In Situ Caries Efficacy of Fluoride Toothpastes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for In Situ Caries Efficacy of Fluoride Toothpastes (NCT00708097), a randomized phase 3 study. The primary outcome is Percentage Surface Microhardness Recovery (%SMHR) of Sound Enamel Specimens Exposed to NaF Toothpaste (1450ppmF) and NaF Toothpaste (1400ppmF).

2. Clinical Question

The study was designed to assess the effect of fluoride dentifrices on enamel with artificial caries lesions in an in situ model. Its results therefore address measured changes in enamel specimens under the study conditions. They do not, by themselves, establish whether the toothpaste formulations change clinical caries outcomes in patients.

3. Trial Design

The registry describes a completed, Phase 3, randomized, single-masked crossover study sponsored by GlaxoSmithKline. Enrollment was 57. The study began in 2008-04, and primary completion was recorded as 2008-07. The record identifies caries as the condition and describes the model as an in situ caries model.

Five toothpaste formulations are described: NaF toothpaste containing 1450ppmF with carbopol, NaF toothpaste containing 1400ppmF, reference NaMFP/NaF toothpaste containing 1450ppmF, NaF toothpaste containing 675ppmF, and fluoride-free placebo toothpaste containing 0ppmF. The primary endpoint compares the 1450ppmF and 1400ppmF NaF toothpastes in sound enamel specimens. Secondary outcomes include surface microhardness recovery in sound and demineralized specimens, and fluoride uptake in both specimen types.

A crossover design means participants receive more than one study condition in a planned sequence, so comparisons may draw on observations from the same participant. That is a general description of the design; the registry record does not report the treatment sequence, the timing or length of any washout, or how the analysis handled within-participant correlation. The masking is described as single, but who was masked is not reported in the registry record.

Although 57 participants were enrolled, the efficacy group sizes vary by treatment. The registry states that dropout led to differences in the number of participants per treatment group. The primary population included randomized participants with at least one post-baseline efficacy assessment, and missing data were not imputed. Further details about recruitment, eligibility criteria, allocation concealment, and the original design assumptions 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: Two days before the start of each treatment period, participants received a professional dental cleaning of their natural teeth, then brushed at home with the fluoride-free toothpaste.

Period 1

Reasons for not completing this period:

Period 2

Reasons for not completing this period:

Period 3

Period 4

Analysis population for the primary outcome, as defined in the registry: Intent to Treat (ITT) population: All randomized participants who had at least one post baseline efficacy assessment. Missing data was not imputed. Due to drop outs there were differences in the number of participants (N) per treatment group.

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 Percentage Surface Microhardness Recovery (%SMHR) of Sound Enamel Specimens Exposed to NaF Toothpaste (1450ppmF) and NaF Toothpaste (1400ppmF), measured over Baseline to 14 days.

The primary endpoint is percentage surface microhardness recovery (%SMHR) of sound enamel specimens, assessed from baseline to 14 days. Surface microhardness testing was used to assess the mineral status of partially demineralized enamel specimens. A Wilson 2100 Hardness tester measured the lengths of indentations in the enamel. The record explains that longer indentations relative to baseline indicate softening or demineralization, while shorter indentations indicate rehardening or remineralization of the enamel surface.

The percentage is calculated from indentation values at baseline, after intra-oral exposure on Day 14, and after in-vitro demineralization on Day 14, using the stated formula (D-R)/(D-B)*100. It expresses the recovery measure relative to the specimen’s baseline and demineralized measurements. In general, a continuous endpoint such as this preserves information about the measured range of responses rather than reducing the result to a yes-or-no event.

This endpoint is useful for the stated model because it quantifies a physical enamel property linked in the registry description to softening and rehardening. It is nevertheless a specimen-level laboratory measure in an in situ model. It is not a direct measure of new cavities, progression of clinical caries, symptoms, or other patient-level outcomes. Its meaning should therefore be kept within the scope of the experimental model and the assessment period recorded.

7. Results Posted in the Registry

• NaF Toothpaste (1400ppmF): mean 26.32 versus 28.31 on control; difference -1.99 (95% CI -17.30 to 13.32), standardized effect d = -0.05; z-test p = 0.799.

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 SMHR of Sound Enamel Specimens Exposed to NaF Toothpaste (1450ppmF), NaF Toothpaste (1400ppmF), NaMFP/NaF Toothpaste (1450ppmF), NaF Toothpaste (675ppmF) and Placebo Toothpaste (0ppmF)

• 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:
  • NaF/Carbopol Toothpaste(1450ppmF): mean 28.31 versus 53.38 on control; difference -25.07 (95% CI -40.40 to -9.74), standardized effect d = -0.64; z-test p = 0.001.
  • NaF Toothpaste(1400ppmF): mean 26.32 versus 53.38 on control; difference -27.06 (95% CI -42.35 to -11.77), standardized effect d = -0.68; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for NaF/Carbopol Toothpaste(1450ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for NaF Toothpaste(1400ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Percentage SMHR of Demineralized Enamel Specimens Exposed to NaF Toothpaste (1450ppmF), NaF Toothpaste (1400ppmF), NaF Toothpaste (675ppmF), NaMFP/NaF Toothpaste(1450ppmF) and Placebo Toothpaste (0ppmF)

• 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:
  • NaF Toothpaste(1450ppmF): mean 34.62 versus 11.10 on control; difference 23.52 (95% CI 15.73 to 31.31), standardized effect d = 1.18; z-test p = <0.001.
  • NaF Toothpaste(1400ppmF): mean 36.06 versus 11.10 on control; difference 24.96 (95% CI 17.21 to 32.71), standardized effect d = 1.24; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for NaF Toothpaste(1450ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for NaF Toothpaste(1400ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Enamel Fluoride Uptake (Sound Enamel Specimens)

• 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:
  • NaF Toothpaste (1450ppmF): mean 648.06 versus 425.59 on control; difference 222.47 (95% CI 92.51 to 352.43), standardized effect d = 0.67; z-test p = <0.001.
  • NaF Toothpaste (1400ppmF): mean 650.07 versus 425.59 on control; difference 224.48 (95% CI 95.63 to 353.33), standardized effect d = 0.67; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for NaF Toothpaste (1450ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for NaF Toothpaste (1400ppmF) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes broaden the assessment beyond the primary comparison of sound enamel with two NaF formulations. They include surface microhardness recovery for sound and demineralized specimens across multiple formulations, plus fluoride uptake in both specimen types. These are related measures of enamel response, but they are not interchangeable endpoints: microhardness recovery and fluoride uptake quantify different properties, and sound and demineralized specimens represent different starting conditions.

Across the posted secondary results, the patterns vary by specimen type and outcome. For sound specimens, the listed SMHR summaries do not form a simple pattern in which every fluoride toothpaste has a higher value than placebo. For demineralized specimens, the summary pattern differs, with fluoride toothpaste values above the placebo value. Fluoride uptake summaries also show higher values for the fluoride toothpaste groups than for placebo. These broad patterns provide context for the primary result, but they do not change what the primary comparison tested.

The sponsor-reported secondary ANOVA comparisons are not uniformly statistically clear: some confidence intervals include zero and some do not. These analyses address contrasts beyond the primary question, and the registry record lists multiple outcomes and analyses. Unless a multiplicity strategy was prespecified and applied, unadjusted secondary p-values carry less weight than a single prespecified primary test. A nominally small p-value among many comparisons can arise by chance, and the registry record does not report a multiplicity adjustment plan.

The secondary findings should therefore be treated as supporting descriptive evidence about this model, not as independent confirmation of a clinical effect. They may help identify patterns for further study, but they do not establish an overall ranking of formulations or a patient-level benefit. Their interpretation is also limited by the same missing information about analysis models, missing observations, and crossover handling that applies to the primary endpoint.

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

• Percentage SMHR of Sound Enamel Specimens Exposed to NaF Toothpaste (1450ppmF), NaF Toothpaste (1400ppmF), NaMFP/NaF Toothpaste (1450ppmF), NaF Toothpaste (675ppmF) and Placebo Toothpaste (0ppmF)
• Percentage SMHR of Demineralized Enamel Specimens Exposed to NaF Toothpaste (1450ppmF), NaF Toothpaste (1400ppmF), NaF Toothpaste (675ppmF), NaMFP/NaF Toothpaste(1450ppmF) and Placebo Toothpaste (0ppmF)
• Enamel Fluoride Uptake (Sound Enamel Specimens)
• Enamel Fluoride Uptake (Demineralized Specimens)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a percentage surface microhardness recovery measure. For continuous outcomes, a group mean summarizes the average measured response, while a standard error describes the estimated precision of that mean. The posted results identify the group estimates as least-squares means and their dispersion as standard errors. A least-squares mean is a model-based mean; it can reflect adjustment in the analysis rather than a simple arithmetic average. Details of the model are not reported in the registry record.

What a mean difference describes

A mean difference expresses the contrast between two group means in the units of the outcome. The recalculated primary contrast is -1.99 for 1400ppmF relative to 1450ppmF, based on the posted group summaries. Its negative direction reflects that contrast ordering. The registry's adjusted result is reported as an adjusted mean difference of 1.98 for the named groups. Since the sponsor result is adjusted and the recalculation is a simple unadjusted calculation, the estimates need not match. Their interpretation depends on the contrast direction and analysis specification.

In general, a confidence interval gives a range of values compatible with the observed data and the statistical model at the stated confidence level. It conveys uncertainty about the estimated contrast; it is not a range that contains a fixed proportion of individual observations. For the primary recalculation, the 95% confidence interval runs from -17.3 to 13.32. The sponsor's 95% interval runs from -5.00 to 8.97. Each includes zero, meaning that a zero difference is among the values compatible with the respective analysis at that confidence level. The registry's p-values are also reported, but a p-value does not measure the size or clinical importance of an effect.

Why adjustment and crossover structure matter

In a crossover study, an individual may contribute observations under more than one treatment condition. In general, this can allow treatment comparisons to use within-person information, but the analysis needs to account appropriately for the repeated measurements and design. The registry records the crossover model and reports ANOVA results, but does not report the treatment sequences, any washout, the model terms, or the method used to account for within-participant correlation. It is therefore not possible to reconstruct the sponsor's adjusted result from the registry record alone.

Randomization is intended to reduce systematic differences in treatment assignment. Masking can reduce some forms of bias in assessment or conduct, depending on who is masked. The record describes the study as randomized and single-masked, without reporting allocation concealment or specifying who was masked. These labels provide design context but do not resolve the missing analysis details.

Reading the recalculated statistics

The recalculated mean differences and confidence intervals are unadjusted summaries formed from the posted means and standard errors. They are not the sponsor's model-based ANOVA. The pooled standard deviations used for standardized effect summaries are approximate, because the source data are least-squares means with standard errors rather than raw individual observations. Such calculations can offer a descriptive scale for comparing observed contrasts, but they cannot reproduce the trial's original model or correct for features not represented in the summaries.

In general, over-dispersion means that observed variability is greater than a model assumes. When that occurs, standard errors can be underestimated and inferential results can appear more precise than warranted. The registry record does not report a finding of over-dispersion or enough model diagnostics to assess it in this study. The point is relevant as a general model-checking concept, not as a claim about these data.

The posted p-values for recalculations are unadjusted. They do not account for the many outcomes and contrasts listed in the registry. Multiple testing can increase the chance of at least one apparently notable result when many hypotheses are examined. No multiplicity adjustment strategy is reported in the registry record, so secondary p-values should be read as exploratory rather than as a collection of confirmatory findings.

Missingness is also consequential. The primary population included randomized participants with at least one post-baseline efficacy assessment, and missing data were not imputed. In general, an analysis restricted to participants with observed follow-up can be biased if missingness is related to treatment or outcome. The registry record does not report the reasons for missing observations or enough detail to assess their pattern. The differing group sizes are noted in the record, but the information does not establish the effect of missingness on the estimates.

Finally, a failure to detect a difference is not the same as evidence of equivalence. Equivalence requires a prespecified margin defining a sufficiently small difference, along with an analysis designed to assess that margin. Neither an equivalence margin nor an equivalence hypothesis is reported. The primary finding is therefore best described as an uncertain comparison of the observed means, not a demonstration that the two formulations are identical.

11. Confidence Intervals and Precision

Precision is reflected in the spread of plausible values around an estimate. The primary unadjusted recalculation has a 95% confidence interval from -17.3 to 13.32 around the mean difference of -1.99. The interval includes zero and values on both sides of zero, so the data summarized by this calculation do not pinpoint the direction of the difference. The sponsor's adjusted estimate has a 95% confidence interval from -5.00 to 8.97. That interval also includes zero, though its range is not identical to the recalculated interval because it comes from a different analysis.

In general, a confidence interval is not a guarantee that the true value lies within its particular endpoints. It describes the long-run performance of an interval procedure under its assumptions. Nor does a confidence interval show the distribution of individual specimen results. It focuses on uncertainty in an estimated quantity, here a group contrast.

The primary group means have standard errors of 5.54 and 5.51, based on group sizes of 50 and 52. These standard errors describe precision of the estimated means, not the variation among all individual specimens. Because the registry reports least-squares means and standard errors rather than raw observations, the available summary statistics limit how fully the observed distribution can be assessed.

The confidence intervals are compatible with a range of possible differences. They do not establish equivalence, and the registry record does not report a clinically meaningful difference threshold for this specimen measure. The estimate's precision should be interpreted in relation to the in situ endpoint and the study's short baseline-to-14-day assessment, rather than as a direct statement about patient-level outcomes.

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 sample-size figures are planning illustrations based on the observed effects, not the trial's original design calculations. For these continuous outcomes, figures are participants per arm, not total events. The stated assumptions are two-sided significance level 0.05, equal allocation between the two arms, and a normal approximation. Each row considers an effect equal to the stated fraction of the observed effect. The original trial's sample-size assumptions are not reported in the registry record.

The 80% power column gives the illustrated number of participants per arm for a design with 80% power to detect the specified effect under those assumptions. The 90% power column gives the corresponding planning number for 90% power. In general, power is the probability of detecting an effect of a specified size when that effect is present, conditional on the assumptions used for planning. A higher target power generally requires more participants than a lower target when other assumptions are held fixed.

Rows planning for smaller effects require more participants because a smaller difference is harder to distinguish from random variation. For example, the rows labelled 75% and 50% of the observed effect represent progressively smaller planning targets than the row using 100% of the observed effect. These are not recommendations for a new study without further design work; they show how the calculated requirements change under the listed assumptions.

These figures are not observed power. They do not describe the probability that the completed trial's result is correct, and they should not be read as an explanation for the result. They are illustrative calculations anchored to the observed effect and the stated assumptions, rather than a record of the study's prespecified planning process.

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 = -1.990; pooled standard deviation σ = 39.460.
2. n = 2 × (1.96 + 0.842)² × 39.460² / -1.990² = 6,172.3, rounded up to 6,173 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 <- -1.99                      # observed difference in means
sigma <- 39.4600310694252                     # 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(708097)
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 = 6,173 (80% power), 8,263 (90% power)
 75% of observed effect: n per arm = 10,973 (80% power), 14,690 (90% power)
 50% of observed effect: n per arm = 24,690 (80% power), 33,052 (90% power)
power.t.test: 6173.3 
Simulated power at n = 6,173 per arm: 0.796 (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 ANOVA result as an adjusted mean difference of 1.98, with a 95% confidence interval from -5.00 to 8.97 and p=0.576. The simple recalculation from the posted summary statistics gives a mean difference of -1.99, with a 95% confidence interval from -17.3 to 13.32 and unadjusted p=0.799. Both intervals include zero, but the point estimates and interval widths differ.

The recalculation is unadjusted and uses the posted least-squares means and standard errors. The sponsor's estimate is explicitly an adjusted mean difference from ANOVA. The registry record does not give sufficient model details to reproduce that adjusted analysis or identify all sources of the difference. The contrast direction may also be expressed differently across results. Accordingly, the recalculated value is a descriptive check on the posted summaries, not a replacement for the sponsor's reported model-based analysis.

For interpreting the trial's primary analysis, the sponsor's own reported result takes precedence over a simple recalculation. It should still be read with its confidence interval and with the limitations of the registry record: details of the model, crossover handling, and missing-data pattern are not reported. The recalculation is useful for transparency about the summary statistics, but it cannot resolve those missing details.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and several secondary outcomes, with numerous treatment comparisons and analyses. Testing many contrasts creates a multiplicity issue: even when no true differences exist, the chance of seeing at least one apparently notable result increases as more tests are considered. In general, researchers may address this by prespecifying a hierarchy of tests or applying an adjustment method. The registry record does not report a multiplicity strategy or indicate which secondary comparisons, if any, were confirmatory.

The recalculated p-values are unadjusted. They should not be interpreted as if each were the sole test in the study. The sponsor's reported ANOVA p-values are also listed for selected comparisons, but the registry record does not describe how the larger set of outcomes and comparisons was handled. This limits the strength of claims based on isolated secondary results.

The primary analysis population is described as intent to treat, defined in this record as all randomized participants with at least one post-baseline efficacy assessment. Missing data were not imputed. The registry notes differences in the number of participants per treatment group due to dropouts. This definition is narrower than a population containing every randomized participant regardless of follow-up, so readers should note the record's specific operational definition rather than assume a conventional handling of missing outcomes.

In general, missing data can affect an estimate if the probability of an observation being missing is related to treatment or the unobserved outcome. Without information on reasons for dropout, timing of missingness, or sensitivity analyses, the direction and size of any potential bias cannot be assessed. These details are not reported in the registry record.

The record describes randomization, single masking, and a crossover model, but does not report the allocation concealment method, who was masked, treatment sequence, washout details, or the terms included in the ANOVA. These omissions matter for evaluating risks of bias and determining how well the analysis reflects the design. They do not demonstrate that the analysis was inappropriate; they limit independent assessment from the registry information.

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 non-serious adverse event terms

The registry lists serious adverse events affecting 0 participants in each treatment group and 0 overall. The numbers at risk are 50 for NaF toothpaste containing 1450ppmF, 52 for 1400ppmF, 53 for NaMFP/NaF containing 1450ppmF, 53 for NaF containing 675ppmF, and 51 for placebo; the overall group lists 57. No serious adverse event terms are listed.

For other adverse events, the registry reports 3 affected participants in the 1450ppmF NaF group, 1 in the 1400ppmF NaF group, 3 in the NaMFP/NaF group, 4 in the 675ppmF NaF group, and 1 in the placebo group; 10 are listed overall. These are participant counts in the record. The information does not establish whether events were caused by a toothpaste, how severe or long-lasting they were, or whether the same participant contributed events under multiple crossover conditions.

Deaths are not reported in the registry record. The available safety counts should be read as the posted adverse-event summary, not as a complete assessment of comparative safety or a basis for medical decisions.

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

The primary result describes percentage surface microhardness recovery in sound enamel specimens exposed to two NaF toothpaste formulations over baseline to 14 days. The sponsor's adjusted estimate and the simple unadjusted recalculation both have confidence intervals that include zero. The posted data therefore do not establish a clear difference for this primary comparison under the reported analyses.

This conclusion is not the same as saying the formulations are equivalent. Equivalence requires a defined margin and an analysis aimed at showing that any difference falls within that margin. Neither an equivalence margin nor an equivalence analysis is reported in the registry record. The results also do not prove that there is no difference; the confidence intervals leave room for a range of possible contrasts.

The study evaluates artificial enamel lesions in an in situ model. Its measurements concern enamel specimens and a short assessment interval, rather than clinical caries incidence, disease progression in patients, symptoms, or long-term outcomes. A measured change in specimen microhardness cannot automatically be translated into a change in clinical outcomes.

Secondary measures provide additional information about microhardness recovery and fluoride uptake in sound and demineralized specimens. Their patterns are not a substitute for the primary result, and multiple comparisons make isolated unadjusted findings less conclusive. The registry record also leaves important analytic details unreported, including treatment sequences, model terms, and the pattern of missing data.

The most defensible reading is therefore limited: the registry reports a randomized crossover enamel-specimen study, and its primary comparison did not yield a statistically clear difference in the posted adjusted analysis. The uncertainty and the model context should remain central when interpreting that result.

18. Why This Trial Matters Statistically

This trial illustrates why the endpoint and study model matter as much as the numerical result. The primary measure is surface microhardness recovery in enamel specimens, not a direct count of caries lesions in participants. Treating a laboratory or in situ measurement as though it were a clinical outcome would overstate what the data address.

It also demonstrates the distinction between an adjusted sponsor analysis and a recalculation from summary statistics. The sponsor reports an ANOVA-based adjusted contrast, while the recalculation uses posted means and standard errors without adjustment. Both can be informative for different purposes, but they do not necessarily estimate the same quantity in the same way. The registry's limited model detail prevents full reconstruction of the sponsor analysis.

The crossover label raises a further statistical consideration. In general, repeated treatment observations from the same participant may be correlated, and analysis should reflect that structure. The registry identifies a crossover model but does not report sequence, washout, or the model's handling of repeated observations. This is a useful reminder that a design label alone is not enough to assess an analysis.

Finally, the study has multiple secondary endpoints and comparisons, variable group sizes, and no imputation of missing data. These features affect interpretation and precision. The trial is therefore a case study in keeping conclusions aligned with the prespecified primary endpoint, distinguishing exploratory secondary patterns from confirmatory evidence, and treating uncertainty as part of the result rather than as a footnote.

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 central limitation is the endpoint's scope. Surface microhardness recovery and fluoride uptake were assessed in enamel specimens in an in situ model. The registry does not report clinical caries incidence or another participant-level clinical endpoint, so these results cannot establish effects on clinical disease outcomes.

The assessment timeframe is baseline to 14 days. Longer-term outcomes are not reported in the registry record. The record also does not report enough detail to evaluate how treatment sequence, possible carryover, or washout was handled in the crossover design. These are important design considerations, but the missing information does not prove that carryover occurred.

There is a difference between the sponsor's adjusted ANOVA results and simple recalculations from summary statistics. Because the registry record does not provide the full model specification or raw participant-level data, the adjusted analysis cannot be independently reproduced from the posted summaries. The recalculated standard deviations are approximate, based on least-squares means and standard errors, and should not be mistaken for raw-data estimates.

Group sizes differ across treatment conditions, and the record attributes this to dropout. Missing data were not imputed, while the stated intent-to-treat population required at least one post-baseline efficacy assessment. Reasons for dropout, missingness patterns, and sensitivity analyses are not reported. The extent to which missing data could affect the results is consequently uncertain.

Multiple outcomes and analyses are listed, but a multiplicity plan is not reported. This limits the confirmatory weight of unadjusted secondary findings. Other details absent from the record include allocation concealment, who was masked, the treatment sequences, and the original sample-size justification. These limitations constrain independent appraisal and the range of conclusions warranted by the posted results.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: In Situ Caries Efficacy of Fluoride Toothpastes — NCT00708097.
• Registry record last updated 2015-01-01; status last verified 2014-12.
• Results first posted 2013-06-24.
• 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

NCT00708097 is a completed, Phase 3, randomized, single-masked crossover study sponsored by GlaxoSmithKline. It enrolled 57 participants and evaluated fluoride toothpastes using an in situ model with artificial caries lesions. The primary endpoint was percentage surface microhardness recovery in sound enamel specimens from baseline to 14 days, comparing NaF toothpaste containing 1450ppmF with NaF toothpaste containing 1400ppmF.

The sponsor-reported adjusted mean difference was 1.98, with a 95% confidence interval from -5.00 to 8.97 and p=0.576. The unadjusted recalculation gives a mean difference of -1.99, with a 95% confidence interval from -17.3 to 13.32 and p=0.799. Both intervals include zero. Secondary outcomes covered microhardness recovery and fluoride uptake in sound and demineralized specimens. Missing data were not imputed, and the registry record does not report a multiplicity strategy or several details of the crossover analysis.