Utilizing Protein During Weight Loss to Impact Physical Function and Bone: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Utilizing Protein During Weight Loss to Impact Physical Function and Bone (NCT03819478), a randomized na study. The primary outcome is 18-month Change From Baseline in Total Hip Trabecular Volumetric Bone Mineral Density Assessed by Computed Tomography (CT) in mg/cm^3.

2. Clinical Question

3. Trial Design

This randomized, parallel-group study is registered as an ancillary study to the UPLIFT trial. The registry describes triple masking, but does not identify which participant, staff, or assessor roles were masked. The listed phase is NA. The recorded enrollment is 187, and the study status is completed. The registry dates are 2017-05-03 for study start and 2022-04-28 for primary completion.

The study evaluates diet composition during a 6-month behavioral weight-loss intervention and a subsequent 12-month follow-up, with bone phenotypes as the focus. Exercise and behavioral weight-loss interventions are listed for months 0–6. The arms are named RecProt, 6-mo HiProt, and 18-mo HiProt. RecProt is described as receiving a lower-protein, higher-carbohydrate diet and a carbohydrate supplement during the initial phase. The 6-mo HiProt description specifies a higher-protein, lower-carbohydrate diet and protein supplement during that phase, followed by follow-up. The 18-mo HiProt description specifies the higher-protein, lower-carbohydrate diet during weight loss and follow-up, but its full supplement description is incomplete in the registry record.

The registry summary describes the work as an ancillary evaluation of diet composition within UPLIFT, rather than a study of diet composition in isolation from behavioral weight loss and exercise. The sponsor is Wake Forest University Health Sciences. Details such as recruitment sites, eligibility criteria, the allocation sequence method, and who was masked are not reported in the registry record. The registry record also does not report the original sample-size justification or the planned primary analysis in sufficient detail to reconstruct the trial’s design assumptions.

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.

Overall Study

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.

6. Primary Endpoint

The registry lists the primary outcome as 18-month Change From Baseline in Total Hip Trabecular Volumetric Bone Mineral Density Assessed by Computed Tomography (CT) in mg/cm^3, measured over Change from Baseline to Month 18.

The primary endpoint is the change from baseline to month 18 in total hip trabecular volumetric bone mineral density, measured in mg/cm^3. It is described as a least-squares mean, with standard error as the reported measure of dispersion. The measure comes from baseline and 18-month quantitative CT scans, calibrated using a Mindways Model 3 phantom.

Volumetric bone mineral density describes mineral density within a defined volume of bone. The trabecular specification identifies the bone compartment of interest, while the total hip location makes the outcome a regional imaging phenotype. In general, imaging measures such as this can be used to characterize bone structure and composition in a way that differs from a direct clinical outcome such as a fracture. The registry lists fractures among the study conditions, but the primary endpoint is not fracture incidence.

The endpoint is used to assess whether the dietary strategies are associated with a difference in bone change during weight loss and follow-up. It does not, by itself, establish what a measured change means for an individual’s symptoms, function, or future clinical events. Those questions would require evidence beyond this imaging outcome. The registry record does not report a threshold for a clinically important change in total hip trabecular vBMD.

7. Results Posted in the Registry

• 6-mo HiProt: mean -1.77 versus -0.91 on control; difference -0.86 (95% CI -4.66 to 2.94), standardized effect d = -0.10; z-test p = 0.657.
• 18-mo HiProt: mean -0.03 versus -0.91 on control; difference 0.88 (95% CI -2.98 to 4.74), standardized effect d = 0.10; z-test p = 0.655.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

6-month Change From Baseline in Percentage of Adipose Tissue Within the Lumbar Spine Bone Marrow as Assessed by Dual-energy Computed Tomography Algorithm

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • HiProt: mean 6.45 versus 7.42 on control; difference -0.97 (95% CI -11.88 to 9.94), standardized effect d = -0.03; z-test p = 0.862.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between HiProt and the control. The interval gives the range of effects that remain compatible with the data.

18-month Change From Baseline in Percentage of Adipose Tissue Within the Lumbar Spine Bone Marrow as Assessed by Dual-energy Computed Tomography Algorithm

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • 6-mo HiProt: mean -9.96 versus -0.20 on control; difference -9.76 (95% CI -20.14 to 0.62), standardized effect d = -0.42; z-test p = 0.065.
  • 18-mo HiProt: mean -5.17 versus -0.20 on control; difference -4.97 (95% CI -15.54 to 5.60), standardized effect d = -0.21; z-test p = 0.357.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 6-mo HiProt and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between 18-mo HiProt and the control. The interval gives the range of effects that remain compatible with the data.

6-month Change From Baseline in Femoral Bone Strength in kN

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • HiProt: mean 0.10 versus -0.02 on control; difference 0.12 (95% CI -0.09 to 0.33), standardized effect d = 0.20; z-test p = 0.259.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between HiProt and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several aspects of bone and marrow phenotypes, including lumbar spine marrow adiposity, femoral bone strength and cortical thickness, and vertebral volumetric bone mineral density. The available detailed summaries show a mixed pattern rather than a single consistent direction across measures and follow-up times. Some comparisons have estimates on either side of zero, and the intervals for the re-calculated comparisons include zero. These outcomes provide context for the primary total hip trabecular vBMD result, but they do not turn it into evidence about a broader clinical outcome.

The registry lists nine secondary outcomes, while detailed numerical results are available in the record for only some of them. Detailed numerical summaries for the remaining listed measures are not reported in the registry record. The available results include secondary changes measured at month 6 and month 18, and the sponsor-reported analyses are labeled ANCOVA. The re-calculated secondary contrasts are unadjusted. Where those methods yield different p-values, the distinction in method should remain explicit.

Secondary outcomes are generally less definitive than a prespecified primary outcome, particularly when several endpoints, time points, and comparisons are examined. In general, testing many outcomes can increase the chance of finding an apparently notable result by chance. The registry record does not report whether a multiplicity procedure was planned or used. Secondary results are therefore best viewed as additional evidence to interpret cautiously, not as independent confirmation of the primary result or as a basis for claims about outcomes that were not measured.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• 6-month Change From Baseline in Percentage of Adipose Tissue Within the Lumbar Spine Bone Marrow as Assessed by Dual-energy Computed Tomography Algorithm
• 18-month Change From Baseline in Percentage of Adipose Tissue Within the Lumbar Spine Bone Marrow as Assessed by Dual-energy Computed Tomography Algorithm
• 6-month Change From Baseline in Femoral Bone Strength in kN
• 18-month Change From Baseline in Femoral Bone Strength in kN
• 18-month Change From Baseline in Femoral Cortical Thickness Assessed by Computed Tomography (CT) in mm
• 6-month Change From Baseline in Femoral Cortical Thickness Assessed by Computed Tomography (CT) in mm
• 18-month Change From Baseline in Vertebral Volumetric Bone Mineral Density in the Lumbar Spine Assessed by Computed Tomography (CT) in mg/cm^3
• 6-month Change From Baseline in Vertebral Volumetric Bone Mineral Density in the Lumbar Spine Assessed by Computed Tomography (CT) in mg/cm^3

9. Statistical Interpretation

10. Statistical Methods Explained

What does a least-squares mean represent?

A least-squares mean is an estimate of an outcome mean from a statistical model. In general, it can reflect adjustment for variables included in that model, rather than simply being the raw arithmetic average of observed values. The registry identifies the primary outcome summaries as least-squares means, but does not report enough model details to determine which covariates were included in the primary outcome model. The arm summaries should therefore be interpreted as model-based estimates as labeled in the record, not assumed to be unadjusted raw means.

The standard error describes the estimated uncertainty in a mean estimate. In general, a smaller standard error indicates greater precision for that estimate, while a larger standard error indicates less precision. A standard error is not the same thing as the spread of individual participant measurements. The registry does not post the individual participant data or a distribution of the observed measurements, so the within-arm variability cannot be directly inspected from the record.

What does the re-calculation compare?

The re-calculated comparisons are simple, unadjusted contrasts based on the posted arm summaries. They express the difference in estimated mean change between a high-protein arm and RecProt. The estimates are not the same as an ANCOVA coefficient unless the model and adjustment structure make them equivalent; that information is not reported. The re-calculation uses an approximate pooled standard deviation derived from the posted standard errors and sample sizes, as indicated in the analysis record. It is therefore an approximation from summary information, not a reanalysis of individual participant records.

In general, an unadjusted comparison does not account for baseline values or other covariates. ANCOVA, by contrast, is a model-based method that can compare groups while adjusting for variables specified in the model. Adjustment can affect the estimated contrast and its uncertainty. The exact ANCOVA specification, including its covariates and handling of repeated or incomplete measurements, is not reported in the registry record. The two approaches answer related but not necessarily identical statistical questions.

How should a confidence interval be read?

A confidence interval gives a range of estimates compatible with the data and the method used to construct it. In general, a wider interval indicates less precision, while a narrower interval indicates more precision. A 95% confidence interval is not a guarantee that the true value lies in the particular calculated interval; rather, the confidence level describes the long-run behavior of the interval procedure under its assumptions. For the re-calculated primary contrasts, the intervals include zero and also allow differences in either direction. This makes the estimated magnitude uncertain, rather than establishing that the groups are identical.

A confidence interval and a p-value convey related but different information. A p-value summarizes how inconsistent the observed data are with a specified null hypothesis under the analysis assumptions. It does not measure the probability that the null hypothesis is true, and it does not quantify clinical importance. The registry does not post confidence intervals for its ANCOVA results, so the p-values alone do not show the range of effect sizes compatible with those adjusted analyses.

Why distinguish adjusted and unadjusted results?

The sponsor-reported primary analyses are labeled ANCOVA. The posted re-calculations are explicitly unadjusted, based on summary statistics, and use a different reference-arm framing from one of the sponsor-reported contrasts. Differences between their p-values or estimates can arise because the methods, covariates, comparison groups, or analysis populations differ. The registry record does not provide enough detail to identify every source of difference. These results should be presented as distinct analyses rather than blended into one estimate.

Randomization is intended, in general, to support fair comparisons by balancing prognostic characteristics across groups on average. It does not ensure that every measured or unmeasured characteristic is identical in the groups, especially in a study with limited analyzed counts. Baseline values, balance summaries, and the allocation procedure are not reported in the record used for these results. The presence of random allocation is important context, but does not remove the need to examine how the outcomes were analyzed and who contributed data.

What assumptions affect interpretation?

Every statistical comparison relies on assumptions about the outcome, the model, and the available data. The re-calculation uses an approximate pooled standard deviation and a normal approximation for planning illustrations. Those are not descriptions of the sponsor’s ANCOVA model. General concerns include whether the outcome distribution and variance are adequately represented, whether observations are independent as the analysis assumes, and whether missing outcome data could affect comparisons. The registry record does not report diagnostic checks or sensitivity analyses addressing these issues.

For continuous outcomes, over-dispersion means that variability is greater than a chosen model or reference distribution anticipates. In general, if such extra variability is ignored, standard errors may be too small and confidence intervals may be too narrow. The record does not establish that over-dispersion occurred for this outcome; it is a general consideration when assessing model adequacy. Because individual-level distributions and model diagnostics are not reported, this concern cannot be evaluated directly.

The primary outcome is one of several registered measures, and the record reports multiple analyses. In general, testing multiple outcomes or contrasts increases the chance that at least one result will appear unusual by chance if no adjustment is made. Whether multiplicity adjustment was planned or used is not reported in the registry record. The posted p-values should therefore be read as the results of their listed comparisons, without assuming that they account for all other outcome tests.

Finally, statistical evidence, precision, and practical importance are separate considerations. An estimate can be imprecise even when a point estimate appears large, and a precise estimate can still be of limited practical importance if the effect is small. The registry record does not identify a clinically important threshold for the primary measure. General interpretation should therefore focus on the estimates, their uncertainty, and the outcome’s scope rather than on a threshold that is not documented.

11. Confidence Intervals and Precision

The primary re-calculated contrasts have wide confidence intervals relative to their point estimates. For 6-mo HiProt versus RecProt, the estimated mean difference is -0.86 mg/cm^3, and the 95% confidence interval extends from -4.658 to 2.938. For 18-mo HiProt versus RecProt, the estimate is 0.88 mg/cm^3, with a 95% confidence interval from -2.975 to 4.735. Each interval includes zero, and each allows values on both sides of zero. The posted summaries therefore do not pin down a narrow range for the underlying difference.

In general, a confidence interval describes uncertainty in an estimated contrast, not the range of individual participant outcomes. Its width depends on factors such as the variability of the outcome and the amount of information available. The primary arm summaries each have a reported sample size of 38, but the registry record does not provide individual-level data or the full model specification needed to assess precision under the sponsor’s ANCOVA directly.

The registry’s ANCOVA results are reported as p-values without corresponding estimates or confidence intervals. A p-value alone cannot show whether the estimated difference is small, large, or precisely estimated. The re-calculated intervals help describe uncertainty for the unadjusted contrasts, but they should not be treated as confidence intervals from the sponsor’s adjusted analysis. The registry record does not report a clinical importance threshold for this endpoint, so statistical precision cannot be translated into a judgment about clinical importance.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are stated as participants per arm for 80% power and participants per arm for 90% power. They are not total sample sizes and are not event counts. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The rows for 100%, 75%, and 50% of the observed effect illustrate how the planning figures change when the assumed target effect is smaller.

In plain language, the 80% power column gives the planned number per arm for a design that would have an 80% chance of detecting the specified effect under the stated assumptions, if that effect were present and the assumptions held. The 90% power column applies the same idea with a higher chance of detection. These are design probabilities, not guarantees for a particular study. Smaller effects are harder to distinguish from random variation, so planning for a smaller fraction of the observed effect requires more participants per arm.

The figures are planning illustrations based on the observed effect and are not the trial’s original design justification. The original sample-size assumptions and justification are not reported in the registry record. These figures are not observed power; they do not measure the probability that the completed trial’s result is correct or that a treatment effect exists. They also do not replace interpretation of the estimates and confidence intervals for the completed study.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.860; pooled standard deviation σ = 8.447.
2. n = 2 × (1.96 + 0.842)² × 8.447² / -0.860² = 1,514.5, rounded up to 1,515 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.86                      # observed difference in means
sigma <- 8.44727174891396                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3819478)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 1,515 (80% power), 2,028 (90% power)
 75% of observed effect: n per arm = 2,693 (80% power), 3,605 (90% power)
 50% of observed effect: n per arm = 6,059 (80% power), 8,111 (90% power)
power.t.test: 1515.5 
Simulated power at n = 1,515 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry’s own reported primary comparisons use ANCOVA. The sponsor-reported tests compare RecProt with 18-mo HiProt and 6-mo HiProt with 18-mo HiProt. The record gives p-values for those comparisons but does not provide the ANCOVA coefficients, confidence intervals, covariates, or full model specification. The reported p-values do not indicate strong evidence against the null hypothesis under the listed tests, but they do not give a direct measure of the size or precision of a difference.

The simple re-calculations instead compare each high-protein arm with RecProt using posted summary statistics, without covariate adjustment. Their estimated differences and confidence intervals are useful as transparent, approximate unadjusted contrasts, but they are not substitutes for the sponsor’s model-based analysis. Different comparison framing and adjustment can yield different results. The record does not provide enough model detail to explain all differences between the reported and re-calculated analyses.

For a secondary outcome, the sponsor-reported ANCOVA test for femoral bone strength has a p-value below 0.05, while the simple unadjusted re-calculation has a different p-value. This distinction illustrates why results from different analyses should not be treated as interchangeable. The registry does not report a discrepancy flag, and the adjusted and unadjusted results should be described according to their respective methods rather than collapsed into one conclusion.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures and 15 analyses. That is relevant because interpreting several outcomes and comparisons raises a multiplicity question: in general, the more hypotheses tested, the greater the chance that one or more results may appear notable by chance. The registry record does not report whether the analyses were organized into a prespecified testing hierarchy, whether any adjustment for multiple comparisons was planned or applied, or how the listed analyses relate to the full analysis plan.

The record identifies the primary outcome and provides sponsor-reported ANCOVA tests, but it does not report the full statistical analysis plan. The exact covariates, model diagnostics, handling of repeated measurements, and rules for selecting analysis populations are not reported in the registry record. It is also not reported whether interim analyses were planned or performed. No interim-analysis schedule or stopping boundary should be inferred from the completed status alone.

Missing data are another important consideration for a longitudinal outcome measured at baseline and month 18. The primary arm summaries report 38 participants per arm, but those figures do not by themselves establish how many participants were randomized, how many contributed measurements at each visit, or why any observations might be absent. The registry record does not describe missing-data amounts, reasons for missingness, imputation, or sensitivity analyses.

The analysis population is not reported in enough detail to determine whether the sponsor used an intention-to-treat population, a per-protocol population, or another defined set of participants. The record also does not describe treatment adherence or how participants who did not follow the assigned dietary strategy were handled analytically. These omissions limit assessment of the robustness and generalizability of the reported comparisons. They are uncertainties about what is documented, not evidence that particular analyses or data problems occurred.

16. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: atrial flutter (1.6%), myocardial infarction (1.6%), gastroenteritis (1.6%), abdominal pain (1.6%), anaphylaxis (1.6%).

The serious adverse event summaries report 8 affected participants among 62 at risk in RecProt, 5 among 62 at risk in 6-mo HiProt, and 5 among 63 at risk in 18-mo HiProt. The record lists zero deaths in each arm. These are participant counts for the adverse-event reporting groups, not the primary endpoint analysis counts. The groups at risk for safety reporting therefore should not be conflated with the primary outcome sample sizes.

The listed serious adverse event terms each have a reported frequency of 1.6%: atrial flutter, myocardial infarction, gastroenteritis, abdominal pain, anaphylaxis, and lung infection. The record does not provide enough information in these summaries to assess event relatedness, timing, duration, recurrence, or comparative risk. The presence of a listed event does not by itself establish that it was caused by an intervention.

These summaries describe the posted safety counts; they do not establish that one arm is safer than another. No adjusted comparative safety analysis is reported in the registry record. Adverse-event counts should be interpreted with the reporting definitions and follow-up period in mind, which are not fully described in the available registry entries.

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

The result addresses a specific statistical question: how the assigned dietary strategies relate to change in an imaging-based measure of total hip trabecular vBMD from baseline to month 18. The posted arm estimates and unadjusted contrasts describe the observed pattern and its uncertainty. The re-calculated confidence intervals include zero and extend to values on either side, so the summaries do not identify a precise between-arm difference.

This does not establish that the strategies are equivalent. A confidence interval that includes zero can still be compatible with meaningful differences in either direction, depending on the interval and on what difference would matter. The registry does not report an equivalence or non-inferiority margin, or a threshold for clinically important change. Equivalence cannot be concluded merely because a test does not provide strong evidence against a null hypothesis.

It also does not establish that protein supplementation has, or lacks, an effect on fractures, bone strength, physical function, symptoms, or other outcomes. The primary measure is a CT-derived bone phenotype. Some secondary measures address other bone features, but the available summaries do not provide a complete account of every listed secondary endpoint. Results for one measure should not be generalized automatically to unmeasured or incompletely reported outcomes.

The interventions are embedded within a broader weight-loss and exercise program. The registry record does not isolate dietary composition from every other component of the intervention package, and the arm descriptions do not fully document all details for the 18-month high-protein group. Therefore, the result should be read as an outcome comparison within this study context, rather than as a general estimate for protein supplementation in all older adults.

Finally, registry summaries are not a substitute for a full account of the analysis plan, participant flow, missing data, adherence, and model specification. Those details are not reported in the registry record. The outcome estimates can be described, but their interpretation must remain bounded by the information documented for the study.

18. Why This Trial Matters Statistically

This trial is statistically interesting because it separates an imaging-based primary outcome from the broader claims often attached to nutrition and bone health. The primary question concerns a specific anatomical site, bone compartment, measurement method, and follow-up point. Keeping those features explicit helps prevent a result for total hip trabecular vBMD from being treated as a result for every aspect of skeletal health.

It also illustrates the importance of distinguishing model-based sponsor analyses from transparent summary-statistic re-calculations. ANCOVA can account for covariates specified in a model, whereas the posted re-calculation is unadjusted and approximate. Neither result should be silently substituted for the other. Seeing both approaches makes the role of analytical choices and incomplete reporting clearer.

The uncertainty around the primary contrasts matters as much as their point estimates. Intervals that span values on either side of zero show why a single estimate or p-value cannot communicate the full range of plausible differences. The registry’s lack of confidence intervals for the ANCOVA results makes complete precision assessment harder, while the available unadjusted intervals have their own limitations.

The study also provides a useful example of how multiple outcomes and time points complicate interpretation. The secondary measures broaden the biological picture, but they do not automatically strengthen the primary result, especially when multiplicity handling and a complete set of detailed outcome summaries are not reported. For biostatisticians and clinical researchers, the central lesson is to separate the observed estimates, the analysis methods, and the scope of inference.

19. 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 registry record leaves important design and analysis details unresolved. Specific eligibility criteria, detailed participant characteristics, recruitment sites, the allocation procedure, and the roles included in triple masking are not reported. The full arm description for 18-mo HiProt is incomplete, so the precise supplement regimen for that arm cannot be reconstructed from the record.

The sponsor’s ANCOVA coefficients and confidence intervals are not reported, and the model covariates and analysis population are not specified. The re-calculated comparisons are approximate unadjusted analyses from summary statistics, not a reanalysis of participant-level data. This limits direct comparison between the two sets of results. Baseline balance, individual outcome distributions, and model diagnostics are also not reported.

The registry does not provide missing-data amounts or methods, reasons for incomplete follow-up, adherence information, or sensitivity analyses. Although the record lists multiple outcome measures and analyses, multiplicity procedures are not described. The original sample-size justification, interim analyses, and full statistical analysis plan are not reported in the registry record.

Detailed numerical summaries are available for only some of the listed secondary measures. The record also does not specify a clinically important threshold for the primary imaging outcome. These limitations do not invalidate the posted estimates, but they constrain conclusions about precision, robustness, clinical importance, and the extent to which the findings apply beyond the analyzed participants and study setting.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Utilizing Protein During Weight Loss to Impact Physical Function and Bone — NCT03819478.
• Registry record last updated 2024-01-19; status last verified 2021-12.
• Results first posted 2024-01-19.
• 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.

23. Record Summary

NCT03819478 is a completed randomized, parallel-group ancillary study of dietary protein and carbohydrate composition during weight loss in obese older adults. The registry lists triple masking and an enrollment of 187. The study evaluates bone phenotypes during a 6-month weight-loss intervention and 12-month follow-up, with behavioral weight loss and exercise listed during months 0–6.

The primary outcome is 18-month change from baseline in total hip trabecular volumetric bone mineral density measured by quantitative CT. The posted arm summaries and unadjusted re-calculations show uncertain contrasts, with confidence intervals that include zero. The registry also reports ANCOVA p-values without the corresponding model estimates or confidence intervals. Several secondary bone outcomes are listed, but detailed numerical summaries are not available for all of them.

Important details about the original sample-size justification, analysis population, missing data, multiplicity, and full analysis plan are not reported in the registry record. The results should be interpreted as outcome summaries within the combined dietary, behavioral weight-loss, and exercise study context.