Minocycline for the Treatment of Decreased Mental Function in HIV-Infected Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Minocycline for the Treatment of Decreased Mental Function in HIV-Infected Adults (NCT00361257), a randomized phase 2 study. The primary outcome is Change in Cognitive Performance Compared to Baseline.

2. Clinical Question

The clinical question is therefore a comparison of changes between randomized groups, rather than a test of whether either group changed from baseline on its own. The primary endpoint is continuous: each participant’s change is represented on a z-score scale, and the group summaries describe mean change and its variability. The registry also lists secondary outcomes covering a global deficit score, a clinical global impression, and several cognitive domains. Those measures provide additional perspectives, but the primary question remains the between-group comparison on NPZ-8 at week 24.

The record gives an enrollment figure of 107. The posted primary outcome summaries are based on 41 participants in the minocycline arm and 48 in the matching-placebo arm, while the primary analysis description says that the analysis was conducted using an intention-to-treat approach with missing week-24 outcomes imputed by multiple regression. The registry does not report detailed eligibility criteria, so the precise boundaries of the population to which the result applies are not reported in the registry record.

3. Trial Design

This was a phase 2, randomized, placebo-controlled, double-blind, parallel-group study. It was registered as a study of minocycline for decreased mental function in adults with HIV infection. The registry reports enrollment of 107 and identifies HIV infections as the condition. The study status is terminated; the registry record does not give a reason for termination, so no conclusion about why it ended is warranted.

The two arms were minocycline and matching placebo. The minocycline arm received 100 mg orally every 12 hours; the placebo arm was described as taken orally every 12 hours. Random allocation supports a comparison between groups by reducing the influence of systematic differences at assignment, while a placebo comparator and double masking are design features intended to limit differences in care and assessment related to knowledge of assignment. Those features do not by themselves guarantee that all bias is absent.

The primary outcome was change in cognitive performance compared with baseline, measured at baseline and week 24. Its measure, NPZ-8, combines eight neuropsychological test components into an average of age- and education-adjusted z-scores. The registry also lists secondary cognitive and clinical measures. It reports 24 outcome measures and 19 analyses, and 23 secondary outcomes are posted in the outcome information. The record does not explain the relationship among these counts or provide a complete account of every planned analysis.

The primary outcome description distinguishes the descriptive summaries from the inferential analysis: descriptive statistics use observed data, whereas the analysis was conducted under an intention-to-treat approach and missing week-24 values were imputed using multiple regression. The precise randomization procedures, allocation concealment details, original sample-size assumptions, and detailed statistical analysis plan are not reported in the registry record. The registry gives a study start of 2007-03 and primary completion of 2010-01; those dates describe the record and do not explain the conduct or reasons for termination.

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.

Step 1: Minocycline vs. Placebo

Step 2: Minocycline(Open Label) for All

Analysis population for the primary outcome, as defined in the registry: The descriptive statistics above were based on the observed data; however, the statistical analysis was conducted by the ITT analysis and the missing outcomes at week 24 were imputed based on multiple regression imputations.

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 Cognitive Performance Compared to Baseline, measured over At baseline and week 24.

The primary endpoint is change in cognitive performance from baseline to week 24, expressed as the NPZ-8 score at week 24 minus the NPZ-8 score at baseline. NPZ-8 is the average of age- and education-adjusted z-scores from eight neuropsychological test components: Grooved Pegboard with the dominant and non-dominant hands, Choice Reaction Time, Sequential Reaction Time, Timed Gait, Trail Making Parts A and B, and Symbol Digit. The registry describes the endpoint in z-score units and reports means and standard deviations.

A composite measure can bring several related test results into one outcome, making the primary comparison less dependent on any single test component. Age- and education-adjusted scores are intended to express performance relative to those characteristics, while averaging the components produces an overall measure of the specified cognitive test battery. This endpoint is useful for assessing change in test performance across a broad set of measured functions. It does not directly measure every lived or functional consequence of cognitive impairment, and the registry does not report a threshold for a clinically important change.

The comparison of group mean changes addresses whether average change differed between minocycline and matching placebo over the specified follow-up. The endpoint’s interpretation depends on the scale, the completeness and handling of outcome data, and the analysis model. The record states that missing week-24 outcomes were imputed for the intention-to-treat analysis, while the descriptive statistics were based on observed data. It does not report further details about the imputation model or assumptions.

7. Results Posted in the Registry

• Minocycline: mean 0.12 versus 0.17 on control; difference -0.05 (95% CI -0.34 to 0.24), standardized effect d = -0.07; z-test p = 0.734.

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.

Change in Global Deficit Z-Score (GDS)

• 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 deviations.
• Result:
  • Minocycline: mean 0.11 versus 0.09 on control; difference 0.02 (95% CI -0.21 to 0.25), standardized effect d = 0.04; z-test p = 0.864.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Minocycline and the control. The interval gives the range of effects that remain compatible with the data.

Change in Cognitive Gross Motor Function Domain Z-Score

• 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 deviations.
• Result:
  • Minocycline: mean 0.22 versus 0.08 on control; difference 0.14 (95% CI -0.81 to 1.09), standardized effect d = 0.06; z-test p = 0.772.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Minocycline and the control. The interval gives the range of effects that remain compatible with the data.

Change in Fine Motor Function Domain Z-Score

• 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 deviations.
• Result:
  • Minocycline: mean 0.27 versus 0.05 on control; difference 0.22 (95% CI -0.09 to 0.53), standardized effect d = 0.30; z-test p = 0.164.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Minocycline and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes broaden the assessment beyond the overall NPZ-8 measure. They include a global deficit score, investigator clinical global impression, and domains covering motor, psychomotor, information-processing, and memory functions. Across the secondary analyses, estimates are not uniformly directed, and the posted comparisons do not form a clear, consistent pattern of separation between minocycline and matching placebo. Some measures have estimates favoring one arm on the numerical scale, but direction on a score is not automatically equivalent to clinical benefit unless the scale’s interpretation is established.

This overall pattern is broadly consistent with the primary result in one important respect: the secondary outcomes do not turn the uncertain primary comparison into a clear, robust finding across measures. They may provide context about which aspects of cognitive performance were measured, but they do not replace the primary endpoint or independently resolve its uncertainty. The registry posts 23 secondary outcomes and reports 19 analyses, while the broader outcome listing identifies 24 outcome measures. The record does not explain the differences among those counts or establish which analyses were planned as confirmatory.

Unadjusted secondary results carry less weight because many outcomes and comparisons create more opportunities for chance findings. Without a reported multiplicity strategy, individual p-values cannot be interpreted as if only one question had been tested. The registry does not report whether adjustments were made, whether secondary analyses were prespecified in a detailed plan, or whether these outcomes were intended to support confirmatory conclusions. They are best read as exploratory evidence about the range of measured outcomes, with the primary comparison remaining central.

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

• Change in Global Deficit Z-Score (GDS)
• Change in Investigator's Clinical Global Impression Score (ICGIS)
• Change in Cognitive Gross Motor Function Domain Z-Score
• Change in Fine Motor Function Domain Z-Score
• Change in Psychomotor Function Domain Z-Score
• Change in Fine Motor/Nonverbal Function Domain Z-Score
• Change in Information Processing Function Domain Z-Score
• Change in Verbal Memory Domain Z-Score

9. Statistical Interpretation

10. Statistical Methods Explained

What does a mean difference describe?

For a continuous outcome such as a z-score, a group mean is the average observed value for that arm. The posted descriptive statistics give the mean change and standard deviation in each arm. A mean difference compares the two arm means, with the direction determined by which arm is subtracted from the other. In the recalculation, the contrast is minocycline minus matching placebo. A negative estimate therefore means the observed mean change was lower in the minocycline arm on this subtraction convention; it does not, by itself, establish a clinical advantage or disadvantage.

The standard deviation describes how spread out individual observations are around a group mean. It is not the uncertainty in the mean itself. General statistical explanation: the standard error describes how much an estimate would tend to vary across repeated samples under a specified model. The uncertainty around a difference depends on both groups’ variability and the amount of information in each group. The registry provides group standard deviations and standard errors for the posted summaries, while the inferential result for the primary endpoint is reported from a linear regression model.

How should the confidence interval be read?

A confidence interval gives a range of estimates compatible with the data and the specified statistical procedure. In general, a 95% confidence interval is constructed by a method that would cover the target quantity in 95% of repeated samples if the model assumptions hold. It is not a statement that there is a 95% probability that this particular interval contains the true value. For the primary reported regression result, the estimate is 0.064, and the 95% confidence interval extends from -0.258 to 0.386. Because the interval includes zero and values in both directions, the estimate is not precise enough to settle the direction or size of the difference.

The unadjusted recalculation uses the posted group means and standard deviations to contrast the two arms. It is a simple comparison, not the registry’s regression model. Its mean difference is -0.05 and its unadjusted p-value is 0.734. The confidence interval for that recalculation also includes zero. The two analyses are related because they address the primary outcome comparison, but their estimates differ because the analyses are not identical. The registry does not give enough detail about the regression model to determine all reasons for the difference.

What is the role of the p-value?

A p-value assesses how unusual data at least as discrepant as those observed would be under a specified null model and its assumptions. It does not state the probability that the null model is true, the probability that a result occurred “by chance,” or the probability that a treatment effect exists. Here, the registry reports p = 0.651 for the regression analysis, and the unadjusted recalculation reports p = 0.734. Both are consistent with substantial uncertainty about the between-group contrast. The p-value should be read alongside the estimate and its confidence interval rather than as a stand-alone verdict.

Why distinguish the analysis from the descriptive summaries?

The primary outcome description says the descriptive statistics are based on observed data, but the statistical analysis was conducted using an intention-to-treat approach and missing week-24 outcomes were imputed based on multiple regression. An intention-to-treat approach aims to analyze participants according to their randomized assignments, regardless of later treatment adherence, although the record does not detail the operational rules used. Imputation replaces missing outcome values with estimates informed by a model, allowing an analysis to include participants whose endpoint values are missing. Its validity depends on assumptions about why values are missing and how well the imputation model represents the observed data.

General statistical explanation: missingness can be related to observed information, unobserved outcomes, or both. Multiple regression imputation can account for measured predictors in estimating missing values, but it cannot remove uncertainty about unobserved outcomes or guarantee that the imputation assumptions are correct. The registry does not report the variables in the imputation model, the number of imputations, diagnostics, or sensitivity analyses. Those details are not reported in the registry record.

Why can over-dispersion matter?

General explanation: over-dispersion occurs when the observed variability is greater than a statistical model assumes. If a model understates variability, standard errors can be too small and confidence intervals too narrow, making apparent precision misleading. Continuous outcomes are often analyzed with methods that model their variation, but whether over-dispersion is relevant depends on the outcome distribution and the fitted model. The registry identifies standard deviations for the group summaries and linear regression for the reported primary analysis, but does not report diagnostics or a specific assessment of over-dispersion. It would therefore be inappropriate to claim that over-dispersion was present or absent in this trial.

What can be learned from the secondary analyses?

Secondary measures cover global deficit, clinical impression, and cognitive domains. They can help describe whether the primary result is accompanied by a similar pattern across related outcomes. However, assessing many outcomes creates a multiplicity problem: even when there is no real effect, multiple separate tests can produce apparently notable results. General statistical explanation: adjustment procedures can control error rates across a family of tests, but they may also change the interpretation and precision of individual comparisons. The registry does not report a multiplicity strategy, so the posted secondary p-values are best treated as exploratory rather than as independent confirmatory evidence.

The planning illustrations use the observed effect as a starting point and consider smaller fractions of that effect. They assume a two-sided significance level of 0.05, equal allocation, and a normal approximation. General statistical explanation: a sample-size calculation is a prospective planning exercise that estimates the group size needed to have a specified chance of detecting an assumed effect under stated conditions. It is not a measure of what the completed trial achieved. The original trial’s own sample-size justification and assumptions are not reported in the registry record.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data estimate a comparison. The primary regression estimate is 0.064, with a 95% confidence interval from -0.258 to 0.386. This interval includes zero, as well as negative and positive values, so the result does not sharply distinguish the direction or size of a possible group difference. The interval’s width is a reminder that the estimate is uncertain, even though it is centered on a modest positive value.

The unadjusted recalculation gives a mean difference of -0.05 and a confidence interval that also includes zero. Its unadjusted p-value is 0.734. The posted group means are 0.12 for minocycline and 0.17 for matching placebo, with standard deviations of 0.71 and 0.67, respectively. These are descriptive observed-data summaries; they do not carry the same model specification as the reported regression analysis.

Confidence intervals are more informative than a binary distinction based on whether a p-value crosses a conventional threshold. A wide interval can include both no difference and differences that might matter, while a narrow interval can rule out a wider range of possibilities. The registry does not define a clinically important NPZ-8 difference, so the practical importance of values within the interval cannot be judged from a threshold reported in the record. The result is best characterized as inconclusive about the magnitude and direction of the primary comparison, rather than as proof of equality.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrative estimates based on the observed effect, not the trial’s own original design assumptions. For these continuous outcomes, the listed figures are participants per arm, not total events. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect set at the stated fraction of the observed effect. The registry does not report the assumptions used for the original trial’s sample-size justification.

The 80% power column describes the number of participants per arm estimated to give an 80% chance of detecting the assumed effect under the stated calculation assumptions, if that effect and the model assumptions are correct. The 90% power column uses a higher target chance of detection and therefore gives a larger planned group size for the same assumed effect. These are design concepts: power is calculated before a study from an assumed effect, variability, significance level, allocation, and analysis approach.

For the primary outcome, planning for 100% of the observed effect gives 2,979 participants per arm for 80% power and 3,987 participants per arm for 90% power. Planning for 75% of the observed effect gives 5,295 and 7,088 participants per arm, respectively; planning for 50% gives 11,913 and 15,948. Smaller effects are harder to distinguish from random variation, so they require more participants under the same assumptions. These values are not observed power and do not describe the probability that the completed trial detected its effect.

The planning display should be read as a sensitivity illustration, not as a recommendation or a reconstruction of the original design. It shows how assumed effect size and desired power change the planned group size when the other assumptions are held fixed. The trial’s own target effect, power target, and original sample-size rationale are not reported in the registry record.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.050; pooled standard deviation σ = 0.689.
2. n = 2 × (1.96 + 0.842)² × 0.689² / -0.050² = 2,978.0, rounded up to 2,979 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.05                      # observed difference in means
sigma <- 0.688679395905546                     # 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(361257)
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 = 2,979 (80% power), 3,987 (90% power)
 75% of observed effect: n per arm = 5,295 (80% power), 7,088 (90% power)
 50% of observed effect: n per arm = 11,913 (80% power), 15,948 (90% power)
power.t.test: 2979 
Simulated power at n = 2,979 per arm: 0.797 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry’s reported primary analysis is a linear regression result: slope 0.064, with a 95% confidence interval from -0.258 to 0.386 and p = 0.651. This is the sponsor-reported inferential analysis of the primary endpoint. Its interval includes zero and spans effects in both directions, so the result remains uncertain about the comparative change.

The recalculated estimate is an unadjusted difference between the posted observed-data means: -0.05 z-score, with an unadjusted p-value of 0.734. It is a simple re-analysis of summary statistics, not a reproduction of the regression model or its intention-to-treat analysis with imputed outcomes. The methods therefore differ, and the estimates and p-values need not match exactly. The registry record does not provide enough model detail to recreate the sponsor’s regression analysis from the posted summaries.

The record does not flag a discrepancy between the results. That does not make the two procedures identical; it means the recalculation should be interpreted as a transparent, unadjusted comparison alongside the reported analysis, not as a replacement for it. Both analyses leave the primary comparison imprecise and neither supplies evidence that resolves the possibility of effects in either direction. The reported model result is the relevant registry analysis; the recalculation is useful for understanding the raw group-mean contrast.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when investigators examine several outcomes, time points, subgroups, or statistical contrasts. If each is assessed separately without accounting for the full set of tests, the chance of at least one seemingly notable result can increase. This matters particularly for secondary outcomes, where a result may look persuasive in isolation even though many related questions have been examined. General statistical explanation: multiplicity procedures can control a chosen error rate across a defined family of comparisons, but their use and interpretation depend on how that family is defined.

The registry lists numerous secondary measures and reports multiple outcome and analysis counts, but it does not report a multiplicity adjustment or a hierarchy assigning confirmatory status to secondary endpoints. It is therefore not possible to establish from the record whether the posted secondary p-values were adjusted, which outcomes formed any testing family, or how their results were intended to be interpreted. The primary comparison can be distinguished from the wider set of secondary results, but the record does not provide a full prespecified analysis plan.

For missing data, the primary outcome description says descriptive statistics were based on observed data, while the statistical analysis used an intention-to-treat approach and imputed missing week-24 outcomes based on multiple regression. This indicates that missing outcomes were addressed in the inferential analysis rather than simply omitted from it. The registry does not report the imputation predictors, model specification, number of imputed data sets, or sensitivity analyses to alternative assumptions.

Missing outcomes can affect validity if the reasons for missingness relate to outcomes in ways not captured by the imputation model. Multiple regression imputation can use observed information, but the method does not make missing values known, and its conclusions depend on assumptions. The registry does not describe the amount or pattern of missingness, reasons for missing assessments, or checks of the imputation assumptions. Detailed participant-level analysis populations and the original statistical plan are also not reported in the registry record. These gaps limit independent assessment of how robust the primary estimate is to analysis choices.

16. Safety Results

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

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: blood amylase increased (1.9%), vocal cord paralysis (1.9%), basal cell carcinoma (1.9%), tooth infection (1.9%), suicidal ideation (1.9%).

The serious adverse event summaries report 7 affected participants among 52 at risk in the minocycline group and 3 among 55 at risk in the matching-placebo group. For other adverse events, the record reports 18 affected participants among 52 at risk with minocycline and 34 among 55 at risk with matching placebo. These are counts and at-risk denominators as recorded; they should not be interpreted as proof that treatment caused or prevented any event.

The record lists serious-event terms including blood amylase increased, vocal cord paralysis, basal cell carcinoma, tooth infection, suicidal ideation, and blood triglycerides increased, each reported as 1.9%. The listing does not establish that these terms are exhaustive or provide enough context to compare event types by treatment arm. Death counts are not reported in the registry record.

Safety summaries require caution because event attribution, timing, severity, recurrence, and the relationship to study treatment are not described in detail in these figures. The reported counts are a descriptive account within the recorded at-risk groups, not a complete causal assessment or a substitute for clinical interpretation of individual events.

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

The posted results do not establish a clear difference between minocycline and matching placebo in mean change in NPZ-8 at week 24. The sponsor-reported regression estimate is 0.064, with a 95% confidence interval from -0.258 to 0.386 and p = 0.651. The unadjusted recalculation gives a mean difference of -0.05 and p = 0.734. Both analyses leave meaningful uncertainty, and neither proves that the groups are equivalent.

The interval around the reported analysis includes zero and values on both sides. It therefore does not settle whether the comparative change was positive or negative, nor does the registry identify a clinically important difference that could be used to judge the interval against a practical threshold. The result is about a composite neuropsychological score, not every dimension of cognition, daily functioning, or quality of life. The outcome was assessed at baseline and week 24; longer-term effects are not addressed by this primary comparison.

The study’s objective concerned minocycline for decreased mental function associated with HIV, with the brief summary referring to effects sometimes caused by anti-HIV drugs. The result should not be generalized beyond the studied population, intervention, comparator, outcome, and follow-up without additional evidence. Detailed eligibility criteria and other population characteristics are not reported in the registry record.

The trial’s randomized, double-blind, placebo-controlled design supports a structured comparison, but design features do not remove uncertainty from incomplete outcomes, model assumptions, or imprecision. The primary analysis used imputation for missing week-24 outcomes, and the record does not report enough detail to assess those assumptions fully. The appropriate conclusion is limited: the posted evidence does not provide a precise estimate of the comparative effect on NPZ-8, and it does not justify a broad claim about cognitive benefit or lack of benefit.

18. Why This Trial Matters Statistically

This trial is statistically instructive because the primary question concerns change in a continuous composite measure, and the record presents both observed-data summaries and an intention-to-treat analysis with imputed missing outcomes. These are distinct layers of evidence. Group means help describe the observed results; a model-based analysis addresses the comparison under stated assumptions. Keeping them separate prevents a simple recalculation from being mistaken for the trial’s full inferential analysis.

The primary estimate is modest relative to its uncertainty. Its confidence interval spans zero and effects in both directions, illustrating why a point estimate alone is not enough to characterize evidence. A p-value also cannot replace the estimate and interval: it does not tell readers the probability that the intervention works, and it does not communicate the range of effects compatible with the analysis.

The trial also highlights the importance of planning assumptions. The illustrative sample-size calculations change substantially when the assumed effect is reduced, because smaller differences are harder to distinguish from variation. Those calculations are planning illustrations based on the observed effect and stated assumptions, not evidence about what the original trial was designed to detect. The original design justification is not reported in the registry record.

Finally, the registry includes many secondary outcomes, making multiplicity and missing-data reporting important to interpretation. Without a stated testing strategy or detailed imputation information, readers should distinguish the primary analysis from exploratory patterns and avoid over-reading individual secondary results. These issues are broadly relevant to trials that assess several related cognitive domains using both observed summaries and model-based analyses.

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.

Several trial-specific limitations constrain interpretation. The registry marks the study as terminated but does not report why it was terminated. The enrollment figure is 107, whereas the posted primary outcome summaries show 41 participants for minocycline and 48 for matching placebo. The registry does not explain the relationship between enrollment and the observed-data outcome groups, so the counts should not be used to infer reasons for missing outcome data or participant flow.

The primary outcome description states that missing week-24 outcomes were imputed using multiple regression for the intention-to-treat analysis, but the variables, model details, number of imputations, missingness patterns, and sensitivity analyses are not reported. These omissions make it difficult to assess how strongly the primary result depends on assumptions about missing outcomes. The descriptive statistics use observed data, so they and the imputed analysis do not represent identical data handling.

The registry does not report detailed eligibility criteria, baseline characteristics, adherence, treatment discontinuations, or the specific randomization and allocation-concealment procedures. It also does not give a full statistical analysis plan, explain the relationship among the counts of outcomes and analyses, or report a multiplicity strategy for the secondary measures. The original sample-size justification and the target effect assumed for the trial are not reported in the registry record.

Finally, the primary endpoint is a composite test score at week 24. The record does not provide a clinically important change threshold, and the result does not directly establish effects on everyday function or longer-term cognition. These limitations do not invalidate the posted findings, but they restrict the level of detail and certainty with which the comparison can be interpreted.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Minocycline for the Treatment of Decreased Mental Function in HIV-Infected Adults — NCT00361257.
• Registry record last updated 2016-02-05; status last verified 2016-01.
• Results first posted 2011-07-26.
• 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

The registry describes a randomized, double-blind, parallel-group phase 2 study of minocycline versus matching placebo in adults with HIV infection and cognitive impairment. Minocycline was given orally at 100 mg every 12 hours; the placebo was also taken orally every 12 hours. Enrollment was 107, and the study status is terminated.

The primary endpoint was change from baseline to week 24 in NPZ-8, an average of age- and education-adjusted z-scores from eight neuropsychological test components. The reported linear regression result was 0.064, with a 95% confidence interval from -0.258 to 0.386 and p = 0.651. The unadjusted recalculation was -0.05, with p = 0.734. Both leave substantial uncertainty about the comparative change.

The primary outcome description says missing week-24 outcomes were imputed using multiple regression for the intention-to-treat analysis, while descriptive statistics were based on observed data. Details of the imputation assumptions, multiplicity strategy, original sample-size justification, and reason for termination are not reported in the registry record.