This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Human Mesenchymal Stromal Cells For Acute Respiratory Distress Syndrome (START) (NCT02097641), a randomized phase 2 study. The primary outcome is Numbers of Patients Occurred Pre-specified Infusion Associated Events Occurring Within 6 Hours of Study Infusion.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02097641 |
| Phase | Phase 2 |
| Condition | Respiratory Distress Syndrome, Adult |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Triple |
| Enrollment | 60 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Michael A. Matthay |
| Results posted | Yes |
2. Clinical Question
Population
People with acute respiratory distress syndrome (ARDS) were the target population. The registry record describes the study as a Phase 2a trial in patients with ARDS and lists respiratory distress syndrome, adult, as the condition. It does not report baseline characteristics or further eligibility details in the record summarized for these results.
Intervention
The intervention was a single intravenous dose of allogeneic bone marrow-derived human mesenchymal stromal cells (hMSCs). The arm description states a dose of 10 million cells/kg predicted body weight, infused over approximately 60-80 minutes.
Comparator
The comparator was a single intravenous dose of Plasma-Lyte A, identified as placebo. The infusion was described as taking approximately 60-80 minutes. Both arms were part of the same randomized, parallel-group trial.
Primary question
The primary endpoint counted participants who experienced a pre-specified infusion-associated event within 6 hours of study infusion. The trial therefore asked whether any participants met one or more specified event criteria during that period. The registry reports zero such participants in both arms, with 40 participants in the hMSC arm and 20 in the placebo arm.
This is a short-term, event-count question focused on safety around the infusion, rather than a direct measure of longer-term respiratory recovery. Mortality and several physiological and organ-function outcomes were also included as secondary outcomes. The registry record does not report patient characteristics that would allow readers to assess how the enrolled participants compare with other ARDS populations.
3. Trial Design
START is identified as a randomized, parallel-group Phase 2 trial with planned enrollment of 60 participants. The registry lists the study as completed. It records a start date of 2014-03-15 and a primary completion date of 2017-03-09. The trial summary describes it as Phase 2a, double-blind, placebo-controlled, and multicenter. The structured record lists triple masking; the registry record does not explain the relationship between that label and the double-blind wording in the summary.
The study compared one intravenous infusion of allogeneic bone marrow-derived hMSCs with one intravenous infusion of Plasma-Lyte A placebo. The hMSC arm description specifies a dose of 10 million cells/kg predicted body weight. Both descriptions give an infusion duration of approximately 60-80 minutes. The result counts include 40 participants in the hMSC arm and 20 in the placebo arm. The registry record does not explain the allocation ratio or why the observed arm counts differ.
The primary endpoint was a count of participants with pre-specified infusion-associated events within 6 hours. Listed event criteria included specified increases in vasopressor support, serious new cardiac rhythm events, certain worsening oxygenation requirements, and clinical findings consistent with transfusion incompatibility or infection. The endpoint was therefore designed to capture a defined set of acute events around the infusion, not every possible adverse event or later clinical outcome.
Secondary outcomes included measures of oxygenation, lung injury, organ dysfunction, mortality, ventilator-free days, and non-pulmonary organ-failure-free days. The registry record does not report a sample-size justification, a full statistical analysis plan, or details of how participants were assigned to analysis populations. It also does not report the methods used to address missing outcome data.
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: Three patients who were randomly assigned to receive human mesenchymal stem cells, were excluded: 2 did not meet baseline stability criteria; 1 patient's PaO2:FiO2 \> 200 mmHg.
Overall Study
| Stage | Human Mesenchymal Stem Cells | Plasma-Lyte A |
|---|---|---|
| STARTED | 40 | 20 |
| COMPLETED | 40 | 20 |
| NOT COMPLETED | 0 | 0 |
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.
| Characteristic | Unit | Human Mesenchymal Stem Cells (hMSCs) | Plasma-Lyte A (Placebo) | Total |
|---|---|---|---|---|
| Participants analysed | 40 | 20 | 60 | |
| Age, Categorical: <=18 years | Participants | 0 | 0 | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 30 | 12 | 42 |
| Age, Categorical: >=65 years | Participants | 10 | 8 | 18 |
| Age, Continuous | years | 55 (SD 17) | 55 (SD 20) | 55 (SD 18) |
| Sex: Female, Male: Female | Participants | 23 | 10 | 33 |
| Sex: Female, Male: Male | Participants | 17 | 10 | 27 |
| Ethnicity (NIH/OMB): Hispanic or Latino | Participants | 4 | 2 | 6 |
| Ethnicity (NIH/OMB): Not Hispanic or Latino | Participants | 31 | 16 | 47 |
| Ethnicity (NIH/OMB): Unknown or Not Reported | Participants | 5 | 2 | 7 |
| Race (NIH/OMB): American Indian or Alaska Native | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Asian | Participants | 2 | 2 | 4 |
| Race (NIH/OMB): Native Hawaiian or Other Pacific Islander | Participants | 3 | 0 | 3 |
| Race (NIH/OMB): Black or African American | Participants | 4 | 1 | 5 |
| Race (NIH/OMB): White | Participants | 27 | 13 | 40 |
| Race (NIH/OMB): More than one race | Participants | 0 | 0 | 0 |
| Race (NIH/OMB): Unknown or Not Reported | Participants | 4 | 4 | 8 |
| Region of Enrollment: United States | participants | 40 | 20 | 60 |
| Cause of Acute Respiratory Distress Syndrome (ARDS): Sepsis with pneumonia | Participants | 19 | 12 | 31 |
| Cause of Acute Respiratory Distress Syndrome (ARDS): Sepsis without pneumonia | Participants | 5 | 2 | 7 |
| Cause of Acute Respiratory Distress Syndrome (ARDS): Pneumonia without sepsis | Participants | 11 | 5 | 16 |
| Cause of Acute Respiratory Distress Syndrome (ARDS): Aspiration only | Participants | 4 | 1 | 5 |
| Cause of Acute Respiratory Distress Syndrome (ARDS): Other | Participants | 1 | 0 | 1 |
| Arterial pressure | mm Hg | 75 (SD 10) | 76 (SD 9) | 75 (SD 10) |
| Taking vasopressors at the time of infusion | Participants | 24 | 9 | 33 |
| Sequential Organ Failure Assessment (SOFA) | units on a scale | 8.1 (SD 3.3) | 6.9 (SD 2.7) | 7.7 (SD 3.2) |
| Acute Physiologic Assessment and Chronic Health Evaluation (APACHE) III score | units on a scale | 104 (SD 31) | 89 (SD 33) | 99 (SD 32) |
| Minute ventilation | L/min | 11.1 (SD 3.2) | 9.6 (SD 2.4) | 10.6 (SD 3.0) |
| Respiratory rate | breaths/min | 27.8 (SD 6.6) | 24.5 (SD 6.3) | 26.7 (SD 6.6) |
| Tidal volume | mL/kg predicted body weight | 6.3 (SD 0.9) | 6.1 (SD 0.7) | 6.2 (SD 0.9) |
6. Primary Endpoint
The registry lists the primary outcome as Numbers of Patients Occurred Pre-specified Infusion Associated Events Occurring Within 6 Hours of Study Infusion, measured over 6 hours.
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The primary endpoint counts participants who experienced at least one pre-specified infusion-associated event within 6 hours of study infusion. The event criteria cover acute changes in vasopressor support, new serious cardiac rhythm events, specified worsening hypoxaemia requiring changes in oxygen and ventilator support, and clinical findings consistent with transfusion incompatibility or infection. The registry defines the outcome as a participant count, rather than a severity score or a time-to-event measure.
A short observation window makes this endpoint specific to acute events occurring near the infusion. A defined list of criteria can make event counting more reproducible than an open-ended judgment about whether a participant seemed worse. At the same time, the endpoint does not summarize all aspects of safety: events outside the listed criteria or outside the 6-hour window are not captured by this primary measure. The trial also posted serious adverse event summaries and longer-term secondary outcomes, which address different aspects of the study experience.
The registry reports zero participants meeting the primary endpoint in the hMSC arm and zero in the placebo arm. It also reports a risk difference of 0.0 with an interval from 0.0 to 0.0. These values describe the recorded endpoint results; they do not establish that the underlying risks are identical. With no observed events in either group, the result gives no observed between-arm contrast for this endpoint, while the relatively small arm counts limit what can be learned about uncommon events. The registry record does not report a sponsor-conducted inferential analysis for this endpoint.
7. Results Posted in the Registry
- Human Mesenchymal Stem Cells (hMSCs): 0.0% (0/40) versus 0.0% (0/20) on control; risk difference 0.0 percentage points (95% CI 0.0 to 0.0); unadjusted two-proportion z-test p = n/a.
Risk difference: Human Mesenchymal Stem Cells (hMSCs)
95% CI: 0.0 to 0.0 percentage points
Recalculated from the posted summary statistics
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.
| Secondary outcome | Type | Effect measure | Estimate (95% CI) | Planning size, 80% power |
|---|---|---|---|---|
| Number of Patients Death to Day 28 | Binary Endpoint | Risk difference | -15.0 pts (-36.1 to 6.1) | 121 per arm |
| Mortality to Day 60 | Binary Endpoint | Risk difference | -12.5 pts (-36.7 to 11.7) | 215 per arm |
Number of Patients Death to Day 28
- Endpoint type: Binary: the number of participants with the event, compared as proportions.
- Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
- Result:
- Plasma-Lyte A: 15.0% (3/20) versus 30.0% (12/40) on control; risk difference -15.0 percentage points (95% CI -36.1 to 6.1), risk ratio 0.50 (0.16–1.57); unadjusted two-proportion z-test p = 0.206.
- Interpretation:
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Plasma-Lyte A and the control. The interval gives the range of effects that remain compatible with the data.
Mortality to Day 60
- Endpoint type: Binary: the number of participants with the event, compared as proportions.
- Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
- Result:
- Plasma-Lyte A: 25.0% (5/20) versus 37.5% (15/40) on control; risk difference -12.5 percentage points (95% CI -36.7 to 11.7), risk ratio 0.67 (0.28–1.57); unadjusted two-proportion z-test p = 0.333.
- Interpretation:
- The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Plasma-Lyte A and the control. The interval gives the range of effects that remain compatible with the data.
The secondary outcomes cover several distinct domains: oxygenation, lung injury, organ dysfunction, mortality, ventilator-free days, and non-pulmonary organ-failure-free days. These outcomes add context to the acute infusion-event endpoint because they address clinical status beyond the first 6 hours. However, each outcome measures a different aspect of illness and has its own observation period and interpretation; they should not be treated as interchangeable measures of one underlying effect.
Numerical results are available in the registry record for mortality to day 28 and mortality to day 60. Both posted risk-difference intervals span values on either side of zero, and both unadjusted p values are above 0.05. This pattern is inconclusive rather than evidence of equivalence. Numerical results for the other secondary outcomes are not reported in the registry record. The record therefore does not support a complete assessment of whether findings across physiological and organ-function outcomes were consistent with one another.
The primary endpoint and the mortality outcomes address different questions. The primary endpoint concerns a specified set of events within 6 hours of infusion; mortality outcomes concern whether participants died by later time points. The absence of primary events does not predict or explain the later mortality findings, and the mortality comparisons do not change what was observed for the primary endpoint.
Secondary results generally carry less weight than a pre-specified primary result when many outcomes are assessed. In general, evaluating multiple outcomes increases the chance that at least one comparison will appear unusual by chance. The mortality p values are unadjusted, and the registry record does not report a multiplicity plan. This makes isolated secondary findings exploratory, particularly when estimates are imprecise. The registry record also does not report a complete analysis plan or numerical results for all listed secondary outcomes, limiting the ability to judge the full pattern.
Secondary outcomes registered for this trial, for reference (only those above were recalculated):
- PaO2:FiO2 Change From Baseline to Day 3
- Lung Injury Score From Baseline to Day 3
- Oxygenation Index Change From Baseline to Day 2
- SOFA Score Change From Baseline to Day 3
- Number of Patients Death to Day 28
- Mortality to Day 60
- Number of Ventilator-free Days to Day 28
- Non-pulmonary Organ-failure-free Days to Day 28
9. Statistical Interpretation
For the primary endpoint, the registry records no participants with a pre-specified infusion-associated event within 6 hours in either randomized group. The reported arm counts are 40 for hMSCs and 20 for Plasma-Lyte A. The posted risk difference is 0.0, with the recorded interval also running from 0.0 to 0.0. This is a description of the observed counts, not evidence that the two groups have exactly the same underlying event risk.
When an outcome has no observed events in either arm, the study cannot estimate a nonzero observed risk difference. That absence of events is reassuring as a description of what was recorded during the specified window, but it does not rule out uncommon events or establish equivalence. The registry record does not provide an inferential analysis for the primary endpoint that would characterize uncertainty using a method suitable for zero-event data. The posted interval of 0.0 to 0.0 should therefore be read as the reported comparison, not as proof of perfect precision.
The posted secondary mortality results show numerical differences between the arms, but the associated unadjusted comparisons are imprecise. For mortality to day 28, the risk-difference estimate is -0.15 and the reported interval spans values on both sides of zero; the unadjusted p value is 0.206. For mortality to day 60, the estimate is -0.125, its interval also spans values on both sides of zero, and the unadjusted p value is 0.333. These results do not provide clear statistical evidence of a between-group difference. They also do not establish that the groups have the same mortality risk.
The comparison direction for those mortality results is stated in the registry as Plasma-Lyte A relative to hMSCs. The risk-ratio intervals also include 1 for both time points. Point estimates alone should not be treated as conclusive: the intervals show substantial uncertainty, and the group sizes are limited. The registry record lists additional secondary outcomes, but numerical results for those outcomes are not reported in the record summarized for these results.
These are results from a Phase 2a study, not a definitive estimate of all possible benefits or harms. The record does not report a sample-size justification or a sponsor analysis for the primary endpoint. The mortality p values are unadjusted; no multiplicity procedure is reported. Consequently, the findings are best read as descriptive and exploratory, with particular caution around claims based on individual secondary comparisons.
10. Statistical Methods Explained
Counting an event. The primary outcome is a count of participants meeting at least one of the pre-specified event criteria within 6 hours. A participant count is distinct from a count of every event: a person with more than one qualifying event would still contribute as a participant with an event under this outcome definition. The registry reports zero participants meeting the definition in both arms. The record does not give event-level detail or explain how overlapping events were handled.
Comparing risks. A risk difference expresses the difference between the proportions of participants with an event in two groups. A value of 0 means the observed proportions are the same. For the primary endpoint, the recorded risk difference is 0.0. With no observed events in either arm, that point estimate contains no observed contrast to distinguish the groups. The method used to form the posted interval is not described in the registry record, so the interval should not be interpreted as a complete account of uncertainty around rare or unobserved events.
Confidence intervals. In general, a confidence interval describes the range of values compatible with the data under a specified statistical method. A wide interval indicates limited precision; an interval spanning the no-difference value indicates that the data are compatible with no difference as well as with effects in either direction. For the posted mortality outcomes, the risk-difference intervals span values on both sides of zero. The registry record does not report sponsor-conducted inferential analyses for these outcomes. The displayed estimates and intervals should therefore be understood as posted results and comparisons, not as a substitute for a reported analysis plan.
Risk ratios and odds ratios. A risk ratio compares the event risks between groups as a ratio, while an odds ratio compares the odds. A ratio of 1 corresponds to equal risks or odds under the respective measure. The registry includes both measures for the mortality comparisons. Their intervals include 1, consistent with substantial uncertainty about the direction and size of a difference. These measures are alternative ways to express the same binary-outcome comparison; they do not remove the limitations created by small groups or limited event counts.
Unadjusted p values. A p value assesses how unusual the observed data would be under a specified null model and analysis procedure. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of an effect. The mortality p values are explicitly unadjusted. The registry record does not report a multiplicity adjustment or a full statistical analysis plan. They should therefore be interpreted cautiously, especially because multiple outcomes are considered.
Planning illustrations. The registry includes sample-size planning illustrations based on the observed mortality effects, using a two-sided significance level of 0.05, equal allocation, and a normal approximation. The illustrations show participants per arm for 80% and 90% power when planning for 100%, 75%, or 50% of the observed effect. For day 28 mortality, the corresponding 80% planning figures are 121, 228, and 540 participants per arm; for day 60 mortality they are 215, 393, and 906. The record also gives the corresponding 90% figures. These are planning illustrations, not the trial's original design justification and not observed power. Because they are based on effects observed in a small study, they should not be treated as stable estimates of the sample size needed for a future trial.
Analysis populations and missing data. The arm counts are reported for the outcomes, but the registry record does not describe whether all randomized participants were included in each analysis, how missing measurements were treated, or whether any participants were excluded after randomization. It does not report imputation methods, sensitivity analyses, or reasons for missingness. These details matter because an outcome comparison can depend on which participants contribute data and how absent outcomes are handled.
How to read the methods overall. The posted counts and simple comparisons communicate the observed results, but the registry record does not supply a complete account of the statistical procedures behind them. The primary endpoint's zero-event result is particularly dependent on appropriate methods for rare events. For mortality, the wide intervals and unadjusted p values call for emphasis on uncertainty rather than a binary conclusion. General statistical principles do not fill gaps in the registry's description of the trial's own analysis.
11. Confidence Intervals and Precision
The primary endpoint has zero recorded events in both groups. The risk difference is reported as 0.0, and its posted interval runs from 0.0 to 0.0. Although the displayed interval has no width, that should not be mistaken for certainty about the true event risks. The registry record does not describe the interval method, and zero observed events leave limited information about how frequently an uncommon event might occur.
The mortality comparisons are more informative about precision because their intervals are not collapsed to a single value. For day 28 and day 60 mortality, each risk-difference interval spans both negative and positive values. This means the posted data are compatible with a lower risk in one group, little or no difference, or a higher risk, depending on the value within the interval. The breadth of those intervals reflects substantial uncertainty in the comparisons. The risk-ratio intervals also include 1 for both outcomes.
The group sizes are 40 and 20, and the mortality event counts are limited. With results based on groups of this size, a modest change in the number of participants with an event can materially affect an estimate. That is a general feature of binary outcomes in small studies, not a separate result from the trial. Estimates from such studies can be unstable, and future studies may observe a different magnitude or direction.
Precision is not the same as statistical significance. The unadjusted mortality p values are 0.206 and 0.333, while their intervals include the no-difference values. These quantities both signal uncertainty but answer different statistical questions. Neither a nonsignificant p value nor an interval that includes no difference establishes that the groups are equivalent. The record does not report a formal equivalence or non-inferiority design.
The most defensible reading is therefore descriptive: no primary events were recorded during the specified window, while the secondary mortality estimates remain imprecise. The registry record does not report a primary-endpoint interval method or a sponsor analysis that would support stronger precision claims.
12. Multiplicity, Missing Data and Analysis Populations
The registry identifies multiple outcomes, including the primary infusion-associated event count and secondary measures of oxygenation, lung injury, organ dysfunction, mortality, and days without particular forms of organ support or failure. Considering multiple outcomes raises the general issue of multiplicity: as more hypotheses are tested, the chance of at least one apparently notable result can increase even if there are no underlying effects. The record does not report whether the investigators used a hierarchy of outcomes, adjusted significance thresholds, or another method to address multiplicity.
The posted mortality p values are unadjusted. They should not be read as if they accounted for all outcomes examined. The registry record does not report the number of statistical tests conducted, a multiplicity procedure, or a complete statistical analysis plan. It also lists no sponsor-reported analyses for the displayed results. As a result, the intended inferential role of the secondary comparisons is not fully described.
Missing-data handling is also not reported in the registry record. The outcome counts give arm sizes, but the record does not explain whether every randomized participant contributed to every outcome, whether any observations were missing, or how missing values were handled. It does not report an intention-to-treat definition, a per-protocol population, imputation, sensitivity analyses, or reasons for missingness. These omissions prevent assessment of whether the analyzed participants and observed outcomes differed from the full randomized groups.
These gaps do not change the recorded event counts, but they limit how confidently the comparisons can be interpreted. Missingness can affect an estimate if participants without recorded outcomes differ from those with outcomes. Multiplicity can affect how much weight should be placed on isolated secondary findings. Without details on either issue, emphasis should remain on the descriptive results and their uncertainty, rather than on a definitive claim based on one comparison.
13. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Human Mesenchymal Stem Cells | 17 | 40 | 42.5% |
| Plasma-Lyte A | 11 | 20 | 55.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Human Mesenchymal Stem Cells | 15 of 40 |
| Plasma-Lyte A | 5 of 20 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Human Mesenchymal Stem Cells | 2 of 40 |
| Plasma-Lyte A | 2 of 20 |
Most frequent serious adverse event terms
| Event term | Organ system | Human Mesenchymal Stem Cells | Plasma-Lyte A |
|---|---|---|---|
| Death due to multi-organ failure | General disorders | 8 of 40 | 2 of 20 |
| Death due to worsening respiratory failure | Respiratory, thoracic and mediastinal disorders | 3 of 40 | 0 of 20 |
| Pulmonary embolism | Respiratory, thoracic and mediastinal disorders | 1 of 40 | 2 of 20 |
| Death due to cardiac arrest | Cardiac disorders | 1 of 40 | 1 of 20 |
| Lower extremity weakness | Nervous system disorders | 0 of 40 | 1 of 20 |
| Hospital readmission due to post-surgical pain | Surgical and medical procedures | 0 of 40 | 1 of 20 |
| Acute renal failure | Renal and urinary disorders | 1 of 40 | 0 of 20 |
| Hospital readmission due to pyelonephritis | Renal and urinary disorders | 0 of 40 | 1 of 20 |
| Acute leukemia | Blood and lymphatic system disorders | 0 of 40 | 1 of 20 |
| Diffuse alveolar hemorrhage | Blood and lymphatic system disorders | 1 of 40 | 0 of 20 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Human Mesenchymal Stem Cells | Plasma-Lyte A |
|---|---|---|---|
| Hepatitis | Hepatobiliary disorders | 1 of 40 | 0 of 20 |
| Fecal and urinary incontinence | Renal and urinary disorders | 0 of 40 | 1 of 20 |
| Deep venous thrombosis | Vascular disorders | 1 of 40 | 0 of 20 |
| Acute anemia | Blood and lymphatic system disorders | 0 of 40 | 1 of 20 |
Serious events with the highest arm-level rate: death due to multi-organ failure (20.0%), pulmonary embolism (10.0%), death due to worsening respiratory failure (7.5%), lower extremity weakness (5.0%), hospital readmission due to post-surgical pain (5.0%).
The registry's serious adverse event summary reports 17 participants affected among 40 in the hMSC group and 11 among 20 in the Plasma-Lyte A group. The summary separately reports other adverse events affecting 2 participants in each group. It also records 15 deaths in the hMSC group and 5 in the placebo group. These are group-level counts; the registry record does not establish that the events were caused by the assigned infusion.
The listed serious event terms include death due to multi-organ failure, pulmonary embolism, death due to worsening respiratory failure, lower extremity weakness, and hospital readmissions for post-surgical pain and pyelonephritis. The registry gives percentages alongside these terms, but the denominator and counting basis for each listed percentage are not explained in the record. The percentages should therefore be read as reported labels, not used to infer individual risk or causation.
Serious adverse event summaries are broader than the primary endpoint. The primary endpoint concerns a defined set of events within 6 hours; the serious adverse event summary does not use that same restriction. Neither summary alone determines whether an event was related to treatment. The registry record does not provide attribution details or a complete account of event timing and overlap.
14. What the Result Does, and Does Not, Mean
The primary result means that the registry recorded no participants meeting the pre-specified infusion-associated event definition within 6 hours in either randomized group. It does not mean that no adverse events occurred over the full study period: the registry separately reports serious adverse events, other adverse events, and deaths. Those safety summaries cover a broader scope than the short-window primary endpoint.
The result also does not prove that the hMSC and placebo groups have identical risks. Zero observed events in both groups leave uncertainty about uncommon events, especially with 40 participants in one arm and 20 in the other. The reported risk difference of 0.0 is an observed comparison, not a guarantee that the underlying risks match. The registry record does not report a suitable inferential analysis for this zero-event primary result.
The mortality comparisons show numerical estimates, but their intervals include the no-difference value and their unadjusted p values are 0.206 for day 28 and 0.333 for day 60. These results do not establish a mortality difference. They also do not demonstrate equivalence or rule out an important difference, because the estimates are imprecise and the registry does not report an equivalence or non-inferiority framework.
More broadly, the trial's secondary outcomes address clinical measures beyond infusion safety, but most of their numerical results are not reported in the registry record. This limits the conclusions that can be drawn about the overall pattern of outcomes. The trial is a Phase 2a study, and the record does not report a sample-size justification, full analysis plan, missing-data methods, or multiplicity approach. The posted results therefore support a cautious account of what was observed, not a definitive conclusion about all potential effects or harms.
No treatment recommendation follows from these statistics. Whether a result is clinically important depends on context beyond the estimates alone, and the registry record does not provide enough information to make that judgment. The findings are most useful as a description of observed outcomes and a guide to the uncertainty that a larger, fully specified study would need to address.
15. Why This Trial Matters Statistically
This trial illustrates a central statistical challenge in early clinical research: a primary safety endpoint can have no observed events, yet that result does not provide strong evidence that the event is impossible or that risks are equal. The zero counts describe the trial experience within a specified window. They also leave little observed information with which to distinguish the groups, especially for uncommon outcomes.
The study also shows why estimates and their precision matter alongside point results. The mortality estimates differ numerically, but their intervals are broad and cross the no-difference value. A point estimate alone can invite an overly simple conclusion; the interval reminds readers that the data are compatible with a range of possibilities. The unadjusted p values do not resolve that uncertainty or establish equivalence.
Another statistical lesson concerns the separation between a primary endpoint and multiple secondary outcomes. The primary endpoint targets acute events around infusion, whereas secondary outcomes cover later mortality and other aspects of illness. A result on one cannot substitute for results on another. When multiple secondary comparisons are considered without a reported multiplicity plan, isolated findings deserve particular caution.
Finally, the registry illustrates how incomplete reporting affects interpretation. The record does not report a primary-endpoint analysis method, a sample-size justification, missing-data methods, a full analysis plan, or a multiplicity procedure. Those details are important for evaluating precision and the intended role of each comparison. The posted counts remain informative descriptions, but a fuller statistical account would be needed to support stronger inferential claims.
16. 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 endpoint has zero recorded events in both arms, and the registry does not report the method used to calculate its interval. This makes the displayed interval difficult to interpret as a measure of uncertainty for a rare or zero-event outcome. The result is limited to the specified criteria and 6-hour window and cannot describe all adverse events during later follow-up.
The observed arm counts are 40 and 20. The registry record does not explain the difference in group sizes or provide baseline characteristics that would help readers evaluate the enrolled population and group comparability. It also does not report detailed eligibility criteria in the results information summarized for this trial.
The registry lists serious adverse events and deaths, but does not report whether events were judged related to the study product, how event overlap was handled, or the denominator basis for percentages attached to serious event terms. These summaries cannot by themselves establish causation. The primary endpoint and the broader adverse event summaries have different scopes and should not be conflated.
For mortality, the reported comparisons are unadjusted and imprecise. The registry does not report a multiplicity approach, missing-data handling, analysis populations, or a complete statistical plan. Numerical results for most listed secondary outcomes are not reported in the registry record. These omissions limit assessment of consistency across outcomes and of potential bias from missing measurements or selective emphasis.
The registry record also does not report the rationale used to choose the enrollment target. The planning illustrations based on observed mortality effects are not a substitute for the original design justification. Given the Phase 2a setting and limited group sizes, the posted results should be treated as exploratory descriptions rather than precise estimates of uncommon outcomes.
17. Related Tutorials
18. Related Statistical Calculators
19. Sources
- ClinicalTrials.gov: Human Mesenchymal Stromal Cells For Acute Respiratory Distress Syndrome (START) — NCT02097641.
- Registry record last updated 2019-04-10; status last verified 2019-03.
- Results first posted 2019-04-10.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
20. Record Summary
START, also titled Human Mesenchymal Stromal Cells For Acute Respiratory Distress Syndrome, is a completed randomized Phase 2 study of allogeneic bone marrow-derived hMSCs in ARDS. The registry describes a parallel-group, placebo-controlled trial and records triple masking, while its brief summary uses double-blind wording. Enrollment was 60, with 40 participants in the hMSC arm and 20 in the Plasma-Lyte A arm.
The primary endpoint counted participants meeting specified infusion-associated event criteria within 6 hours. The registry reports zero in both arms. Secondary outcomes included mortality and physiological and organ-function measures; numerical mortality comparisons are posted for day 28 and day 60, while numerical results for other listed outcomes are not reported in the registry record. The record does not report a full statistical analysis plan, missing-data methods, or a multiplicity approach.