Comparison of Liraglutide Versus Placebo in Weight Loss Maintenance in Obese Subjects: SCALE - Maintenance: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparison of Liraglutide Versus Placebo in Weight Loss Maintenance in Obese Subjects: SCALE - Maintenance (NCT00781937), a randomized phase 3 study. The primary outcome is Mean Percentage Change in Fasting Body Weight From Baseline.

2. Clinical Question

The question is specifically about maintaining or extending weight loss after an initial diet period, rather than comparing weight change from the start of that run-in. Week 0 is the primary endpoint’s baseline. The registry record identifies week 0 and week 56 as the primary timeframe and states that weighing followed fasting since midnight, with only water consumed. The reported primary results therefore describe average percentage change during the main trial period in the analysis population, not the entire period from screening through follow-up.

3. Trial Design

The registry record describes a phase 3, randomized, double-blind, parallel-group, placebo-controlled trial with 422 enrolled participants. It was conducted in North America. The trial ran through a 12-week run-in, a 56-week main trial period, and a 12-week follow-up period. The run-in extended from week -12 to week 0; the main trial covered weeks 0 to 56, and follow-up covered weeks 56 to 68. The record lists a start date of 2008-10-30, primary completion on 2010-09-01, and a completed status.

During the run-in, screened subjects followed a low calorie diet. Randomisation was limited to subjects who lost at least 5% of screening body weight. The two randomized groups were liraglutide 3.0 mg and placebo. For the main trial, liraglutide was injected subcutaneously once daily; the placebo description also states once-daily subcutaneous administration for 56 weeks. The arm descriptions refer to standard energy-restricted instructions, although the registry wording for those instructions is incomplete. Further detail about the exact diet or other behavioural support is not reported in the registry record.

The primary analysis population is described as a Last Observation Carried Forward, Full Analysis Set. It includes randomized subjects who were exposed to at least one dose of trial product and had at least one post-randomisation assessment of any efficacy endpoint. The primary outcome results list 194 participants in the liraglutide group and 188 in the placebo group. Those endpoint analysis counts do not by themselves describe all enrollment or the participants followed at every visit.

The registry record lists 26 outcome measures and 24 analyses. Its posted secondary outcomes cover weight-loss and weight-regain thresholds, maintenance of run-in weight loss, change in fasting weight, and change in blood pressure. The original protocol’s sample-size assumptions, detailed allocation procedure, sequence generation, concealment procedures, and any interim-analysis plan 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.

Before assignment: Subjects who lost at least 5% of screening body weight after 4 weeks and up to 12 weeks during the run-in were randomised in a 1:1 manner to receive either liraglutide 3.0 mg, or placebo for 56 weeks.

Main Trial Period Through Week 56

Reasons for not completing this period:

Follow up Period (Week 56-68)

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Last Observation Carried Forward, Full Analysis Set (includes all randomised subjects exposed to at least 1 dose of trial product and with at least 1 post-randomisation assessment of any efficacy endpoint.)

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 Mean Percentage Change in Fasting Body Weight From Baseline, measured over Week 0, week 56.

The primary endpoint is mean percentage change in fasting body weight from baseline, with baseline identified as week 0 and the follow-up measurement at week 56. It is a continuous outcome: each participant’s weight change is expressed as a percentage, and the reported group results are least-squares means. A negative value represents a reduction from the week 0 weight. The registry record specifies that participants were weighed after fasting since midnight the previous night, consuming only water.

Using week 0 as the baseline makes the endpoint relevant to the trial’s maintenance question. Participants entered randomisation only after the low calorie diet run-in and the required initial weight loss. Thus, the primary comparison addresses subsequent average weight change during the 56-week main trial, rather than the weight loss achieved during the run-in. The endpoint does not alone describe how outcomes varied among individuals: a group mean can coexist with a range of individual responses.

Least-squares means are model-based group estimates, commonly used to compare groups while accounting for specified covariates. The registry record identifies the primary parameter as a least-squares mean and reports the sponsor’s between-group analysis as ANCOVA, but the covariates and full model specification are not reported in the registry record. The primary analysis population also uses last observation carried forward for missing follow-up values. That approach affects how the endpoint is represented for participants without an observed week 56 measurement; it does not turn carried-forward values into actual week 56 observations.

7. Results Posted in the Registry

• Lira 3.0 mg: mean -6.11 versus -0.05 on control; difference -6.06 (95% CI -7.85 to -4.27), standardized effect d = -0.68; z-test p = <0.001.

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.

Percentage of Subjects Who Lost More Than 10% of Fasting Body Weight From Week 0

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Lira 3.0 mg: 26.3% (51/194) versus 6.4% (12/188) on control; risk difference 19.9 percentage points (95% CI 12.8 to 27.0), risk ratio 4.12 (2.27–7.47); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Lira 3.0 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Subjects With Weight Regain (Fasting) More Than or Equal to 5% From Week 0

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Lira 3.0 mg: 1.9% (4/207) versus 17.5% (36/206) on control; risk difference -15.5 percentage points (95% CI -21.1 to -10.0), risk ratio 0.11 (0.04–0.31); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Lira 3.0 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage of Subjects With Weight Regain (Fasting) More Than or Equal to 10% From Week 0

• Endpoint type: Binary: the percentage of participants with the event (counts rebuilt from the posted percentages and group sizes).
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Lira 3.0 mg: 0.0% (0/207) versus 2.9% (6/206) on control; risk difference -2.9 percentage points (95% CI -5.2 to -0.6); unadjusted two-proportion z-test p = 0.013.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically lower for Lira 3.0 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists secondary outcomes involving further weight loss, weight regain, maintenance of run-in loss, fasting weight change, and blood pressure. Taken together, the posted weight-related results generally align in direction with the primary average comparison: they describe more favorable distributions of weight-loss and weight-regain outcomes in the liraglutide group than in the placebo group. These endpoints provide complementary views of the weight outcome, but they are not substitutes for the primary mean change and should not be collapsed into a single overall effect.

Threshold outcomes address whether a participant crossed a particular boundary, while continuous outcomes retain information about the amount of change. Measures of maintained run-in loss also relate to a different reference point from the primary endpoint, which compares change from week 0. The endpoint definitions and timeframes matter when interpreting whether a secondary result is consistent with the primary question. Results at week 56 do not establish persistence of those outcomes during the subsequent follow-up period.

The registry record lists 26 outcome measures, 24 analyses, and 23 posted secondary outcomes. With numerous outcomes, the chance of obtaining at least one apparently strong result can rise if each is tested separately. In general, multiplicity adjustment or a prespecified testing hierarchy can help control that issue. Such a strategy is not reported in the registry record. Accordingly, unadjusted secondary p-values carry less confirmatory weight than the primary analysis and should be treated as supporting or exploratory evidence rather than as a collection of independent confirmations.

The pattern across outcomes can be useful for understanding whether the primary mean comparison is accompanied by changes in threshold-based outcomes. It does not, by itself, establish a single unified treatment effect across all measures. Nor does it resolve missing-data concerns, model assumptions, or the restricted population created by the run-in eligibility criterion. The secondary results should be interpreted alongside their individual definitions and analysis methods, not as proof beyond the primary endpoint.

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

• Percentage of Subjects Who Lost More Than 10% of Fasting Body Weight From Week 0
• Percentage of Subjects With Weight Regain (Fasting) More Than or Equal to 5% From Week 0
• Percentage of Subjects With Weight Regain (Fasting) More Than or Equal to 10% From Week 0
• Percentage of Subjects With Greater Than 50% of Fasting run-in Weight Loss Maintained From Week 0
• Percentage of Subjects With Greater Than 75% of Fasting run-in Weight Loss Maintained From Week 0
• Change From Baseline in Fasting Weight
• Change From Baseline in Fasting Weight for Subjects Completing the Main Trial Period and Entering the Follow-up Period
• Change From Baseline in Blood Pressure

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous because it measures percentage change in body weight rather than classifying participants into outcome categories. The posted group estimates are least-squares means, and the sponsor’s reported comparison uses ANCOVA. In general, analysis of covariance estimates a group contrast while adjusting for covariates included in the model. This can improve precision or account for baseline differences when the model is appropriately specified. The registry record does not report which covariates were used, how they were selected, or the model’s other detailed assumptions, so the sponsor’s estimate cannot be independently reconstructed from the posted summaries alone.

What does the recalculated comparison represent?

The recalculated mean difference is a simple, unadjusted comparison of the posted group summaries. Its estimate is -6.06, with a 95% confidence interval from -7.848 to -4.272 and an unadjusted p-value < 0.001. The calculation uses the group estimates and standard errors to form an approximate pooled standard deviation; the approximation is identified in the analysis information. It does not include the sponsor’s ANCOVA adjustment and should not be described as the trial’s full statistical model. The reported sponsor analysis is the registry’s own model-based treatment contrast.

A standard error describes the estimated sampling variability of an estimate. It is not a measure of how different individual participants’ outcomes were. The primary arm results report standard errors of 0.66 and 0.63 for the least-squares means. A confidence interval is formed to express uncertainty about an estimate, under the method’s assumptions. A narrower interval indicates greater precision for that estimate; it does not guarantee that the model is unbiased or that its assumptions are correct.

How should the p-values be understood?

A p-value is calculated under a specified null hypothesis and statistical model. In general, a small p-value indicates that data at least as incompatible with the null as those observed would be unusual if the null and model assumptions held. It is not the probability that the treatment contrast is due to chance, and it does not measure the size or clinical importance of an effect. The sponsor reports p < 0.001 for the primary ANCOVA comparison, while the simple recalculation reports p < 0.001. These are results from distinct analyses, with the sponsor’s analysis using ANCOVA and the recalculation using posted summary statistics.

The registry’s primary analysis population combines two features: inclusion in the Full Analysis Set and a Last Observation Carried Forward approach. The set includes randomized participants exposed to at least one dose and with at least one post-randomisation efficacy assessment. In general, last observation carried forward fills a later missing measurement with that participant’s last available observation. This is a missing-data assumption, not an observed value at the later visit. If outcomes would have changed after the last measurement, carrying the earlier value forward may not represent the unobserved outcome. The record does not report sensitivity analyses that assess how results might change under other missing-data assumptions.

Why can continuous and binary outcomes tell different stories?

The continuous primary endpoint uses the mean percentage change across the analysis group. Binary secondary outcomes classify participants according to whether they cross a threshold, such as a specified amount of further loss or regain. These summaries capture different features of the outcome distribution. A mean may change even when the proportion crossing a particular threshold changes little; conversely, a threshold proportion can shift without fully describing the average change. In general, dichotomising a continuous measure discards information about how far each participant is from the threshold.

For binary outcomes, the posted analyses include proportions and, for some comparisons, risk differences or odds ratios. A risk difference describes the difference between group proportions; an odds ratio compares the odds of the outcome between groups. These are not interchangeable effect measures and their numerical values have different interpretations. The sponsor’s reported secondary analyses use logistic regression for two outcomes, with odds ratios. The registry record does not provide a full account of all covariates or model details for those regressions. Unadjusted recalculations based on event counts and group totals are descriptive comparisons and do not reproduce a regression model.

What does over-dispersion mean in general?

For count outcomes, over-dispersion means that observed variability is greater than expected under a simple model such as a Poisson model. Accounting for it can matter because standard errors and confidence intervals may otherwise be too small. This trial’s primary endpoint is continuous, and the posted primary analysis is ANCOVA; over-dispersion is not a stated issue for that comparison. The registry record does not report a count-based primary outcome or a specific over-dispersion analysis.

The posted analyses include multiple outcomes and comparisons. When many hypotheses are tested, the chance of at least one small p-value under a collection of null hypotheses can increase. In general, multiplicity procedures adjust interpretation or testing to account for a defined family of comparisons. Whether the trial used such procedures, and which outcomes would have belonged to any testing hierarchy, is not reported in the registry record. The secondary p-values should therefore be read as unadjusted evidence unless the registry explicitly reports otherwise.

11. Confidence Intervals and Precision

The sponsor’s primary treatment contrast is -6.06, with a 95% confidence interval from -7.50 to -4.62. The interval lies below zero, indicating that values consistent with the sponsor’s model and its assumptions are on the negative side of the comparison. The interval communicates uncertainty about the estimated average contrast; it is not a range containing a stated proportion of individual participant outcomes. The trial’s actual individual changes may be more variable than the uncertainty around the group contrast suggests.

The recalculated unadjusted interval is -7.848 to -4.272. It is based on posted summary estimates and an approximate standard deviation, whereas the sponsor interval comes from ANCOVA. The differences in interval limits reflect the use of different methods and assumptions; they should not be treated as competing measurements of the same fitted model. The recalculated estimate is -6.06, matching the sponsor’s reported central contrast, but matching point estimates do not make the uncertainty calculations identical.

In general, confidence intervals help readers judge both the direction and precision of an estimate. A narrow interval conveys more precision than a wide interval, but precision is distinct from freedom from bias. The interval’s interpretation relies on the method used and its assumptions, including the analysis population and handling of missing data. Because the primary analysis uses last observation carried forward, the interval does not account for every possible uncertainty about the unobserved measurements unless such uncertainty is incorporated into the method.

The group standard errors for the primary least-squares means are 0.66 for liraglutide and 0.63 for placebo. Standard errors describe uncertainty in estimated means, not the spread of observed participant-level weights. The approximate pooled standard deviation used in recalculation is 8.924, but the registry does not provide the underlying individual data needed to assess the full distribution or check model assumptions directly.

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 illustrations based on the observed effect, not the trial’s original design calculation. For the continuous primary outcome, the figures are participants per arm. The stated 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 the stated fraction of the observed effect. The registry record does not report the original protocol’s sample-size assumptions or justification.

The 80% power column gives the planned number of participants per arm that, under the stated assumptions and effect size, would yield an 80% chance of rejecting the null hypothesis if that effect were the true one. The 90% power column describes the corresponding planning target for a 90% chance. These are prospective design concepts, not descriptions of what happened in the completed trial.

The rows express planning for 100%, 75%, and 50% of the observed effect. As the assumed effect becomes smaller, more participants are needed to distinguish it from zero with the same target power and significance level. The displayed curve or other visual representation should be read as the relationship between the assumed effect and the planned sample size under those same assumptions, not as a forecast of the trial’s actual result.

These figures are not observed power. Observed power calculated after a trial largely restates the effect estimate and its uncertainty; it does not provide an additional independent measure of evidence. Planning illustrations also do not establish that the original trial was designed around the observed effect, and they do not account for design features or analysis details not included in the stated assumptions.

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 = -6.060; pooled standard deviation σ = 8.924.
2. n = 2 × (1.96 + 0.842)² × 8.924² / -6.060² = 34.0, rounded up to 35 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 <- -6.06                      # observed difference in means
sigma <- 8.92411747074068                     # 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(781937)
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 = 35 (80% power), 46 (90% power)
 75% of observed effect: n per arm = 61 (80% power), 82 (90% power)
 50% of observed effect: n per arm = 137 (80% power), 183 (90% power)
power.t.test: 35 
Simulated power at n = 35 per arm: 0.799 (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 reports the sponsor’s primary analysis as an ANCOVA treatment contrast of -6.06, with a 95% confidence interval from -7.50 to -4.62 and p < 0.001. The unadjusted recalculation from posted summary statistics gives a mean difference of -6.06, a 95% confidence interval from -7.848 to -4.272, and p < 0.001. The estimates have the same reported central value, while their intervals and p-values differ because the calculations use different methods.

The recalculation is not an attempt to reproduce the ANCOVA. It uses the available group summaries and an approximate standard deviation, without the sponsor model’s covariate adjustment. The registry record does not report the ANCOVA covariates or enough detail to rebuild the full analysis. The sponsor’s estimate is therefore the relevant result for the analysis it reports; the recalculation offers a transparent, simple comparison based on the posted summaries.

Both analyses indicate a negative contrast, and the sponsor interval does not include zero. Their agreement in direction is informative, but it does not show that all modeling choices would yield the same precision or that unreported assumptions have been checked. The record does not flag a discrepancy between the sponsor analysis and the recalculation.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists several outcomes and analyses, but does not report whether a multiplicity adjustment was used, what family of hypotheses would have been adjusted, or whether a hierarchy determined the order of testing. Multiplicity matters because repeated testing can increase the probability of at least one small p-value even when the tested effects are absent. Without a reported plan, the secondary p-values are best understood as unadjusted results, and their evidential weight is lower than that of a designated primary endpoint.

The primary analysis population is identified as a Full Analysis Set that includes randomized subjects exposed to at least one dose and with at least one post-randomisation efficacy assessment. This definition differs from including every enrolled person regardless of exposure or assessment. The primary results also use Last Observation Carried Forward. In general, that method carries a participant’s last observed measurement forward when a later value is missing. It assumes that the carried-forward value is an acceptable representation of the missing outcome, an assumption that may not hold if the participant’s condition would have changed.

The registry record does not report the number of missing primary endpoint measurements, the reasons for missingness, or sensitivity analyses using alternative missing-data methods. It also does not report whether any other imputation or missing-data approach was used for the secondary outcomes. These omissions limit assessment of how robust the findings are to unobserved outcomes. Last observation carried forward can produce a complete analysis dataset, but completion of the dataset is not the same as eliminating uncertainty due to missing data.

The record does not report the original sample-size justification, interim analyses, a detailed statistical analysis plan, or the covariates in the primary ANCOVA. It also does not report a formal assessment of whether model assumptions were met. These are limitations of what can be evaluated from the registry record; they do not establish that a procedure was or was not performed. The reported analysis remains interpretable as an ANCOVA result for the stated population, while the unreported details constrain independent appraisal.

16. Safety Results

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

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: appendicitis (1.0%), cholelithiasis (0.9%), sepsis (0.5%), cardiac failure (0.5%), hidradenitis (0.5%).

The posted safety counts list 9 serious adverse events among 212 participants at risk in the liraglutide group and 5 among 210 at risk in the placebo group. Other adverse events are reported for 177 of 212 participants at risk in the liraglutide group and 163 of 210 in the placebo group. These counts should be read as the registry’s reported participant counts, not as a measure of severity for every event or as proof that treatment caused an event. The registry record does not report deaths.

The serious adverse event terms listed with percentages include appendicitis at 1.0%, cholelithiasis at 0.9%, sepsis at 0.5%, cardiac failure at 0.5%, hidradenitis at 0.5%, and aortic aneurysm at 0.5%. The record does not provide enough detail in this summary to establish the timing, duration, outcome, or treatment attribution for each listed event. Serious adverse event counts also do not capture every aspect of tolerability. These figures describe the posted safety information and should be interpreted in light of the trial’s follow-up, analysis, and reporting definitions.

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

The primary result estimates the average percentage change in fasting body weight from week 0 to week 56 in the stated analysis population. The sponsor-reported contrast is -6.06, and its 95% confidence interval is from -7.50 to -4.62. The result describes a group-level comparison after a low calorie diet run-in and randomisation to liraglutide 3.0 mg or placebo. It does not mean that each participant experienced the average change, nor does it describe weight change during the run-in.

The trial’s eligibility condition is important to the scope of the finding: randomisation followed a run-in, and subjects had to lose at least 5% of screening body weight. The registry’s stated target included obese non-diabetic subjects and overweight subjects with medical problems such as hypertension or dyslipidaemia. The result should not automatically be generalized to people who did not complete the run-in or meet the randomisation criterion, or to populations and settings not described in the registry record.

The primary endpoint is at week 56, even though the trial includes a 12-week follow-up period. The primary contrast does not establish what happened during follow-up or after it. Likewise, an average weight result does not answer every question about health, function, quality of life, or long-term outcomes. Those results are not established by this endpoint, and further details are not reported in the registry record.

A statistical difference and a clinically important difference are not identical concepts. The reported confidence interval and p-value describe statistical uncertainty under the analysis method; they do not, on their own, specify how important the contrast is to an individual or establish a benefit-risk balance. This interpretation does not amount to medical advice. The posted results should be read as evidence about the defined outcome, treatment comparison, timeframe, and analysis population, with the stated limitations around missing data, multiplicity, and incomplete model details.

18. Why This Trial Matters Statistically

This trial illustrates why the baseline and analysis period must be read carefully. The primary comparison begins at week 0, after a 12-week diet run-in, rather than at screening. Participants entered randomisation only after meeting the run-in weight-loss criterion. The resulting estimate therefore addresses change during the subsequent main trial among a selected group, not the combined effects of the run-in and randomized treatment from the original screening point.

It also shows the distinction between a model-based sponsor analysis and a simple recalculation from summary statistics. Both report a primary contrast of -6.06, but the confidence intervals and p-values differ because the ANCOVA and unadjusted recalculation do not use the same analysis approach. Presenting those distinctions helps readers see that an effect estimate is connected to its model, population, and assumptions rather than being a context-free number.

The registry record includes both continuous and threshold-based weight outcomes. These provide different statistical perspectives: the primary mean describes average change, while secondary classifications describe proportions crossing specified thresholds. Their broad alignment can inform interpretation, but the number of outcomes makes multiplicity and the status of secondary p-values important. The missing reporting of multiplicity procedures and detailed model specification limits how strongly the secondary findings can be treated as confirmatory.

Finally, the stated Last Observation Carried Forward approach makes missing-data assumptions part of the primary result. Knowing the analysis population and imputation method is essential to understanding what the estimate represents. The case demonstrates how a clear headline contrast can coexist with unanswered questions about missing observations, model details, and generalisability.

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 randomized population was selected after a 12-week low calorie diet run-in, with randomisation restricted to subjects who lost at least 5% of screening body weight. This limits direct generalisation to people who would not meet that criterion or who did not participate in a similar run-in. The target population described in the registry record includes obese non-diabetic subjects and certain overweight subjects with medical problems; broader applicability is not established by the posted primary result.

The primary analysis uses Last Observation Carried Forward, but the registry record does not report how many week 56 measurements were missing, why they were missing, or whether alternative missing-data analyses were conducted. The method requires assumptions about the unobserved values, and the posted summary does not show how sensitive the result is to those assumptions. The primary analysis population also excludes randomized participants without exposure or a post-randomisation efficacy assessment, as specified in its definition.

The registry record does not provide the covariates or full specification for the sponsor’s ANCOVA, the original sample-size justification, the detailed statistical analysis plan, or information about interim analyses. It also does not report whether multiplicity procedures were used for the many listed outcomes. These gaps limit independent review of the model and the confirmatory status of secondary findings. The simple recalculation cannot fill those gaps because it is based on summary statistics and an approximate standard deviation.

The primary outcome is measured at week 56, while the trial also includes follow-up through week 68. The primary comparison alone does not answer whether the result persisted during follow-up. Finally, the posted safety information gives participant counts and selected serious-event terms but does not provide all clinical details needed to assess timing, attribution, or outcomes of individual events. These limitations describe the information available in the registry record and do not imply that unreported procedures were absent.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Comparison of Liraglutide Versus Placebo in Weight Loss Maintenance in Obese Subjects: SCALE - Maintenance — NCT00781937.
• Registry record last updated 2017-11-01; status last verified 2017-09.
• Results first posted 2011-11-28.
• 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

NCT00781937 is a completed phase 3, randomized, double-blind, parallel-group trial comparing liraglutide 3.0 mg with placebo after a low calorie diet run-in. The registry record reports 422 enrolled participants and a design consisting of a 12-week run-in, a 56-week main trial, and a 12-week follow-up. Randomisation followed a loss of at least 5% of screening body weight.

The primary endpoint was mean percentage change in fasting body weight from week 0 to week 56. The sponsor reported an ANCOVA treatment contrast of -6.06, with a 95% confidence interval from -7.50 to -4.62 and p < 0.001. The primary analysis population used Last Observation Carried Forward and included randomized, exposed participants with at least one post-randomisation efficacy assessment.

Secondary outcomes cover weight-loss and weight-regain thresholds, maintenance of run-in loss, fasting weight change, and blood pressure. The registry record does not report several details needed for a full independent appraisal, including the original sample-size justification, full model specification, multiplicity procedures, and sensitivity analyses for missing data.