Utility of Presepsin in Distinguishing Between Sepsis and SIRS: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Utility of Presepsin in Distinguishing Between Sepsis and SIRS (NCT02052895), a registered phase not stated study. The primary outcome is Area Under the Receiver Operating Characteristic Curve (ROC-AUC) of the Plasma Presepsin Concentration for Discriminating Between SIRS and Sepsis.

2. Clinical Question

3. Trial Design

The registry describes NCT02052895, “Utility of Presepsin in Distinguishing Between Sepsis and SIRS,” as an exploratory, prospective observational study in critically ill patients. Its stated purpose was to evaluate diagnostic accuracy: whether presepsin levels could discriminate between sepsis and SIRS when patients presented with critical illness compatible with either condition. The record lists enrollment as 226 and status as completed. It lists Mochida Pharmaceutical Company, Ltd. as sponsor, a start of 2014-01, and primary completion of 2015-07.

The record distinguishes three groups. The Sepsis/SIRS group comprises patients with sepsis or SIRS. The Control group comprises patients without SIRS, sepsis, or end stage renal disease. The End Stage Renal Disease group comprises patients with end stage renal disease, without SIRS or sepsis. These descriptions identify clinically different groups; they should not be treated as randomized arms or as interchangeable comparison groups. Allocation is not stated, and this is not a randomized comparison.

The primary diagnostic analysis used plasma presepsin levels on Day 0. Sepsis/SIRS adjudication was made on Day 7 or the day of discharge, with the timeframe described as up to 7 days. The primary analysis population description says that out of 196 Sepsis/SIRS patients, 6 were excluded due to inclusion or exclusion criteria violation and 4 were excluded due to insufficient data for Sepsis/SIRS adjudication. The record does not report further details about how adjudication was conducted.

The listed primary outcome is an ROC-AUC for presepsin to distinguish SIRS from sepsis, compared with the ROC-AUC for procalcitonin. Although the record describes one outcome measure and one analysis, the numerical diagnostic results are not reported in the registry record. No additional details about allocation procedures, blinding, or the conduct of clinical assessments are reported in the registry record.

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: Out of 196 Sepsis/SIRS patients, 6 were excluded due to inclusion/exclusion criteria violation and 4 were excluded due to insufficient data for Sepsis/SIRS adjudication

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: Subjects who didn't meet inclusion/exclusion criteria were excluded.

6. Primary Endpoint

The registry lists the primary outcome as Area Under the Receiver Operating Characteristic Curve (ROC-AUC) of the Plasma Presepsin Concentration for Discriminating Between SIRS and Sepsis, measured over Up to 7 days.

The primary endpoint is the area under the receiver operating characteristic curve (ROC-AUC) for plasma presepsin concentration in distinguishing sepsis from SIRS. The measurement used was plasma presepsin on Day 0. The reference classification was Sepsis/SIRS adjudication made on Day 7 or the day of discharge. The registry describes the outcome as a continuous measure in probability units, with a 95% confidence interval, and identifies the parameter as a least-squares mean.

In general, an ROC curve describes how a diagnostic measure separates two classifications across possible decision thresholds. Its area summarizes discrimination across those thresholds rather than reporting performance at one selected threshold. An AUC nearer the midpoint of the probability scale reflects less separation than an AUC nearer the upper end, but its practical meaning depends on the intended clinical context and the quality of the reference classification. The AUC does not, by itself, specify a threshold, sensitivity, specificity, or the consequences of classification errors.

The registry says the presepsin ROC-AUC was compared with that of procalcitonin. It does not report the numerical AUCs, their confidence intervals, or a numerical comparison between the markers. Thus, the endpoint definition is available, but the precision and magnitude of the observed diagnostic performance are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is an area under a receiver operating characteristic curve. In general, an ROC curve examines the trade-off between correctly identifying members of one classification and incorrectly assigning members of the other as a decision threshold changes. The area condenses that threshold-dependent information into a single discrimination summary. This can be useful when the purpose is to compare how well a continuous marker ranks or separates two groups without committing to one threshold. It does not identify a clinically preferred threshold or describe the consequences of using one.

For this study, the record names plasma presepsin concentration on Day 0 as the marker and Sepsis/SIRS adjudication on Day 7 or the day of discharge as the classification. It describes the primary outcome as ROC-AUC for distinguishing SIRS from sepsis and says that this was compared with procalcitonin. The numerical values for either marker and the result of their comparison are not reported in the registry record. A statistical method can be explained from the endpoint definition, but its actual output cannot be inferred from the definition alone.

What does a confidence interval communicate?

In general, a confidence interval expresses uncertainty associated with an estimate under the statistical procedure used. It is not a description of the spread of individual measurements, and it does not guarantee that the true value lies within the particular interval. Its width can help readers judge precision: more uncertainty is generally reflected by a wider interval. The registry identifies a 95% confidence interval for the primary outcome, but its numerical limits are not reported in the registry record. Therefore, readers cannot judge the reported precision from the registry entry.

The registry identifies the primary parameter as a least-squares mean. This label indicates an estimate expressed as a mean under an analysis framework, but the record does not describe the model or adjustment variables used to produce it. It would be inappropriate to infer a specific regression model, covariate adjustment, or transformation. The endpoint is described as continuous and in probability units, but the analysis details that would clarify how the least-squares mean relates to the ROC-AUC are not reported in the registry record.

Why does the reference classification matter?

Diagnostic accuracy is assessed against a classification treated as the reference for the analysis. Here, the record says Sepsis/SIRS adjudication was made on Day 7 or the day of discharge. In general, the accuracy estimate depends not only on the marker but also on how the reference classification is defined and applied. If classification is uncertain, inconsistent, or made with knowledge of the marker, that can affect interpretation. The registry record does not report the adjudication criteria, who made the adjudications, whether assessors were blinded to the marker, or how disagreement was handled. No assumptions about those procedures should be made.

The timing also matters in principle. The marker was measured on Day 0, while adjudication was made later or at discharge. This design asks whether an initial measurement discriminates between later assigned classifications. The record does not provide information about changes in clinical status between measurement and adjudication, and it does not describe repeated marker measurements for the primary endpoint. These details cannot be inferred from the stated timeframe.

How should group comparisons be understood?

The study is observational, and allocation is not stated. In general, observational groups are formed by clinical characteristics rather than random assignment. Differences between groups may therefore reflect differences in their underlying characteristics as well as the marker under study. The record describes a Sepsis/SIRS group, a control group without SIRS, sepsis, or end stage renal disease, and an end stage renal disease group without SIRS or sepsis. Those groups serve different descriptive purposes and are not a randomized, like-for-like control comparison.

The primary diagnostic question is specifically about distinguishing sepsis from SIRS, not about distinguishing either condition from the separate control or end stage renal disease groups. General diagnostic performance can change with the mix of patients and the difficulty of separating the target classifications. The record does not provide subgroup-specific AUC estimates or further details of the analyzed case mix, so no subgroup conclusion can be drawn.

What can planning calculations tell readers?

In general, sample-size planning calculations explore how many participants per arm might be needed under stated assumptions to detect an effect of a chosen size with a chosen testing approach. The registry’s planning description says the illustration uses the observed effect, a stated fraction of that effect, two-sided significance level 0.05, equal allocation between two arms, and a normal approximation. Such calculations are planning illustrations, not evidence that the study was designed around those assumptions, and they are not observed power. The registry does not report that these were the trial’s own design assumptions.

For a continuous endpoint, planning results expressed as participants per arm are different from event totals used for time-to-event outcomes. The record states that its time-to-event planning tables use total events, while other planning outputs use participants per arm. That distinction is general: event-driven designs depend on events accrued, whereas other calculations may be expressed in participant counts. No numerical planning outputs or effect fractions are reported in the registry record, so no specific planning value can be interpreted.

Over-dispersion is a general issue for data whose variability exceeds what a simple model assumes. When relevant, ignoring it can make uncertainty appear smaller than it is. The registry record does not identify an over-dispersion model for this continuous ROC-AUC endpoint, and no such method should be attributed to the analysis. More broadly, methods should be interpreted only to the extent they are described: the record identifies the endpoint, parameter label, and confidence level, but it does not report the detailed computational procedure.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and one analysis. Those counts describe what is listed in the record; by themselves, they do not establish whether the investigators considered additional endpoints, analyses, or comparisons during the study. The primary endpoint description says presepsin ROC-AUC was compared with procalcitonin ROC-AUC, but the record does not report the statistical test used for that comparison or whether any multiplicity adjustment was applied. A correction procedure, testing hierarchy, or prespecified handling of multiple comparisons is not reported in the registry record.

Multiplicity matters in general when several hypotheses or outcomes are tested, because the chance of at least one chance finding can increase as the number of tests grows. Adjustments or a prespecified hierarchy can address that issue, but the record does not provide enough information to say whether either approach was used. The presence of one listed outcome and one listed analysis should not be taken as evidence that no other analyses occurred.

The record gives a specific account of exclusions from the Sepsis/SIRS population: out of 196 patients, 6 were excluded due to inclusion or exclusion criteria violation and 4 were excluded due to insufficient data for adjudication. It does not report an imputation procedure, a missing-data model, or a sensitivity analysis for missing classifications or marker values. The analysis population beyond the stated exclusions is not further detailed in the record. The reasons for the data insufficiency are not reported in the registry record.

In general, missing data can affect diagnostic estimates if availability is related to the marker, the reference classification, or clinical status. Excluding participants with insufficient information is one possible way an analysis population may be defined, but its effect cannot be judged without further detail. The record does not report whether excluded participants differed from analyzed participants or how the exclusions affected uncertainty. No assumptions about those matters are warranted.

11. 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 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: death (13.7%), gastrointestinal hemmorage (0.5%), respiratory failure (0.5%), hemorrhage (0.5%).

The registry lists serious adverse events for the three groups. It reports 26 serious adverse events among 190 at risk in the Sepsis/SIRS group, and 0 among 15 at risk in the Control group and 0 among 15 at risk in the End Stage Renal Disease group. For other adverse events, the record lists 0 in each group, with the same group-specific at-risk counts. These are descriptive registry figures; the study is observational and the groups are not a randomized, like-for-like comparison.

The record also lists serious adverse event terms including death at 13.7%, Gastrointestinal hemmorage at 0.5%, Respiratory failure at 0.5%, and Hemorrhage at 0.5%. The denominator and precise context for these percentages are not reported in the registry record. The group-level death fields are null, so a group-specific death count is not reported there. These entries should not be treated as a complete account of cause, timing, or attribution.

In general, adverse-event counts describe events recorded under the study’s reporting definitions. They do not establish that the study marker caused an event, and differences between clinically distinct observational groups cannot be interpreted as comparative treatment risks. The registry record does not report event definitions, attribution procedures, or additional context needed to interpret the listed terms.

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

The study addresses a diagnostic question: whether plasma presepsin concentration measured on Day 0 can distinguish between sepsis and SIRS in critically ill patients, using adjudication on Day 7 or the day of discharge as the classification. It also identifies procalcitonin as a comparator. This is a clinically relevant statistical question because diagnostic markers are evaluated not only by their measured values but by how well those values separate the classifications of interest.

However, the registry record does not report the primary numerical result. The ROC-AUC estimate, its 95% confidence interval, and the numerical comparison with procalcitonin are not reported in the registry record. Therefore, the record does not support a conclusion about the magnitude of discrimination, the precision of the estimate, or whether one marker had a higher ROC-AUC than the other.

In general, a discrimination measure is not the same as a complete evaluation of a diagnostic strategy. An AUC does not specify a decision threshold, the number of false positive or false negative classifications at that threshold, or the consequences of those errors. It also does not establish whether using the marker improves decisions or outcomes. Those questions require evidence beyond an AUC summary.

The observational design is another important boundary. The registry describes clinically defined groups and does not state allocation. The endpoint concerns classification, not a treatment effect. The study cannot be read from this record as a randomized comparison, and its group descriptions should not be taken as evidence that the groups were comparable on other characteristics. The control and end stage renal disease groups are described separately from the Sepsis/SIRS diagnostic population.

The stated exclusions and lack of reported details about missing-data handling also constrain interpretation. The registry identifies patients excluded because of eligibility violations or insufficient adjudication data, but does not report how those exclusions affected the final estimate. The evidence in this record is therefore sufficient to understand the intended question and broad design, but insufficient to judge the observed diagnostic performance. No clinical action or treatment recommendation follows from the endpoint definition alone.

13. Why This Trial Matters Statistically

This study matters statistically because it illustrates the distinction between defining a diagnostic endpoint and reporting an interpretable diagnostic result. The registry specifies a marker, a reference classification, a timeframe, a comparison marker, and an ROC-AUC outcome. Those elements make the intended analysis understandable. Yet the numerical AUC and its confidence interval are not reported in the registry record, leaving the central estimate and its precision unavailable for assessment.

It also highlights why diagnostic studies need clear descriptions of both the index measure and the reference classification. Here, presepsin was measured on Day 0, while Sepsis/SIRS adjudication was made on Day 7 or the day of discharge. In general, the validity and reproducibility of a diagnostic estimate depend on how the classifications are made, whether assessors are blinded, and how uncertain cases are handled. These procedures are not described in the registry record, so readers cannot evaluate those features from the entry.

The study’s observational character is equally important. Sepsis/SIRS, control, and end stage renal disease groups have different definitions, and allocation is not stated. That makes it essential to keep the diagnostic objective separate from a randomized treatment comparison. The endpoint concerns discrimination between classifications, not an effect of assigning an intervention.

Finally, the record illustrates the role of analysis-population accounting. It identifies exclusions for eligibility violations and insufficient adjudication data, but does not detail missing-data methods or sensitivity analyses. Such information helps readers understand which participants contribute to a diagnostic estimate and whether the estimate may be sensitive to exclusions. The record defines the question and flags some exclusions, but does not report enough numerical and methodological detail to establish the strength or robustness of the diagnostic evidence.

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.

The primary limitation for interpreting the registry record is that it does not report the numerical ROC-AUC for presepsin, its 95% confidence interval, or the comparison with procalcitonin. Without these results, the magnitude and precision of the primary endpoint cannot be evaluated. Although the outcome definition is explicit, its observed value is not reported in the registry record.

The record gives limited information about the reference classification. It states that Sepsis/SIRS adjudication occurred on Day 7 or the day of discharge, but does not report adjudication criteria, assessor roles, blinding, or procedures for resolving disagreement. These details are important for judging how consistently the reference classification was applied and whether knowledge of the biomarker could have influenced assessment.

The analysis population description reports exclusions for eligibility violations and insufficient data for adjudication. It does not explain the reasons for insufficient data, describe missing marker values, or report imputation or sensitivity analyses. The effects of exclusions on the diagnostic estimate therefore cannot be assessed from the registry record.

The study is described as observational, allocation is not stated, and the groups are defined by different clinical characteristics. They should not be interpreted as randomized, comparable arms. The record also does not report adjustment details for the least-squares mean parameter, the statistical test for the presepsin-versus-procalcitonin comparison, or any multiplicity procedure. These omissions limit assessment of the analysis methods and the certainty of conclusions that might otherwise be drawn from a result.

Finally, the registry does not report a selected diagnostic threshold or threshold-specific performance measures. An AUC alone would not supply those details even if its numerical value were available. The record therefore supports a description of the diagnostic aim, but not a complete evaluation of clinical utility.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Utility of Presepsin in Distinguishing Between Sepsis and SIRS — NCT02052895.
• Registry record last updated 2018-08-21; status last verified 2018-08.
• Results first posted 2016-09-27.
• 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

NCT02052895 is listed as a completed, exploratory, prospective observational study sponsored by Mochida Pharmaceutical Company, Ltd. The registry reports enrollment of 226, a start of 2014-01, and primary completion of 2015-07. The study evaluated plasma presepsin measured on Day 0 for distinguishing sepsis from SIRS, with Sepsis/SIRS adjudication on Day 7 or the day of discharge. Procalcitonin was the stated comparator. The registry describes separate Sepsis/SIRS, Control, and End Stage Renal Disease groups and reports exclusions from the Sepsis/SIRS population for eligibility violations and insufficient adjudication data. The primary endpoint was ROC-AUC with a 95% confidence interval. Its numerical result and the comparison with procalcitonin are not reported in the registry record.