Effects of Low Dose Naltrexone in Fibromyalgia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effects of Low Dose Naltrexone in Fibromyalgia (NCT00568555), a randomized na study. The primary outcome is Percent Change in Pain Scores Between Baseline to End of Placebo Treatment and Between Baseline to End of LDN Treatment..

2. Clinical Question

The registry’s brief summary frames the study as an assessment of LDN for symptoms of fibromyalgia and Gulf War Syndrome, noting anecdotal reports and a proposed role in regulating natural pain-reducing systems. Those statements describe the study rationale, not evidence established by the posted results. The statistical question is narrower: what average percentage change was recorded within each treatment condition among the completers, and how precisely were those means estimated? Because this was a crossover trial, participants received both treatments in an assigned order. A direct comparison would need to account for the paired observations and the treatment sequence; the registry record does not report that comparison.

3. Trial Design

This was a randomized, triple-masked crossover trial sponsored by Stanford University. The registry lists an enrollment of 53 and gives the study status as completed. The recorded start was 2007-06, and primary completion was 2010-01. The phase is listed as NA. The registry describes two sequences: LDN first and then placebo, or placebo first and then LDN. Thus, the treatment conditions were administered within a crossover design rather than as permanently separate treatment groups.

The registry’s primary endpoint population is all completers. The posted outcome results list 28 participants for the LDN condition and 28 for the placebo condition. These are condition-specific outcome counts in a crossover study, not evidence that the conditions represent unrelated groups. The record does not explain the relation between enrollment, completion, sequence assignment, and the outcome counts, and it does not report participant flow or reasons for non-completion.

The primary outcome timeframe is described as baseline to the end of placebo treatment (2 weeks + 4 weeks) and baseline to the end of LDN treatment (2 weeks + 12 weeks). The registry says baseline pain was calculated by averaging daily pain scores over the 2 week baseline period. Treatment-condition pain scores were calculated by averaging daily scores during the final 3 days of each condition. Further details of the crossover schedule, including washout procedures or possible carryover handling, are not reported in the registry record.

Randomization and triple masking are recorded, but the methods used to generate or conceal the allocation and the identities of the masked parties are not reported in the registry record. The statistical plan, including any prespecified model for the crossover data, is also not reported in the registry record. Those omissions matter when interpreting how the posted condition-specific means relate to a treatment comparison.

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: Individuals were assessed for eligibility via web screening and phone screening prior to enrollment. Further eligibility (blood tests, questionnaires) assessed after enrollment; ineligible consented participants were withdrawn prior to assignment to groups.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All completers

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.

Population: 2 participants in the low dose naltrexone group did not have data to analyze (1 withdrew by patient choice, 1 had a malfunction with the data collection device) 1 participant in the placebo condition did not have data to analyze (withdrew by patient choice)

6. Primary Endpoint

The registry lists the primary outcome as Percent Change in Pain Scores Between Baseline to End of Placebo Treatment and Between Baseline to End of LDN Treatment., measured over Baseline to end of placebo (2 weeks + 4 weeks) and baseline to end of LDN (2 weeks + 12 weeks).

The primary endpoint is percentage change in pain scores from baseline to the end of each treatment condition. Pain was measured with a Visual Analogue Scale ranging from 0 to 100, where 0 means no pain and 100 means the worst pain imaginable. The registry defines the percentage change as baseline pain minus endpoint pain, divided by baseline pain, then multiplied by 100. Under that definition, a positive value represents a lower endpoint score than baseline, while a negative value would represent a higher endpoint score.

Baseline pain was calculated by averaging daily pain scores across the 2 week baseline period. For each treatment condition, the endpoint score was the average daily pain score during the final 3 days of that condition. The timeframe compares baseline with the end of placebo treatment (2 weeks + 4 weeks) and baseline with the end of LDN treatment (2 weeks + 12 weeks). The registry reports the mean percentage change and 95% confidence interval for each condition.

A pain scale is used to turn a subjective experience into a repeated, numerical measure that can be summarized across participants and time points. Expressing change relative to baseline can describe the direction and magnitude of change in relation to each participant’s starting score. However, percentage change can be sensitive to baseline values, and a condition-specific change from baseline is not itself the same as a randomized comparison between treatments. The registry record does not report a direct treatment contrast for this endpoint.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it summarizes pain scores on a numerical scale and reports the mean percentage change. The registry gives a 95% confidence interval for each condition mean. In general, a mean condenses observations into an average, while its confidence interval communicates uncertainty in estimating the underlying average. It does not describe the full spread of individual participant values. The record does not report a standard deviation or another measure of participant-level dispersion for this endpoint.

What does the confidence interval describe?

In general, a confidence interval is constructed by a statistical procedure that would capture the target quantity at its stated rate over repeated studies if the procedure’s assumptions held. For this outcome, the posted intervals are attached to the separate condition means: 28.8 with an interval from 19.5 to 38.1 for LDN, and 18.0 with an interval from 7.2 to 28.8 for placebo. These intervals convey uncertainty about the means, not a range of likely outcomes for an individual. They also are not a direct interval for the difference between treatments.

For a randomized comparison, the key quantity would ordinarily be a contrast between treatment conditions, with an uncertainty estimate for that contrast. In a crossover study, the same participant contributes information under both conditions, so the paired relationship is relevant. The uncertainty of a paired contrast depends on how the two observations vary together, not only on the separate means and their separate confidence intervals. That paired information is not reported in the registry record. Consequently, a contrast and its confidence interval cannot be recovered from the posted summaries alone without making assumptions that are not justified by the record.

Why does the crossover design matter?

In general, a crossover design can compare treatment conditions within participants, which may account for stable differences between participants. It also requires attention to treatment order, period effects, and possible carryover from an earlier condition. Randomization of sequence helps assign order, but does not by itself show how the analysis handled order or period. The record describes the two sequences and identifies the trial as randomized and triple-masked. It does not report whether the primary analysis included sequence or period terms, how carryover was assessed, or whether a washout interval was used.

The outcome definitions also use different treatment timeframes: baseline to end of placebo treatment (2 weeks + 4 weeks) and baseline to end of LDN treatment (2 weeks + 12 weeks). The registry’s description gives those timeframes but does not explain the rationale for their durations or how the difference in duration was handled analytically. The measure is based on averages of daily scores, with baseline averaged across the 2 week baseline period and condition endpoints averaged during the final 3 days. Repeated daily measurements may reduce reliance on a single day, but the record does not state whether daily observations were analyzed directly or only used to form the described averages.

What can be inferred from the posted summaries?

The condition means and confidence intervals permit descriptive reporting of the observed summaries. They do not supply the participant-level paired values, the variance of within-participant differences, or the model output needed for a formal treatment comparison. The registry indicates that no analyses are listed and notes that values are missing or not estimable for a two-arm comparison of this continuous endpoint. Accordingly, no recalculated unadjusted comparison is available. The record does not report a sponsor analysis for the primary contrast either.

General statistical practice distinguishes a descriptive estimate from a hypothesis test. A confidence interval around a mean can help show its precision, but it cannot be used alone to claim that a randomized treatment contrast is statistically distinguishable from no difference. A p-value, if reported, would also need to correspond to the relevant treatment contrast and analysis population. No such p-value is reported for the primary comparison in the registry record.

What assumptions affect a continuous outcome analysis?

General methods for continuous outcomes often rely on assumptions about the distribution of errors, the independence or covariance structure of observations, and the appropriateness of the chosen summary measure. For crossover data, the within-participant covariance structure is central to precision. Percentage change can also behave differently at different baseline values because baseline appears in its denominator. The record does not report diagnostic checks, transformations, model assumptions, or sensitivity analyses. Without those details, the posted means and intervals can be described, but the suitability of a particular inferential model cannot be assessed.

The statistical plan and sample-size justification are not reported in the registry record. The reported confidence intervals should not be mistaken for a complete account of the planned analysis, and no conclusion about the study’s design power can be drawn from them. Any sample-size illustrations elsewhere are planning exercises based on stated assumptions, not observed power and not evidence of the trial’s prespecified design.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists the pain outcome as the primary measure and names four secondary outcomes: sleep quality, fatigue, pressure pain threshold, and heat pain sensitivity. The secondary outcomes are expressed as percentage changes across baseline and the ends of the placebo and LDN conditions. The record lists 5 outcome measures in total. When several outcomes are examined, the chance of at least one apparently notable result can increase; general approaches to multiplicity include prespecifying a primary outcome or adjusting interpretation across a family of tests. Whether such a procedure was planned or used is not reported in the registry record.

The record identifies the primary endpoint population as all completers. It does not report an intention-to-treat analysis, a modified intention-to-treat definition, or the treatment of participants with incomplete outcome data. Participant flow, reasons for missingness, and any imputation or other missing-data method are not reported in the registry record. These omissions limit assessment of whether the completer results could differ from results for all randomized participants.

The posted outcome counts are 28 for LDN and 28 for placebo. In a crossover trial, these are condition-level counts and should not be added to infer the number of unique participants. The registry enrollment is 53, but it does not give a reconciliation between enrollment, completion, and the reported condition counts. It also does not report whether all completers contributed observations to both conditions or how partial records were handled.

Because no analyses are listed and the registry notes that values are missing or not estimable for a two-arm comparison of the continuous endpoint, the posted summaries do not establish how multiplicity, missing data, or the crossover structure were handled in inferential analyses. Those aspects of the statistical plan are not reported in the registry record.

11. 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 4% reporting threshold.

Most frequent non-serious adverse event terms

The posted serious adverse event figures report zero affected participants in the LDN condition and zero in the placebo condition. Each condition lists 28 participants at risk for serious adverse events. The registry does not list any serious adverse event terms among the top serious terms. These are the figures recorded for the study; they do not establish that no serious events occurred outside the way events were collected or reported.

For other adverse events, the registry reports 20 affected participants among 28 at risk in the LDN condition and 7 among 28 at risk in the placebo condition. These counts are descriptive. The record does not report event-specific details, severity, timing, overlap among affected participants, or an analysis comparing safety between conditions. Because the study used a crossover design, the same participant may contribute safety information under both conditions; the record does not explain how that repeated exposure was accounted for.

Deaths are recorded as null for both conditions, which means a death count is not reported in the registry record; it should not be interpreted as a confirmed count of zero. The registry also does not report whether adverse events were judged related to treatment. The reported counts alone do not establish causation or quantify comparative risk.

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

The main descriptive result is that mean pain percentage change from baseline was 28.8 during LDN and 18.0 during placebo among the reported completer outcome population. The 95% confidence intervals were 19.5 to 38.1 for LDN and 7.2 to 28.8 for placebo. The endpoint definition makes positive change correspond to a lower endpoint pain score than baseline. These summaries describe the condition-specific results; they are not a direct estimate of the effect of LDN relative to placebo.

The registry record does not report a treatment contrast, its confidence interval, or a primary comparison p-value. The separate means cannot replace that comparison, especially in a crossover design where the relationship between each participant’s results under the two conditions affects the precision of the contrast. It is therefore not possible to determine from these summaries how precisely the treatment conditions differed, or whether a formal comparison met a statistical threshold.

The results also apply to the outcome population identified as all completers, rather than necessarily to all enrolled participants. The record does not provide enough detail about completion, missing outcomes, the assigned sequences, or the analysis model to determine how those factors might affect interpretation. The findings should be read as registry-reported summaries rather than as a complete account of the trial’s inferential analysis.

These data do not establish that LDN does or does not work for fibromyalgia. They report average within-condition changes, but the primary randomized comparison is absent from the posted results. Nor do the summaries show how individual participants responded, whether effects varied by sequence, or whether the results generalize beyond the study population. The registered rationale and the posted numerical outcome are distinct: a proposed mechanism or anecdotal report is not confirmed by the outcome summaries.

13. Why This Trial Matters Statistically

This trial is statistically notable because it uses a randomized crossover design for a continuous symptom measure. In such a design, treatment conditions are observed within participants, making the paired comparison central to the analysis. A clear report of the paired treatment contrast, its uncertainty, and how sequence and period were handled would allow readers to connect the design to the primary question. The registry provides condition-specific means and confidence intervals but does not report that direct comparison.

The outcome also illustrates the distinction between change from baseline within each condition and a randomized treatment effect. Both are useful descriptions, but they answer different questions. A positive mean percentage change describes average movement from baseline under a condition; it does not show how that condition compares with the alternative. The distinction is especially important when readers see results for both treatment conditions and might otherwise treat them as independent groups.

The record is also an example of why outcome reporting needs to include the analysis population, missing-data handling, and details of the inferential method. The primary population is identified as all completers, but the participant flow and reasons for missing outcomes are not reported. The registry lists secondary measures, while the multiplicity plan is not reported. Such information affects how readers assess precision, selection, and the evidential weight of multiple outcomes.

Finally, the safety figures illustrate that counts have different meanings depending on how they are recorded. Serious adverse events are listed as zero affected in each condition, while other adverse events are reported as counts. Null death entries do not state a confirmed zero. Careful distinctions between zero, unreported, and not estimable are important for interpreting registry data without overstating what the record establishes.

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.

The primary limitation is that the registry reports separate condition means and confidence intervals but no direct treatment comparison for the primary endpoint. Since the study is a crossover trial, an inferential comparison should reflect the paired observations. The registry record does not report the paired treatment contrast, its uncertainty, or the statistical model used. It also notes that values are missing or not estimable for a two-arm comparison of this continuous endpoint.

The analysis population is described as all completers, but participant flow and reasons for non-completion are not reported. The relationship between the enrollment of 53 and the posted condition counts of 28 is not explained, and the record does not specify how missing condition measurements were handled. These gaps prevent a full assessment of selection and missing-data effects.

The record identifies randomization and triple masking but does not describe the allocation procedure or which parties were masked. It gives the treatment sequences but does not report how the analysis addressed sequence, period, or possible carryover. Washout details, if any, are not reported in the registry record. The treatment timeframes differ as recorded: baseline to end of placebo treatment (2 weeks + 4 weeks) and baseline to end of LDN treatment (2 weeks + 12 weeks); the rationale and analytic handling of those timeframes are not reported.

The registry provides no participant-level data, standard deviations, or breakdown by sequence, so response variability and paired covariance cannot be examined. It lists secondary outcomes, but no multiplicity procedure is reported. The protocol’s statistical plan and sample-size justification are not reported in the registry record. These limitations mean the posted results support descriptive interpretation of the reported means and intervals, but not a complete assessment of the primary treatment comparison.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Effects of Low Dose Naltrexone in Fibromyalgia — NCT00568555.
• Registry record last updated 2015-10-21; status last verified 2015-09.
• Results first posted 2015-10-21.
• 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 completed randomized, triple-masked crossover trial sponsored by Stanford University, with an enrollment of 53. Participants received LDN and placebo in an assigned sequence. The primary outcome was mean percentage change in pain scores from baseline to the end of each condition, measured using a Visual Analogue Scale from 0 to 100. The primary population is listed as all completers.

The posted mean percentage change was 28.8 for LDN (95% confidence interval, 19.5 to 38.1) and 18.0 for placebo (95% confidence interval, 7.2 to 28.8). The registry does not report a direct treatment comparison or its statistical test. Serious adverse events are recorded as affecting zero participants in each condition; other adverse events are reported for 20 participants in the LDN condition and 7 in the placebo condition. Further details of the primary analysis, missing-data handling, and participant flow are not reported in the registry record.