Prevention of Adult Caries Study (PACS): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Prevention of Adult Caries Study (PACS) (NCT00357877), a randomized phase 3 study. The primary outcome is Total Net D12FS Caries Increment (Total of Non-cavitated Lesions (D1), Cavitated Lesions (D2) and Sound Surfaces (S)).

2. Clinical Question

3. Trial Design

The Prevention of Adult Caries Study (PACS), registered as NCT00357877, was a Phase 3 randomized, quadruple-masked, parallel-group trial. It was sponsored by Tufts University and is marked completed in the registry record. The listed study start is 2006-07, and primary completion is listed as 2009-08. The trial enrolled 983 participants with caries, dental, as the condition of interest.

Participants were assigned to a placebo coating or a 10% w/v chlorhexidine acetate coating. The active coating contained chlorhexidine; the placebo coating contained the other ingredients without chlorhexidine. Both were applied topically by a dental professional supragingivally to the full dentition. The registry lists 493 participants in the placebo group and 490 in the active group for the primary outcome.

The primary outcome timeframe ran from the visit at randomization to the 13-month follow-up visit. The registry describes the main analysis population as intention to treat and states that missing values were handled using multiple imputation with 8 datasets generated via Markov Chain Monte Carlo sampling. The registry record does not report the trial’s original sample-size justification, the target effect used in its planning, or its planned power assumptions. Interim analyses and other details of the original statistical plan are not reported in the registry record.

The randomized, masked, parallel design provides the framework for comparing the two assigned coatings. The posted result is an average difference between groups on the stated outcome; it does not describe the outcome for every participant or establish that all participants had the same course. Interpretation also depends on the endpoint’s construction, the missing-data approach, and the precision of the estimated difference.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intention to treat (ITT) sample. Multiple imputation with 8 datasets imputed via Markov Chain Monte Carlo sampling was used to handle missing data.

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 Total Net D12FS Caries Increment (Total of Non-cavitated Lesions (D1), Cavitated Lesions (D2) and Sound Surfaces (S)), measured over (V1) to the 13 month follow-up visit.

The primary endpoint is total net D12FS caries increment, expressed in weighted increment units/13 months. It sums weighted transitions in tooth-surface status between randomization and the 13-month follow-up visit, combining root and coronal surfaces. The registry describes disease progression as a positive contribution and reversal as a negative contribution. For example, a transition from a sound surface to a non-cavitated lesion, or from a non-cavitated lesion to a cavitated lesion, has weight 1; a transition from sound directly to cavitated has weight 2; and a transition from a non-cavitated lesion back to sound has weight -1. No change, transitions involving missing or unscorable surfaces, and impossible transitions have weight 0. Incident fillings and crowns were scored like incident D2 lesions.

This construction is intended to represent both the direction and extent of changes in tooth-surface status, rather than simply counting a single type of lesion at follow-up. The study description notes that its duration was too short to observe progression from D2, cavitated lesions, to D3, cavitated lesions involving dentin; D2 and D3 were therefore treated equivalently for analysis. The endpoint is a continuous measure, so the reported group summaries are means and their standard errors.

Using a defined weighted measure gives the comparison a consistent outcome scale across participants and transitions. It also means that the result is specific to the endpoint’s scoring rules: it is not a direct count of all dental events or a measure of every possible aspect of oral health.

7. Results Posted in the Registry

• Active: mean 2.68 versus 2.43 on control; difference 0.25 (95% CI -0.64 to 1.14), standardized effect d = 0.04; z-test p = 0.581.

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.

Cumulative Net D12FS Caries Increment

• 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:
  • Active Dental Coating (CHX): mean 5.88 versus 5.53 on control; difference 0.35 (95% CI -0.83 to 1.53), standardized effect d = 0.04; z-test p = 0.560.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Dental Coating (CHX) and the control. The interval gives the range of effects that remain compatible with the data.

Total Crude D12FS Caries Increment

• 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:
  • Active Dental Coating (CHX): mean 5.96 versus 6.47 on control; difference -0.51 (95% CI -1.27 to 0.25), standardized effect d = -0.08; z-test p = 0.190.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Dental Coating (CHX) and the control. The interval gives the range of effects that remain compatible with the data.

Cumulative Crude D12FS Caries Increment

• 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:
  • Active Dental Coating (CHX): mean 10.72 versus 11.39 on control; difference -0.67 (95% CI -1.85 to 0.51), standardized effect d = -0.07; z-test p = 0.265.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Active Dental Coating (CHX) and the control. The interval gives the range of effects that remain compatible with the data.

Across the three posted secondary endpoints, the estimated differences do not present a single consistent direction. The two cumulative measures have higher estimated means in the active group, while the total crude measure has a lower estimated mean in the active group. The reported confidence intervals for these comparisons include zero. Taken together, the secondary results do not provide a clear, consistent pattern that changes the overall reading of the primary result.

The secondary outcomes assess related but distinct summaries of caries increment. Cumulative measures track outcomes across the listed visits, whereas total measures refer to the stated follow-up period. Crude and net measures also differ in how they represent transitions, including whether reversals are reflected in the net score. Their results therefore complement the primary endpoint but are not interchangeable with it.

These outcomes should be interpreted as supporting analyses, not as independent confirmation of a specific effect. The registry record lists three secondary endpoints and reports analyses for them. It does not report a multiplicity adjustment or a prespecified hierarchy for interpreting the secondary tests. When several outcomes are examined, the chance of finding an apparently notable result by chance can increase. General statistical explanation: adjustment procedures or a prespecified testing hierarchy can address that issue, but no such procedure is described in the registry record.

For that reason, unadjusted secondary p-values carry less weight than a result from a single, prespecified primary comparison, especially when they are considered together after the results are known. The secondary outcomes may help describe the broader pattern, but they should not be selected individually as decisive evidence without accounting for the number and relationship of analyses. Their overall pattern is best considered alongside the primary result and the uncertainty in all of the estimates.

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

• Cumulative Net D12FS Caries Increment
• Total Crude D12FS Caries Increment
• Cumulative Crude D12FS Caries Increment

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous, and the registry reports group means, standard errors, a mean difference, a confidence interval, and a p-value. The reported trial analysis used linear regression and labelled its estimate “Mean Difference (Net).” The registry record does not give further details about the regression model, such as which covariates were included or how the model handled any features of the outcome distribution. Those details are not reported in the registry record.

What does the mean difference represent?

The mean difference compares the average endpoint value in the active group with the average in the placebo group. The reported direction is active minus placebo: 0.25 means the active-group mean was estimated to be higher by 0.25 weighted increment units/13 months. A negative value would indicate a lower active-group mean. This is a group-level contrast, not a description of how much any one participant changed or an estimate of the number of participants whose outcome improved or worsened.

Because the endpoint combines weighted transitions, the difference is expressed in weighted increment units rather than as a simple count of affected teeth or surfaces. Its meaning depends on the scoring rules for progression, reversal, and other transitions. The endpoint definition also treats incident fillings and crowns as incident D2 lesions for scoring.

How should standard errors be understood?

A standard error describes the sampling uncertainty of an estimated mean. General statistical explanation: if a study were repeated under comparable conditions, an estimated mean would vary from sample to sample; the standard error represents the typical scale of that variation under the model. It is not the spread of individual participant outcomes. The primary group means are reported with standard errors of 0.32 in both groups.

For a difference between two independent groups, uncertainty in each group contributes to uncertainty in the contrast. The posted recalculation uses the group means and standard errors to produce an unadjusted comparison. It is not the same as fitting the trial’s regression model, and it cannot reproduce model details that are not reported in the registry record.

What does a confidence interval convey?

General statistical explanation: a confidence interval gives a range of estimates compatible with the observed data under the analysis assumptions. A wider interval indicates less precision than a narrower interval. If the interval includes zero for a difference, the data are compatible with no mean difference as well as with values on either side of zero. This does not prove that the true difference is zero. The reported primary interval is 95% and extends from -0.60 to 1.11; the unadjusted recalculation extends from -0.637 to 1.137.

A confidence interval is not a range containing a fixed proportion of individual outcomes, nor does it express the probability that the true effect lies in the particular interval after the data have been observed. It is interpreted through the statistical procedure that generated it. The registry identifies linear regression for the reported analysis but does not provide the complete model specification.

What does the p-value address?

General statistical explanation: a p-value measures how incompatible the observed data are with a specified null model, assuming that model and the analysis assumptions hold. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of a difference. The registry reports p = 0.56 for the primary regression analysis. The simple unadjusted recalculation gives p = 0.581. These values are close but arise from analyses that are not identical.

A p-value should be read alongside the estimate and interval. In this trial, the interval communicates that uncertainty covers values below and above zero, while the point estimate is 0.25. That combination is more informative than treating a thresholded p-value as a stand-alone verdict.

How were missing values handled?

The primary analysis is identified as intention to treat. The registry states that multiple imputation with 8 datasets imputed via Markov Chain Monte Carlo sampling was used to handle missing data. General statistical explanation: multiple imputation creates several completed versions of a dataset, analyzes each, and combines the results to account for uncertainty associated with missing values. The method depends on assumptions about the relationships between observed and missing data. The registry record does not report the imputation variables, diagnostics, or sensitivity analyses used to examine those assumptions.

The registry’s description gives the missing-data method for the primary endpoint. Whether the same population and imputation procedure were used for every secondary analysis is not reported in the registry record. Consequently, the primary missing-data description should not automatically be extended to other outcomes.

Why can outcome variability matter?

General statistical explanation: greater variability in individual outcomes makes group means less precise, all else equal. Over-dispersion is one possible issue when outcome variation exceeds what a chosen model expects; in that setting, standard errors and confidence intervals can be too small if the model does not account for the extra variation. The registry describes a continuous weighted increment and reports standard errors, but it does not report a dispersion assessment or a specific adjustment for over-dispersion. No conclusion about whether over-dispersion affected the reported inference can be drawn from the record alone.

11. Confidence Intervals and Precision

The primary estimate is an active-minus-placebo mean difference of 0.25 weighted increment units/13 months. The registry’s reported 95% confidence interval runs from -0.60 to 1.11. This range includes zero, which represents no mean difference, and also includes values on either side. In practical statistical terms, the posted result does not pin down the direction of the difference with precision.

The interval’s lower end is compatible with a lower mean in the active group, while its upper end is compatible with a higher mean. The point estimate sits within that range, but it should not be read as the only plausible value. The interval communicates uncertainty that is not visible from the point estimate alone. It also does not show where individual participant outcomes fall; it concerns uncertainty in the estimated group difference.

The unadjusted re-calculation gives a 95% confidence interval from -0.637 to 1.137. The recalculated interval is close to the registry’s reported interval, which is -0.60 to 1.11. The difference in endpoints reflects the use of an unadjusted calculation from posted summary statistics rather than the trial’s regression analysis. The registry’s own reported analysis remains the primary account of the trial result.

Standard errors for the primary group means are 0.32 for placebo and 0.32 for active. These quantify uncertainty in the respective means, not the spread of participant outcomes. The confidence interval for the contrast is the more direct measure of precision for the comparison between groups.

General statistical explanation: a confidence interval that spans zero can be consistent with a range of effects, not only with no effect. Precision and statistical significance are related but distinct: an interval can be used to see both the range of compatible estimates and whether it includes zero. Here, the interval leaves uncertainty about direction and magnitude. It should not be taken as proof of equivalence or as proof that a clinically important difference is absent; the registry record does not state an equivalence margin or a clinically important threshold.

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 participants per arm, not total events. They are illustrations based on the observed effect, not the trial’s original design assumptions. 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 set to the stated fraction of the observed effect. The trial’s own sample-size justification and original planning assumptions are not reported in the registry record.

For the primary endpoint, planning for 100% of the observed effect gives 12,642 participants per arm for 80% power and 16,923 participants per arm for 90% power. Planning for 75% of the observed effect gives 22,474 participants per arm for 80% power and 30,086 participants per arm for 90% power. Planning for 50% of the observed effect gives 50,565 participants per arm for 80% power and 67,692 participants per arm for 90% power.

In plain language, the 80% power column represents the calculated number of participants per arm under which the planning model would detect the chosen effect in 80% of repeated studies, if its assumptions hold. The 90% power column uses a higher chance of detecting that chosen effect and therefore gives a larger required sample size. For each power target, a smaller effect requires more participants because it is harder to distinguish from random variation.

These calculations are planning illustrations, not a measure of the trial’s observed power. They use an effect estimated from the trial itself, so they should not be treated as independent evidence that the observed result is reliable or as a retrospective verdict on the trial’s design. They also do not establish the sample size needed for a future study without consideration of its own design, assumptions, outcome, and analysis plan.

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 = 0.250; pooled standard deviation σ = 7.094.
2. n = 2 × (1.96 + 0.842)² × 7.094² / 0.250² = 12,641.1, rounded up to 12,642 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 <- 0.25                      # observed difference in means
sigma <- 7.09436887430944                     # 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(357877)
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 = 12,642 (80% power), 16,923 (90% power)
 75% of observed effect: n per arm = 22,474 (80% power), 30,086 (90% power)
 50% of observed effect: n per arm = 50,565 (80% power), 67,692 (90% power)
power.t.test: 12642.1 
Simulated power at n = 12,642 per arm: 0.803 (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 reported primary analysis used linear regression and gives a mean difference of 0.25, a 95% confidence interval from -0.60 to 1.11, and p = 0.56. The simple unadjusted re-calculation from the posted group summaries gives a mean difference of 0.25, a 95% confidence interval from -0.637 to 1.137, and p = 0.581. The estimates are close, and both intervals include zero.

The re-calculation is not a replacement for the trial’s analysis. It is a straightforward comparison based on the posted summary statistics and does not reproduce any regression features that may have been used in the reported analysis. The registry record identifies linear regression but does not give enough detail to establish whether covariate adjustment or other model choices account for the small differences in interval limits and p-values.

The two analyses support the same broad reading: the estimated difference is small relative to the uncertainty in the comparison, and the interval includes values on both sides of zero. The reported regression analysis should be used as the trial’s own result; the recalculated result is a transparent, unadjusted check rather than an alternative sponsor analysis.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary endpoint and three secondary endpoints, with four outcome measures and four analyses posted. The registry record does not report whether the testing plan included a multiplicity adjustment, a hierarchy for testing the outcomes, or another procedure to control the interpretation of multiple comparisons. This information matters because assessing several outcomes can make an isolated small p-value less persuasive than a result from a single prespecified test.

Multiplicity is not evidence that any particular finding is false. Rather, it is a feature of the analysis plan that affects how the collection of results should be interpreted. Without a reported adjustment or testing hierarchy, the secondary p-values are most appropriately treated as exploratory or supportive rather than as stand-alone confirmatory evidence. The primary result remains the main reported comparison, and its confidence interval includes zero.

For the primary outcome, the registry states that the analysis used the intention-to-treat sample and that missing data were handled by multiple imputation with 8 datasets generated via Markov Chain Monte Carlo sampling. General statistical explanation: imputation can preserve participants in an analysis while representing uncertainty about missing values, but it relies on assumptions about how missingness relates to observed and unobserved information. The imputation variables, assumptions, and sensitivity analyses are not reported in the registry record.

The primary analysis population is stated, but the registry record does not fully describe analysis populations for every secondary outcome. Whether the same imputation approach was applied to those outcomes, how estimates across imputed datasets were combined, and whether any other missing-data analyses were performed are not reported in the registry record. The record also does not report interim analyses, the original sample-size justification, or a full statistical analysis plan. Those omissions limit the ability to assess prespecification and robustness beyond the information described for the primary endpoint.

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: angina pectoris (0.4%), pneumonia (0.4%), uterine leiomyoma (0.4%), kidney infection (0.4%), acute myocardial infarction (0.2%).

The registry reports serious adverse events for 31 of 493 participants in the placebo group and 28 of 490 participants in the active group. Other adverse events are reported for 87 of 493 participants in the placebo group and 84 of 490 participants in the active group. These are group-level counts of participants identified in the adverse-event record; they do not by themselves describe the event’s timing, duration, severity beyond the serious classification, or relationship to the coating.

The serious-event terms listed include angina pectoris, pneumonia, uterine leiomyoma, and kidney infection, each reported as 0.4%, and acute myocardial infarction and asthma, each reported as 0.2%. The registry’s list does not specify a group for each percentage in the information available, so the percentages should not be read as direct group-specific comparisons.

Death counts are not reported in the registry record. The event figures are descriptive safety information and do not establish that an event was caused by either coating. The registry record does not provide enough information to infer causality or to compare risk statistically from these counts alone.

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

The primary result estimates an average difference in a defined weighted caries-increment measure over the period from randomization to the 13-month follow-up. The active group’s reported mean was 2.68 and the placebo group’s mean was 2.43. The registry’s regression estimate for the active-minus-placebo difference was 0.25, with a 95% confidence interval from -0.60 to 1.11 and p = 0.56. The interval includes zero and values in both directions, so the result leaves uncertainty about the direction and size of the mean difference.

This does not show that the groups were equivalent. Demonstrating equivalence would require a defined equivalence margin and an analysis designed to assess it; neither is reported in the registry record. Nor does the result prove that no difference exists. It indicates that the posted analysis did not estimate the difference precisely enough to rule out a range of possibilities represented by the confidence interval.

The finding is specific to the endpoint’s weighted scoring rules, the two assigned coatings, the trial population, and the 13-month follow-up. It does not directly describe longer-term outcomes, every type of dental decay, all possible benefits or harms, or the experience of an individual participant. The registry describes an intention-to-treat primary analysis and a multiple-imputation approach for missing data, but the assumptions and sensitivity analyses for that procedure are not reported.

The secondary outcomes provide related information but do not establish a consistent direction across all measures, and the registry record does not report multiplicity handling. Safety counts are descriptive and do not establish causation. Overall, the posted results support a careful statement about an uncertain group-average comparison, not a categorical claim about whether the coating works or does not work.

18. Why This Trial Matters Statistically

This trial illustrates why a clinical comparison should be read through its estimate and uncertainty, not only through a p-value or the direction of the group means. The primary point estimate is positive, but its confidence interval includes zero and extends to values on both sides. The data therefore leave uncertainty about the direction and magnitude of the mean difference. Reporting the interval makes that uncertainty visible.

The endpoint also shows how outcome construction affects interpretation. The primary measure combines weighted progression and reversal across tooth surfaces, and treats incident fillings and crowns as incident D2 lesions. Such a measure can capture more than a single end-of-study lesion count, but it requires readers to understand the weights and the transitions included. Its results are meaningful on the defined scale, not automatically on every other measure of caries.

The record also offers a practical example of the distinction between a trial’s reported analysis and a simple unadjusted re-calculation. The estimates are close, yet their intervals and p-values are not identical. A summary-statistic re-calculation can serve as a check, but it cannot recreate unreported details of a regression model or substitute for the analysis the trial reported.

Finally, the posting identifies multiple secondary outcomes and a missing-data method for the primary analysis, while leaving other planning and analysis details unreported. This makes the boundary between what can be assessed and what remains uncertain important. Statistical interpretation depends not only on the numerical result, but also on endpoint definitions, missing-data assumptions, multiplicity handling, and clear reporting of the analysis plan.

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 details of the original design and analysis unclear. The trial’s original sample-size justification, target effect, and planned power assumptions are not reported in the registry record. The planning figures based on the observed effect are therefore not evidence about what the investigators planned before the trial or whether the enrolled sample met an original design target.

The primary endpoint is a weighted composite of transitions in tooth-surface status. Its interpretation depends on the scoring rules, including the treatment of D2 and D3 lesions as equivalent for analysis, the weights for progression and reversal, and the scoring of incident fillings and crowns. The record notes that study duration was too short to observe progression from D2 to D3. The 13-month timeframe consequently does not answer questions about longer-term progression.

The registry identifies linear regression for the reported primary analysis but does not report the full model specification, including covariates or other modeling choices. It also does not report diagnostic assessments of the outcome distribution or sensitivity analyses that would help assess robustness. The primary missing-data procedure is described as multiple imputation using 8 datasets and Markov Chain Monte Carlo sampling, but imputation variables, assumptions, diagnostics, and sensitivity analyses are not reported in the registry record.

Multiplicity handling for the secondary outcomes is not reported. The registry record also does not provide enough detail to determine whether secondary analyses used the same analysis population and missing-data procedure as the primary outcome. These reporting gaps constrain assessment of the secondary findings and of the robustness of the overall analysis.

Safety information is presented as participant counts and a list of serious-event terms with percentages. Death counts are not reported in the registry record, and the posted event information does not establish causality. These limitations call for an interpretation confined to the posted outcomes, timeframes, and analyses.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Prevention of Adult Caries Study (PACS) — NCT00357877.
• Registry record last updated 2017-06-14; status last verified 2017-05.
• Results first posted 2017-06-14.
• 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

NCT00357877, the Prevention of Adult Caries Study (PACS), is a completed Phase 3 randomized, quadruple-masked, parallel-group trial sponsored by Tufts University. It enrolled 983 participants and compared a 10% w/v chlorhexidine acetate dental coating with a placebo coating. The primary endpoint was total net D12FS caries increment from randomization to the 13-month follow-up.

The registry reports an active-minus-placebo mean difference of 0.25 weighted increment units/13 months, with a 95% confidence interval from -0.60 to 1.11 and p = 0.56. The interval includes zero and values in both directions. The primary analysis is described as intention to treat, with missing data handled through multiple imputation using 8 datasets generated via Markov Chain Monte Carlo sampling.

Three secondary endpoints and adverse-event counts are also posted. The registry record does not report the original sample-size justification, multiplicity procedure, or full statistical model specification.