Acute back pain is a common presenting complaint in the emergency department that leads to a great deal of resource utilization. The differential diagnosis is long and most cases are caused by benign pathology that will resolve on its own. Imaging is over-used and rarely helps. This article presents an algorithmic approach using red flags in the history and physical examination that will help physicians better identify the small of patients with serious conditions that, if untreated, will result in significant neurological damage.
Key points
- •
Of emergency department patients with acute nontraumatic back pain, physicians must understand how to diagnose the minority with serious causes while simultaneously treating the large majority with conservative measures.
- •
Distinguishing patients with simple back pain versus those with serious causes is based on careful history, physical examination, and in some cases inflammatory markers.
- •
In patients with acute back pain and new abnormalities in the neurologic examination of the lower extremities, rapid MRI to diagnose the specific lesion is critical to improving outcomes in this group of patients.
Introduction
Scope of the Problem
Back pain is common and costly, with a lifetime prevalence of 80% to 90% and rapidly increasing health-related expenditures. Adults with acute nontraumatic back pain account for 2% to 3% of emergency department (ED) visits. Although most have benign, self-limited causes, around 5% of patients have serious pathology that, if not rapidly diagnosed and treated, can result in poor outcomes because of neurologic damage. The role of the emergency physician is to identify this subset from among a large majority of patients who often require no more than a history and physical examination. Overall, the quality of evidence on this subject as it specifically relates to ED patients is weak and recommendations in the article are mostly based on guidelines, expert opinion, and clinical experience. This article reviews the most recent literature and guideline revisions in the ED evaluation and management of atraumatic back pain, and is an updated and expanded version of a 2015 review published in the Annals of Emergency Medicine .
Differential Diagnosis
Patients with acute, nontraumatic low back pain are broadly divided into three categories: (1) benign, self-limited musculoskeletal causes; (2) spinal pathologies that can cause severe neurologic disability because of spinal cord or cauda equina damage; and (3) other abdominal or retroperitoneal processes that can present with back pain. We refer to these groups as simple, serious, and nonspine causes of back pain, respectively.
Simple musculoskeletal causes include degenerative spine disease, muscular or ligamentous injury, and most acute disk herniations. Sciatica, the presenting symptom of lumbosacral radiculopathy, is characterized by pain in the back radiating to the leg. This specific entity carries an estimated lifetime prevalence of 3% to 5% in adults. These patients may have severe pain but have normal neurologic examinations, except for some patients with a monoradiculopathy. Making a specific anatomic diagnosis (eg, ligamentous strain vs disk herniation) is neither helpful nor necessary because the initial management is identical and the outcomes are almost always excellent.
Although self-limited musculoskeletal causes of back pain cause the most presentations for this complaint in the primary care setting, one must consider more serious potential spinal pathologies in patients presenting to the ED with back pain. Among the serious causes of back pain, the most common include metastatic epidural tumor, spinal epidural abscess (SEA), epidural hematoma, and central disk herniation ( Box 1 ).
Benign Causes
Muscular and ligamentous strains and sprains
Isolated sciatica (posterolateral disk herniation)
Spinal stenosis
Serious Causes
Cancer related
Epidural metastatic disease
Intradural metastatic disease
Intramedullary tumor
Infection related
Spinal epidural abscess
Vertebral osteomyelitis
Infectious diskitis
Spinal epidural hematoma
Giant (central) disk herniation causing cauda equina syndrome
Nonspine-related causes a
Aortic disease: dissection, aneurysm, ulceration, and aortitis
Genitourinary disease: ureteral colic, renal infarction and tumor, prostatitis
Gastrointestinal causes: pancreatitis and pancreatic cancer, penetrating peptic ulcer, cholecystitis, and cholangitis
Retroperitoneal hemorrhage
Systemic infections including endocarditis, psoas abscess, and other localized abscess
a These lists include the more common or more dangerous causes in each category; they are not meant to be encyclopedic.
It is crucial to remember that although new neurologic physical findings strongly suggest serious disease, the converse is not true. Patients with any of the common serious causes can present with normal neurologic examinations. Not surprisingly, these patients are more likely to be misdiagnosed. Failure to consider a serious diagnosis is the most common cause of misdiagnosis of patients with back pain with serious causes.
Finally, in the diagnostic work-up of back pain, one must consider the life-threatening nonspinal etiologies (eg, aortic aneurysm, cholangitis, retroperitoneal hematoma; see Box 1 ). This article does not focus on the subset of patients with nonspinal etiologies of back pain; however, it is incumbent on the emergency physician to keep a broad differential and appropriately tailor diagnostic strategy in patients without a convincing musculoskeletal cause of back pain. For example, it is well known that a significant minority of patients with acutely symptomatic abdominal aortic aneurysms present with isolated back pain and that ED ultrasound is a fast and sensitive diagnostic test. Each of the conditions in Box 1 needs to be considered.
Introduction
Scope of the Problem
Back pain is common and costly, with a lifetime prevalence of 80% to 90% and rapidly increasing health-related expenditures. Adults with acute nontraumatic back pain account for 2% to 3% of emergency department (ED) visits. Although most have benign, self-limited causes, around 5% of patients have serious pathology that, if not rapidly diagnosed and treated, can result in poor outcomes because of neurologic damage. The role of the emergency physician is to identify this subset from among a large majority of patients who often require no more than a history and physical examination. Overall, the quality of evidence on this subject as it specifically relates to ED patients is weak and recommendations in the article are mostly based on guidelines, expert opinion, and clinical experience. This article reviews the most recent literature and guideline revisions in the ED evaluation and management of atraumatic back pain, and is an updated and expanded version of a 2015 review published in the Annals of Emergency Medicine .
Differential Diagnosis
Patients with acute, nontraumatic low back pain are broadly divided into three categories: (1) benign, self-limited musculoskeletal causes; (2) spinal pathologies that can cause severe neurologic disability because of spinal cord or cauda equina damage; and (3) other abdominal or retroperitoneal processes that can present with back pain. We refer to these groups as simple, serious, and nonspine causes of back pain, respectively.
Simple musculoskeletal causes include degenerative spine disease, muscular or ligamentous injury, and most acute disk herniations. Sciatica, the presenting symptom of lumbosacral radiculopathy, is characterized by pain in the back radiating to the leg. This specific entity carries an estimated lifetime prevalence of 3% to 5% in adults. These patients may have severe pain but have normal neurologic examinations, except for some patients with a monoradiculopathy. Making a specific anatomic diagnosis (eg, ligamentous strain vs disk herniation) is neither helpful nor necessary because the initial management is identical and the outcomes are almost always excellent.
Although self-limited musculoskeletal causes of back pain cause the most presentations for this complaint in the primary care setting, one must consider more serious potential spinal pathologies in patients presenting to the ED with back pain. Among the serious causes of back pain, the most common include metastatic epidural tumor, spinal epidural abscess (SEA), epidural hematoma, and central disk herniation ( Box 1 ).
Benign Causes
Muscular and ligamentous strains and sprains
Isolated sciatica (posterolateral disk herniation)
Spinal stenosis
Serious Causes
Cancer related
Epidural metastatic disease
Intradural metastatic disease
Intramedullary tumor
Infection related
Spinal epidural abscess
Vertebral osteomyelitis
Infectious diskitis
Spinal epidural hematoma
Giant (central) disk herniation causing cauda equina syndrome
Nonspine-related causes a
Aortic disease: dissection, aneurysm, ulceration, and aortitis
Genitourinary disease: ureteral colic, renal infarction and tumor, prostatitis
Gastrointestinal causes: pancreatitis and pancreatic cancer, penetrating peptic ulcer, cholecystitis, and cholangitis
Retroperitoneal hemorrhage
Systemic infections including endocarditis, psoas abscess, and other localized abscess
a These lists include the more common or more dangerous causes in each category; they are not meant to be encyclopedic.
It is crucial to remember that although new neurologic physical findings strongly suggest serious disease, the converse is not true. Patients with any of the common serious causes can present with normal neurologic examinations. Not surprisingly, these patients are more likely to be misdiagnosed. Failure to consider a serious diagnosis is the most common cause of misdiagnosis of patients with back pain with serious causes.
Finally, in the diagnostic work-up of back pain, one must consider the life-threatening nonspinal etiologies (eg, aortic aneurysm, cholangitis, retroperitoneal hematoma; see Box 1 ). This article does not focus on the subset of patients with nonspinal etiologies of back pain; however, it is incumbent on the emergency physician to keep a broad differential and appropriately tailor diagnostic strategy in patients without a convincing musculoskeletal cause of back pain. For example, it is well known that a significant minority of patients with acutely symptomatic abdominal aortic aneurysms present with isolated back pain and that ED ultrasound is a fast and sensitive diagnostic test. Each of the conditions in Box 1 needs to be considered.
Risk stratification
In evaluating a patient with back pain, the major decision point often comes down to whether a patient requires advanced imaging as part of a diagnostic work-up for serious causes of back pain. This decision should not be made lightly: routine imaging has not been shown to change outcomes in patients with back pain. Additionally, the choice to pursue advanced imaging may have downstream effects on ED resource use and throughput. In most cases, this decision can be made solely with a thorough history and physical examination, after which those with simple back pain may be safely discharged from the ED.
Choosing who requires advanced imaging is usually based on the assessment of red flags or risk factors for serious causes of back pain ( Fig. 1 ). The evidence for many of these red flags is weak. The best validated red flags for fracture include older age, prolonged corticosteroid use, severe trauma, and presence of an abrasion/contusion. The probability of fracture is higher when multiple red flags are present. In 2013, Downie and colleagues performed a systemic review of 14 studies to identify evidence-based red flags associated with malignancy and fracture. The authors found that the only well-supported red flag for spinal malignancy was a history of malignancy. The posttest probability of a malignant cause was much higher in the ED compared with the primary care setting, further highlighting the skewed acuity of patients presenting to the ED. Red flags classified as “minor risk” by the American College of Physicians guideline, including old age, unexplained weight loss, and failure to improve after 1 month, had posttest probability point estimates less than 3% in this review for various spinal pathology including fractures. Other frequently cited red flags, including thoracic pain, nonmechanical pain, and weight loss, were considered uninformative for serious spinal cause of back pain based on the included studies.
Well-validated red flags for serious causes of back pain include abnormal neurologic physical examination findings (ie, new ataxia and saddle anesthesia). A 2014 retrospective study of 206 patients presenting to a UK teaching hospital with the chief complaint of atraumatic back pain demonstrated that bowel/bladder dysfunction and saddle anesthesia had likelihood ratios greater than 2.45 and 2.11, respectively, for spinal cord compression. Taken together, the two had a likelihood ratio of 3.46.
It is possible that combinations of red flags and interpreting them in the context of an individual patient could improve their utility. Patients with no red flags and normal neurologic examinations are at extremely low risk for serious causes of back pain. Patients with new hard neurologic findings (including a sensory level and saddle anesthesia) are at high risk for serious causes. Patients with the presence of historic red flags but normal neurologic examinations are at intermediate risk for cord compression or cauda equina syndrome. This group is more difficult to identify, and often requires a more diligent, careful history. To highlight the importance of historical context, primary (intradural) tumors can present with symptoms suggestive of an uncomplicated lumbosacral radiculopathy, but with the subtle element of gradually increased pain over time and with recumbency, which should lead one to consider the possibility of a neoplastic cause.
Diagnostic evaluation
Biomarkers
Routine laboratory testing is not useful. For example, elevated white blood cell counts are only found in two-thirds of patients with SEA. Neither the percentage of neutrophils nor presence of immature forms (bands) are sensitive enough to make the diagnosis of SEA. Inflammatory markers, such as erythrocyte sedimentation rate (ESR) or C-reactive protein (CRP), are highly sensitive, although nonspecific markers, for epidural abscess and to a lesser extent for cancer. Timing is variable between the two markers: CRP levels increase faster at the onset of inflammation and return to normal before ESR. From a sensitivity standpoint, inflammatory biomarkers may lend themselves to increased posttest probability for SEA. An ESR of greater than 20 mm has a sensitivity approaching 100% for epidural abscess but with poor specificity. As would be expected, when one increases the threshold definition of a positive ESR, the specificity improves but at the price of lowered sensitivity.
A single institution study of patients with suspected SEA compared clinical outcomes before and after implementation of a guideline using ESR (cutoff of 20 mm) and CRP. Diagnostic delays dropped from 84% to 10% and the proportion of patients with motor abnormalities at time of diagnosis fell from 82% to 19%. ESR performed better than CRP in this small study. Using ESR for patients with possible metastatic cancer to the spine has a lower sensitivity (78% using a cutoff of 20 mm). Because of poor specificity, ESR and CRP are not recommended in patients without red flags and because of poor sensitivity, they are not useful in patients where disk herniation or epidural hematoma are the main diagnostic considerations.
Imaging
Better quality evidence underlies the recommendations against routine imaging of patients with simple back pain, although the source is mostly from primary care populations. In actual ED practice, more than 30% of patients with nontraumatic back pain are imaged, possibly reflecting the skewed acuity. Over time, there has been a strong trend toward computed tomography (CT) or MRI. A meta-analysis of 1804 patients from six studies comparing clinical outcomes in patients who received no imaging with those with any imaging (spine plain films or MRI) found no difference in outcomes between imaged and nonimaged patients. Another study randomized 380 patients with back pain (whose physicians had ordered plain films) to plain films versus MRI. Use of MRI did not improve outcomes, but costs trended higher in the MRI group, in part because of increased numbers of procedures based on abnormal MRI findings that are often incidental; fully 52% of asymptomatic individuals with no history of back pain have disk bulges and 27% have disk protrusions. Similar data exist for lumbar disk herniations.
Neurologic dysfunction does not always follow a linear progression of “compression” caused by increasing mass effect. Patients with SEA can abruptly decompensate because of cord infarction from vascular thrombosis. Patients with metastatic spine disease can also deteriorate abruptly because of vertebral collapse with acute compression or acute vascular ischemia caused by compression on a feeding spinal artery. Therefore, even in neurologically intact patients, there is urgency for the MRI if SEA is the target diagnosis and also if cancer is the major concern. The key point that bears emphasis is that neurologic dysfunction can occur abruptly in neurologically normal patients with abscess or tumor.
In patients with red flags, plain radiographs should not be used to distinguish simple from serious back pain because negative plain film are insufficient to exclude serious pathology and positive studies require follow-up MRI.
Management of simple back pain
Application of heat is weakly recommended. Rapid resumption of ordinary activity leads to faster, better outcomes than bed rest. If the patient gets relief from bed rest, very short duration (2 days) results in faster recovery than longer courses. Acutely, exercise is not recommended.
Guidelines and expert opinion recommend nonnarcotic analgesics. It is important to acknowledge that despite numerous studies and recommendations, there are few data about what works for ED patients with acute back pain in the first days. It is not known if overall, ED patients have more severe pain than the primary care patients for which these guidelines were developed. In practice, emergency physicians commonly (61% in a large national sample) use opioids. Recently, Friedman and colleagues compared functional outcomes and pain control among patients randomized to 10-day treatment of (1) naproxen and placebo, (2) naproxen and cyclobenzaprine, or (3) naproxen and oxycodone/acetaminophen. The authors found no improvement in functional outcome with the addition of either a muscle relaxant or opiate analgesic in the short-term management of back pain (1-week follow-up). Oral steroids do not help unselected patients with acute back pain, but the subset of patients with an acute radiculopathy may benefit.
One of the most important aspects of treating the patient with simple musculoskeletal back pain is adequate counseling. Our current practice is to recommend follow-up with a primary care physician within 1 to 2 weeks and give careful instructions about symptoms for which to return sooner (red flags).
Patient satisfaction seems to be more related to the perception that a careful history and physical examination have been done and that the provider has clearly explained the diagnosis and care plan rather than to being imaged.
Related posts:
Full access? Get Clinical Tree
