Safety and Efficacy Study of Isolagen Therapy in the Treatment of Nasolabial Fold Wrinkles: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy Study of Isolagen Therapy in the Treatment of Nasolabial Fold Wrinkles (NCT00655356), a randomized phase 3 study. The primary outcome is Subject Wrinkle Assessment Responders.

2. Clinical Question

3. Trial Design

NCT00655356 is a completed Phase 3 study titled “Safety and Efficacy Study of Isolagen Therapy in the Treatment of Nasolabial Fold Wrinkles.” The registry lists randomized allocation, triple masking, and a parallel model. It records 218 participants enrolled, with the study starting in 2006-11 and primary completion in 2008-07. The sponsor is Castle Creek Biosciences, LLC.

The intervention groups were Autologous Human Fibroblasts (azficel-T), described in the results as Autologous Fibroblasts, and placebo. The condition is bilateral nasolabial fold wrinkles. The registry record gives no descriptions for the individual arm details, so practical features of the treatment and comparator beyond their names are not available in this record.

The primary outcome was Subject Wrinkle Assessment Responders. Its stated assessment period was baseline, before the first treatment, and six months after final treatment. The registry also lists Subject Wrinkle Assessment Responders and Evaluator Wrinkle Severity Assessment Responders among secondary outcomes; details of the secondary results are not needed to interpret the primary comparison.

Randomization and masking describe important elements of trial design, but the record does not report the sequence-generation procedure, concealment process, or which people were masked. It also does not report the trial's original sample-size justification or other original design assumptions. The present estimates describe the posted outcome data and do not fill in those design details.

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: Patients were enrolled and biopsied for manufacture of study product. All randomized patients were included in the Intent to Treat (ITT) population.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis population was the ITT population, defined as all randomized subjects.

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 Subject Wrinkle Assessment Responders, measured over Baseline (prior to first treatment) and 6 months post final treatment.

The primary endpoint counts participants who met a responder definition on the Subject Wrinkle Assessment. The scale was a five point assessment of the wrinkles of the lower part of the face, ranging from -2, Very Dissatisfied, to +2, Very Satisfied. A two point improvement from baseline at six months after final treatment qualified a participant as a responder. The posted results give 50 responders among 110 participants assigned to Autologous Fibroblasts and 19 among 108 assigned to placebo.

A responder endpoint turns a graded rating into a participant-level event: each person either meets the stated improvement threshold or does not. This makes it possible to compare the proportion meeting a prespecified type of meaningful change, rather than comparing only average scale scores. The threshold and the participant's own assessment are central to what the outcome represents. It captures the subject's judgment of wrinkle change according to this scale; it is not interchangeable with an evaluator rating or a continuous measure of wrinkle severity.

Using a defined endpoint also makes the comparison easier to communicate, but it discards information about the size of change among people who do not cross the threshold and among those who do. The registry record does not report additional details about validation or selection of the responder threshold.

7. Results Posted in the Registry

• Autologous Fibroblasts: 45.5% (50/110) versus 17.6% (19/108) on control; risk difference 27.9 percentage points (95% CI 16.1 to 39.6), risk ratio 2.58 (1.64–4.08); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

What does the participant count represent?

The endpoint is binary: a participant either met the responder definition or did not. The reported values are numbers of participants, not mean scores or counts of repeated wrinkle assessments. The active group has 50 responders among 110 participants, and the placebo group has 19 responders among 108 participants. Comparing the proportions accounts for the fact that the group denominators differ. The event definition remains important: the comparison concerns a two point improvement on the subject's scale at the specified follow-up, not any improvement whatsoever.

What is the risk difference?

The risk difference expresses the contrast between the two event proportions on an absolute scale. Its recalculated estimate is 0.279, and its 95% confidence interval runs from 0.161 to 0.396. A risk difference of zero corresponds to equal event proportions; values above zero indicate a larger observed proportion in the intervention group under the stated group ordering. This measure keeps the contrast in terms of the outcome probability rather than converting it into a relative comparison. It is unadjusted, meaning it is calculated directly from the posted group counts rather than from a model controlling for additional variables.

How do relative measures differ?

The risk ratio compares the event proportion in the intervention group with the event proportion in the comparator group. The recalculated risk ratio is 2.584, with a confidence interval from 1.637 to 4.079. A value of one represents equal event proportions on this scale. The odds ratio compares odds rather than probabilities; its estimate is 3.904, with a confidence interval from 2.097 to 7.267. A value of one similarly represents equal odds. Odds and probabilities are different quantities, so the odds ratio should not be read as though it were a risk ratio, especially when the event is not rare.

What does a confidence interval communicate?

A confidence interval describes uncertainty around an estimated effect under the assumptions of the method used. It is not a range containing a known fixed proportion of individual participant outcomes. For the risk difference, the interval from 0.161 to 0.396 does not include zero, which indicates that the estimate is distinguishable from no difference under the conventional two-sided 0.05 testing framework used for the planning illustration. The interval still allows a range of effect sizes; the point estimate should not be treated as perfectly precise.

How should the reported p-values be understood?

The unadjusted recalculation reports p < 0.001. The registry's reported analysis gives p <.00001 using the Cochran-Mantel-Haenszel method. A p-value describes how incompatible the observed data are with a specified null hypothesis under the analysis assumptions; it is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. The p-values are both small, but they come from differently described analyses and should not be compared as if they were calculated from an identical procedure.

Why can adjustment matter?

In general, the Cochran-Mantel-Haenszel method combines comparisons across strata, which can account for categorical stratification in an analysis. The record does not report the strata used, whether adjustment variables were specified in advance, or the detailed model implementation. It is therefore not possible to reconstruct the sponsor's calculation from the method label alone. The directly recalculated risk difference, risk ratio, and odds ratio are unadjusted summaries of the posted group counts.

Why does over-dispersion matter in some outcomes?

General statistical context: for count outcomes, a model that assumes a particular variance can understate uncertainty when the observed variation is greater than that assumption allows; this is called over-dispersion. The primary outcome in this trial is binary participant response, not a rate of recurrent events or a count of multiple events per person. The registry does not report a dispersion parameter for this endpoint, and no over-dispersion adjustment is part of the stated primary comparison.

All these methods describe a group-level comparison. They do not establish that the intervention caused each observed response, explain why a particular participant did or did not respond, or measure outcomes beyond those recorded. Interpretation also depends on how randomization, masking, outcome ascertainment, and missing assessments were handled, and several of those implementation details are not reported in the registry record.

10. Confidence Intervals and Precision

The unadjusted risk-difference estimate is 0.279, with a 95% confidence interval from 0.161 to 0.396. The interval is wholly on the positive side of zero, supporting a higher observed responder proportion in the Autologous Fibroblasts group for this endpoint. Its width also signals that the exact size of the group difference is not known with certainty. The estimate is more informative when read together with the interval than when read as a point value alone.

The relative summaries have their own uncertainty intervals: the risk ratio is 2.584, with a confidence interval from 1.637 to 4.079, and the odds ratio is 3.904, with a confidence interval from 2.097 to 7.267. These intervals are expressed on different effect scales, so their endpoints should not be compared directly with those for the risk difference. Each interval describes uncertainty about its own estimated measure.

The registry's Cochran-Mantel-Haenszel analysis lists a 95% confidence level but does not report an effect estimate or interval limits. A reported confidence level alone does not reveal the precision of that sponsor analysis. The intervals available for interpretation are the recalculated unadjusted intervals, based on the posted group counts. They do not reflect adjustment for unreported factors, uncertainty arising from missing outcome data, or possible limitations in outcome ascertainment.

11. 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 participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. For planning for 100% of the observed effect, the figures are 43 participants per arm for 80% power and 57 participants per arm for 90% power. For planning for 75% of the observed effect, they are 72 and 96 participants per arm, respectively. For planning for 50% of the observed effect, they are 149 and 199 participants per arm, respectively.

The calculations are planning illustrations based on the observed effect, not the trial's original design justification. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set to the stated fraction of the observed effect. The trial's own original sample-size assumptions are not reported in the registry record.

In plain terms, 80% power describes the probability, under the planning assumptions and for the specified effect, of detecting an effect at the stated significance level. The 90% power column uses a higher target probability of detection, and therefore calls for more participants for each listed effect. Smaller effects require more participants to detect than larger effects under the same assumptions. These figures are planning illustrations, not observed power; they do not describe a probability that the trial's result is correct or a retrospective measure of what the completed study achieved.

12. 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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.176; treatment proportion p₁ = 0.455; average p̄ = 0.315.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6571 = 1.2878.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6268 = 0.5275.
4. n = (1.2878 + 0.5275)² / (0.2786)² = 42.5, rounded up to 43 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
p0 <- 0.175925925925926                          # control proportion
p1 <- 0.454545454545455                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(655356)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 43 (80% power), 57 (90% power)
 75% of observed effect: n per arm = 72 (80% power), 96 (90% power)
 50% of observed effect: n per arm = 149 (80% power), 199 (90% power)
power.prop.test: 42.5 
Simulated power at n = 43 per arm: 0.815 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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.

13. Analyses Reported by the Sponsor

The registry's own reported comparison used a Cochran-Mantel-Haenszel analysis and reported p <.00001. It does not provide an effect estimate or confidence interval limits for that analysis. The recalculated risk difference, risk ratio, odds ratio, confidence intervals, and p-value are unadjusted calculations from the posted group counts. They give interpretable effect-size summaries, but they do not reproduce the sponsor's analysis method.

The registry record does not identify the strata or other implementation details for the Cochran-Mantel-Haenszel analysis, so the reported result cannot be independently reconstructed from the method label and group totals alone. The sponsor analysis and unadjusted recalculation should be considered related but distinct summaries. The registry does not flag a discrepancy between the posted result and the recalculation. The sponsor's reported analysis remains the trial's own inferential result, while the recalculated estimates offer a transparent unadjusted view of the posted counts.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists Subject Wrinkle Assessment Responders and Evaluator Wrinkle Severity Assessment Responders as secondary outcomes, alongside the primary Subject Wrinkle Assessment endpoint. It also records four outcome measures and two analyses. These details indicate that the posted results include more than a single outcome or analysis, but they do not establish how many comparisons were planned or tested across the full protocol.

Multiplicity arises when several outcomes, time points, treatment contrasts, or analyses are considered. In general, repeated testing can increase the chance of obtaining at least one small p-value by chance if no adjustment or prespecified hierarchy is used. The registry record does not report a multiplicity adjustment, testing hierarchy, or whether the primary outcome was protected from other comparisons. It is therefore not possible to determine from this record how multiplicity was handled.

The primary analysis population is described as the ITT population, defined as all randomized subjects. However, the record does not report the number of randomized subjects with missing primary outcome data, the reasons for missingness, or the method used to handle missing assessments. It does not state whether a missing assessment was counted as nonresponse, imputed, excluded, or treated another way. The posted counts should be interpreted without assuming an undocumented missing-data rule.

The record also does not provide a full statistical analysis plan, detailed inclusion rules for each analysis, or information about protocol deviations. Randomization supports a comparison based on assigned groups when implemented and analyzed appropriately, but the available record does not document all aspects needed to evaluate those procedures. These omissions limit conclusions about the exact estimand and the robustness of the primary comparison to analytic choices.

15. 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: menorrhagia (2.0%), spinal column stenosis (1.0%), staphylococcal infection (1.0%), dysmenorrhoea (1.0%), endometriosis (1.0%).

The registry reports serious adverse events affecting 8 of 98 participants at risk in the Autologous Fibroblasts group and 7 of 99 participants at risk in the placebo group. It also reports other adverse events affecting 32 of 98 and 33 of 99 participants, respectively. These are counts of participants with events and their listed denominators; they are not rates per unit of follow-up time. The record does not explain why the safety denominators differ from the primary outcome denominators.

The listed serious-event terms and percentages are Menorrhagia, 2.0%; Spinal Column Stenosis, 1.0%; Staphylococcal Infection, 1.0%; Dysmenorrhoea, 1.0%; Endometriosis, 1.0%; and Muscular Weakness, 1.0%. The record does not attach these terms to a particular treatment group in the information reported for these terms, so they should not be attributed to either group on this basis.

Deaths are not reported in the registry record. These summaries describe reported event counts; they do not establish whether an event was caused by treatment or provide enough detail to compare event severity, timing, duration, or recurrence.

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

The posted primary results show more participants meeting the defined Subject Wrinkle Assessment responder criterion in the Autologous Fibroblasts group than in the placebo group. The unadjusted risk difference and its confidence interval quantify the size and uncertainty of that observed contrast. The sponsor-reported Cochran-Mantel-Haenszel analysis also reports a small p-value. Together, these results describe statistical evidence of a group difference in this particular endpoint at the stated assessment time.

The finding is specific to a subject-rated two point improvement on the five point scale, assessed six months after final treatment relative to baseline. It does not describe the average amount of wrinkle change, establish that every participant improved, or tell how participants who did not meet the responder threshold changed. It is also distinct from the evaluator-rated secondary endpoint, for which the details needed to interpret a result are not included in the primary comparison.

A statistically detectable contrast is not, by itself, a complete judgment of clinical importance. The record does not provide context for the responder threshold's meaning to participants, nor does the primary estimate summarize safety or all outcomes. The serious and other adverse-event counts should be considered separately, with the limits of their reporting in mind. The outcome contrast does not determine the cause of individual events or settle an overall benefit-harm assessment.

The record describes a randomized, triple-masked parallel trial, but several implementation details and analytic decisions are not reported. Missing-outcome handling, multiplicity procedures, the sponsor analysis strata, and the original sample-size justification are not reported in the registry record. Accordingly, the results support interpretation of the posted comparison but cannot answer every question about robustness, generalizability, or the full trial protocol.

17. Why This Trial Matters Statistically

This trial offers a useful example of why an endpoint definition matters as much as the name of a scale. The primary outcome was not simply a wrinkle score: it classified participants according to whether their subject-rated score improved by a specified amount from baseline at a defined follow-up. The resulting counts make the clinical question concrete, while also narrowing the inference to participants who cross that threshold.

The results also illustrate why absolute and relative measures should be read together. A risk difference expresses the contrast in event proportions, while risk and odds ratios express relative contrasts on different scales. Confidence intervals then show uncertainty for each effect measure. Looking only at a p-value would conceal much of this information, including the estimated size and precision of the group difference.

A second statistical lesson is that a registry may report a sponsor analysis without giving all the details needed to reproduce it. Here, the named Cochran-Mantel-Haenszel method and p-value are available, but the corresponding effect estimate, interval limits, and analysis strata are not. The unadjusted recalculation adds an effect-size perspective, but it does not replace or reproduce the sponsor's procedure.

Finally, sample-size illustrations based on an observed effect are useful for understanding how target effect size and desired power affect planning, but they are not evidence about the trial's original design or achieved power. Distinguishing prospective planning from retrospective inference prevents a common misreading of power calculations. Taken together, endpoint definition, effect scale, uncertainty, analysis transparency, and planning assumptions shape a careful statistical reading of this trial.

18. 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.

Several trial-specific details needed for a full assessment are not reported in the registry record. The record labels the study randomized and triple masked but does not identify who was masked, describe sequence generation or allocation concealment, or explain the procedures used to maintain masking. Arm descriptions are blank, limiting detail about the intervention and placebo beyond their names.

The record does not report the original sample-size calculation or its assumptions. The planning illustrations based on the observed effect cannot be used to infer what the investigators planned before enrollment. The sponsor's analysis is identified as Cochran-Mantel-Haenszel, but the strata, detailed implementation, effect estimate, and confidence interval limits are not reported. This prevents a complete reconstruction of the sponsor analysis from the record.

Although the primary population is defined as all randomized subjects, missing outcome counts and handling methods are not reported. The record also does not describe multiplicity control, the analysis hierarchy, or whether other analyses were prespecified. These omissions make it difficult to assess sensitivity to missing data and repeated testing.

For safety, the record gives serious and other adverse-event counts with denominators, but does not report deaths, explain the safety denominators, or provide detailed timing, duration, severity, or group attribution for the listed serious-event terms. The primary endpoint is a subject assessment and is threshold-based; the posted comparison therefore does not convey the full distribution of scores or all aspects of treatment experience. These limits apply to interpretation of the registry record and do not establish that any unreported procedure or outcome was absent from the trial.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Safety and Efficacy Study of Isolagen Therapy in the Treatment of Nasolabial Fold Wrinkles — NCT00655356.
• Registry record last updated 2012-03-13; status last verified 2012-02.
• Results first posted 2012-03-13.
• 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.

22. Record Summary

NCT00655356 is a completed randomized, triple-masked, parallel Phase 3 trial sponsored by Castle Creek Biosciences, LLC. It enrolled 218 participants with bilateral nasolabial fold wrinkles and compared Autologous Human Fibroblasts (azficel-T), described in the outcome results as Autologous Fibroblasts, with placebo.

The primary outcome was Subject Wrinkle Assessment Responders at six months after final treatment. The posted counts were 50 of 110 participants in the Autologous Fibroblasts group and 19 of 108 in the placebo group. The unadjusted risk difference was 0.279, with a 95% confidence interval from 0.161 to 0.396. The registry's reported Cochran-Mantel-Haenszel analysis gave p <.00001.

The record does not report important details including missing-data handling, multiplicity procedures, analysis strata, and the original sample-size justification. The posted estimates describe the defined primary endpoint and should be interpreted with these reporting limits in view.