor comparison to interpret.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint of change from baseline in average systolic blood pressure at 8 weeks. It also lists two secondary outcomes: change from baseline in serum LDL-cholesterol at 8 weeks and body weight at 8 weeks. Multiple outcomes and treatment comparisons can raise multiplicity concerns in general, because examining several hypotheses can increase the chance of finding an apparently notable result by chance. The registry record does not report a multiplicity adjustment, a testing hierarchy, or which comparisons were designated as primary beyond its identification of the primary endpoint.

The record states that analysis followed the intention-to-treat principle and that all participants with data were included. This describes the stated analysis population, but it does not explain how missing endpoint measurements were handled, whether any imputation was used, or how participants without data were treated. The method of handling missing observations is not reported in the registry record. The enrollment figure is 352, while the outcome summaries list treatment-specific counts of 284 for soy protein, 286 for milk protein, and 287 for carbohydrate. Because this is a crossover study, these treatment-specific counts should not be added to infer a number of distinct participants.

The registry lists no analysis entries, although it does give arm-level outcome summaries. The continuous two-arm comparison is marked missing or not estimable. Details about the statistical model, the handling of incomplete crossover periods, and any sensitivity analyses are not reported in the registry record. These omissions limit assessment of whether the reported summaries reflect the intended comparison and how robust the findings are to missingness or analysis choices.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

For soy protein supplementation, the registry lists 352 participants at risk for serious adverse events and 0 affected; it also lists 352 at risk for other adverse events and 0 affected. For milk protein supplementation, the corresponding figures are 352 at risk and 0 affected for serious adverse events, and 352 at risk and 0 affected for other adverse events. For carbohydrate supplementation, the record likewise lists 352 at risk and 0 affected for each category. Deaths are listed as 0 for all three treatment conditions, and no top serious adverse-event terms are listed.

These are registry counts, not proof that no adverse events occurred outside the definitions, observation periods, or reporting practices represented in the record. Because the trial uses a crossover design and the same at-risk count appears for each treatment condition, the figures should not be added to estimate a number of distinct participants. The registry does not report further detail about adverse-event definitions, ascertainment, severity categories beyond the listed serious and other groupings, or the time periods covered by these counts. The numbers therefore describe what the record reports, while leaving those reporting details unresolved.

12. What the Result Does, and Does Not, Mean

The posted results show treatment-specific average changes in systolic blood pressure: -1.5 mmHg for soy protein, -1.8 mmHg for milk protein, and 0.5 mmHg for carbohydrate, each with its own 95% confidence interval. These estimates describe average change within the treatment conditions. They do not directly quantify how much the average change differed between soy protein and carbohydrate, milk protein and carbohydrate, or soy protein and milk protein. The registry record says a two-arm comparison for this continuous endpoint is missing or not estimable.

That distinction matters especially for a crossover design. Participants received multiple supplements in different sequences, so treatment comparisons should account for the relationship between measurements from the same participant and for the treatment periods. Treatment-specific summaries alone do not provide the information needed to estimate a paired contrast and its precision. No direct contrast result should be inferred from the apparent separation between the reported means or from whether individual confidence intervals overlap.

The posted estimates also do not describe the range of responses among individuals. A mean can summarize the central tendency while leaving variation in individual outcomes unreported. The registry gives no standard deviations for this endpoint. It also does not report enough analysis detail to assess period effects, sequence effects, or carryover, and missing-data handling beyond the stated inclusion of participants with data is not reported in the registry record.

Accordingly, these results can be read as descriptive summaries of average changes under each supplement condition, with reported uncertainty around those means. They do not establish a comparative effect, the probability that a treatment is beneficial, or whether an observed change matters clinically. The registry record does not report the treatment contrast or its uncertainty needed to answer the central comparative question. No medical conclusion should be drawn from the treatment-specific summaries alone.

13. Why This Trial Matters Statistically

This trial illustrates why the design and estimand matter when reading clinical results. In a parallel-group trial, a reader may often focus on the difference between group means. In a crossover trial, each participant may contribute measurements under several conditions, and the comparison depends on within-participant information. The separate means for each supplement do not reveal the covariance between a participant’s outcomes across periods, which can affect the estimated precision of a treatment contrast.

The record reports useful details about endpoint construction and measurement: the change is based on averaged readings across baseline and 8-week visits, the instrument and measurement procedures are described, and observers are described as masked. This makes the outcome definition more transparent than a label alone. At the same time, detailed measurement procedures cannot substitute for a reported treatment contrast or a clear description of the analysis model.

The case also demonstrates the importance of distinguishing a confidence interval for a treatment-specific mean from an interval for a comparative effect. Readers may be tempted to compare the three mean changes directly or judge the contrasts by the overlap of separate intervals. Neither approach gives the uncertainty of the crossover comparison. The registry identifies that comparison as not estimable, so the available summaries leave the principal comparative inference unresolved.

Finally, the record’s intention-to-treat statement, outcome counts, and adverse-event figures are informative but incomplete without details of missing observations and analysis methods. These are not unusual concerns in clinical registry interpretation; they are general statistical issues made visible by this example. The results support careful description of the reported averages, not reconstruction of an unreported comparison.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The central limitation is that no estimable two-arm comparison is reported for the continuous primary endpoint. The treatment-specific means and confidence intervals cannot supply the missing direct contrast, particularly in a crossover design where measurements within a participant are related. The registry record does not provide a contrast estimate, its confidence interval, or a test result.

The record does not describe the statistical model used for the endpoint, including whether it accounted for treatment period, sequence, within-participant correlation, or possible carryover. It does not report the trial’s sample-size justification, interim analyses, or a detailed statistical plan. The masking field says triple, while the brief summary says double-blind; the record does not explain the difference.

Although the analysis is described as intention-to-treat and includes all participants with data, the registry does not explain how missing values or incomplete treatment sequences were handled. The outcome counts differ by treatment condition, but the reason is not reported. Standard deviations and individual-level distributions are also not reported, limiting assessment of variation and preventing reconstruction of comparative uncertainty.

The registry’s adverse-event figures list zero affected participants in the reported categories, but definitions and ascertainment details are not reported in the registry record. The data should therefore be interpreted as registry-reported counts, not as comprehensive evidence about every possible event. The record also does not explain the difference between its triple-masking designation and double-blind wording in the summary, nor does it provide further information about adherence to the supplement periods.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Clinical Trial of Protein and Blood Pressure — NCT00107744.
• Registry record last updated 2023-04-12; status last verified 2018-06.
• Results first posted 2018-07-05.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT00107744 is a completed randomized crossover trial titled “Clinical Trial of Protein and Blood Pressure,” sponsored by Tulane University. The registry lists 352 for enrollment, a start of 2003-09, and primary completion of 2008-04. The population described in the summary comprises persons with prehypertension or stage-1 hypertension. Three sequences used daily soy protein, milk protein, and carbohydrate supplementation for 8 weeks per supplement period.

The primary endpoint was change from baseline in average systolic blood pressure at 8 weeks. The registry reports treatment-specific means and 95% confidence intervals, but marks a two-arm comparison for this continuous endpoint as missing or not estimable. Secondary outcomes listed are serum LDL-cholesterol change and body weight at 8 weeks. The record states that analysis followed the intention-to-treat principle and included participants with data, but does not report missing-data methods or a detailed statistical model. Serious adverse events, other adverse events, and deaths are listed as zero in each treatment condition.