Fluid Balance in the ICU - Interventions to Minimize Fluids in Patients With Septic Shock: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Fluid Balance in the ICU - Interventions to Minimize Fluids in Patients With Septic Shock (NCT02473718), a randomized na study. The primary outcome is Cumulative Fluid Administered.

2. Clinical Question

The question is about a measurable aspect of treatment: cumulative administered study fluid by a specified point in the ICU stay. It is not, by itself, a question about whether the protocol changes survival, ventilation complications, or another patient outcome. Those outcomes require their own results and analysis.

The registry describes daily fluid intake and output assessment as part of the intervention. It also says ultrasound would measure changes in the diameter of the inferior vena cava with respiration. Details beyond those descriptions, including precisely how clinicians used those measurements to make treatment decisions, are not reported in the registry record.

3. Trial Design

The study is listed as randomized, with a parallel-group model and no masking. It compares the fluid minimization group with the usual care group. The registry record lists Washington University School of Medicine as sponsor, reports enrollment of 82, and marks the study as completed. The listed phase is NA. The start is recorded as 2014-01 and primary completion as 2014-12.

The intervention is named the fluid minimization protocol, with ultrasound also listed as an intervention. The brief description says the protocol was intended to assess daily intake and output and determine whether it could decrease the overall amount of fluid received during the first five days in the ICU. It further describes daily ultrasounds and blood draws for fluid-balance evaluation. The ultrasound measure is the change in inferior vena cava diameter with respiration.

The primary endpoint reported is cumulative fluid administered by day 3. The endpoint includes crystalloid boluses, continuous infusions, and colloid fluids. The registry gives results as medians with inter-quartile ranges. Its population note states that five patients in the fluid minimization group and two in the usual care group died before the first time-point analysis at day three; one patient in the usual care group was transferred out of the ICU before that analysis.

The registry also lists ventilator days, rate of renal replacement therapy, and mortality among secondary outcomes. Results are posted for renal replacement therapy and mortality. Results for ventilator days and other listed outcomes are not reported in the registry record. The record does not report the detailed randomization procedure, allocation concealment, the protocol's full decision rules, or a statistical analysis plan.

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

Analysis population for the primary outcome, as defined in the registry: Five patients in the fluid minimization group and two patients in the usual care group died prior to the first time point analysis at day three. One patient in the usual care group was transferred out of the ICU prior to the first time point analysis at day 3.

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 Cumulative Fluid Administered, measured over Day 3.

The primary endpoint is cumulative fluid administered by day 3, expressed in mL of study fluid. It comprises the cumulative volume of crystalloid boluses, continuous infusions, and colloid fluids administered by that time. It therefore measures an amount of fluid delivered over the period up to the endpoint, rather than a single fluid balance measurement or a patient-centered outcome.

The registry reports the endpoint using the median and inter-quartile range. In the fluid minimization group, the median was 4417 mL and the inter-quartile range was 3139 to 6528. In the usual care group, the median was 4110 mL and the inter-quartile range was 2702 to 10000. The median describes the midpoint of the observed values, while the inter-quartile range describes the middle half of their distribution. These summaries are useful when a distribution may be skewed or include unusually large values, since the median is less influenced by extreme observations than the mean.

The endpoint is relevant to the stated study aim of reducing fluid administration. However, a group-level difference in administered volume does not by itself establish a difference in complications, survival, or other outcomes. The registry's note about deaths and transfer before the day-three analysis is important context: the primary summaries do not represent all randomized participants at that time point. The exact method for handling those missing endpoint measurements is not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a median with an inter-quartile range convey?

The median is the central value after observations are ordered. The inter-quartile range marks the interval from the lower quartile to the upper quartile and contains the middle half of the observed values. In general, medians and inter-quartile ranges describe the location and spread of a distribution without relying on a normal, symmetric shape. They can be informative when values are skewed, as cumulative fluid volumes may be. These summaries do not, on their own, quantify uncertainty about a difference between groups.

For the day-three primary endpoint, the registry reports a median and an inter-quartile range for each group. It does not report a confidence interval or test for a comparison of those distributions. A confidence interval for a treatment contrast would describe the range of effect estimates compatible with the data under the method used. Without that interval or a reported test, the observed group summaries should be read descriptively rather than as a statistically established treatment effect.

How should the binary secondary outcomes be read?

Renal replacement therapy and mortality are posted as counts of participants. The registry reports 17 of 41 participants in the fluid minimization group and 16 of 41 in usual care for renal replacement therapy. For mortality, it reports 23 of 41 and 20 of 41, respectively. The associated recalculated comparisons express the difference as a risk difference: the event risk in the fluid minimization group minus the event risk in usual care. A positive estimate means the observed event risk is higher in the fluid minimization group in that comparison; it does not establish that the underlying population risk is higher.

The renal replacement therapy risk difference is 0.024, with a confidence interval from -0.188 to 0.237; the unadjusted p-value is 0.822. The mortality risk difference is 0.073, with a confidence interval from -0.142 to 0.289; the unadjusted p-value is 0.507. These are recalculated, unadjusted results based on the posted summary counts. They are not reported sponsor analyses. The intervals include zero, so the estimates are compatible with no risk difference as well as with differences in either direction. A p-value is not the probability that a treatment has no effect, and a non-small p-value does not demonstrate equivalence.

In general, a confidence interval communicates uncertainty more fully than a point estimate alone. A broad interval indicates that the estimate is not precise enough to rule out a range of potentially meaningful effects. The secondary intervals are broad and cross zero. Their interpretation should therefore emphasize uncertainty, not simply whether a p-value crosses a conventional threshold.

What does “unadjusted” mean in this context?

General explanation: an unadjusted comparison is calculated without accounting for covariates or other factors through a model-based adjustment. The posted secondary comparisons use the outcome counts and group sizes directly. The registry record does not report an adjusted analysis for these endpoints, nor does it report the sponsor's own analyses. The recalculated estimates can help describe the contrast in the posted data, but they do not substitute for a prespecified analysis if one existed.

Random allocation is intended to make treatment groups comparable on average, but it does not guarantee balance in every characteristic in a particular trial. An adjusted analysis could address prespecified prognostic factors, but whether one was planned or appropriate cannot be determined from this record. The detailed statistical plan and covariate strategy are not reported in the registry record.

Why is a planning sample size not observed power?

The registry includes planning illustrations based on fractions of the observed effects. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between the arms, and a normal approximation. These are planning scenarios for future study design, not evidence about the power of the completed trial after its results are observed. They should not be interpreted as an estimate of the probability that the study detected a true effect.

General explanation: a sample-size calculation asks how much information a hypothetical study would need to detect a specified effect under stated assumptions. The answer depends on the effect chosen, the significance level, allocation, and the statistical method. If the effect used for planning is based on an observed estimate, its uncertainty also matters: an unstable observed effect can yield a misleading planning target. The illustrations are therefore best treated as scenarios, not as a validation of the original design. The trial's own sample-size justification is not reported in the registry record.

What statistical information is absent?

The registry record reports zero analyses and does not provide a primary comparison test, primary confidence interval, or secondary sponsor analyses. It does not report the method for the primary endpoint, missing-data procedures, an analysis population definition, or whether any interim analyses were conducted. These details affect how an estimate should be interpreted and how well it answers the randomized study question. Their absence is a limitation of the record, not evidence that a particular method was or was not used.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists eight outcome measures and identifies several secondary outcomes, including ventilator days, renal replacement therapy, and mortality. It posts results for renal replacement therapy and mortality, while results for ventilator days and other listed outcomes are not reported in the registry record. The record lists zero analyses and no sponsor-reported secondary analyses. It does not report whether a multiplicity strategy was planned or used.

General explanation: when investigators assess multiple outcomes or make multiple comparisons, the chance of at least one apparently unusual result can increase. A prespecified approach might limit that risk by identifying a primary outcome, setting a testing hierarchy, or adjusting the interpretation of multiple tests. Without a reported plan, it is not possible to determine whether the secondary p-values were intended as confirmatory, how many comparisons were considered, or whether any adjustment was applied. The unadjusted recalculations should consequently be interpreted cautiously.

Missingness is relevant to the primary result. The record says five patients in the fluid minimization group and two in usual care died before the first time-point analysis at day three, and one in usual care was transferred out of the ICU before that analysis. It does not say how these events were handled in the endpoint analysis, whether any other day-three measurements were unavailable, or whether imputation or another missing-data method was used. The endpoint summaries therefore cannot be assumed to represent a complete intention-to-treat analysis.

The analysis populations are not explicitly defined in the registry record. In particular, it does not explain whether the primary endpoint population was based on randomized assignment, endpoint availability, or another criterion. The record also does not report interim-analysis procedures, stopping rules, or a detailed statistical analysis plan. These are unreported details; their absence from the registry record does not establish whether they were absent from the study conduct.

11. 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 0% reporting threshold.

The serious adverse event figures report zero serious events in each group, with 41 at risk in each group. The other adverse event figures also report zero events in each group, again with 41 at risk in each group. These are the registry's posted adverse-event counts, not a complete account of all possible safety outcomes.

Deaths are listed separately: 23 in the fluid minimization group and 20 in usual care. The death counts should not be read as contradicting the zero adverse-event entries. Adverse-event reporting categories and mortality are distinct entries in the registry record, and a zero count in an adverse-event category does not mean that no deaths occurred. The record does not provide a detailed explanation of how events were classified or the follow-up interval for these figures.

Zero reported events in an outcome category does not prove that the risk is absent. Safety interpretation depends on the definitions used, the time period covered, the completeness of collection, and the number of participants observed. Further details about adverse-event ascertainment and the relationship between event reporting and deaths are not reported in the registry record. These counts should be read as the registry's recorded figures, not as a definitive comparison of safety.

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

The primary result describes cumulative fluid administered by day three in the two randomized groups. The fluid minimization group had a median of 4417 mL, and the usual care group had a median of 4110 mL. The reported inter-quartile ranges indicate substantial spread in both groups. These descriptive values do not establish that the protocol changed fluid administration: the registry record does not report an inferential comparison or uncertainty interval for the primary endpoint.

The secondary results also do not support a firm conclusion about renal replacement therapy or mortality. Their recalculated unadjusted confidence intervals include zero and permit effects in either direction. The findings are therefore imprecise rather than evidence of equivalence or proof of no effect. No sponsor-reported analyses are listed for those outcomes.

The result does not show whether fluid minimization reduced mechanical-ventilation complications, because the relevant posted results are not reported in the registry record. It also does not establish an effect on survival or overall safety. The adverse-event counts are limited to the registry categories and do not replace a full clinical assessment of benefits and harms.

General explanation: a randomized design can support a causal comparison when the study is conducted and analyzed according to appropriate methods, but randomization alone does not resolve missing outcomes, imprecise estimates, or incomplete reporting. Here, the day-three endpoint note identifies deaths and a transfer before assessment, while methods for handling those observations are not reported. The interpretation must therefore remain tied to the posted summaries and their documented limitations.

In practical statistical terms, the registry record offers a descriptive snapshot of fluid administration and selected secondary outcomes. It does not provide enough detail to confirm the primary contrast statistically, to assess the planned analysis strategy, or to draw strong conclusions about downstream clinical outcomes.

13. Why This Trial Matters Statistically

This trial is statistically informative as an example of the difference between reporting an endpoint and reporting a comparative analysis. The primary measure is expressed with medians and inter-quartile ranges, which describe the observed distributions. Without a confidence interval, test, or stated method for comparing those distributions, the reader cannot judge the precision of a primary treatment contrast from the registry record alone.

The primary population note also illustrates why an endpoint's analysis set matters. Deaths and an ICU transfer occurred before the day-three assessment. When an endpoint is unavailable after such events, the way those observations are handled can influence both the estimand—the precise treatment effect being targeted—and the analysis. The registry does not report those methods, so the reader cannot determine how the reported summaries relate to the randomized comparison.

The secondary results illustrate another general issue: counts can be converted into interpretable effect estimates, but recalculation does not supply the missing prespecified analysis or account for multiple outcomes. The confidence intervals convey considerable uncertainty, while unadjusted p-values alone would give an incomplete view. In this case, the registry lists multiple outcomes but does not describe a multiplicity plan.

Finally, planning illustrations based on observed effects are not observed power. They can be useful for considering hypothetical future designs under stated assumptions, but they do not show that the completed study was adequately powered or that its results are reliable. The trial is a useful reminder that careful interpretation depends on the endpoint, analysis population, missing-data handling, precision, and reporting—not just the randomized label or a point estimate.

14. Limitations

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

Several limitations specific to this trial are apparent from the registry record. The primary endpoint is reported descriptively, with no reported confidence interval, hypothesis test, or comparative analysis method. This prevents assessment of the precision of the primary contrast. The day-three population note identifies deaths before assessment and an ICU transfer before assessment, but the record does not explain how these cases were handled or define the analysis population.

The registry reports a median and inter-quartile range for cumulative fluid volume, but does not explain the exact statistical method used to compare groups or whether the endpoint distribution was assessed before choosing that summary. Additional details about the measurement process, including how fluid records were verified and how missing observations were managed, are not reported in the registry record.

For secondary outcomes, recalculated unadjusted comparisons are available for renal replacement therapy and mortality, but no sponsor-reported analyses are listed. The record does not report a strategy for multiplicity, an analysis hierarchy, or a detailed statistical analysis plan. Results for ventilator days and other listed outcome measures are not reported in the registry record, limiting assessment of the study's stated interest in ventilation-related outcomes.

The record also does not provide details about allocation concealment, the randomization procedure, the intervention's full implementation rules, or the duration and completeness of adverse-event collection. The adverse-event figures list zero events in the reported categories while mortality is separately listed; event definitions and classification details are not reported. These gaps constrain interpretation, but they do not establish that the unreported procedures were not conducted.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Fluid Balance in the ICU - Interventions to Minimize Fluids in Patients With Septic Shock — NCT02473718.
• Registry record last updated 2024-07-24; status last verified 2024-07.
• Results first posted 2018-03-29.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

NCT02473718 is a completed randomized, unmasked, parallel-group study sponsored by Washington University School of Medicine. It enrolled 82 and compares a fluid minimization protocol with usual care in patients with septic shock. The registry lists the start as 2014-01 and primary completion as 2014-12.

The primary endpoint is cumulative crystalloid boluses, continuous infusions, and colloid fluids administered by day 3, reported as medians and inter-quartile ranges. The record notes deaths and an ICU transfer before the first day-three analysis. It also lists ventilator days, renal replacement therapy, and mortality as secondary outcomes, with results posted for renal replacement therapy and mortality.

The registry does not report a primary inferential comparison, detailed analysis methods, missing-data procedures, or a multiplicity plan. The posted secondary comparisons are unadjusted recalculations from the reported counts, and no sponsor-reported analyses are listed. Interpretation should account for those reporting gaps and the uncertainty in the secondary estimates.