Safety and Efficacy of Technosphere® Insulin Inhalation Powder and Lantus® Compared to Humalog® and Lantus® Over 16-Weeks: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Safety and Efficacy of Technosphere® Insulin Inhalation Powder and Lantus® Compared to Humalog® and Lantus® Over 16-Weeks (NCT00700622), a randomized phase 3 study. The primary outcome is Change From Baseline in HbA1c to Week 16.

2. Clinical Question

The trial was randomized, parallel-group, and open-label. It enrolled 130 participants, while the primary outcome lists available Week 16 data for 52 in the Technosphere insulin plus insulin glargine group and 58 in the insulin lispro plus insulin glargine group. Those outcome counts describe the posted endpoint analysis, not necessarily every participant’s trial experience. The registry record does not explain the difference between enrollment and the endpoint analysis counts.

3. Trial Design

The registry record describes a phase 3, randomized, multicenter, open-label, parallel-group clinical trial in people with type 1 diabetes. The comparison was Technosphere insulin inhalation powder combined with Lantus, or insulin glargine, versus Humalog, or insulin lispro, combined with Lantus. Treatment continued over a 16-week period. The record lists the trial status as terminated, with a start date of 2008-05 and a primary completion date of 2009-11. The reason for termination is not reported in the registry record.

The stated objective was to compare the combinations on HbA1c. The primary endpoint was change from baseline in HbA1c to Week 16, measured as a percentage of total hemoglobin. Results are presented as least-squares mean changes with standard errors. The primary outcome population is described as intent to treat with available data at Week 16. The endpoint lists 52 participants in the Technosphere insulin plus insulin glargine arm and 58 in the insulin lispro plus insulin glargine arm.

Randomization supports a planned comparison between groups, while the open-label design means masking was absent. The registry record does not report the randomization procedure, allocation concealment, or any steps taken to limit the influence of unmasked treatment assignment on outcome assessment or participant behavior. It also does not report the protocol’s planned sample-size justification or the reason the trial was terminated. Those details cannot be inferred from the posted outcome counts or the endpoint estimates.

The record lists one outcome measure and one analysis, and no posted secondary outcomes. The sponsor’s reported analysis is an ANCOVA estimate classified as non-inferiority or equivalence. A separate unadjusted recalculation uses the posted arm summaries. These approaches answer related but not identical statistical questions: the registry’s ANCOVA estimate is the sponsor’s analysis, while the recalculation is a descriptive comparison based on summary statistics rather than an adjusted model.

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: 2 week screening period prior to randomization - 276 Screened / 138 Eligible. 130 subjects were randomized; 138 screen failures and 8 discontinued prior to randomization.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to Treat with Available Data at Week 16

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 From Baseline in HbA1c to Week 16, measured over Baseline to Week 16.

The primary endpoint is change in HbA1c from baseline to Week 16, expressed as a percentage of total hemoglobin. HbA1c is a continuous measure, and the registry record reports the arm results as least-squares mean changes. The listed values are -0.1 for Technosphere insulin plus insulin glargine and -0.03 for insulin lispro plus insulin glargine. The accompanying spread values are standard errors, not standard deviations.

Change from baseline uses each participant’s baseline measurement as a reference and evaluates the value at the specified follow-up. In general, an endpoint based on change can describe movement over the treatment period while accounting for the starting measurement. Least-squares means are model-based adjusted means; the registry record does not identify the adjustment variables used to produce these means. It gives the outcome population as intent to treat with available data at Week 16, so the endpoint label does not mean that every randomized participant necessarily contributed an observed Week 16 value.

HbA1c was selected as the trial’s primary measure because the stated study objective concerned comparison on HbA1c. The record describes the endpoint and timeframe but does not report the clinical threshold used to interpret a change, a non-inferiority margin, or additional details about how the endpoint was collected. The numerical estimates therefore describe the posted Week 16 comparison; their clinical interpretation depends on criteria not reported in the registry record.

7. Results Posted in the Registry

• Insulin Lispro + Insulin Glargine: mean -0.03 versus -0.10 on control; difference 0.07 (95% CI -0.16 to 0.30), standardized effect d = 0.11; z-test p = 0.558.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a continuous endpoint and gives each treatment arm’s least-squares mean change with a standard error. It also reports an ANCOVA mean difference and a 95% confidence interval. A separate unadjusted calculation estimates the difference from the posted arm summaries. These methods have related aims—describing the comparison between groups—but their estimates may differ because they use different information and assumptions.

What does the ANCOVA result represent?

The sponsor’s reported method is analysis of covariance, or ANCOVA. In general, ANCOVA models an outcome while accounting for one or more other variables. In clinical trials, it is often used to compare groups while adjusting for baseline measurements or other prespecified covariates. The registry record does not say which covariates were included, how the model was specified, or how any missing observations were handled. The reported result is a net mean difference of -0.07, with a 95% confidence interval from -0.31 to 0.17. The result is classified as a non-inferiority or equivalence analysis, but no margin is reported.

The arm values are described as least-squares means. In general, a least-squares mean is a model-based mean adjusted for variables in the model, rather than necessarily the raw arithmetic average of the observed values. The model details needed to reproduce those adjusted means are not reported in the registry record. The associated standard errors, 0.087 and 0.082, describe uncertainty in the estimated arm means; they are not measures of participant-to-participant variability such as standard deviations.

What does the unadjusted recalculation do?

The recalculated contrast uses the posted arm-level estimates and standard errors to form an unadjusted comparison. It reports a mean difference of 0.07 when the contrast is expressed as the insulin lispro plus insulin glargine mean minus the Technosphere insulin plus insulin glargine mean. Its 95% confidence interval is -0.164 to 0.304, and its unadjusted p-value is 0.558. The calculation uses an approximate pooled standard deviation of 0.626, as indicated by the recalculation details. Because it is reconstructed from summary statistics rather than the original participant-level model, it cannot reproduce any unreported ANCOVA adjustments.

In general, an unadjusted comparison treats the posted group summaries as the basis for the estimate and does not control for covariates. A covariate-adjusted estimate can differ from an unadjusted one, even when both address the same treatment contrast. The exact contrast direction also matters: reversing the order of the groups reverses the sign of a difference. The sponsor’s net estimate has the opposite sign from the recalculated contrast, and the record does not explicitly state the sponsor’s contrast orientation. The estimates should therefore be compared with attention to direction, model, and interval rather than by sign alone.

What do the confidence interval and p-value say?

In general, a confidence interval is a range of values compatible with the estimate under the method’s assumptions. Its width conveys precision: a narrower interval represents a more concentrated range, while a wider interval leaves more uncertainty about the size and direction of the effect. The sponsor’s 95% interval runs from -0.31 to 0.17, while the recalculated interval runs from -0.164 to 0.304. Both include zero. The intervals do not establish that the true difference is exactly zero; they indicate that the data and method do not rule out zero at the stated confidence level.

The recalculated p-value of 0.558 is an unadjusted test result. In general, a p-value describes how unusual data at least as incompatible with a null hypothesis would be if that null hypothesis and the statistical model were correct. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. The sponsor analysis has no p-value posted in this record, so its reported estimate should be interpreted through its confidence interval and stated analysis type rather than by attributing the recalculated p-value to it.

Why does over-dispersion matter in general?

For count outcomes, over-dispersion means that the observed variability exceeds what a simple count model expects. A model that ignores it can understate uncertainty and produce confidence intervals that are too narrow. This endpoint is continuous rather than a count, so over-dispersion is not an issue identified for the posted primary analysis. More generally, statistical methods should match the scale and distribution of the outcome; the registry record does not give enough model detail to assess assumptions beyond the methods it names.

What remains unknown about the analysis?

The registry record does not report the covariates in the ANCOVA, the non-inferiority or equivalence margin, the planned sample-size justification, or the details of the statistical analysis plan. It also does not report whether interim analyses were planned or conducted. Those details are important for judging whether the analysis followed a prespecified plan and how a non-inferiority or equivalence conclusion was intended to be made. The posted statistics permit a description of the point estimates and their uncertainty, but not a reconstruction of those unreported design and analysis decisions.

10. Confidence Intervals and Precision

Precision concerns how narrowly an estimate is located by its confidence interval. For the sponsor-reported ANCOVA, the net mean difference is -0.07 and its 95% confidence interval extends from -0.31 to 0.17. The unadjusted recalculation reports a mean difference of 0.07 with a 95% confidence interval from -0.164 to 0.304. Both intervals cross zero, so each is compatible with no difference as well as with nonzero differences in either direction, on its respective contrast scale and under its method.

The interval is not a range containing every plausible value with certainty. In general, a 95% confidence interval is produced by a procedure that would cover the target parameter in 95% of repeated applications under its assumptions. It does not mean there is a 95% probability that the fixed treatment effect lies inside this particular interval. Interpretation also depends on the contrast direction, outcome definition, model, and any relevant clinical or non-inferiority margin.

The posted interval widths leave uncertainty about the size of the group difference. Neither interval alone demonstrates equivalence or non-inferiority, because the registry record does not report the margin against which the interval would need to be assessed. The sponsor’s interval comes from ANCOVA; the recalculated interval is unadjusted and based on summary statistics. They are not interchangeable measures of precision, and the registry record does not provide enough model detail to explain their differences fully. The confidence intervals describe statistical uncertainty, not the probability that one treatment combination is preferable or an assessment of clinical importance.

11. 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 participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. The calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning scenario targets a stated fraction of the observed effect: 100%, 75%, or 50%. These are planning illustrations based on the observed effect, not the trial’s original sample-size assumptions; those original assumptions are not reported in the registry record.

For planning at 100% of the observed effect, the estimate is 1,255 participants per arm for 80% power and 1,680 participants per arm for 90% power. At 75% of the observed effect, the corresponding figures are 2,231 and 2,987 participants per arm. At 50% of the observed effect, they are 5,020 and 6,720 participants per arm. In plain language, the 80% power column gives the planning count associated with an 80% chance of detecting the specified effect under the calculation’s assumptions, if that effect is present. The 90% column uses a higher target chance of detection and therefore requires a larger planned group size.

Smaller target effects require more participants because a smaller signal is harder to distinguish from random variation. The planning calculation is not observed power. It is a forward-looking illustration using the observed estimate as its starting point, not a measure of the probability that this completed trial found an effect.

12. 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 = 0.070; pooled standard deviation σ = 0.626.
2. n = 2 × (1.96 + 0.842)² × 0.626² / 0.070² = 1,254.8, rounded up to 1,255 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 <- 0.07                      # observed difference in means
sigma <- 0.625851508835052                     # 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(700622)
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 = 1,255 (80% power), 1,680 (90% power)
 75% of observed effect: n per arm = 2,231 (80% power), 2,987 (90% power)
 50% of observed effect: n per arm = 5,020 (80% power), 6,720 (90% power)
power.t.test: 1255.8 
Simulated power at n = 1,255 per arm: 0.798 (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.

13. Analyses Reported by the Sponsor

The registry’s own reported analysis is an ANCOVA mean difference of -0.07, with a 95% confidence interval from -0.31 to 0.17. Its p-value is not reported in the registry record. The unadjusted recalculation gives a mean difference of 0.07, a 95% confidence interval from -0.164 to 0.304, and an unadjusted p-value of 0.558. The recalculated contrast is oriented as insulin lispro plus insulin glargine minus Technosphere insulin plus insulin glargine; the sponsor’s contrast orientation is not stated explicitly. Reversing the order of a contrast reverses its sign, so the signs should not be compared without considering orientation.

The recalculation uses posted summary statistics and an approximate pooled standard deviation; it does not reproduce the sponsor’s ANCOVA. The registry record does not give the sponsor model’s covariates or other details needed to recreate that analysis. The record marks no discrepancy, but the recalculation remains an unadjusted check rather than a substitute for the sponsor’s model-based result. The sponsor analysis should be read as the registry-reported primary analysis, and the recalculation as a separate estimate with its own assumptions.

The sponsor result is labeled non-inferiority or equivalence, but its margin and decision criterion are not reported in the registry record. The estimate and interval can be described, but the intended non-inferiority or equivalence conclusion cannot be independently evaluated without that information.

14. Multiplicity, Missing Data and Analysis Populations

The record lists one outcome measure and one analysis, and no secondary outcomes are posted. This limits what can be assessed about multiplicity. In general, multiplicity arises when several outcomes, time points, treatment contrasts, or statistical tests are examined; repeated testing can increase the chance of at least one result appearing unusual by chance. The registry record does not report a multiplicity-adjustment procedure, or whether other analyses were planned but not posted. The absence of posted secondary results is not evidence that no additional analyses were planned or conducted.

The primary outcome population is described as intent to treat with available data at Week 16. The registry record lists 52 participants for the Technosphere insulin plus insulin glargine endpoint and 58 for the insulin lispro plus insulin glargine endpoint. The enrolled count is 130. The record does not explain how participants without available Week 16 data were handled, why the endpoint counts differ from enrollment, or whether any imputation or sensitivity analyses were performed. It also does not report the amount or pattern of missing outcome data.

In general, an intention-to-treat approach aims to preserve the randomized group comparison by analyzing participants according to assigned group. The phrase “with available data at Week 16” indicates that the posted endpoint is based on available observations, but does not itself explain the missing-data strategy or establish how every randomized participant was represented in the analysis. The record does not provide enough detail to assess whether missingness could have affected the estimate.

The analysis population and missing-data procedures matter because a complete-case estimate may be affected if availability of outcome measurements is related to treatment assignment or the outcome. This is a general statistical concern, not a claim that such a pattern occurred in this trial. The registry record does not report the protocol’s planned analysis population in greater detail, the handling of withdrawals, or sensitivity analyses addressing missing observations. Consequently, the posted estimate can be interpreted for the stated available-data population, while the robustness of that estimate to missing data remains uncertain.

15. 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: hypoglycemia (3.1%), delayed recovery from anesthesia (1.5%), intentional overdose (1.5%), hypoglycemic seizure (1.5%), suicide attempt (1.5%).

The adverse-event listing reports serious adverse events affecting 1 of 65 participants at risk in the Technosphere insulin plus insulin glargine group and 3 of 65 at risk in the insulin lispro plus insulin glargine group. Other adverse events are reported for 42 of 65 and 34 of 65, respectively. These are separate safety summaries from the primary HbA1c endpoint counts; the record does not explain the relationship between the safety populations and the endpoint populations.

The most frequently listed serious adverse event terms are Hypoglycemia at 3.1%, Delayed recovery from anesthesia at 1.5%, Intentional overdose at 1.5%, Hypoglycemic seizure at 1.5%, and Suicide attempt at 1.5%. The registry record does not specify in this listing whether the terms overlap or how each percentage was calculated. Deaths are not reported in the registry record.

These figures are descriptive counts and percentages. They do not, by themselves, establish that one regimen caused more or fewer events: the record does not provide a comparative safety analysis, event timing, severity details beyond the serious-event classification, or a statistical uncertainty estimate for these comparisons. The event terms and participant counts should be read as the posted safety information, not as a complete assessment of safety or a basis for treatment decisions.

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

The posted primary outcome describes mean change in HbA1c from baseline to Week 16 for the two treatment combinations. The arm estimates are -0.1 for Technosphere insulin plus insulin glargine and -0.03 for insulin lispro plus insulin glargine. The sponsor’s ANCOVA estimate is -0.07, with a 95% confidence interval from -0.31 to 0.17. The unadjusted recalculation gives 0.07 with a 95% confidence interval from -0.164 to 0.304. Both intervals include zero, and both leave uncertainty about the size and direction of the difference.

These findings do not establish that the combinations are identical. Nor does the recalculation’s p-value of 0.558 establish that there is no difference. A failure to distinguish an effect from zero is not proof of equivalence. Equivalence and non-inferiority require a prespecified margin and a suitable analysis; although the sponsor result is categorized as non-inferiority or equivalence, the margin is not reported in the registry record. The posted statistics therefore do not permit an independent judgment about whether that criterion was met.

The estimates also do not show that either treatment combination is clinically preferable. The registry record does not provide a threshold for a clinically important HbA1c change, and statistical uncertainty is not the same as clinical importance. The endpoint concerns change over the stated treatment period and in the available-data Week 16 population; it does not, on its own, describe outcomes beyond that period or outcomes not posted as secondary measures.

The trial was open-label and is listed as terminated. The reason for termination, original sample-size justification, missing-data procedures, and details of the ANCOVA model are not reported in the registry record. Those limitations constrain interpretation. The most defensible reading is a description of the reported estimates and their uncertainty, without extending them into claims about treatment benefit, equivalence, or longer-term effects.

17. Why This Trial Matters Statistically

This trial illustrates why a point estimate should not be read in isolation. The arm means and the reported between-group estimates are accompanied by confidence intervals, which make uncertainty visible. Both posted intervals include zero and extend across values on either side of it. That pattern does not prove no difference; it indicates that the estimates do not precisely distinguish a direction on their respective contrast scales.

It also shows why analysis type matters. The registry reports an ANCOVA result, while the recalculation uses arm-level summaries and an approximate pooled standard deviation. The two calculations are not identical methods, and their contrast signs depend on group order. A reader who sees only one estimate or one p-value could miss these distinctions. The record’s classification of the sponsor result as non-inferiority or equivalence adds another important point: those claims require a margin, which is not reported in the registry record.

The endpoint population also matters. The trial enrollment is listed as 130, while the posted Week 16 primary outcome uses available-data group counts of 52 and 58. The record does not explain the difference or the missing-data approach. This makes transparency about populations and outcome availability central to evaluating how well the posted estimate represents the planned comparison.

Finally, the sample-size illustrations show the relationship between target effect size and planning requirements under stated assumptions. Smaller effects require more participants for the same power target. Those figures are planning scenarios based on the observed effect, not the original design rationale or a measure of observed power. Together, these features make the trial useful for understanding the interpretation of adjusted and unadjusted estimates, confidence intervals, non-inferiority claims, and incomplete reporting.

18. 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 lists the trial as terminated but does not report the reason. It gives a start date of 2008-05 and a primary completion date of 2009-11, but does not describe whether termination affected recruitment, follow-up, or analysis. The registry enrollment is 130, while the primary endpoint lists available-data counts of 52 and 58; the record does not explain the difference or report the missing-data pattern.

The trial was open-label, with no masking. The registry record does not report whether endpoint assessment or data analysis included safeguards against possible influence from awareness of assigned treatment. It also does not provide details of the randomization procedure or allocation concealment. These omissions prevent a fuller appraisal of design implementation from the registry information alone.

The sponsor’s ANCOVA is not fully described: covariates, model specification, and missing-data handling are not reported. The sponsor analysis is labeled non-inferiority or equivalence, but the margin and decision rule are absent. As a result, the posted interval cannot be used to independently verify that type of conclusion. The unadjusted recalculation is based on summary statistics and an approximate pooled standard deviation, so it does not resolve the missing model details.

The record lists one outcome measure and no secondary outcomes, but does not state whether additional outcomes or analyses were planned. Multiplicity procedures and interim analyses are not reported in the registry record. The planning sample-size assumptions shown for illustration are based on the observed effect; the original design justification is not reported. Finally, the primary endpoint covers change to Week 16, and the registry record does not post longer-term results. These limitations define what can be concluded from the posted results without implying that unreported procedures were or were not undertaken.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Safety and Efficacy of Technosphere® Insulin Inhalation Powder and Lantus® Compared to Humalog® and Lantus® Over 16-Weeks — NCT00700622.
• Registry record last updated 2014-10-16; status last verified 2014-10.
• Results first posted 2014-10-16.
• 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.

22. Record Summary

This phase 3 randomized, open-label, parallel-group trial compared Technosphere insulin inhalation powder plus insulin glargine with insulin lispro plus insulin glargine in people with type 1 diabetes. The trial is listed as terminated, with a 16-week treatment period. Its primary endpoint was change in HbA1c from baseline to Week 16.

The posted least-squares mean changes were -0.1 and -0.03. The sponsor-reported ANCOVA mean difference was -0.07, with a 95% confidence interval from -0.31 to 0.17. An unadjusted recalculation reported a mean difference of 0.07, with a 95% confidence interval from -0.164 to 0.304 and p = 0.558. The recalculation and sponsor analysis differ in method and reported contrast direction.

The sponsor result is classified as non-inferiority or equivalence, but the margin is not reported in the registry record. The reason for termination, missing-data procedures, and original sample-size justification are also not reported. The posted estimates describe the Week 16 results but do not independently establish equivalence, non-inferiority, or clinical importance.