Does a DNR order mean no CPR? No. A do-not-resuscitate order specifically prohibits cardiopulmonary resuscitation—chest compressions, rescue breaths, defibrillation, and intubation—when a patient is in cardiac arrest. It does not prohibit oxygen, pain medication, antibiotics, surgery, dialysis, or any other treatment. The order is narrow by design: it governs one intervention at one moment, not the overall plan of care.
That narrowness is the source of most confusion. Families often hear "DNR" and assume it means "do nothing." Clinicians sometimes document it without confirming the patient understands what remains available. And because a DNR order is written to activate at the moment of arrest, it interacts with everything that happens before that moment—which is where the real clinical complexity lives.
What a DNR order actually authorizes
A DNR order is a conditional instruction: if the heart stops or breathing ceases, do not begin resuscitation. It is written in advance, signed by a physician, and entered into the medical record. The condition it addresses—clinical death—is binary. Either the patient is in arrest or they are not.
The mechanism matters because CPR is not a single act. It is a protocol with discrete components: chest compressions to generate forward blood flow, positive-pressure ventilation, defibrillation for shockable rhythms, vascular access, and often endotracheal intubation. A DNR order withholds the entire protocol as a unit. It does not authorize withholding oxygen for a patient who is breathing, fluids for someone who is hypotensive, or antibiotics for pneumonia. A patient with a DNR order who develops sepsis receives full antibiotic therapy and ICU-level supportive care unless a separate order limits those treatments.
This is why the phrase "comfort-focused care" and "DNR" are not synonyms. Comfort-focused care is a broader treatment philosophy. DNR is a single procedural instruction. Confusing the two leads families to fear that signing a DNR order will cause their relative to be neglected—a fear that is clinically unfounded but emotionally rational, because the conversation often happens in a crisis.
One professional caveat deserves emphasis: a DNR order does not automatically mean "do not intubate." Intubation for respiratory failure in a patient who is not in arrest is a separate decision. Many institutions now use combined orders—DNR/DNI—but the two are distinct, and conflating them causes errors in both directions.
Why CPR success rates reframe the decision
The cultural image of CPR—a dramatic save, a return to normal life—diverges sharply from measured outcomes. In-hospital cardiac arrest survival to discharge in the United States has improved over recent decades and generally falls in the range of roughly one in four to one in five patients, depending on the registry and case mix. Out-of-hospital cardiac arrest survival is lower, and survival with good neurological function is lower still.
These figures are population averages, and they conceal enormous variation. A 40-year-old with a witnessed ventricular fibrillation arrest in a monitored hospital bed has a fundamentally different prognosis than an 85-year-old with metastatic cancer whose arrest is unwitnessed and unwitnessed by staff for several minutes. The DNR conversation is, at its core, an attempt to estimate where a specific patient sits relative to those averages—and to decide whether the likely outcome aligns with what that patient values.
The mechanism behind poor outcomes is ischemic time. Brain tissue begins to sustain irreversible injury within roughly four to six minutes without circulation, and defibrillation success declines with each minute of delay. CPR buys time, but it does not reverse the underlying cause. For patients whose arrest reflects end-stage disease, the underlying cause is not reversible, and the resuscitation succeeds only in prolonging the dying process.
DNR versus other directives and code statuses
The terminology is where errors multiply. A DNR order is not the same as a living will, a POLST form, or a health care proxy designation. Each carries different legal weight and different scope.
| Document or order | What it governs | Who completes it | Activation trigger |
|---|---|---|---|
| DNR order | Withholds CPR at arrest | Physician, per patient/surrogate consent | Cardiac or respiratory arrest |
| DNI order | Withholds intubation | Physician, per patient/surrogate consent | Respiratory failure |
| POLST / MOLST | Portable medical orders across settings | Clinician with patient | Any setting, immediately actionable |
| Living will | General values and preferences | Patient, often notarized | When patient lacks capacity |
| Health care proxy | Names a decision-maker | Patient | When patient lacks capacity |
| Code status "full code" | Full resuscitation attempted | Physician order | Cardiac or respiratory arrest |
The practical distinction: a living will states values but is not an actionable medical order. A POLST form is an actionable order that travels with the patient across care settings—home, ambulance, nursing facility, hospital. A DNR order is typically setting-specific unless embedded in a POLST. This is why patients who complete a hospital DNR order may still receive CPR from emergency medical services at home unless a portable order exists. The mechanism of activation differs, and so does the outcome.
The counterintuitive detail: a patient can be a full code and still have a valid living will refusing prolonged life support. The code status governs the arrest moment; the living will governs the trajectory. They can point in opposite directions, and clinicians are expected to honor both.
Edge cases that change the calculus
Perioperative DNR orders are the classic edge case. Many institutions require automatic suspension of DNR orders during surgery and anesthesia, on the reasoning that anesthesia and surgical stress can precipitate arrest from causes that are immediately reversible—anesthetic overdose, vagal response, hemorrhage from a correctable source. Professional societies have moved toward requiring individualized reconsideration rather than automatic suspension, but practice varies. A patient with a DNR order undergoing surgery should have an explicit preoperative conversation about what will happen if arrest occurs on the table.
DNR in pregnancy raises separate questions. When a pregnant patient with a DNR order arrests, the decision may involve emergent delivery to save the fetus—a procedure that can itself be life-sustaining for the pregnant patient. Institutions vary in how they reconcile the DNR order with fetal rescue obligations, and there is no single national standard.
Pediatric DNR orders are rarer and follow different consent pathways, generally requiring both parental consent and physician authorization, with heightened scrutiny. And DNR orders do not automatically travel across state lines or between facilities; a patient transferred to a new hospital may arrive as a full code unless the order is re-entered, which is a documented source of unwanted resuscitation.
Another edge case: the patient who arrests because of a readily reversible cause—opioid overdose, profound hypoglycemia, anaphylaxis. In these situations, CPR and naloxone or glucose are highly effective, and some clinicians argue a DNR order should be reconsidered in real time. This is ethically contested. The order says what it says; whether to override it for a reversible cause is a judgment that should be made in advance, in writing, not in the moment.
Misconceptions that cause harm
The most consequential misconception is that DNR means "do not treat." Studies of surrogate decision-makers consistently find that families overestimate what DNR restricts. The correction is straightforward: DNR restricts CPR only. Everything else—analgesia, oxygen, antibiotics, transfusions, surgery, dialysis—remains permissible and is often appropriate.
A second misconception: DNR means the patient has given up. In practice, DNR orders are frequently written for patients who are receiving aggressive concurrent treatment—chemotherapy, dialysis, major surgery—but who have decided that if their heart stops, they do not want the specific intervention of CPR. The order is about one procedure, not about hope.
A third: verbal DNR requests are sufficient. They are not. Without a written, physician-signed order or a valid portable form, emergency responders are generally obligated to initiate CPR. The mechanism is legal liability and protocol design: EMS cannot verify a family's verbal statement in the field. This is why POLST forms exist—to convert a verbal preference into an actionable order that travels.
A fourth: DNR orders are permanent. They are not. Patients can revoke them at any time, and clinicians should reassess them when prognosis changes or when the patient regains capacity. An order written during an acute crisis may not reflect values once the crisis resolves.
Practical actions for patients and families
The concrete steps matter more than the definitions:
Ask for the order in writing. Request a copy of the DNR order or POLST form and confirm it is in the medical record before any transfer between facilities. Clarify scope explicitly. Ask: "Does this order withhold anything besides CPR?" The answer should be no, and it should be documented. Separate DNR from DNI. If the patient does not want intubation for respiratory failure either, that requires a separate order. Confirm portability. For patients at home or in a nursing facility, request a POLST form so EMS can act on the order without initiating CPR. Revisit after any major change. New diagnosis, new prognosis, or recovery of capacity should trigger a fresh conversation. Document the conversation, not just the checkbox. A signed order without a documented discussion is a legal document with no underlying consent to verify.
Red flags that warrant immediate clarification: a DNR order written without patient or surrogate signature; a DNR order that a family believes restricts pain medication; a patient transferred between facilities with no confirmed code status; and a verbal DNR request in a patient being discharged home without a portable order.
What the evidence cannot settle
The evidence base is strong on mechanisms and on population-level outcomes. It is weak on individual prediction. No calculator reliably tells a specific patient whether CPR will succeed, and the registries that generate survival statistics are subject to selection and reporting bias. This is a genuine limit, not a hedge: the DNR decision is a values judgment informed by probabilities, not a calculation.
The other limit is communication. The order is only as good as the conversation behind it and the documentation in front of it. A DNR order that the patient does not understand, that the family misinterprets, or that does not travel with the patient is an order that will fail at exactly the moment it is supposed to work. The narrow scope of DNR—no CPR, and nothing more—is precisely why it must be paired with a broader conversation about what care the patient does want.