Effects of Dapagliflozin on Kidney Function (Glomerular Filtration Rate) in Subjects With Type 2 Diabetes: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Effects of Dapagliflozin on Kidney Function (Glomerular Filtration Rate) in Subjects With Type 2 Diabetes (NCT00976495), a randomized phase 2 study. The primary outcome is Adjusted Percent Change From Baseline in Glomerular Filtration Rate (GFR) at Week 12 (Modified Last Observation Carried Forward [MLOCF]).

2. Clinical Question

The trial’s stated purpose was to evaluate dapagliflozin’s effects on kidney function as assessed by glomerular filtration rate. Its primary endpoint captures change in this measure over the double-blind period. The posted results report group means and a recalculated unadjusted comparison of dapagliflozin with placebo. A separate recalculation compares hydrochlorothiazide with placebo. These estimates describe the observed data and their statistical precision; they do not by themselves answer questions beyond the endpoint, analysis population, and follow-up specified in the registry record.

3. Trial Design

This was a randomized, double-masked, parallel-group Phase 2 study. The registry records an enrollment of 154 and a completed status. The study began in 2009-10, and primary completion was recorded as 2010-11. The listed treatments were dapagliflozin, placebo, and hydrochlorothiazide. The posted arm results identify dapagliflozin 10 mg and hydrochlorothiazide 25 mg.

A parallel-group design assigns participants to a treatment group and evaluates the groups during the same study period. Randomization is intended, in general, to make treatment groups comparable on average at the time of allocation. Masking is intended, in general, to reduce the influence of knowledge about assigned treatment on participant experience and study assessment. The registry does not report details about who was masked or how randomization was carried out.

The primary endpoint was assessed from baseline to Week 12. The registry defines baseline as the last assessment before the first dose of double-blind medication, with a date-based alternative when timing details were unavailable. Measurements were obtained at the randomization visit and Week 12 in the double-blind period by a central laboratory. The primary population was restricted to randomized participants who received study medication and had nonmissing HbA1c values at baseline and Week 12, using the MLOCF approach described in the record.

The original allocation ratio, eligibility criteria beyond the condition and study-title description, duration of treatment beyond the stated endpoint timing, and protocol-specific sample-size rationale are not reported in the registry record. The overall enrollment should not be read as the number contributing to each primary comparison; the posted primary arm counts are smaller and differ between groups.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: All randomized participants who received study medication and had nonmissing HbA1c values at baseline and Week 12 (MLOCF)

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: All randomized participants who received at least 1 dose of study medication

6. Primary Endpoint

The registry lists the primary outcome as Adjusted Percent Change From Baseline in Glomerular Filtration Rate (GFR) at Week 12 (Modified Last Observation Carried Forward [MLOCF]), measured over From Baseline to Week 12.

The primary endpoint was adjusted percent change from baseline in glomerular filtration rate at Week 12, with the Week 12 value handled using modified last observation carried forward (MLOCF). Glomerular filtration rate is the kidney-function measure named in the study’s purpose. Expressing change as a percentage of baseline describes how the measure changed relative to each participant’s starting value, rather than only reporting a difference in the original measurement scale.

The endpoint is suited to the study’s stated exploratory question because it measures change in the same kidney-function quantity the study set out to assess. The registry specifies that baseline was generally the last assessment before the first dose of double-blind medication. If a Week 12 measurement was unavailable, the last available post-baseline measurement obtained on or after Day 23 was used, regardless of rescue medication. Those rules matter to interpretation: the reported Week 12 outcome may include carried-forward measurements rather than only measurements collected precisely at Week 12.

The registry identifies the primary measure as a continuous outcome and reports adjusted means with standard errors. The analysis population is described as randomized participants who received medication and had nonmissing baseline and Week 12 HbA1c values, with MLOCF. The reason for incorporating HbA1c availability into this endpoint’s analysis population is not reported in the registry record. Nor is a detailed description of the adjustment model reported there. The endpoint therefore provides a defined summary of the posted results, but the registry record does not give enough methodological detail to reconstruct every aspect of the sponsor’s primary analysis.

7. Results Posted in the Registry

• Dapagliflozin 10 mg: mean -10.76 versus -2.92 on control; difference -7.84 (95% CI -13.36 to -2.32), standardized effect d = -0.81; z-test p = 0.005.
• Hydrochlorothiazide 25 mg: mean -3.40 versus -2.92 on control; difference -0.48 (95% CI -6.01 to 5.05), standardized effect d = -0.05; z-test p = 0.865.

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.

Adjusted Mean Change From Baseline in 24-Hour Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])

• 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 errors.
• Result:
  • Dapagliflozin 10 mg: mean -3.28 versus -0.88 on control; difference -2.40 (95% CI -7.12 to 2.32), standardized effect d = -0.29; z-test p = 0.319.
  • Hydrochlorothiazide 25 mg: mean -6.55 versus -0.88 on control; difference -5.67 (95% CI -10.27 to -1.07), standardized effect d = -0.68; z-test p = 0.016.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dapagliflozin 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Hydrochlorothiazide 25 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Adjusted Mean Change From Baseline in Daytime (0900 to 2100 Hours) Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])

• 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 errors.
• Result:
  • Dapagliflozin 10 mg: mean -5.93 versus -1.55 on control; difference -4.38 (95% CI -9.36 to 0.60), standardized effect d = -0.50; z-test p = 0.085.
  • Hydrochlorothiazide 25 mg: mean -6.83 versus -1.55 on control; difference -5.28 (95% CI -10.12 to -0.44), standardized effect d = -0.60; z-test p = 0.032.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dapagliflozin 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Hydrochlorothiazide 25 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Adjusted Mean Change From Baseline in Nighttime (0100 to 0600 Hours) Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])

• 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 errors.
• Result:
  • Dapagliflozin 10 mg: mean -0.45 versus -0.61 on control; difference 0.16 (95% CI -5.43 to 5.75), standardized effect d = 0.02; z-test p = 0.955.
  • Hydrochlorothiazide 25 mg: mean -7.76 versus -0.61 on control; difference -7.15 (95% CI -12.60 to -1.70), standardized effect d = -0.73; z-test p = 0.010.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Dapagliflozin 10 mg and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Hydrochlorothiazide 25 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The secondary outcomes concern change in 24-hour, daytime, and nighttime ambulatory systolic blood pressure at Week 12. Across these measures, the recalculated dapagliflozin comparisons with placebo do not show the same degree of separation across all time periods: estimates are more negative for daytime and 24-hour measurement, while the nighttime estimate is close to zero. The hydrochlorothiazide comparisons are more consistently negative relative to placebo across the three periods. These broad patterns describe the direction of the posted summaries, not a single combined blood-pressure effect.

The secondary outcomes provide context alongside the primary kidney-function endpoint, but they measure a different clinical quantity. They do not establish the mechanism or explain the primary result. The registry also reports sponsor comparisons for some secondary contrasts, and those values are based on final-value mean differences. The recalculated contrasts concern adjusted mean changes from baseline. The measures and contrasts should therefore be interpreted according to their own definitions rather than merged into one conclusion.

All recalculated secondary p-values are unadjusted. When several outcomes or comparisons are examined, the chance of observing at least one apparently unusual result can increase, even if no corresponding effect exists. This is a general multiplicity concern. The registry record does not report a correction for multiple testing or a confirmatory hierarchy for the secondary outcomes. Accordingly, individual secondary findings carry less weight than a result supported by a prespecified testing strategy, and they are best treated as exploratory descriptions of these data rather than independent confirmation of the primary result.

The registry lists three secondary outcomes, each measured over the baseline-to-Week 12 period and using last observation carried forward. The posted secondary summaries do not establish whether patterns persist beyond that assessment or whether they apply to other participant groups. The sample-size illustrations for secondary comparisons are also conditional on their observed effects and assumptions, not evidence that the trial was designed to detect those outcomes.

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

• Adjusted Mean Change From Baseline in 24-Hour Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])
• Adjusted Mean Change From Baseline in Daytime (0900 to 2100 Hours) Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])
• Adjusted Mean Change From Baseline in Nighttime (0100 to 0600 Hours) Ambulatory Systolic Blood Pressure (ASBP) at Week 12 (Last Observation Carried Forward [LOCF])

9. Statistical Interpretation

10. Statistical Methods Explained

What does an adjusted mean summarize?

The registry labels the primary outcome as an adjusted mean change, expressed as percent change from baseline. An adjusted mean is a model-based group summary that accounts for variables or structure specified in the analysis model. The registry does not identify the covariates or give the model specification, so the exact adjustment cannot be reconstructed from the record. The recalculated comparison instead uses the posted group summaries and is described as an unadjusted re-calculation. It should therefore be distinguished from the sponsor’s adjusted analysis.

How should the mean difference be read?

A mean difference is the contrast between two group means on the outcome scale. Here, a negative difference for dapagliflozin versus placebo means the posted dapagliflozin mean is more negative. It does not mean that every participant had a change in that direction, because a group mean does not describe every individual result. The recalculation is based on the posted summary statistics; it is not a refit of participant-level data or an adjusted treatment model.

What does the standard error describe?

The registry reports standard errors as the dispersion measure for the group means. In general, a standard error describes the estimated uncertainty in a sample estimate, rather than the spread of individual participant measurements. It helps convey how precisely a group mean has been estimated. The posted standard errors are used in the simple unadjusted recalculations. The registry does not provide the individual observations or enough information to check assumptions about their distribution.

What are the interval and p-value for?

A confidence interval gives a range of values compatible with the data and the method used, under the assumptions of that method. A narrower interval indicates greater precision than a wider interval on the same scale. Whether an interval includes zero is one way to see whether the data are compatible with no difference under that interval procedure. A p-value summarizes how unusual the data, or more extreme data, would be under a specified null hypothesis and statistical model. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect.

The recalculated comparisons give unadjusted p-values and interval bounds. The registry’s sponsor analyses give 95% confidence intervals but no p-values for the listed primary comparisons. The analysis method for those sponsor comparisons is not reported in the registry record. Thus the reported and recalculated statistics should not be treated as interchangeable just because they concern related groups and outcomes.

How does the missing-measurement rule work?

MLOCF means modified last observation carried forward. In general, a carried-forward approach uses an earlier available measurement in place of a missing later measurement, according to specified rules. The registry says that if Week 12 data were unavailable, the last available post-baseline measurement on or after Day 23 was used, regardless of rescue medication. This defines how some missing Week 12 values were represented in the analysis. It does not demonstrate that the unobserved Week 12 value would have been the same as the carried-forward value. The reason for choosing the rule, sensitivity analyses, and assumptions about missingness are not reported in the registry record.

Why does over-dispersion matter in general?

For a continuous outcome, variability among participants affects the uncertainty of the estimated mean difference. In general, greater variability relative to the effect being estimated makes an interval wider and can require more participants for a planned level of power. The posted group standard errors and recalculated pooled standard deviations describe variability for these illustrations, but the registry does not report individual-level distributions or diagnostic checks. Over-dispersion is especially a concern when a model assumes less variability than is present in the data; it can make standard errors too small and inference too confident. The registry does not report whether over-dispersion was assessed or relevant to its sponsor analyses.

What distinguishes the trial analysis from the recalculation?

The registry reports adjusted outcomes and sponsor comparisons, while the recalculated contrasts use posted summary means and standard errors to form simple, unadjusted comparisons. The registry does not provide a full statistical analysis plan, model details, covariate set, or all information needed to reproduce its adjusted estimates. The recalculation is useful for describing the direction and approximate precision of a contrast from the posted summaries, but it should not be used to override a sponsor result that relies on a different parameter or analysis method. No discrepancy flag is recorded for this study, yet the different endpoint summaries and contrast definitions still require careful distinction.

11. Confidence Intervals and Precision

Precision is reflected by the width of an interval around an estimate: a relatively narrow interval restricts the range of values compatible with the data more than a wide interval does. For the recalculated dapagliflozin-placebo primary contrast, the interval extends from -13.36 to -2.321. Its bounds lie on the negative side of zero, and the associated unadjusted p-value is 0.005. The estimate is therefore more clearly separated from zero than the hydrochlorothiazide-placebo recalculation.

For hydrochlorothiazide versus placebo, the recalculated interval runs from -6.008 to 5.048, crossing zero and spanning negative and positive values. The associated unadjusted p-value is 0.865. This indicates limited precision for the direction and size of that contrast in these posted data; it does not establish equivalence. A conclusion of equivalence would require an appropriate design and prespecified margin, neither of which is reported in the registry record.

The primary arm counts used in the recalculated contrasts were 25.0 for placebo, 22.0 for dapagliflozin 10 mg, and 26.0 for hydrochlorothiazide 25 mg. These are not the full enrollment count of 154. Smaller analysis groups generally provide less information than larger groups, all else equal, but precision also depends on variability and the analysis method. The reported standard errors give some information about uncertainty in each group mean; they do not reveal the full participant-level distributions.

The sponsor’s primary analyses report 95% confidence intervals for final-value mean differences. They are not the same contrasts as the recalculated percent-change comparisons, so their bounds should be interpreted in their own stated context. The registry does not report all details needed to assess model-based precision, missing-data sensitivity, or the effect of covariate adjustment. These limitations are important when comparing interval widths or using them to judge how stable an estimate may be in other settings.

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 participants per arm, not total events. They are illustrations based on the observed effect, not the trial’s original design calculation. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; each scenario asks what would be needed to detect the stated fraction of the observed effect. The original trial’s design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the estimated participants per arm under those assumptions for an 80% chance of detecting the specified effect if that effect were the true one. The 90% power column uses a higher target chance of detection and therefore generally calls for more participants per arm. These are planning quantities, not guarantees for a future study. In particular, the figures are not observed power and do not describe the probability that the completed trial’s finding is true.

Rows for 100%, 75%, and 50% of the observed effect ask about progressively smaller effects. Smaller effects are harder to distinguish from random variation, so the illustrations require more participants per arm. This relationship is visible in the listed planning figures for both primary comparisons. When the observed effect is close to zero, planning directly from it can produce extremely large sample-size illustrations; such values should not be mistaken for a recommended or feasible design. They are conditional calculations based on the stated assumptions, not a new estimate of the study’s effect or a replacement for a protocol-based sample-size justification.

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 = -7.840; pooled standard deviation σ = 9.686.
2. n = 2 × (1.96 + 0.842)² × 9.686² / -7.840² = 24.0, rounded up to 24 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 <- -7.84                      # observed difference in means
sigma <- 9.68629119336533                     # 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(976495)
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 = 24 (80% power), 33 (90% power)
 75% of observed effect: n per arm = 43 (80% power), 58 (90% power)
 50% of observed effect: n per arm = 96 (80% power), 129 (90% power)
power.t.test: 25 
Simulated power at n = 24 per arm: 0.782 (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 sponsor analyses report mean differences in final values. For placebo versus dapagliflozin, the reported estimate is -8.07, with a 95% confidence interval from -13.34 to -2.49. For dapagliflozin versus hydrochlorothiazide, the reported estimate is -7.62, with a 95% confidence interval from -12.87 to -2.06. The registry does not report p-values or identify the analysis method for these comparisons.

The recalculated primary comparisons instead use posted adjusted percent changes from baseline and compare dapagliflozin or hydrochlorothiazide with placebo. The difference in stated parameter and comparison means the estimates are not direct duplicates. The sponsor’s second reported contrast also compares dapagliflozin with hydrochlorothiazide, rather than comparing hydrochlorothiazide with placebo. Differences in estimates should not be explained by assuming an undocumented model or adjustment.

The registry record does not mark a discrepancy between the sponsor analysis and recalculation. The sponsor’s reported results should nevertheless be read as the registry’s own analysis, while the recalculations are simple, unadjusted summaries of the posted arm results. The statistical method and further details of the sponsor analysis are not reported in the registry record, so a full reconciliation is not possible from the posted information.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary endpoint and three secondary outcomes, and the posted analyses include multiple treatment contrasts. Testing several outcomes and comparisons creates a multiplicity issue: if every test is judged separately without adjustment, the chance of at least one nominally unusual result can be greater than for a single prespecified test. This is a general statistical concern, not evidence that any particular posted finding is false. A prespecified hierarchy or multiplicity adjustment can help define how results should be interpreted, but neither is reported in the registry record.

The registry identifies the primary population as randomized participants who received study medication and had nonmissing baseline and Week 12 HbA1c values, with MLOCF. This is narrower than simply referring to all enrolled participants. The arm-specific primary analysis counts are 25.0 for placebo, 22.0 for dapagliflozin 10 mg, and 26.0 for hydrochlorothiazide 25 mg. The record does not explain why HbA1c availability is part of the primary population definition for a GFR endpoint, nor does it provide a complete participant flow accounting for the difference between enrollment and the endpoint counts.

The Week 12 missing-data rule uses the last available post-baseline measurement obtained on or after Day 23 when a Week 12 value was unavailable, regardless of rescue medication. In general, carrying an observation forward is an assumption about how an unobserved value relates to an earlier one; it does not recover the missing measurement. The registry does not report the extent or pattern of missingness, reasons for missing measurements, sensitivity analyses using alternative assumptions, or how rescue medication was handled beyond the stated rule.

The statistical analysis plan, including model details, covariate adjustment, any interim analyses, and the original sample-size rationale, is not reported in the registry record. These omissions limit independent assessment of whether the posted analyses followed a prespecified approach and how robust the results are to analytic choices. They also limit interpretation of the unadjusted secondary p-values as a family of results.

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: localised infection (4.2%), intervertebral disc injury (4.2%), pleuritic pain (4.2%).

The registry reports serious adverse events for 25 participants at risk in the placebo group, 24 in the dapagliflozin 10 mg group, and 26 in the hydrochlorothiazide 25 mg group. The number with a serious adverse event was 0, 3, and 0, respectively. These are counts for the listed groups; they do not, on their own, establish whether a treatment caused an event or allow a broad safety comparison.

The record lists localised infection, intervertebral disc injury, and pleuritic pain among the top serious-event terms, each reported as 4.2%. The registry record does not provide the number of participants for each term in this summary, so the percentages should be read as reported and not converted into inferred counts. Death counts are not reported in the registry record.

Other adverse events were reported for 2 participants in the placebo group, 6 in the dapagliflozin group, and 9 in the hydrochlorothiazide group. The registry excerpt does not provide event-level detail or a fuller account of severity, timing, relatedness, or exposure duration. The group counts should therefore be viewed as descriptive safety information from this study rather than as a complete comparison of risks.

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

The posted primary summaries show different mean percent changes in GFR across the three groups. The unadjusted recalculation for dapagliflozin versus placebo is negative, and its interval excludes zero; the recalculated hydrochlorothiazide-placebo interval includes zero. These observations describe the endpoint as analyzed in the registry record, in the specified primary population and by Week 12. They do not establish what happened at other times or in populations not represented by that analysis.

A difference in group means is not a description of every participant’s experience. It also does not establish why the groups differed, whether the difference is important in clinical practice, or whether it persists. The registry record does not provide a threshold for a clinically meaningful change in GFR, nor does it report longer-term kidney outcomes. The study is identified as exploratory Phase 2, so the results should be read in that setting rather than as a complete assessment of treatment effects.

The recalculated intervals and p-values address statistical uncertainty under their respective methods. They do not measure the probability that a treatment is beneficial or harmful, and they do not resolve limitations due to missing data, the selected analysis population, multiplicity, or unspecified model details. The sponsor analyses use final-value mean differences and should be interpreted separately from the recalculated percent-change contrasts.

The safety counts are also limited to the events and groups summarized in the registry. They do not provide a full account of adverse-event type, timing, attribution, or all possible risks. Taken together, the posted results can inform a careful description of measured group outcomes in this study; they do not justify medical advice or a general claim that a treatment works or does not work.

18. Why This Trial Matters Statistically

This study illustrates why an endpoint definition matters. Percent change from baseline, adjusted mean change, and mean difference in final values are related summaries, but they are not interchangeable. The registry’s sponsor analyses report final-value mean differences, while the simple recalculations use adjusted percent changes from baseline. Careful statistical reading begins by identifying the scale, time point, population, and comparison for each estimate.

It also shows how uncertainty can differ across contrasts. The dapagliflozin-placebo recalculation has an interval on one side of zero, while the hydrochlorothiazide-placebo interval spans zero. The contrast estimates are not equally precise, and the latter interval does not prove the groups are the same. A p-value alone would not communicate the full range of plausible values or the uncertainty about direction and size.

The primary analysis uses a defined subset of participants and a carried-forward rule for missing Week 12 measurements. Such choices affect what the result summarizes and should be considered alongside the numerical estimate. The registry’s incomplete reporting of analysis-model details, multiplicity handling, and original sample-size justification further emphasizes the need to separate what is explicitly documented from what cannot be reconstructed.

Finally, the planning illustrations show how sample-size needs depend on the effect assumed and the variability reflected in the observed data. They are not a retrospective verdict on study quality or observed power. Together, these features make the trial useful for statistical learning about endpoint scales, group contrasts, precision, missing data, and the limits of interpreting exploratory results.

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 does not report the randomization method, allocation ratio, masking details, or protocol-specific sample-size justification. It also does not provide the statistical analysis plan, covariates in the adjusted model, or a detailed account of sponsor-analysis methods. These gaps prevent complete reproduction of the sponsor’s results and limit assessment of how analytic decisions may affect the estimates.

The primary analysis population is defined by receipt of study medication and nonmissing HbA1c values at baseline and Week 12, in addition to randomization. The reason for this HbA1c requirement is not reported in the registry record. The primary arm counts are smaller than the enrollment total, and a full account of participant disposition for the endpoint is not reported. This makes it difficult to assess how the analyzed groups relate to all randomized participants.

The MLOCF rule is described, but the amount and pattern of missing GFR data, reasons for missingness, and sensitivity analyses are not reported. The registry also does not report whether rescue medication use was examined in alternative analyses; the stated rule says carried-forward measurements were used regardless of rescue medication. Missing-data assumptions may matter when interpreting a Week 12 estimate.

The study is exploratory Phase 2 and includes a primary endpoint, three secondary outcomes, and multiple comparisons. Multiplicity handling is not reported in the registry record. The study has no longer-term outcome results in the registry record for this analysis, and the posted endpoint is limited to Week 12. These design and reporting limits mean the estimates should be understood as study-specific summaries rather than broad or definitive evidence.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Effects of Dapagliflozin on Kidney Function (Glomerular Filtration Rate) in Subjects With Type 2 Diabetes — NCT00976495.
• Registry record last updated 2017-03-08; status last verified 2017-01.
• Results first posted 2017-03-08.
• 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

NCT00976495 is a completed randomized, double-masked, parallel-group Phase 2 study of dapagliflozin, placebo, and hydrochlorothiazide in type 2 diabetes. Enrollment was 154. The primary endpoint was adjusted percent change in GFR from baseline to Week 12, using MLOCF and a specified analysis population.

The recalculated unadjusted dapagliflozin-placebo contrast was -7.84, with an interval from -13.36 to -2.321 and p=0.005. The hydrochlorothiazide-placebo recalculation was -0.48, with an interval from -6.008 to 5.048 and p=0.865. Sponsor-reported final-value comparisons are distinct analyses. Interpretation is limited by unreported model and design details, missing-data information, and multiplicity handling.