Restraining all vented patients is unhelpful and mean
It's time to take a closer look at this increasingly routine practice
“MR. JOHNSON! IF YOU KEEP DOING THAT, WE’RE GOING TO HAVE TO TIE YOU DOWN!”
Restraining the wrists of mechanically ventilated patients with cloth wrist cuffs strapped to bedrails has become a standard practice in many, if not most ICUs in the U.S. This has been defended as necessary for patients (for their safety), nurses (overstressed by attending to fidgety patients), or the hospital (reducing liability claims).
None of these purported benefits has ever been demonstrated to be true. Until 2026, the use of physical restraints in critically ill patients had not been examined in any randomized trial.
In small observational studies, most patients who self-extubated were restrained. (This confounded data point doesn’t prove restraints don’t prevent self-extubation, because caregivers may have been more likely to restrain agitated patients they thought might self-extubate, who then did. It doesn’t tell us if or how often restraints prevented self-extubation in the many others who did not succeed at the task.)
Restraints have also been associated with higher sedative use, longer lengths of stay, and delirium or disorientation—all of which likewise have obvious potential for confounding with the need for restraints. A 2021 meta-analysis of low-quality studies suggested an association between restraint use and PTSD.
A mere six decades after mechanical ventilation and physical restraints entered wide use in developed nations’ ICUs, a randomized trial was performed to test whether the practice provided any benefit.
The R2D2 Trial
Among 405 mechanically ventilated adults in 10 French ICUs, patients were randomized to a restrictive strategy (wrist restraints avoided unless severe agitation occurred, defined as RASS ≥3) or a liberal strategy (in which wrist restraints were applied by default and reassessed daily).
In the “low-use” (unrestrained) group, wrist-straps were used on at least one day in 36% of patients (i.e., 64% of vented patients went completely restraint-free throughout their ICU stay). Most of these were from caregiver noncompliance with the protocol — only 3% of patients were recorded as requiring restraints for severe agitation.
There was no difference in the primary endpoint of “days alive without coma or delirium at 14 days”. This is a muddy endpoint to parse in this unblinded, dynamic trial, due to the complex interactions between restraints, agitation, delirium, and sedation levels.
What everyone wants to know instead is, did the unrestrained patients self-extubate more often? They did not: 18 patients self-extubated in the unrestrained group (9.2%) vs 17 in the restrained (8.5%).
Nor did patients remove devices like I.V.s or urinary catheters more often (two devices in the unrestrained group vs. one in the restrained group). They did not have more agitation, either.
The second most salient question: did the restrained patients receive more sedation (to “snow” them pre-emptively or reactively, to reduce the perceived risk for self-extubation or other self-harm)? They did not. (See eFigures 3-7 in the supplementary appendix for the closely comparable dosing of multiple sedatives.)
The unrestrained patients got up and moved ever so slightly more in the first 14 days, but at two weeks, mobilization was nearly identical between groups.
Mortality and 90-day functional, cognitive, and psychological outcomes were also similar. About 4% (absolute) more patients in the restrained group developed PTSD at 90 days, but this was not statistically significant.
Pressure ulcers were reported in 30 unrestrained patients (15.3%) and 34 restrained patients (17.0%), nonsignificant.
But Let’s Also Show Some Restraint
It would be premature to draw broad conclusions about the non-utility (or disutility) of restraints outside the scope of this trial performed exclusively in French ICUs.
Many in the U.S. will raise a skeptical eyebrow at the report that fewer than 1% (only three patients out of ~400) removed their I.V.s, nasogastric tubes, urinary catheters, or other devices. Anecdotally speaking, critically ill patients in many U.S. ICUs seem to be significantly more unruly than this.
The caregivers participating in the trial were doubtless influenced by the knowledge that their behavior was being observed and the outcomes recorded. This may have resulted in extra vigilance toward reduction of adverse events in the intervention (low-restraint) arm.
The accepted way to account for this and other potential secular (over time) changes in clinician behavior is to randomly start the intervention in each of the ICUs over several months or a year, in a cluster-randomized design. That wasn’t done here.
Should Restraints Be So Routine in the ICU?
Per the CMS regulatory code, restraints may only be “imposed to ensure the immediate physical safety of the patient, a staff member, or others,” and never “as a means of coercion, discipline, convenience, or retaliation by staff. [They] must be discontinued at the earliest possible time.”
One hopes that restraints are virtually always imposed with some motivation to ensure ventilated patients’ safety. But there exists an uncomfortably large overlap between a concern for patient safety and a desire to maximize care teams’ convenience.
The R2D2 trial raises real doubts about whether restraints are required to maintain the safety of most mechanically ventilated patients.
Hundreds of thousands of people are physically restrained during their ICU stays every year around the world, without any good evidence to support the practice. A meta-analysis (albeit of low-quality studies) suggested that restraint use may be associated with the development of PTSD.
Now that the benefits for restraints have been called into question, conducting a properly powered cluster-randomized trial (in thousands of patients) would seem to be the bare minimum we should do to ethically justify this common but often disturbing and conflicted act of care.
References
Restrictive vs Liberal Physical Restraint Strategies in Critically Ill Patients. The Journal of the American Medical Association. 2026. Sonneville R, Couffignal C, Sigaud F, et al.
Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU. Critical Care Medicine. 2018. Devlin JW, Skrobik Y, Gélinas C, et al.
Physical Restraints and Post-Traumatic Stress Disorder in Survivors of Critical Illness. A Systematic Review and Meta-Analysis. Annals of the American Thoracic Society. 2021. Franks ZM, Alcock JA, Lam T, et al.
Unplanned Endotracheal Extubations in the Intensive Care Unit: Systematic Review, Critical Appraisal, and Evidence-Based Recommendations. Anesthesia and Analgesia. 2012. da Silva PS, Fonseca MC.
Non-Pharmacological Interventions for Minimizing Physical Restraints Use in Intensive Care Units: An Umbrella Review. Frontiers in Medicine. 2022. Cui N, Yan X, Zhang Y, et al.
Revisiting Unplanned Endotracheal Extubation and Disease Severity in Intensive Care Units. PloS One. 2015. Chuang ML, Lee CY, Chen YF, Huang SF, Lin IF.
Critical care nurses’ experiences of physical restraint in intensive care units: A qualitative systematic review and meta‐synthesis. Journal of Clinical Nursing. 2023. Lao Y, Chen X, Zhang Y, et al.SR
As an ICU survivor this hits home for me. I was thankfully never restrained, and was fully cognizant of what was going on around me. They did a good job keeping sedation down for me but this indeed can cause PTSD and PICS.
I’m glad this is starting to get real attention.
Small quibble: I hope that no one is tying restraints to bedrails, which is well known to cause injuries if the rail is dropped suddenly. All modern beds and stretchers have tie points that should be used if restraints are needed.