Neurological Impacts of Artificial Sweeteners in the Context of Diet Sodas: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Neurological Impacts of Artificial Sweeteners in the Context of Diet Sodas (NCT04048681), a randomized na study. The primary outcome is Changes in Brain Response to Food Cues in the Insula.

2. Clinical Question

3. Trial Design

The registry describes a randomized crossover study with no masking. The crossover model means the study was structured around participants receiving more than one condition, allowing a design to compare responses across conditions within participants. In a crossover design, this structure can help account for stable differences between participants. However, the registry record does not report the actual treatment sequence, whether every participant completed every condition, or how the analysis handled within-participant measurements.

The recorded conditions were Diet Soda, Soda, and Carbonated Water. Their descriptions specify a 12oz can of Diet Coke, a 12oz can of Coke, and a 12oz can of unflavored carbonated water, respectively. The sponsor was Beth Israel Deaconess Medical Center. The study is listed as completed, with a start date of 2019-08-06 and a primary completion date of 2019-12-20. The phase is listed as NA.

Enrollment is recorded as 5. For the primary outcome, the arm-level result counts are 5 for Diet Soda, 5 for Soda, and 4 for Carbonated Water. These counts describe the numbers attached to the posted arm summaries; they do not establish why the counts differ. The registry record does not report whether the difference reflects missing measurements, incomplete participation, or another reason.

The registry lists a secondary outcome of Neurocognitive Testing (Stop Signal Reaction Time). It does not report the allocation sequence, the crossover schedule, washout duration, masking procedures beyond NONE, or a detailed statistical analysis plan. Those details should not be inferred from the design label alone.

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: This is a cross-over trial, and participants were assigned to receive diet soda, regular soda, and carbonated water in a random order.

Overall Study

Reasons for not completing this period:

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 Changes in Brain Response to Food Cues in the Insula, measured over 1 hour.

The primary endpoint is titled “Changes in Brain Response to Food Cues in the Insula.” It uses functional magnetic resonance imaging BOLD signals recorded while participants view food and non-food cues. The outcome is expressed as a z-score representing the magnitude of activation in the insula to food cues compared with non-food cues. The registry explains that the score represents the estimated response magnitude in that area relative to the overall mean: positive values indicate values above the mean, and negative values indicate values below the mean. The endpoint timeframe is 1 hour.

This measure addresses a brain response to a food-cue task, rather than directly measuring appetite, food intake, or a clinical health outcome. It was selected to examine the study’s stated question about whether artificial sweeteners in diet soda may alter the brain’s response to food cues. A brain imaging response can be useful for testing a mechanistic question because it quantifies activity associated with a defined task. It should not, by itself, be interpreted as a direct measure of a participant’s appetite or as evidence of a later behavioral or health effect.

The registry reports means and standard errors for the arm-level z-scores. It does not report confidence intervals or a formal estimate of the contrast between beverage conditions. The reported summaries therefore describe the endpoint values in each arm but do not establish the precision of a treatment comparison.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant’s response is represented by a numerical z-score, and the posted arm summaries use means. For a continuous outcome, the mean describes the average value in a group or condition. It can be informative about the central level of the recorded measure, but it does not show the full distribution of participant-level responses. A mean can be affected by unusually high or low values, particularly when the number of observations is small.

What does the standard error describe?

The registry reports standard error as the measure of dispersion accompanying each mean. A standard error describes the precision of an estimated mean: it reflects how much the mean estimate would be expected to vary across repeated samples under a statistical model. It is not the same as the standard deviation, which describes variation among individual observations. A standard error also does not directly describe the uncertainty in a difference between conditions.

For a treatment contrast, uncertainty depends on the variability of the contrast itself. In a crossover study, measurements from the same participant may be correlated. The degree and direction of that correlation can affect the precision of a within-participant comparison. Arm-level means and standard errors do not reveal the correlation between each participant’s measurements across conditions. Therefore, they are insufficient to recover a paired contrast or its standard error.

Why does the crossover label matter?

A crossover design is intended to compare conditions while accounting for repeated measurements from participants. A suitable analysis typically needs to respect the design, including which conditions each participant received and in what sequence. The registry record does not report the sequence, the number of completed periods for each participant, or the statistical model used. It lists no analyses for the primary outcome. Without those details, a reader cannot verify whether the posted arm summaries represent all periods, complete cases, or another analysis set, and cannot determine how the study’s design was reflected in an inferential comparison.

Randomization concerns how participants or treatment sequences are assigned. It can help limit systematic differences in assignment, but it does not itself provide a treatment-effect estimate or guarantee that the reported analysis accounts for repeated measurements. The registry lists allocation as randomized and masking as none. It does not give further details about the randomization procedure or any steps taken to reduce bias in imaging assessment or other study procedures.

What can be learned from an interval?

A confidence interval is a range of values compatible with the data and the assumptions of a statistical model, expressed at a chosen confidence level. It communicates both the estimated effect and its uncertainty. For this study, the registry does not report a confidence interval for a between-condition contrast. The arm-level standard errors cannot be treated as if they were confidence intervals, nor can they be used alone to infer the interval for a paired comparison.

A p-value, when reported, evaluates how compatible the observed data are with a specified null hypothesis under a statistical model. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. The registry record reports no primary-outcome p-value or formal analysis. Thus, no significance claim can be drawn from the posted means alone.

Why can over-dispersion matter?

Over-dispersion is a general concern for data whose variability is greater than a statistical model expects. It can make standard errors too small if the model fails to account for the extra variation. This concept is most often discussed for outcomes such as counts, where a model may assume a particular relationship between the mean and variance. The primary endpoint in this trial is continuous, and the registry record does not report a model or an over-dispersion assessment. There is therefore no basis for claiming that over-dispersion was present or that it affected the reported results. More generally, the chosen model should reflect the type and structure of the outcome data.

How should planning illustrations be read?

Sample-size illustrations based on an observed effect are planning scenarios, not observed evidence about the trial’s power. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation, with the effect to detect set as a stated fraction of the observed effect. These assumptions do not describe the trial’s actual design justification, which is not reported in the registry record. An illustration based on the observed effect also does not show that the study had a particular probability of detecting a true effect. No sample-size justification or prospective power calculation is reported in the registry record.

In this case, the registry provides no primary-outcome analysis from which to establish an observed comparative effect. The arm summaries alone do not supply the paired treatment contrast required for a design-aligned planning calculation. Any planning figures should therefore be understood only according to the assumptions attached to them, not as evidence that the trial did or did not have adequate power.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Diet Soda", "Soda", "Carbonated Water")
n   <- c(5, 5, 4)
m   <- c(0.25, -0.9, -0.01)
sp  <- c(0.3, 0.3, 0.6)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Diet Soda  5  0.250  0.671  -0.583  1.083
CARM  Soda  5  -0.900  0.671  -1.733  -0.067
CARM  Carbonated Water  4  -0.010  1.200  -1.919  1.899

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures overall, including the primary imaging outcome and the secondary outcome Neurocognitive Testing (Stop Signal Reaction Time). It does not report a multiplicity strategy, such as a hierarchy among outcomes, an adjustment for multiple testing, or a prespecified approach to interpreting multiple comparisons. The record also lists no primary-outcome analyses and does not report a result for the secondary outcome. It is therefore not possible to determine how multiplicity was addressed in the study’s analysis, if at all.

The primary-outcome arm summaries include counts of 5 for Diet Soda, 5 for Soda, and 4 for Carbonated Water. The registry record does not explain the difference in counts. It does not report whether measurements were missing, whether a participant did not complete a condition, or whether another analysis-population rule was used. No imputation method or missing-data sensitivity analysis is reported in the registry record.

The record also does not define an intention-to-treat population, a per-protocol population, or another analysis population for the primary outcome. In a crossover study, it is important to know which participants and periods contributed to each estimate, and how incomplete sequences or unavailable measurements were handled. Those details are not reported in the registry record. The arm-level counts should therefore be read as counts attached to the posted summaries, not as a complete account of participation or missingness.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry reports serious adverse events and other adverse events separately for each condition. For Diet Soda, 0 serious adverse events and 0 other adverse events are recorded among 5 participants at risk in each category. For Soda, 0 serious adverse events and 0 other adverse events are recorded among 5 participants at risk in each category. For Carbonated Water, 0 serious adverse events and 0 other adverse events are recorded among 4 participants at risk in each category. The death count is 0 for each group.

These entries mean that the registry’s posted safety results list no events in the recorded categories during the study. They do not establish that no unrecorded events occurred, and they do not establish that a condition is free of risk. The participant counts are small, and the record does not provide a detailed adverse-event collection method, follow-up duration for safety, or event-level descriptions. No serious adverse event terms are listed.

Zero recorded events should be interpreted as a description of the registry entry, not as a comparative safety analysis. The record does not report a statistical comparison of adverse events between conditions. It also does not provide enough detail to assess how adverse events were solicited, classified, or attributed.

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

The results describe mean insula responses to food versus non-food cues after the listed beverage conditions, using the registry’s z-score measure. The mean is positive for Diet Soda, negative for Soda, and slightly negative for Carbonated Water, with standard errors reported for each arm. These values locate the arm-level summaries relative to the measure’s overall mean as described in the registry.

The posted results do not establish that one beverage condition changed brain response relative to another. A comparison requires an estimate of the contrast and its uncertainty. Neither is reported for the primary outcome. In particular, the arm-level summaries cannot show whether the apparent separation among means is compatible with sampling variability or reflects a reliable within-participant difference. The registry lists no primary-outcome analyses.

The endpoint is a brain-imaging response during a food-cue task. It is not a direct measurement of appetite, food consumption, or a clinical health outcome. Although the study’s brief summary links the research question to appetite, no appetite result is reported in the registry record. The registry lists Stop Signal Reaction Time as a secondary outcome, but does not report its result.

The study was randomized and used a crossover model, but details needed to evaluate the analysis—such as treatment sequence, period handling, and the model used—are not reported in the registry record. The arm counts are also not fully explained. These gaps limit what can be inferred from the descriptive numbers and prevent an independent assessment of the intended within-participant comparison.

Accordingly, the posted information is best read as a small set of descriptive imaging summaries, not as evidence that a beverage did or did not alter appetite or brain activity. It also does not establish the size or precision of a comparative effect. The absence of a reported analysis is an important limit on interpretation, not proof that an effect is absent.

14. Why This Trial Matters Statistically

This study illustrates why the label “crossover” matters when reading clinical trial results. In a crossover design, participants may contribute measurements under multiple conditions, so the relationship between those measurements is central to estimating a treatment contrast. Reporting only arm-level means and standard errors leaves that relationship unknown. A paired analysis can have different precision from a comparison that treats condition summaries as independent, and the appropriate calculation depends on the actual participant-level pattern and study structure.

The registry’s outcome is also a useful example of the distinction between a mechanistic measure and a patient-centered or behavioral outcome. The insula z-score quantifies a response during an imaging task. It may address a question about brain activity, but it is not interchangeable with appetite or food intake. Keeping the endpoint’s meaning clear prevents an imaging result from being interpreted as a direct clinical outcome.

The entry also demonstrates how descriptive results and inferential results answer different questions. Means show the recorded average in each condition; standard errors describe precision of those means. Neither alone gives a treatment contrast, a confidence interval, or a test of a comparative hypothesis. This distinction is especially important when the analysis population and missing-data handling are not described.

Finally, the registry’s small enrollment and lack of a posted primary analysis make transparent reporting of the design and analysis particularly important. Information on condition order, participant-level completion, and the model used would help readers evaluate the result. The posted record allows a limited description of the endpoint summaries, while leaving the central comparative question unresolved.

15. 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 trial’s recorded enrollment is 5, and the outcome summaries are based on counts of 5 for Diet Soda, 5 for Soda, and 4 for Carbonated Water. With so few recorded observations, individual responses may have a substantial influence on an arm mean. The registry does not provide participant-level data or distributions, so the extent of individual variation cannot be assessed from the posted summaries.

The main limitation for interpretation is the absence of a reported comparative analysis. The registry lists no primary-outcome analyses and gives no contrast estimate, confidence interval, or p-value. This is particularly consequential for a crossover study because the within-participant relationship between measurements is needed to assess a paired comparison. The treatment order, period effects, and any washout procedures are not reported in the registry record.

The registry also does not explain why the Carbonated Water summary has a count of 4 while the other condition summaries have counts of 5. It does not identify the analysis population or state how unavailable measurements were handled. The population description is limited to the condition of obesity; other participant characteristics and eligibility details are not reported in the registry record.

The endpoint is a task-related imaging measure, not a direct appetite outcome. The secondary outcome is named, but its result is not reported in the registry record. The record does not describe multiplicity handling, a sample-size justification, or a detailed statistical plan. These omissions restrict assessment of the study’s methods and the conclusions that can be drawn from the descriptive findings.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Neurological Impacts of Artificial Sweeteners in the Context of Diet Sodas — NCT04048681.
• Registry record last updated 2020-02-05; status last verified 2020-01.
• Results first posted 2020-02-05.
• 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.

19. Record Summary

The registry record describes a completed, randomized, unmasked crossover study sponsored by Beth Israel Deaconess Medical Center. It lists enrollment of 5 and a study condition of obesity. The recorded conditions were Diet Soda, Soda, and Carbonated Water, described as 12oz cans of Diet Coke, Coke, and unflavored carbonated water. The study start date is 2019-08-06, and the primary completion date is 2019-12-20.

The primary endpoint was change in insula response to food cues, measured by an fMRI-based z-score at 1 hour. Arm-level means and standard errors are posted, but the registry reports no primary-outcome analysis or comparative estimate. A secondary outcome, Stop Signal Reaction Time, is listed without a reported result. The registry lists 0 serious adverse events, 0 other adverse events, and 0 deaths in each condition. The record does not report the analysis plan, missing-data handling, or sample-size justification.