NEXUS Criteria for Cervical Spine Imaging and Clearance

The NEXUS criteria are a set of five clinical rules designed to help emergency physicians decide whether a trauma patient needs imaging of the cervical spine. Developed in the 1990s and validated in a landmark study of more than 34,000 patients, the tool works by identifying people at such low risk of a neck injury that an X-ray or CT scan can safely be skipped. The idea is simple: if a patient clears all five checkpoints, imaging adds cost and radiation without meaningfully improving safety. In practice, though, the criteria come with real limitations, especially in certain patient groups, and the question of when to trust them and when to image anyway remains one of the more debated topics in emergency medicine.

The Five Criteria and How They Work

A patient can be “cleared” by NEXUS only if all five of the following conditions are met. If any single criterion is present, imaging is recommended.

No midline tenderness: the patient has no pain or tenderness when a clinician presses along the bony midline of the back of the neck.
No focal neurologic deficit: there is no weakness, numbness, or other neurological abnormality suggesting spinal cord or nerve root involvement.
Normal alertness: the patient is fully awake and oriented, not confused, drowsy, or unresponsive.
No intoxication: alcohol or drug impairment is absent, since intoxication can mask pain and make the physical exam unreliable.
No painful distracting injury: there is no other significant painful injury (such as a long-bone fracture or a large burn) that might prevent the patient from noticing neck pain.

The logic behind all five items is the same: they address reasons a patient might have a cervical spine fracture or ligament injury yet not feel or report neck pain. If none of those masking factors apply and the patient simply does not have midline tenderness, the probability of a meaningful injury is extremely low.

Performance in the Original Validation

The study that put NEXUS on the map enrolled 34,069 blunt trauma patients across 21 emergency departments in the United States. Among them, 818 had a confirmed cervical spine injury. The decision instrument correctly flagged all but eight of those injuries, yielding a sensitivity of about 99 percent and a negative predictive value of roughly 99.8 percent. When the researchers narrowed the definition to injuries that were clinically significant, only two patients who would have been cleared by NEXUS actually had such injuries, and just one of those two needed surgery. If clinicians had followed the tool strictly, imaging could have been avoided in about 12.6 percent of all evaluated patients.¹

Those numbers sound reassuring, and they are, but two details deserve attention. First, the specificity was only about 13 percent. That means a large majority of uninjured patients still ended up being flagged for imaging. NEXUS is designed to catch nearly every injury, not to efficiently sort out who truly needs a scan. Second, a sensitivity of 99 percent still means a small number of injuries slip through. For a tool applied to millions of trauma encounters each year, even a one-in-a-hundred miss rate translates to real patients with real fractures going undetected.

How NEXUS Compares to the Canadian C-Spine Rule

The main rival to NEXUS is the Canadian C-Spine Rule, or CCR, which uses a more structured, stepwise algorithm that considers things like age, mechanism of injury, and the ability to rotate the neck. A head-to-head comparison published in the New England Journal of Medicine found the CCR was more sensitive (about 99.4 percent versus roughly 91 percent for NEXUS) and more specific (around 45 percent versus 37 percent). The CCR missed one patient with a significant injury in that study; NEXUS missed 16.²

A later meta-analysis reinforced the pattern. Pooled across multiple studies, the sensitivity of NEXUS for clinically important cervical spine injuries was about 90 percent, with specificity near 40 percent.³ Those pooled numbers are noticeably lower than the original validation’s 99 percent sensitivity, which suggests that real-world performance across diverse settings and populations does not always match the controlled conditions of the landmark trial.

Despite the CCR’s statistical edge, NEXUS remains widely used, partly because it is simpler. The five criteria can be assessed in under a minute without requiring the patient to actively move their neck, which matters for patients who are boarded and collared, heavily medicated, or in too much pain to cooperate with range-of-motion testing. The CCR’s greater accuracy comes at the cost of more steps and more clinical judgment calls.

The Problem With Older Adults

Elderly patients present one of the clearest challenges for NEXUS. Older adults are more prone to cervical spine fractures from relatively low-energy mechanisms like ground-level falls, and their injuries can be subtle on exam. Several studies have found that NEXUS is less reliable in this group.

One study of blunt trauma patients over 65 found that NEXUS had a sensitivity of only about 95 percent for cervical spine fractures, with 21 fracture patients incorrectly classified as low risk. The authors concluded that the criteria are insufficient to safely exclude fractures in older trauma patients.⁴ A separate analysis of severe blunt trauma patients found even lower sensitivity in the elderly group, around 66 percent, with a higher rate of missed injuries than in younger patients.⁵

Not every study reaches the same conclusion. A smaller prospective study that applied modified NEXUS criteria to low-risk elderly fall patients, using signs of head or neck trauma as the only distracting injury and baseline mental status as the threshold for alertness, achieved 100 percent sensitivity, though the confidence intervals were wide given the small number of injured patients.⁶ One single-center study of elderly patients also reported 100 percent sensitivity and specificity, though it too involved a limited number of fracture cases.⁷

The overall picture is that in older adults, particularly those with higher-energy trauma mechanisms or multiple comorbidities, NEXUS should be applied cautiously. Many trauma centers have a low threshold for imaging elderly patients regardless of what the criteria suggest, and that conservatism appears justified by the evidence.

Children and NEXUS

Pediatric patients pose a different kind of challenge. Cervical spine injuries are uncommon in children, but when they occur, the consequences can be devastating. The NEXUS criteria were developed and validated in adults, and their application to children has been studied only in a handful of small trials.

A Cochrane review of triage tools for detecting cervical spine injury in pediatric trauma patients found that the sensitivity of NEXUS across individual studies ranged wildly, from as low as 0.57 in one study to 1.00 in another. The confidence intervals were broad, and there were four false-negative results across the studies reviewed. The authors concluded that the NEXUS criteria are “at best a guide to clinical assessment” in children and that current evidence does not support strict or protocolized adoption of the tool in pediatric trauma care.⁸

Part of the difficulty is practical. Young children may not be able to articulate pain clearly, cooperate with a neurological exam, or hold still long enough for a reliable assessment. The “distracting injury” criterion, already one of the most subjective items in the adult version, becomes even harder to interpret in a screaming toddler with a broken arm. Most pediatric trauma guidelines recommend a much lower threshold for imaging than NEXUS alone would suggest.

The “Distracting Injury” Problem

Of the five NEXUS criteria, the concept of a “painful distracting injury” has drawn the most criticism for vagueness. The original study did not provide a strict definition, leaving it to individual clinicians to decide whether a given injury was painful enough to distract a patient from noticing neck pain. That ambiguity introduces real variability in how the tool gets applied.

A study examining agreement between emergency medicine residents and faculty physicians on each NEXUS component found that the distracting-injury criterion had the lowest inter-rater reliability, with a kappa of just 0.13, which falls in the “slight agreement” range. By comparison, posterior midline tenderness and intoxication both had kappa values around 0.70, indicating substantially better consistency. Altered mental status and focal neurologic deficit fell in between, with kappa values around 0.21 to 0.22.⁹

This matters because a clinical decision rule is only as good as the consistency with which it is applied. If two physicians look at the same patient and disagree about whether a distracting injury is present, one might clear the cervical spine and the other might order a CT. That inconsistency partly explains why pooled sensitivity numbers across studies are lower than the original validation: the tool’s performance depends heavily on who is using it and how they interpret its loosest criterion.

Reducing Unnecessary Imaging and Radiation

The core promise of NEXUS is fewer unnecessary scans. In the original validation, strict application would have eliminated imaging for about one in eight patients. A later prospective study at a single center found that consistent use of the criteria could meaningfully reduce the number of screening cervical spine CT scans performed in the emergency department, avoiding both unnecessary radiation exposure and cost.¹⁰ Both NEXUS and the Canadian C-Spine Rule share this goal of cutting down on imaging in patients who do not need it.¹¹

In a busy emergency department, even a modest reduction in CT volume adds up. Each cervical spine CT delivers a meaningful dose of ionizing radiation to the neck and thyroid, and the financial costs per scan are not trivial. For low-risk patients who meet all five NEXUS criteria, the scan adds essentially no clinical value while exposing them to both radiation and the downstream consequences of incidental findings, which can trigger additional workups and patient anxiety.

Nurses and Prehospital Providers Using the Criteria

NEXUS was designed for emergency physicians, but there has been growing interest in whether nurses and paramedics can safely apply it to speed up patient flow. If a triage nurse can identify low-risk patients and remove a cervical collar before a physician evaluates them, patients spend less time immobilized and emergency departments move more efficiently.

A prospective trial in an emergency department found that triage nurses who applied clinical clearance criteria (including the NEXUS components) to low-risk patients were able to safely remove collars, reducing immobilization time by an average of about 23 minutes with no significant injuries missed.¹² An integrative review of the broader literature found that inter-rater reliability between nurses and physicians for cervical spine clearance was generally high, and that nurses felt confident applying the decision rules.¹³

Reducing collar time is not just about convenience. Prolonged cervical collar use can cause pressure injuries to the skin, particularly in patients who are immobilized for hours while awaiting clearance in a busy trauma center. Patients who fail NEXUS or the Canadian C-Spine Rule because of altered consciousness or intoxication can end up in a collar for extended periods, and that creates a separate set of risks. The prehospital setting introduces additional uncertainty. A matched-pairs analysis of field immobilization decisions found that in fewer than 7 percent of cases were NEXUS or CCR criteria fully documented by EMS providers, and appropriate immobilization decisions were made only about half the time in both physician-staffed and tele-EMS groups.¹⁴ A scoping review focused on low-resource settings found no studies specifically evaluating these tools in such environments, making it difficult to determine how well they translate outside well-equipped emergency departments.¹⁵

NEXUS Beyond the Cervical Spine

The researchers who developed NEXUS for the cervical spine went on to create analogous decision instruments for chest imaging after blunt trauma. The idea is identical: identify low-risk patients who can skip a scan without compromising safety.

The NEXUS Chest instrument for plain radiography was validated in a large multicenter study and achieved a sensitivity of about 99 percent for thoracic injuries requiring clinical intervention, with a negative predictive value above 98 percent. Most of the injuries it missed were clinically minor, and only one patient with a missed finding required a procedure (chest tube placement).¹⁶ A companion tool for selective chest CT was also developed, with the version focused on major injuries reaching a sensitivity of about 99 percent and a negative predictive value near 100 percent in the validation cohort.¹⁷

A modeling study estimated that implementing the NEXUS Chest CT rule in a hypothetical group of 1,000 blunt chest trauma patients would result in about 161 fewer CT scans, a substantial reduction in radiation exposure, and cost savings exceeding $136,000, with only a fractional increase in missed injuries.¹⁸ These chest-focused instruments have not been adopted as widely as the cervical spine criteria, partly because chest CT ordering patterns are more variable across institutions and the consequences of a missed chest injury, while serious, are often caught clinically in ways that a missed cervical fracture may not be.

Why Real-World Adherence Falls Short

Clinical decision rules work best when they are applied consistently, and that is where NEXUS runs into friction. Emergency departments are high-pressure environments where clinicians make dozens of rapid-fire decisions per shift. Even a rule as straightforward as five yes-or-no checkpoints gets shortcut, misremembered, or overridden by gut instinct. The inter-rater reliability data showing near-random agreement on “distracting injury” is one piece of the puzzle. The prehospital documentation findings cited earlier, where criteria were fully documented fewer than 7 percent of the time, are another.

Defensive medicine also plays a role. In medicolegal environments where missing a cervical fracture carries enormous liability, some clinicians image every trauma patient regardless of what any decision rule says. The cost of a CT scan is tangible but small compared to a malpractice judgment, and that calculus quietly undermines the rule’s practical effect on imaging rates. Institutional culture matters too: trauma centers that have embedded NEXUS or the CCR into their electronic health record workflows, with prompts and required documentation, tend to see higher adherence than those relying on clinician memory alone.

For patients, the practical takeaway is that your care in the emergency department after a car accident, fall, or sports collision may or may not involve a cervical spine scan depending on which rule your hospital uses, how strictly the staff applies it, and how risk-averse your specific physician is. If you are alert, sober, pain-free in the neck, neurologically intact, and have no other major painful injury, the evidence says imaging adds very little. Whether you actually skip the scan depends on much more than the evidence.

Intoxication and Altered Mental Status as Complicating Factors

Two of the five NEXUS criteria, intoxication and altered mental status, effectively disqualify a patient from clearance whenever the physical exam cannot be trusted. In busy urban emergency departments, where alcohol and drug intoxication are common among trauma patients, this means a large share of the population automatically fails NEXUS and proceeds to imaging. A study of trauma patients who were underimmobilized by prehospital providers found that all four patients with potential cervical spine injuries who were inadequately managed had intoxication or altered mentation noted in their charts.¹⁹

The challenge extends beyond the initial assessment. An intoxicated patient who arrives in a cervical collar cannot be cleared until they are sober enough for a reliable exam. If that takes several hours, the collar stays on, and the risks of prolonged immobilization, including skin breakdown and patient discomfort, accumulate. Some institutions have adopted protocols that use CT imaging to clear the bony spine in obtunded or intoxicated patients, bypassing the clinical exam entirely, but this trades one set of costs (prolonged immobilization) for another (radiation exposure and scan expense). There is no clean solution for patients who cannot participate in their own exam, and this remains one of the most practically challenging scenarios in trauma care.