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 Remote Ischemic Conditioning in Patients With Acute Stroke (RESIST) (NCT03481777), a randomized na study. The primary outcome is Modified Rankin Scale at 3 Months in Acute Stroke (AIS and ICH).
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03481777 |
| Phase | NA |
| Condition | Stroke, Acute, Ischemic Stroke, Hemorrhagic Stroke |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Double |
| Enrollment | 1500 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Grethe Andersen |
| Results posted | Yes |
2. Clinical Question
Population
People with acute stroke, including acute ischemic stroke and intracerebral hemorrhage, are the broad population of interest. The registry lists stroke conditions that also include transient ischemic attack and other cerebrovascular diagnoses. The primary endpoint description refers to acute stroke patients, while several secondary outcomes are restricted to particular diagnoses or treatment groups. Eligibility criteria and further participant characteristics are not reported in the registry record.
Intervention
Remote ischemic conditioning (RIC) was delivered with an automated device in the hyperacute prehospital phase, as an adjunct to standard treatment. The intervention description specifies five cycles over 50 minutes, with cuff inflation followed by deflation in each cycle, and gives an inflation pressure of 200 mmHg.
Comparator
The comparison was sham remote ischemic conditioning, also delivered with an automated device in the hyperacute prehospital phase. The registry describes the same cycle schedule for the sham procedure, but the sham cuff-pressure description is incomplete in the record. Further details about how the sham was designed to maintain masking are not reported in the registry record.
Primary question
The main question was whether adding RIC to standard treatment was associated with better long-term recovery after acute stroke, assessed by the modified Rankin Scale at three months. The scale captures the degree of disability and dependency, from no symptoms to death, with higher scores indicating worse outcome. The registry specifies a generalized ordinal logistic regression analysis, but does not report its comparative result. The posted arm summaries alone do not establish whether the intervention changed the distribution of disability outcomes.
This question is clinically broad because acute stroke includes different diagnoses and treatment pathways. The registry also lists secondary questions for ischemic stroke, intracerebral hemorrhage, patients receiving reperfusion therapy, early neurological change, symptom remission, and major adverse cardiac and cerebral events. These outcomes can help describe different aspects of recovery and safety, but they do not replace the primary question. Their interpretation depends on their own definitions, populations, follow-up periods, and analyses, several of which are not fully reported in the registry record.
3. Trial Design
RESIST is described as a multicenter, randomized, patient-assessor blinded, sham-controlled study. The registry identifies the design as randomized, double masked, and parallel group. Participants were assigned to remote ischemic conditioning or sham remote ischemic conditioning; the record does not describe the randomization sequence or allocation-concealment procedures. The registry lists an enrollment of 1500. The numbers contributing to the posted primary outcome summaries differ from that enrollment figure, and the reasons for those differences are not reported in the registry record.
The intervention was intended for the hyperacute prehospital phase, before or alongside usual acute-stroke care. The RIC description specifies five cycles over 50 minutes, each with five minutes of cuff inflation followed by five minutes with the cuff deflated. The stated cuff pressure is 200 mmHg. The sham description repeats the cycle schedule but ends before a complete pressure specification is given. The record identifies both interventions as devices; further operational details, adherence information, and the exact distinction between active and sham device procedures are not reported in the registry record.
The primary endpoint was the modified Rankin Scale at three months among acute stroke patients with ischemic stroke or intracerebral hemorrhage. The registry says two independent assessors would evaluate participants by telephone or face to face, with a further blinded assessor involved if assessments disagreed. It describes the primary analysis as generalized ordinal logistic regression. The posted primary results, however, give medians and inter-quartile ranges by arm rather than a comparative estimate from that model. A model result, its uncertainty, and any adjustment details are not reported in the registry record.
The record marks the study completed and lists a start date of 2018-03-15 and a primary completion date of 2023-02-03. It does not report the protocol’s sample-size justification, interim analyses, or detailed statistical analysis plan. Although multiple secondary outcomes are listed, the registry record does not provide a complete account of how each was analyzed or how the analyses were prioritized.
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: The RESIST trial is a prehospital stroke trial. Patients are randomzied to remote ischemic conditioning or sham in the ambualnce before their final diagnosis. Patients who have a stroke (ischemic stroke or intracerebral hemorrhage, n= 902) are the target population and treated further according to randomization and outcome assessed as 3 months modified Rankin Scale . Patients with a non stroke diagnosis (n=531) are not treated further and not followed beyond hospital discharge.
Overall Study
| Stage | Remote Ischemic Condtioning (RIC) | Sham Remote Ischemic Conditioning |
|---|---|---|
| STARTED | 749 | 751 |
| COMPLETED | 436 | 466 |
| NOT COMPLETED | 313 | 285 |
Reasons for not completing this period:
| Reason | Remote Ischemic Condtioning (RIC) | Sham Remote Ischemic Conditioning |
|---|---|---|
| Withdrawal by Subject | 36 | 31 |
| Non stroke = non target population | 277 | 254 |
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 | Remote Ischemic Conditioning | Sham | Total |
|---|---|---|---|---|
| Participants analysed | 436 | 466 | 902 | |
| Age, Continuous | years | 72 (62 to 79) | 73 (62 to 80) | 73 (62 to 79) |
| Sex: Female, Male: Female | Participants | 165 | 170 | 335 |
| Sex: Female, Male: Male | Participants | 271 | 296 | 567 |
| Race and Ethnicity Not Collected | Participants | — | — | 0 |
| Region of Enrollment: Denmark | participants | 436 | 466 | 902 |
| National Institutes of Health Stroke Scale (NIHSS) | points on NIHSS | 5 (2 to 11) | 5 (2 to 10) | 5 (2 to 10) |
| Diagnosis: Acute ischemic stroke | Participants | 349 | 388 | 737 |
| Diagnosis: Intracerebral hemorrhage | Participants | 87 | 78 | 165 |
6. Primary Endpoint
The registry lists the primary outcome as Modified Rankin Scale at 3 Months in Acute Stroke (AIS and ICH), measured over 3 months.
| Element | Registry entry |
|---|---|
| Result type | Median |
| Measure of spread | Inter-Quartile Range |
| Unit | units on a scale |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is the modified Rankin Scale (mRS) at three months in acute stroke, including ischemic stroke and intracerebral hemorrhage. The scale ranges from 0 to 6: a score of 0 means no symptoms and a score of 6 means death. Higher scores represent worse outcome. Between these endpoints, the scale describes increasing limitations and dependency after stroke.
The registry specifies a median and inter-quartile range as descriptive summaries and names generalized ordinal logistic regression as the intended comparative analysis. A median identifies the middle of the observed score distribution; the inter-quartile range describes the central spread. These summaries can convey the typical score and its variability, but they do not show the full distribution or, by themselves, quantify the probability of a better outcome in one group. Because mRS is ordinal, its categories have a meaningful order but the distance between neighboring scores should not automatically be treated as equal.
An ordinal regression model can use the ordering of all scale categories rather than reducing the outcome to a simple good-versus-poor classification. General statistical explanation: such a model can assess whether the outcome distribution tends to shift toward better or worse categories between groups, subject to the model assumptions and the chosen analysis. The registry record does not report the model estimate, confidence interval, or test result. The posted medians are therefore descriptive evidence only and cannot substitute for the planned comparative analysis.
Assessment procedures matter because mRS scoring involves judgment about symptoms and dependency. The registry describes independent telephone or face-to-face assessments and a blinded third assessor in cases of disagreement. It does not report how often disagreements occurred or whether assessment mode affected the results.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The primary outcome summaries show a median mRS of 2 in the RIC group, with an inter-quartile range from 1 to 3, among 436 participants. The sham group median was 1, also with an inter-quartile range from 1 to 3, among 466 participants. Since higher mRS scores indicate worse outcome, the arm medians are descriptively higher in the RIC group. The inter-quartile ranges are the same as reported, but those summaries do not reveal how scores were distributed across all scale categories or how much the groups overlapped.
The registry names generalized ordinal logistic regression for the primary outcome, but no comparative estimate, confidence interval, or p-value for that analysis is posted. Consequently, the central statistical question—whether the overall ordered outcome distribution differed between randomized groups—cannot be answered from the primary summaries alone. The registry record also does not report whether the primary model was adjusted, how missing outcome values were handled, or which participants were included in the analysis. No conclusion about the size or precision of a primary treatment contrast can be drawn without those results.
The posted secondary comparisons include two binary outcomes. For complete remission of symptoms within 24 hours among people with transient ischemic attack, the recorded counts are 96 of 515 in the RIC group and 90 of 536 in the sham group; the unadjusted p-value is 0.432. For early neurological improvement in acute ischemic stroke within 24 hours, the counts are 108 of 425 and 131 of 451, respectively; the unadjusted p-value is 0.227. The reported unadjusted confidence intervals for the risk differences include no difference for both comparisons. These are not adjusted analyses, and the registry does not report a multiplicity procedure for considering several secondary outcomes.
These secondary results point in different descriptive directions: the remission count is higher in the RIC arm, while the early-improvement count is lower. Neither comparison, considered alone, provides strong evidence of a group difference at conventional significance levels, and neither establishes equivalence. A non-significant result is not proof that the groups are identical; the estimate and its uncertainty matter, as do outcome definitions and the analysis population.
The primary result deserves the greatest emphasis because it was identified as the main outcome, but the required inferential result is absent. The descriptive medians and the secondary comparisons should not be combined into an overall claim about benefit or harm. They refer to different scales, populations, and time points, and their uncertainty and analysis details are incomplete. The posted data allow a limited description of outcome values and event counts, not a definitive assessment of the intervention’s comparative effect.
9. Statistical Methods Explained
Randomization is intended to make treatment groups comparable, on average, with respect to both measured and unmeasured factors at the time of assignment. A sham comparison can also help distinguish effects associated with the active procedure from effects related to receiving a device-based procedure or expectations. These are general design principles; the registry record identifies randomization and sham control but does not report the sequence-generation method, concealment method, or details of how masking was maintained.
Why describe an ordinal outcome with an ordinal model?
The modified Rankin Scale has ordered categories, from no symptoms through increasing disability to death. A simple comparison of medians compresses the information into a central value, while a generalized ordinal logistic regression can use the ordering across categories. General statistical explanation: ordinal models are useful when the ordering matters but equal spacing between categories cannot be assumed. Their conclusions depend on the model specification and assumptions. The registry record names this model for the primary endpoint but does not report the estimate, uncertainty interval, model assumptions, or diagnostics. It is therefore not possible to evaluate the fitted model or its robustness from the posted primary summaries.
A confidence interval gives a range of values compatible with the data and the statistical model under its assumptions. It communicates precision as well as direction: a wide interval indicates that a range of effects remains plausible, while an interval concentrated near no difference would indicate greater precision around small effects. For the primary mRS analysis, no confidence interval is reported in the registry record. For the two posted binary secondary comparisons, the reported risk-difference intervals include no difference. Those intervals do not prove that there is no effect; they indicate that the data do not distinguish a nonzero difference from no difference at the stated level of uncertainty.
How should the secondary comparisons be read?
The two binary endpoints are summarized as participant counts in each group and accompanied by unadjusted risk-difference estimates, confidence intervals, and p-values. A risk difference compares the proportions with the event between groups. General statistical explanation: the sign of a risk difference depends on which group is subtracted from which, so the direction must be read with the stated comparison in mind. The registry’s unadjusted p-values are 0.432 for complete remission within 24 hours and 0.227 for early neurological improvement. They are not evidence that the treatment groups are exactly alike, and they do not account for considering multiple outcomes.
The confidence intervals and p-values answer related but different questions. A p-value measures how unusual the data would be under a specified null hypothesis and analysis; it is not the probability that the null hypothesis is true. A confidence interval describes a range of effect estimates compatible with the data and model. In these secondary results, the intervals include no difference, so the estimates are imprecise with respect to whether there is a difference in either direction. No adjusted secondary analyses are reported.
What do the planning illustrations represent?
The registry includes sample-size planning illustrations for the two binary comparisons. Their assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation. The calculations use the observed effect as a starting point and also illustrate planning for 75% and 50% of the observed effect. For the remission outcome, the listed participants per arm for 80% power are 6,687, 11,767, and 26,200 across those effect assumptions; the corresponding figures for 90% power are 8,952, 15,752, and 35,073. For early neurological improvement, the respective figures for 80% power are 2,354, 4,228, and 9,609, and for 90% power are 3,150, 5,660, and 12,864.
These are planning illustrations calculated from observed effects, not the trial’s own design justification and not observed power. General statistical explanation: planning estimates can change substantially when the assumed effect changes, especially when the effect is small. Estimates based on an observed effect can also be unstable, so the illustrations should not be treated as a reliable forecast of what a future study must enroll. The protocol’s sample-size justification is not reported in the registry record.
Why can over-dispersion matter?
General statistical explanation: over-dispersion occurs when observed variation in data is greater than a statistical model expects. If a model underestimates variation, standard errors can be too small and confidence intervals too narrow, making evidence appear more precise than it is. Whether this concern applies depends on the outcome and model. The registry record does not report a model fit assessment or information that would allow over-dispersion to be evaluated for the primary ordinal analysis. It should not be assumed either to have occurred or to have been absent.
Finally, all inferential results depend on the analysis population and handling of missing observations. The endpoint summaries have different participant counts across outcomes, but the registry does not explain the reasons, report missingness by group, or state whether the primary analysis followed an intention-to-treat approach. General statistical explanation: excluding participants or outcomes after randomization can affect comparability if the reasons for missingness relate to prognosis or treatment. Without the analysis-population and missing-data details, the precision and potential bias of the posted comparisons cannot be fully assessed.
10. Multiplicity, Missing Data and Analysis Populations
The registry record does not report how the trial handled multiplicity across its primary and secondary outcomes. This matters because the study lists multiple outcomes, including outcomes at different time points and in different stroke subgroups. General statistical explanation: testing many hypotheses increases the opportunity for chance findings. A prespecified hierarchy, adjustment procedure, or careful distinction between confirmatory and exploratory outcomes can guide interpretation. Without such information, the unadjusted secondary p-values should be treated as exploratory rather than as stand-alone evidence of a definitive effect.
Missing-data methods are also not reported. The primary outcome summaries include 436 participants in the RIC group and 466 in the sham group, while other posted outcomes use different denominators. The registry does not explain whether this reflects outcome eligibility, incomplete assessment, exclusions, or other reasons. Nor does it state how missing mRS assessments or missing binary outcomes were handled. The assessment description explains how disagreement between assessors would be resolved, but it does not describe missing follow-up or imputation procedures.
The analysis populations are not reported in sufficient detail to determine whether the primary comparison included all randomized participants, only those with observed outcomes, or another defined group. The record does not specify whether analyses followed intention-to-treat principles, whether per-protocol or safety populations were used, or how deviations from the assigned procedure were handled. These details affect how directly a result reflects the effect of assignment versus the effect of treatment received.
The registry specifies generalized ordinal logistic regression for the primary outcome, but the corresponding model results and analysis details are absent. The statistical plan, interim analyses, and any stopping rules are not reported in the registry record. The record also lists an enrollment of 1500 but does not report the sample-size justification or reconcile enrollment with the outcome-specific denominators. These omissions prevent a complete assessment of analysis choices, precision, and potential sources of bias.
11. 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 |
|---|---|---|---|
| Remote Ischemic Conditioning (RIC) | 335 | 713 | 47.0% |
| Sham | 325 | 720 | 45.1% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Remote Ischemic Conditioning (RIC) | 41 of 713 |
| Sham | 47 of 720 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 5% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Remote Ischemic Conditioning (RIC) | 54 of 713 |
| Sham | 11 of 720 |
Most frequent serious adverse event terms
| Event term | Organ system | Remote Ischemic Conditioning (RIC) | Sham |
|---|---|---|---|
| SAE | Cardiac disorders | 335 of 713 | 325 of 720 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Remote Ischemic Conditioning (RIC) | Sham |
|---|---|---|---|
| Upper extremity pain during treatment and/or skin petechia | Musculoskeletal and connective tissue disorders | 54 of 713 | 11 of 720 |
Serious events with the highest arm-level rate: sae (47.0%).
The safety results give serious adverse event counts of 335 among 713 participants at risk in the RIC group and 325 among 720 participants at risk in the sham group. The record also reports other adverse event counts of 54 among 713 and 11 among 720, respectively. Death counts are 41 in the RIC group and 47 in the sham group. These are posted counts with the stated at-risk denominators; they are not a substitute for a prespecified comparative safety analysis.
The registry lists SAE at 47.0% among the top serious adverse event terms. This term-level information should be distinguished from the overall serious adverse event counts. The record does not provide further detail in the information reported about event definitions, severity grading, timing, relatedness to the intervention, or statistical comparisons. The groups also have different safety denominators from the primary endpoint denominators, and the reasons are not reported.
General statistical explanation: adverse event counts can describe the burden recorded during a trial, but interpretation of comparative risk requires attention to the population at risk, follow-up duration, event ascertainment, and whether participants can have more than one event. The posted counts alone do not establish that the intervention caused an event or that the groups differ in safety. They should be read as descriptive registry data rather than as a complete causal safety assessment.
12. What the Result Does, and Does Not, Mean
The registry’s primary summaries describe the mRS distribution only through arm medians and inter-quartile ranges. They show a median score of 2 for RIC and 1 for sham, with an inter-quartile range from 1 to 3 in both groups. Since higher scores indicate worse outcome, the median summaries are descriptively less favorable in the RIC group. However, the complete score distributions and the planned ordinal regression result are not reported. A difference between medians alone does not establish a statistically reliable or clinically important difference in the overall outcome.
The two posted secondary comparisons likewise do not establish a clear difference between groups: their unadjusted intervals include no difference, and the p-values are 0.432 and 0.227. Those results do not demonstrate equivalence or rule out effects that the data estimate imprecisely. They also concern distinct binary endpoints and cannot settle the primary question about the ordered three-month disability outcome.
The registry record does not report the primary model estimate, its confidence interval, or a primary p-value. It therefore does not allow readers to judge the size or precision of the primary comparative effect. Nor does it report enough detail about missing outcomes, analysis populations, multiplicity, or the sample-size rationale to evaluate those aspects of the evidence fully. The absence of a posted analysis is not itself evidence that there was no difference; it means the result cannot be assessed from the record’s reported primary statistics.
These findings should not be translated into advice for individual patients or into a claim that RIC does or does not work. The registry describes a randomized comparison and records outcome and safety data, but the central inferential result is not available. The most defensible reading is limited: the reported descriptive primary summaries differ in their medians, the two posted unadjusted secondary comparisons are inconclusive, and the full comparative analysis needed to interpret the trial’s primary question is not reported in the registry record.
13. Why This Trial Matters Statistically
RESIST raises a statistically important challenge: how to compare recovery when the main outcome is an ordered disability scale rather than a simple binary event. The modified Rankin Scale contains multiple levels, and collapsing it into a single threshold would discard distinctions among those levels. An ordinal analysis can retain more of that structure, but its result needs to be reported with an effect estimate and uncertainty so readers can assess direction and precision. The registry identifies an ordinal model, yet does not report the model result.
The design also illustrates why a sham-controlled randomized comparison is informative for a device-based intervention. Randomization is intended to support a fair comparison, and masking can reduce the influence of expectations or assessment knowledge. The registry describes patient-assessor blinding and an independent assessment process, including resolution of disagreements. At the same time, important procedural details about concealment, masking, adherence, and the sham device are not reported, limiting assessment of how these design protections worked in practice.
The outcome program encompasses the overall three-month disability question as well as early neurological changes, symptom remission, subgroup outcomes, and safety events. This breadth can offer a wider view of recovery, but it also creates interpretive demands: endpoints may differ in population and timing, and repeated testing increases the need for a clear multiplicity strategy. The registry does not report such a strategy or a complete set of comparative estimates across the secondary outcomes.
For biostatistical interpretation, the central lesson is the distinction between descriptive summaries and an inferential treatment comparison. Medians and counts describe what was recorded in each group. They do not, on their own, provide a model-based estimate of the difference or its precision. A completed trial record is most useful when it reports the planned analysis, the analysis population, missing-data handling, and uncertainty alongside the descriptive results. Those details are not reported for the primary analysis in this record.
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.
A central limitation is that the primary endpoint has descriptive arm summaries but no reported comparative result from the specified generalized ordinal logistic regression. This prevents assessment of the primary effect estimate, its uncertainty, or whether the analysis supported a distributional difference. The mRS summaries do not show the counts in every category, so the extent and pattern of overlap between groups cannot be evaluated from the medians and inter-quartile ranges alone.
The participant counts vary by outcome. The registry lists enrollment of 1500, while the primary summaries and secondary comparisons use different denominators. The reasons for these differences, the number and pattern of missing assessments, and the analysis populations are not reported in the registry record. It is therefore difficult to determine whether outcome availability differed systematically between assigned groups or whether the posted analyses include all randomized participants.
The study includes broad acute-stroke diagnoses and also lists subgroup-specific outcomes for ischemic stroke, intracerebral hemorrhage, and patients receiving reperfusion therapy. Those populations may have distinct clinical courses, but the record does not provide a complete set of subgroup estimates or information about whether subgroup comparisons were prespecified. The available secondary results cannot establish a common effect across these distinct groups.
Several implementation details are incomplete. The registry describes RIC cycles and pressure, but the sham pressure description is truncated. It does not report adherence, timing variation, treatment deviations, or how the active and sham procedures were distinguished while preserving masking. The record also lacks the detailed statistical plan, multiplicity approach, interim-analysis information, sample-size justification, and full safety analysis methods. These limitations restrict reproducibility and the ability to assess bias, precision, and the strength of the overall evidence.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Remote Ischemic Conditioning in Patients With Acute Stroke (RESIST) — NCT03481777.
- Registry record last updated 2025-01-27; status last verified 2025-01.
- Results first posted 2025-01-27.
- 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.
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.
18. Record Summary
The registry describes RESIST as a completed, multicenter, randomized, double-masked, parallel, sham-controlled study of remote ischemic conditioning in acute stroke. It lists an enrollment of 1500, a start date of 2018-03-15, and a primary completion date of 2023-02-03. RIC was applied with an automated device in the hyperacute prehospital phase; the primary outcome was the modified Rankin Scale at three months.
The posted primary summaries report a median mRS of 2 with an inter-quartile range from 1 to 3 for RIC, and a median of 1 with an inter-quartile range from 1 to 3 for sham. The registry names generalized ordinal logistic regression but does not report its comparative result. Two secondary binary comparisons include unadjusted estimates, intervals, and p-values; both intervals include no difference. The registry record does not report the primary analysis result, missing-data methods, analysis populations, or multiplicity handling, so the central comparative question remains unresolved by the posted inferential results.