Gao NARASD Lithium Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Gao NARASD Lithium Study (NCT02909504), a na phase 4 study. The primary outcome is Change in Expression of ~45 Molecules in Blood Cells of Patients With Bipolar I or II Disorder Being Treated With Lithium.

2. Clinical Question

This design addresses a descriptive and exploratory question about blood-cell measurements during lithium treatment and how those measurements differ across groups defined by symptom change. It does not directly compare lithium with another treatment or with a concurrent untreated group. Because responder status is defined using symptom change observed during the study, any comparison between responders and nonresponders is a subgroup comparison within the enrolled group, not a randomized comparison. The registry record does not report the assessment method for mood symptoms, the threshold rules beyond the stated percentage decrease, or whether any other clinical outcomes were analyzed.

3. Trial Design

The registry describes a Phase 4, open-label, single-group study of lithium for acute treatment of bipolar I or II disorder. It lists 24 participants as enrolled. The brief summary describes an intended enrollment of 50, while the registry enrollment field gives 24; the record does not explain the difference. The study status is terminated. The reason for termination is not reported in the registry record.

The planned treatment period is described as four months, and the primary outcome timeframe is baseline and Week 16. Participants were to receive lithium 300 mg twice daily, with titration in 300 mg increments every 7 days as tolerated to levels > 0.6 mEq/L. The registry does not report individual treatment exposure, adherence, achieved lithium levels, or how many participants completed the planned period.

There was no random allocation, separate control arm, or masking. The study is therefore not a randomized treatment comparison. Its single-group structure means that changes over time cannot by themselves distinguish changes associated with lithium from changes due to other influences, including the natural course of symptoms, concurrent care, or measurement variation. These are general limitations of an uncontrolled before-and-after design, not findings about what occurred for individual participants in this study.

The primary outcome was originally described as measuring about 45 molecules, using fluorescence intensity in mononuclear blood cells. The outcome description says that limitations involving antibody quality and finances restricted the analysis to 28 proteins. The registry record lists one outcome measure but no analysis entries. It does not report the statistical analysis plan, whether any interim analyses were planned or performed, the reason for selecting the recorded enrollment, or a sample-size justification. The dates recorded for study start and primary completion are 2016-09 and 2018-11, respectively; the record does not provide further detail about the timing of termination or data collection.

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: While we planned to analyze \~45 molecules of interest, poor quality of antibodies to some proteins \& financial limitations limited us to only 28 proteins. The number of participants analyzed for protein expression differs from baseline characteristic and varies below as not all subjects had high enough levels of a particular protein to be analyzed.

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 Change in Expression of ~45 Molecules in Blood Cells of Patients With Bipolar I or II Disorder Being Treated With Lithium, measured over Baseline and Week 16.

The primary outcome measures changes in the relative fluorescence intensity of molecules in mononuclear blood cells from baseline to Week 16. The description explains that a photomultiplier tube counts photons and that higher values represent higher fluorescence intensity. The listed unit is fluorescence intensity, and the stated summary parameter is the mean with standard deviation.

The outcome is intended to characterize molecular measurements during the study and compare those changes between participants classified by symptom response. Responders are defined as participants with a >=50% decrease in mood symptom severity between baseline and Week 16; nonresponders did not meet that definition. This classification provides a way to describe whether molecular changes differ across groups with different observed symptom changes. It does not establish that a molecular change causes symptom improvement, or that a particular measurement can predict response in other patients.

The registry describes the original plan as covering about 45 molecules, but says that only 28 proteins were analyzed because of antibody-quality and financial limitations. It also notes that the number of participants analyzed for protein expression varies. The numerical results for individual molecules, including the magnitude and uncertainty of the changes, are not reported in the registry record. The registry does not describe a clinically meaningful threshold for fluorescence intensity or explain how the measured changes would be interpreted in clinical practice.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical descriptions in the registry are limited. The primary outcome is a continuous measurement: fluorescence intensity, summarized as a mean and standard deviation. The registry does not state the test or model used to compare baseline and Week 16, how responder and nonresponder groups were compared, or how uncertainty was quantified. The explanations below are general statistical context, not a description of analyses confirmed to have been used in this study.

What does a mean describe?

General: a mean is an arithmetic summary of the values in a group. It can describe the average measured fluorescence intensity or average change, depending on exactly which quantity is summarized. The registry names the mean as the outcome parameter and describes values as changes between baseline and Week 16. It does not provide the actual means for the molecules.

A mean does not show how values are distributed among participants. It can be affected by unusually high or low observations, and it does not show whether most participants had changes close to the average. The standard deviation describes the spread of individual values around the mean; it is not the uncertainty interval for the mean. The registry does not post molecule-specific standard deviations or participant counts, so neither the group averages nor their observed variability can be reviewed from the record.

What would a confidence interval add?

General: a confidence interval gives a range of values compatible with an estimate under a stated statistical procedure. Its width conveys precision: wider intervals indicate less precise estimates, while narrower intervals indicate greater precision. An interval also helps readers assess whether the data are compatible with a range of effect sizes, rather than focusing only on whether a p-value passes a threshold. No confidence intervals are reported in the registry record, so the precision of the molecular measurements cannot be assessed from the posted information.

How should the responder comparison be understood?

The registry defines responders by a >=50% decrease in mood symptom severity between baseline and Week 16, and defines nonresponders as those who did not experience that decrease. General: dividing participants into groups based on an observed outcome can be useful for describing patterns, but it does not create randomized groups. The groups may differ in ways other than symptom change, and an association between group membership and a molecular measurement does not establish a causal relationship.

Because symptom response and molecular change are both assessed over the same interval, the record does not establish which change came first. The responder classification also depends on the symptom measure, whose specific instrument and assessment procedures are not reported. The registry does not report the numbers of responders and nonresponders or the analysis used to compare them.

What does a single-group design allow?

General: a baseline-to-follow-up design can describe how measurements change over time among participants. Without a concurrent comparison group, it cannot isolate the change attributable to the intervention from other changes occurring over the same period. Randomization and a comparator are common ways to strengthen treatment comparisons, but neither is described for this study: it is recorded as single-group, with allocation not applicable and no masking.

The registry does not report whether a paired analysis was used for repeated measurements, how missing observations were handled, or whether an adjusted model was fitted. It would be inappropriate to assume a particular statistical test based only on the fact that baseline and Week 16 measurements were planned.

Why can multiple measurements matter?

The outcome description says that about 45 molecules were planned and 28 proteins were analyzed. General: examining multiple outcomes creates a multiplicity issue because the chance of obtaining at least one apparently notable result can increase as more tests are conducted. Statistical approaches can account for this, but the registry does not report whether a multiplicity adjustment or other strategy was used. Without the molecule-level results and analysis details, readers cannot determine how multiplicity was addressed or how it affects interpretation.

What do sample-size illustrations mean?

General: a sample-size calculation is a planning tool based on assumptions about the effect to be detected, variability, significance level, and design. An illustration based on an observed effect is not a justification for the study's original enrollment and is not observed power. The stated planning assumptions for illustrations are a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. Those assumptions describe a hypothetical two-arm planning exercise, not the recorded single-group design. The registry does not report a trial-specific sample-size justification or the numerical results of any planning illustration.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome description refers to about 45 molecules, while also stating that antibody-quality and financial limitations restricted the analysis to 28 proteins. Multiple molecular outcomes can raise multiplicity concerns. General: when many outcomes are examined, the probability of at least one apparently notable result can increase, so an analysis plan may specify adjusted significance thresholds or other methods for interpreting multiple tests. Whether any such method was used is not reported in the registry record.

The record lists one outcome measure and no analysis entries. It does not report molecule-specific estimates, test results, confidence intervals, or a prespecified hierarchy among the molecules. The number of participants analyzed for protein expression is described as varying, but the counts for individual proteins are not reported. Therefore, the completeness of the outcome data and the precision of each molecular comparison cannot be evaluated.

General: missing-data methods can influence results when participants or measurements are unavailable. The registry does not report why participant counts varied, how missing measurements were handled, whether participants were excluded from particular analyses, or what analysis population was used. It also does not report whether analyses followed an intention-to-treat approach or another defined population. The record notes the limited number of proteins analyzed, but this is distinct from explaining missing participant-level measurements.

The description defines responder and nonresponder groups by symptom change, but does not give their sizes or report whether all enrolled participants could be classified. No imputation strategy, sensitivity analysis, or other missing-data assessment is reported in the registry record. These omissions limit evaluation of both the molecular comparisons and the possible impact of incomplete data.

11. 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 4% reporting threshold.

Most frequent non-serious adverse event terms

The registry lists 24 participants at risk for serious adverse events and reports zero participants with a serious adverse event. It also lists 24 at risk for other adverse events and reports 20 participants affected in that category. Deaths are reported as zero. No serious adverse-event terms are listed.

These are registry counts, not a detailed description of event types, timing, duration, severity, or relationship to lithium. The record does not state whether the count of 20 refers to participants with at least one event or provide event-level counts; it should therefore be read according to the registry's participant-affected field, not as a count of distinct events. The absence of listed serious-event terms does not provide details about how events were assessed or coded.

General: safety counts are most informative when read alongside their denominators, definitions, follow-up duration, and completeness of reporting. The registry gives the at-risk counts but does not provide a detailed safety narrative or explain how long each participant was observed. These figures describe what is posted in the registry and do not, by themselves, establish the safety profile of lithium more broadly.

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

The registry record describes a study of lithium in participants with bipolar I or II disorder, with planned blood-cell measurements at baseline and Week 16. The main statistical question concerns changes in fluorescence intensity for molecules in mononuclear blood cells and whether those changes differ between participants classified as responders and nonresponders. The posted record does not provide the numerical molecular results, so the direction, magnitude, and precision of any such differences cannot be determined from it.

The design also limits what a result could establish. There was one open-label group, with no randomized comparator. General: a change observed over time in a single group does not by itself show that the intervention caused the change. A comparison between responders and nonresponders would describe an association among participants grouped by their observed symptom changes; it would not prove that a molecular change caused, predicted, or mediated symptom response.

The responder definition is a >=50% decrease in mood symptom severity from baseline to Week 16. The registry does not identify the symptom instrument or describe the assessment process. It also does not report the numbers in the responder categories, the participant count for each molecular measurement, or methods for addressing missing data and multiple outcomes.

The record says that about 45 molecules were intended for analysis but only 28 proteins were analyzed, citing antibody quality and financial limitations. It reports variable participant counts for protein expression. The status is terminated, but the reason is not reported in the registry record; the record also does not explain the relationship between the planned enrollment of 50 in the summary and the enrollment of 24. These details are important context, but they do not reveal the direction or statistical strength of the unreported molecular findings.

In short, the registry documents the study design, intended measurement, and some enrollment and safety information. It does not provide enough statistical results to judge the observed molecular changes or their uncertainty. No conclusion about clinical benefit, lack of benefit, or predictive value can be drawn from the posted information alone.

13. Why This Trial Matters Statistically

This study illustrates a statistical challenge common to biomarker research: the scientific question may involve many measurements, while the strength of interpretation depends on transparent reporting of estimates, uncertainty, analysis populations, and methods. The registry describes a plan involving about 45 molecules and reports that 28 proteins were ultimately analyzed. When multiple molecular outcomes are considered, readers need to know how outcomes were selected and whether multiplicity was addressed. Those analysis details are not reported in the registry record.

The study also shows why design and interpretation must be considered together. It is recorded as open-label and single-group, without a separate comparator. Measurements at baseline and Week 16 can describe change over time, but general statistical reasoning cautions against treating that change as an isolated treatment effect. Similarly, grouping participants by observed symptom response can describe relationships between response status and molecular measurements, but does not establish causation or predictive performance.

Participant counts matter for biomarker analyses because the record says that the number analyzed varied across protein-expression measures. Without molecule-specific denominators, estimates, and uncertainty intervals, readers cannot tell how much information supports each comparison. The record's enrollment of 24 and its summary's stated plan to enroll 50 also make it important to distinguish intended enrollment from registered enrollment, while recognizing that the reason for the difference is not reported.

The statistical importance of this record is therefore chiefly methodological. It highlights the need for complete outcome-level reporting, explicit analysis populations, clear handling of missing observations, and a stated strategy for multiple outcomes. Without those details, a biomarker study's design and measurement goals can be understood, but its empirical findings cannot be evaluated fully.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several limitations specific to this study are evident from the registry record. First, it is a single-group, open-label study with no randomized comparator. That design supports description of measurements over time but does not separate treatment-associated change from other influences. The registry does not report masking or a control condition.

Second, the planned number of molecules and the number ultimately analyzed differ: about 45 were planned, while 28 proteins were analyzed because of antibody-quality and financial limitations. The record does not identify which proteins were omitted or explain whether selection occurred before or after examining study data. It also does not report a multiplicity strategy.

Third, participant counts for protein-expression analyses varied. The registry does not give the count for each molecule, explain the variation, describe how missing measurements were handled, or report the analysis population. As a result, readers cannot assess the completeness or precision of each outcome comparison.

Fourth, the primary outcome description proposes comparisons between responders and nonresponders, defined using symptom change over the same baseline-to-Week-16 interval as the molecular measurements. The record does not give group sizes, the symptom instrument, or the analysis method. Such groups are not randomized, and the described comparison cannot establish causation.

Finally, the record lists an enrollment of 24 while the summary describes a plan to enroll 50, and marks the study terminated without stating why. It does not report the trial-specific sample-size justification, detailed analysis plan, numerical primary results, or confidence intervals. These gaps prevent a full assessment of the study's results and precision.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Gao NARASD Lithium Study — NCT02909504.
• Registry record last updated 2020-07-10; status last verified 2020-06.
• Results first posted 2020-07-10.
• 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.

18. Record Summary

The registry record describes a Phase 4, open-label, single-group study of lithium in bipolar I or II disorder. It lists enrollment of 24 and a terminated status; the reason for termination is not reported in the registry record. The brief summary describes a plan to enroll 50. Treatment was described as lithium 300 mg twice daily, titrated as tolerated, with blood-cell measurements planned at baseline and Week 16.

The primary outcome concerns changes in fluorescence intensity for molecules in mononuclear blood cells and comparisons between responders and nonresponders. The description says that about 45 molecules were planned but 28 proteins were analyzed. Numerical molecule-level results and analysis entries are not reported in the registry record. The safety listing reports zero participants with serious adverse events, 20 participants affected in the other-adverse-event category, and zero deaths, among 24 at risk in each category.