Staccato Loxapine in Agitation (Proof of Concept): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Staccato Loxapine in Agitation (Proof of Concept) (NCT00369577), a randomized phase 2 study. The primary outcome is PANSS-EC Change From Baseline.

2. Clinical Question

The study asked whether a single inhaled dose of loxapine was associated with a different change in agitation scores than inhaled placebo in patients with schizophrenia and acute agitation. It compared two loxapine doses with placebo, with the primary outcome assessed 2 hours after treatment.

The registry also lists changes in BARS, the Clinical Global Impressions-Improvement Scale, and CGI-I responder status as secondary outcomes. These provide additional perspectives on agitation and clinical improvement, but they do not replace the primary endpoint or resolve limitations in the reported analysis.

3. Trial Design

This was a Phase 2, randomized, quadruple-masked, parallel-group, placebo-controlled study. Its official title describes a multicenter, single-dose efficacy and safety design. The registry records the study as completed, with 129 enrolled participants. The brief summary describes a plan to study 120 agitated patients; the registry record does not explain the difference between that planned figure and the recorded enrollment.

Participants were assigned to inhaled placebo, inhaled loxapine 5 mg, or inhaled loxapine 10 mg. Each was described as a single dose. The registry identifies the sponsor as Alexza Pharmaceuticals, Inc. The recorded study start is 2006-08 and the primary completion is 2007-01.

The primary outcome was mean change from baseline in PANSS-EC at 2 hours. Posted outcome groups included 43 participants for placebo, 45 for the 5 mg group, and 41 for the 10 mg group. The registry defines the ITT population as patients who took any study medication and had baseline data and at least one efficacy assessment after treatment. It also states that observations recorded after rescue medication were censored and handled using last observation carried forward.

The record lists four outcome measures, including three secondary outcomes, and eight analyses. It does not report the original statistical analysis plan, the trial’s own sample-size assumptions, the detailed randomization procedure, or the responsibilities covered by quadruple masking. These details should not be inferred from the design labels alone. The posted results are the basis for describing the outcome estimates; the registry’s own analysis is the relevant source for its formal comparisons.

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: Pre-Treatment Period was defined as the period immediately prior to dosing in which screening procedures and inclusion/exclusion criteria were used to evaluate all patients for eligibility to participate in the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The ITT population consisted of all patients who took any study medication and had both baseline data and at least 1 efficacy assessment after the. Any observation recorded after the use of rescue medication was censored and subject to the last observation carried forward algorithm.

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 PANSS-EC Change From Baseline, measured over 2 hours.

The primary endpoint was change from baseline in the Positive and Negative Syndrome Scale-Excited Component, or PANSS-EC, at 2 hours. The scale comprises five agitation-related items: poor impulse control, tension, hostility, uncooperativeness, and excitement. Each item is scored from 1, the minimum, to 7, the maximum, and the total ranges from 5 to 35.

The endpoint was used to quantify change in agitation over the short period after the single study dose. The registry says eligible participants had a PANSS-EC score of at least 14 and a score of at least 4 on at least one of the five items. These entry criteria selected patients with elevated agitation-related scores under the study’s definitions.

Change scores compare the follow-up assessment with baseline. A negative mean change indicates that the average follow-up score was lower than the average baseline score; it does not by itself establish how much of that change is attributable to study treatment. The placebo group also had a negative mean change, so the treatment comparisons must be read against that group rather than by considering each group’s change in isolation.

The endpoint is continuous, and the posted results give group means and standard deviations. Such summaries describe average change and variation among participants, but they do not show the full distribution of individual responses. The registry’s ITT definition and its handling of assessments after rescue medication also affect which observations contribute to the reported endpoint.

7. Results Posted in the Registry

• Inhaled Loxapine 5 mg: mean -6.71 versus -4.98 on control; difference -1.73 (95% CI -3.67 to 0.21), standardized effect d = -0.37; z-test p = 0.081.
• Inhaled Loxapine 10 mg: mean -8.56 versus -4.98 on control; difference -3.58 (95% CI -5.52 to -1.64), standardized effect d = -0.79; z-test p = <0.001.

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.

BARS Change From Baseline After Drug Treatment

• 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:
  • Inhaled Loxapine 5 mg: mean -1.53 versus -0.93 on control; difference -0.60 (95% CI -1.09 to -0.11), standardized effect d = -0.51; z-test p = 0.017.
  • Inhaled Loxapine 10 mg: mean -2.02 versus -0.93 on control; difference -1.09 (95% CI -1.59 to -0.59), standardized effect d = -0.94; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Inhaled Loxapine 5 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Inhaled Loxapine 10 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Clinical Global Impressions-Improvement Scale (CGI-I) After Drug Administration

• 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:
  • Inhaled Loxapine 5 mg: mean 2.53 versus 3.19 on control; difference -0.66 (95% CI -1.09 to -0.23), standardized effect d = -0.65; z-test p = 0.002.
  • Inhaled Loxapine 10 mg: mean 2.28 versus 3.19 on control; difference -0.91 (95% CI -1.37 to -0.45), standardized effect d = -0.86; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Inhaled Loxapine 5 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Inhaled Loxapine 10 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

CGI-I Responders

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Inhaled Loxapine 5 mg: 48.9% (22/45) versus 20.9% (9/43) on control; risk difference 28.0 percentage points (95% CI 9.0 to 47.0), risk ratio 2.34 (1.22–4.49); unadjusted two-proportion z-test p = 0.006.
  • Inhaled Loxapine 10 mg: 62.5% (25/40) versus 20.9% (9/43) on control; risk difference 41.6 percentage points (95% CI 22.3 to 60.9), risk ratio 2.99 (1.59–5.60); unadjusted two-proportion z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Inhaled Loxapine 5 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the proportion with the event is statistically higher for Inhaled Loxapine 10 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes offer additional measures of agitation and improvement at 2 hours. Across these measures, the posted group summaries generally show changes or scores in a direction consistent with the primary pattern. CGI-I is coded so that lower scores indicate better status, while the other outcomes have their own measurement scales. The responder outcome presents counts rather than a continuous mean.

These secondary results complement the primary PANSS-EC comparison by examining related but not identical aspects of the participants’ status. Agreement in broad direction can be informative, but it does not make the outcomes interchangeable. Each scale captures a different construct or summary, and the registry does not report that these secondary measures were designated to confirm the primary finding.

The registry lists Wilcoxon (Mann-Whitney) tests for the continuous secondary measures and Fisher exact tests for responder status. It does not report a multiplicity adjustment. When several outcomes and comparisons are examined, the chance of finding at least one apparently notable result can increase. For this reason, unadjusted secondary p-values should carry less weight than the prespecified primary result and should be interpreted as exploratory evidence unless the analysis plan establishes otherwise. That plan is not reported in the registry record.

The secondary outcomes are therefore best read as contextual evidence about the pattern of measured responses, not as independent proof of a treatment effect. Their limitations include the same short assessment timeframe, the reported handling of observations after rescue medication, and the absence of a reported multiplicity strategy. The registry does not provide enough information to determine how the sponsor ranked or prioritized the secondary outcomes.

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

• BARS Change From Baseline After Drug Treatment
• Clinical Global Impressions-Improvement Scale (CGI-I) After Drug Administration
• CGI-I Responders

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: PANSS-EC change from baseline is summarized by a mean and standard deviation in each group. A mean describes the group’s average change, while the standard deviation describes the spread of individual values around that average. A group mean is not a prediction for every participant, and the standard deviation is not the uncertainty interval for the mean.

What does the mean difference represent?

The unadjusted mean difference compares the average change in a loxapine group with the average change in the placebo group. For the 5 mg comparison, the recalculated difference is -1.73 PANSS units; for the 10 mg comparison, it is -3.58 PANSS units. Negative values indicate a more negative average change in the loxapine group than in placebo. These are simple comparisons of posted summary statistics, not the registry’s adjusted ANCOVA estimates.

An analysis of covariance, or ANCOVA, compares groups while accounting for a measured covariate or covariates. The registry identifies ANCOVA for both primary comparisons, but does not provide the adjusted estimates, confidence intervals, covariates, or other model details. Therefore, the posted sponsor p-values cannot be used to reconstruct the adjusted treatment differences. The record’s formal analysis is more relevant to the trial’s own inference than a separate calculation from group means and standard deviations.

How should a p-value be read?

A p-value describes how unusual data at least as discordant with a specified null hypothesis would be under the assumptions of the analysis. It is not the probability that the null hypothesis is true, the probability that results occurred by chance, or a measure of clinical importance. The registry labels the primary analyses as superiority analyses. A p-value alone does not establish the size or practical importance of a difference.

The recalculated p-values are unadjusted and arise from simple group-summary comparisons. The registry’s primary p-values come from ANCOVA and differ from those recalculations. Since the record flags a discrepancy, the sponsor’s own analysis takes precedence. That does not make the summary-statistic calculations equivalent to the sponsor analysis; they serve as a separate, transparent check using the posted means and standard deviations.

What does a confidence interval add?

A confidence interval describes a range of estimates compatible with the data and the method used. A narrower interval indicates greater precision under that method; a wider interval indicates more uncertainty. The confidence level for the intervals in the posted recalculations is not stated in the registry record, so it should not be assumed. For the 5 mg recalculation, the interval extends across zero. For the 10 mg recalculation, it lies below zero. Neither interval describes all possible sources of bias or uncertainty in the trial.

The recalculated comparisons use pooled standard deviations from the posted group summaries. Standardization can also express a mean difference relative to the observed variability, but such a standardized effect is not a direct clinical unit and is not needed to interpret the PANSS-unit estimates. No additional standardized values are required to understand the reported primary comparisons.

Why do the secondary analyses use different tests?

The registry lists Wilcoxon (Mann-Whitney) tests for the BARS and CGI-I score outcomes, and Fisher exact tests for responder status. A rank-based Wilcoxon test compares the relative ordering of observations between groups and does not directly test a difference in means in the same way as a mean-based model. Fisher exact testing evaluates group differences in a binary outcome using the observed counts. The record does not provide further details about test assumptions, sidedness, or adjustment for multiple comparisons.

The primary outcome’s reported ANCOVA and the separate unadjusted mean-difference recalculation answer related but not identical statistical questions. This distinction matters: an adjusted model can account for specified variables, while the recalculation uses only the posted summary statistics. The registry does not state which covariates were used in ANCOVA, so the adjusted comparison cannot be independently reproduced from the record.

How does the analysis population affect the result?

The record defines the ITT population as participants who took any study medication and had baseline data plus at least one efficacy assessment after treatment. This is the definition reported for this study, and it should not be assumed to include every randomized participant. The record also says assessments after rescue medication were censored and subject to last observation carried forward. That approach replaces a missing later observation with the last recorded value under the stated rule. It can affect estimates if outcomes would otherwise have changed after that observation. The registry does not report how many observations were censored, the amount of missing data, or sensitivity analyses assessing this approach.

11. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimated comparison. The recalculated primary mean differences are accompanied by confidence intervals: -1.73 for 5 mg versus placebo, with limits -3.674 and 0.214; and -3.58 for 10 mg versus placebo, with limits -5.523 and -1.637. The interval for the 5 mg comparison spans zero, while the interval for the 10 mg comparison does not. The confidence level is not reported in the registry record, so these intervals should be interpreted as the posted recalculation intervals without assigning them an unstated level.

The intervals are not guarantees about the true effect, and they do not incorporate every uncertainty that may arise from study conduct, outcome measurement, missingness, or analysis choices. They quantify sampling uncertainty under the recalculation method. The registry’s sponsor analyses are ANCOVA comparisons, but adjusted confidence intervals are not reported. Because the record flags a discrepancy, the sponsor’s own analysis takes precedence for the formal trial analysis.

Group standard deviations indicate substantial variability in change scores relative to the group means. This means that the average change does not describe every participant’s response. The group sizes and the spread of observed values influence precision: in general, more information and less variability tend to narrow an interval, while less information or greater variability tend to widen it. This general statistical principle does not resolve the unreported details of the sponsor model or missing-data handling.

Precision should also be distinguished from importance. An interval can be informative about the range of values compatible with a particular statistical method, but the registry does not supply a threshold for a clinically important PANSS-EC difference. Thus the estimates and intervals describe the size and uncertainty of score differences, not whether those differences meet an unstated clinical criterion.

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 illustrations express required participants per arm for the continuous-outcome comparisons. They are not total study enrollment figures or event counts. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to be detected is expressed as a stated fraction of the observed effect. The original trial’s own sample-size justification and assumptions are not reported in the registry record.

The 80% power column gives the illustrative participants per arm for a design with an 80% probability of detecting the specified effect under those assumptions, if that effect and the model assumptions were correct. The 90% power column gives the corresponding planning figure for 90% power. These are prospective design calculations, not estimates of what the completed study achieved.

Rows based on 100% of the observed effect use the full observed comparison as the planning target. Rows based on 75% or 50% use smaller target effects. Smaller effects are harder to detect with the same design and generally require more participants per arm to attain the same planned power. This relationship explains why the illustrative participant requirements rise as the target fraction decreases.

The values are planning illustrations computed from observed effects, not a reconstruction of the trial’s design. They are not observed power. Observed power is not a useful substitute for reading the estimated effect and its uncertainty: the completed study’s results are better understood through the group estimates, confidence intervals, and the registry’s own analysis.

13. How the Sample Size Is Derived

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

Formula

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

Worked calculation for this trial

1. Observed difference Δ = -1.730; pooled standard deviation σ = 4.674.
2. n = 2 × (1.96 + 0.842)² × 4.674² / -1.730² = 114.6, rounded up to 115 participants per arm for 80% power.

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

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -1.73                      # observed difference in means
sigma <- 4.67409013655611                     # 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(369577)
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 = 115 (80% power), 154 (90% power)
 75% of observed effect: n per arm = 204 (80% power), 273 (90% power)
 50% of observed effect: n per arm = 459 (80% power), 614 (90% power)
power.t.test: 115.6 
Simulated power at n = 115 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 reports ANCOVA for the primary placebo comparisons, with p-values of 0.088 for the 5 mg group and < 0.001 for the 10 mg group. It does not provide the adjusted effect estimates, confidence intervals, or model covariates. The sponsor analysis is therefore reported as recorded, but its full numerical interpretation cannot be reconstructed from the registry entries available.

The recalculated comparisons are unadjusted calculations from the posted group means, standard deviations, and group sizes. For 5 mg, the recalculated mean difference is -1.73 and the unadjusted p-value is 0.081; for 10 mg, the mean difference is -3.58 and the unadjusted p-value is <0.001. These are not replacements for the sponsor’s ANCOVA results and should not be presented as the trial’s adjusted analysis.

The record flags a discrepancy between the posted summary-statistic calculations and the sponsor analysis. Under that circumstance, the sponsor’s own analysis takes precedence. The recalculations can still help readers understand how the posted means relate to an unadjusted group contrast, but they cannot explain why the analyses differ. The registry record does not report enough model detail to identify the cause, so no explanation should be assumed.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and multiple secondary outcomes, with more than one comparison against placebo. It does not report whether the sponsor adjusted for multiplicity, used a hierarchical testing strategy, or otherwise controlled the interpretation of repeated tests. The analysis plan and prespecified order of testing are not reported in the registry record.

Multiplicity matters because repeated statistical tests create more opportunities for an apparently unusual result even when no underlying differences exist. A p-value from a single test does not account for the number of outcomes or comparisons unless the analysis explicitly incorporates that structure. The reported secondary p-values are therefore unadjusted in the information available and should not be treated as if they were multiplicity-controlled.

The registry describes an ITT population as participants who took any study medication and had baseline data and at least one post-treatment efficacy assessment. It also states that observations after rescue medication were censored and handled by last observation carried forward. This description does not establish how every randomized participant was handled, and the record does not report the number of missing observations, the reasons for missingness, or the extent of censoring after rescue medication.

Last observation carried forward uses a participant’s last recorded value in place of a later unavailable value under the stated rule. It can produce biased estimates when the unobserved outcome would have changed, and its effect depends on why data are missing and how outcomes evolve. The registry does not report sensitivity analyses or alternative missing-data methods. Details of the randomization sequence, allocation concealment, model covariates, and the full original statistical analysis plan are also not reported in the registry record. These gaps limit independent assessment of the reported analyses.

16. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: exacerbation of schizophrenia (2.4%), overdose (2.3%), worsening of hypertension (2.2%).

The registry reports one serious adverse event in each group: placebo, 1 among 43 at risk; loxapine 5 mg, 1 among 45 at risk; and loxapine 10 mg, 1 among 41 at risk. It reports zero deaths in each group. These figures describe events recorded during the study; they do not by themselves establish whether an event was caused by the study treatment.

The registry separately records other adverse events affecting 14 participants in the placebo group, 14 in the 5 mg group, and 16 in the 10 mg group. The record also lists serious adverse-event terms with percentages, including exacerbation of schizophrenia, overdose, and worsening of hypertension. It does not provide group attribution for those listed terms in the information, so they cannot be assigned to a treatment group from these figures.

Safety comparisons should be read with caution. The record provides counts and at-risk group sizes but does not give detailed event descriptions, timing, exposure-adjusted rates, or a basis for attributing causality. A small number of reported events cannot establish that the groups have equivalent safety profiles.

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

The posted data show that mean PANSS-EC scores fell from baseline in all three groups by the 2-hour assessment. The loxapine group means were more negative than the placebo mean, with a larger difference for the 10 mg comparison than for the 5 mg comparison. The registry’s own ANCOVA analyses report different p-values for the two comparisons. The record does not provide adjusted effect estimates or confidence intervals for those analyses.

These findings describe measurements in this randomized, single-dose study of patients with schizophrenia and acute agitation. They do not establish how scores changed after the stated assessment time, whether a difference would persist, or how results would apply to other populations or settings. They also do not establish a clinically important threshold, because the registry record does not report one.

The results should not be reduced to whether a p-value is above or below a threshold. The size of the estimated difference, its uncertainty, the analysis method, missing-data handling, and the study’s scope all matter. The unadjusted recalculations are not the same as the sponsor’s ANCOVA analysis, and the discrepancy means the sponsor’s analysis takes precedence for the trial’s formal result.

Secondary outcomes provide context, but the registry does not report multiplicity control or a complete statistical analysis plan. The safety counts likewise do not establish causation or equivalence. Overall, the record describes short-term score changes and reported group comparisons; it does not support broader claims beyond those measured outcomes and analyses.

18. Why This Trial Matters Statistically

This trial illustrates why a group’s change from baseline should not be interpreted in isolation. The placebo group also had a negative mean PANSS-EC change, so the relevant comparison is the difference between groups rather than the change within a loxapine group alone.

It also demonstrates that an unadjusted recalculation and a sponsor-reported adjusted analysis can yield different statistical results. The posted group means and standard deviations permit a simple comparison, while ANCOVA may account for specified covariates. Since the registry does not report the model details and flags a discrepancy, the sponsor’s analysis takes precedence and the recalculation should remain clearly labeled as unadjusted.

The study further highlights the distinction between estimated effect and precision. A mean difference gives an estimated contrast in PANSS units; its interval shows uncertainty under the calculation used. A p-value answers a different, narrower question and does not describe clinical importance.

Finally, the record makes visible issues that are often essential to interpretation but not fully documented: how multiple outcomes were handled, the extent of missing data, and the consequences of censoring observations after rescue medication with last observation carried forward. These are not peripheral technicalities. They shape how confidently readers can interpret the reported comparisons and how far the results can be generalized.

19. Limitations

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

The registry record does not provide the complete statistical analysis plan, ANCOVA covariates, adjusted effect estimates, or adjusted confidence intervals. It flags a discrepancy between the sponsor analysis and simple unadjusted recalculations, without explaining the cause. This prevents independent reconstruction of the formal primary comparisons.

The record does not report multiplicity control despite listing multiple outcomes and comparisons. It also does not report the extent of missing data, the number of observations censored after rescue medication, or sensitivity analyses for the last-observation-carried-forward approach. These omissions leave uncertainty about how missingness and repeated testing affected the reported results.

The primary endpoint is reported at 2 hours after a single dose, so the record does not describe persistence of score changes beyond that assessment. The registry does not report a threshold for clinically important change. The posted group summaries do not show the distribution of individual outcomes, and the available safety counts do not establish causality or provide detailed event context.

The record describes a plan to study 120 patients and records enrollment of 129, but does not explain the difference. The original sample-size justification is not reported. The record also does not provide detailed information on the randomization procedure or the roles covered by quadruple masking. These limitations concern the available registry information and should not be filled in by assumption.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Staccato Loxapine in Agitation (Proof of Concept) — NCT00369577.
• Registry record last updated 2018-01-02; status last verified 2013-06.
• Results first posted 2018-01-02.
• 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

NCT00369577 is a completed Phase 2 randomized, quadruple-masked, parallel-group study of single-dose inhaled loxapine for acute agitation in patients with schizophrenia. It compared inhaled loxapine 5 mg and 10 mg with inhaled placebo. The registry records enrollment of 129 and a primary completion date of 2007-01.

The primary endpoint was mean PANSS-EC change from baseline at 2 hours. Mean changes were -4.98 with placebo, -6.71 with 5 mg, and -8.56 with 10 mg. The registry reports ANCOVA p-values for the placebo comparisons but not adjusted estimates or confidence intervals. Because the record flags a discrepancy, the sponsor’s own analysis takes precedence over unadjusted recalculations.

Secondary outcomes include BARS change, CGI-I, and CGI-I responder status. The registry does not report multiplicity control, the original sample-size assumptions, or full details of missing-data analyses. Safety reporting includes one serious adverse event in each group and zero deaths in each group.