what can history physical examination tell us about low back...what can the history and physical...

What Can the History andPhysical Examination Tell UsAbout Low Back Pain?Richard A. Deyo, MD, MPH; James Rainville, MD; Daniel L. Kent, MD

BACK pain ranks second only to upperrespiratory illness as a symptomatic rea-son for office visits to physicians.1 About70% of adults have low back pain atsome time, but only 14% have an epi-sode that lasts more than 2 weeks. About1.5% have such episodes with featuresof sciatica.2,3 Most causes of back painrespond to symptomatic and physicalmeasures, but some are surgically re-mediable and some are systemic diseases(cancer or disseminated infection) re-

quiring specific therapy, so careful di-agnostic evaluation is important. Fea-tures of the clinical history and physicalexamination influence not only thera-peutic choices but also decisions aboutdiagnostic imaging, laboratory testing,and specialist referral.

ANATOMIC/PHYSIOLOGIC ORIGINSOF FINDINGS IN THE LOW BACK

Low back pain may arise from severalstructures in the lumbar spine, includ-ing the ligaments that interconnect ver-

tebrae, outer fibers of the annulus fi-brosus, facet joints, vertebral perios-teum, paravertebral musculature andfascia, blood vessels, and spinal nerveroots. The causes of low back pain gen¬erated through these structures include(1) musculoligamentous injuries; (2) de¬generative changes in the interverte¬bral disks and facet joints; (3) herniationof the nucleus pulposus of an interver¬tebral disk, with irritation of adjacentnerve roots; (4) spinal stenosis (narrow¬ing of the central spinal canal or the

From the Health Services Research and Develop-ment Field Program, Seattle (Wash) Veterans AffairsMedical Center (Drs Deyo and Kent); the Departmentsof Medicine (Drs Deyo and Kent) and Health Services(Dr Deyo), University of Washington, Seattle; and theDepartment of Rehabilitation Medicine, Tufts UniversitySchool of Medicine, Boston, Mass (Dr Rainville).

Reprint requests to Back Pain Outcome AssessmentTeam, JD-23, University of Washington, Seattle, WA98195 (Dr Deyo).

lateral recesses of the canal in which thenerve roots travel caudally; this usuallyresults from hypertrophie degenerativechanges in the disks, ligamentum fla-vum, and facet joints); (5) anatomicanomalies of the spine, such as scoliosisand spondylolisthesis, which are oftenasymptomatic but may cause pain whenthey are severe; (6) underlying systemicdiseases, such as primary or metastaticcancer, spinal infections, and ankylosingspondylitis; and (7) visceral diseases un¬related to the spine, including diseasesof the pelvic organs, kidneys, gas¬trointestinal tract, and aorta (diagnosisof which will not be discussed in thepresent report).

PREVALENCE OF DISEASES THATPRODUCE LOW BACK PAIN

Up to 85% of patients cannot be givena definitive diagnosis because of weakassociations among symptoms, patho¬logical changes, and imaging results.4·5We assume that many of these cases arerelated to musculoligamentous injury or

degenerative changes.Anatomic evidence of a herniated disk

is found in 20% to 30% of imaging tests(myelography, computed tomography,and magnetic resonance imaging) amongnormal persons.6,7 These herniations are

asymptomatic and result in no clinicaldisease. The proportion of all personswith low back pain who undergo surgeryfor a disk herniation is only about 2%.2

In primary care, about 4% of patientswith back pain will prove to have com¬

pression fractures, 3% have spondylolis¬thesis, and only 0.7% have spinal ma¬

lignant neoplasms (primary or meta¬static).813 Even fewer have ankylosingspondylitis (about 0.3%) or spinal infec¬tions (0.01%).8·14'16 Widespread recogni¬tion of spinal stenosis has occurred onlyin the last 15 years. It is most commonin older adults, but its prevalence is un¬known.

Since a specific cause frequently can¬not be identified, diagnostic efforts areoften disappointing. Instead of seekinga precise cause in every case of backpain, it may be most useful to answerthree basic questions9: (1) Is there a se¬rious systemic disease causing the pain?(2) Is there neurologic compromise thatmight require surgical evaluation? (3) Isthere social or psychological distress thatmay amplify or prolong pain? Thesequestions can generally be answered onthe basis of history and physical exam¬ination alone, and a minority of patientsrequire further diagnostic testing.IS THERE EVIDENCE OFSYSTEMIC DISEASE?Cancer

Malignant neoplasm (primary or met-astatic) is the most common systemicdisease affecting the spine, although itaccounts for less than 1% of episodes oflow back pain. Approximately 80% ofpatients with this diagnosis are over theage of 50 years (Table 1). A previoushistory of cancer has such high speci¬ficity (0.98) that such patients should beconsidered to have cancer until provenotherwise (SpPin [an acronym for whenSpecificity is extremely high, a Positivetest result rules in the target disor¬der]). However, only one third of pa¬tients with an underlying malignant neo¬

plasm have this history (sensitivity, 0.31).Unexplained weight loss, pain durationgreater than 1 month, and failure to im¬prove with conservative therapy are

moderately specific findings. Most pa¬tients with back pain due to cancer re¬

port that pain is unrelieved by bed rest(sensitivity >0.90, SnNout [an acronymfor when Sewsitivity of a symptom or

sign is high, a Negative repsonse rulesout the target disorder]), but the find¬ing is nonspecific.10 In a study of nearly2000 patients with back pain, no cancerwas identified in any patient under age

at University of California - San Francisco on December 23, 2009 www.jama.comDownloaded from

http://jama.ama-assn.org

50 years without a history of cancer,unexplained weight loss, or a failure ofconservative therapy (combined sensi¬tivity, 100%, SnNout).10

The physical examination is less use¬ful than the history for detecting un¬

derlying cancer,10 except in late stages.Since the breast, lung, and prostate arethe most common sources of spinal mé¬

tastases, these organs should be exam¬ined when cancer is suspected.Spinal Infections

Spinal infections usually are blood-borne from other sites, including uri¬nary tract infections, indwelling urinarycatheters, skin infections, and injectionsites for illicit intravenous drugs. One ofthese sites is identified in approximately40% of patients with spinal infections(sensitivity, 0.40).1G

In patients with spinal infections, thesensitivity of fever is disappointing,varying from 0.27 for tuberculous os¬

teomyelitis to 0.50 for pyogenic osteo¬myelitis17 and 0.83 for spinal epiduralabscess.18 Because 2% of patients in pri¬mary care with mechanical low back painhave fever (perhaps due to viral syn¬dromes), specificity for bacterial infec¬tion is approximately 0.98.10 Spine ten¬derness in response to percussion has a

sensitivity of0.86 for bacterial infection,but specificity is poor (0.60).10·19·20Compression Fractures

Although spinal compression fracturesare not "systemic" diseases, they oftenoccur in persons with generalized os¬

teoporosis. Most patients with this prob¬lem do not have a history of identifiabletrauma (sensitivity, 0.30). A person withback pain who is receiving long-termcorticosteroid therapy is considered tohave a compression fracture until provenotherwise (specificity, 0.99, SpPin). Af¬rican-American and Mexican-Americanwomen have only one fourth as manycompression fractures as white women.21As shown in Table 1, age greater than70 years is a relatively specific finding(specificity, 0.96, SpPin).Ankylosing Spondylitis and SpineRange of Motion Measures

Ankylosing spondylitis shares severalhistorical features with other inflamma¬tory arthropathies, such as rheumatoidarthritis. Calin and colleagues22 de¬scribed five screening questions forankylosing spondylitis: (1) Is there morn¬

ing stiffness? (2) Is there improvementin discomfort with exercise? (3) Was theonset of back pain before age 40 years?(4) Did the problem begin slowly? (5)Has the pain persisted for at least 3months?

Using at least four positive answers

Table 1.—Estimated Accuracy of the Medical History in the Diagnosis of Spine Diseases Causing Low BackPain

Disease toBe Detected Source, y Medical History Sensitivity Specificity*

Cancer Deyo and Diehl,'0 1988 Age 2 50 yPrevious history of cancer 0.31

Unexplained weight loss 0.15Failure to improve with a

month of therapy0.31

No relief with bed rest >0.90Duration of pain >1 mo

Age 250 y or history of canceror unexplained weight lossor failure of conservativetherapy

0.501.00

0.980.940.90

0.460.810.60

Spinal Waldvogel and Vasey,'osteomyelitis 1980

Intravenous drug abuse,urinary tract infection,or skin infection

0.40 NA

Compressionfracture

Unpublished datât Age 250 y

Age 270 yTraumaCorticosteroid use

0.84

0.06

0.61

0.960.85

0.995Hemiated

diskDeyo and Tsui-Wu,2 1987;

Spangfort,33 1972Sciatica 0.88

Spinalstenosis

Turner et al,52 1992 Pseudoclaudication

Age 250 y 0.90

NA

0.70

Ankylosingspondylitis

Gran,23 1985 4 out of 5 positive responses§

Age at onset £40 yPain not relieved supineMorning back stiffnessPain duration 23 mo

0.23

1.000.800.640.71

0.82

0.070.490.590.54

*NA Indicates not available.tFrom 833 patients with back pain at a walk-in clinic, all of whom received plain lumbar roentgenograms.^Authors' estimate.§The five screening questions were (1 ) onset of back discomfort before age 40 years? (2) did the problem begin

slowly? (3) persistence for at least 3 months? (4) morning stiffness? and (5) improved by exercise?

to define a positive "test" result, thesensitivity of these questions was 0.95and specificity 0.85,22 although other au¬thors report lower sensitivity.23·24 Whenscreening for a rare disease such as anky¬losing spondylitis, however, the predic¬tive value of a positive test is low. In anindustrial screening program, only 16 of367 persons with positive criteria provedto have ankylosing spondylitis (a pre¬dictive value of 0.04).25 "Inflammatory"symptoms (morningstiffness, night pain,reliefwith exercise) are moderately sen¬sitive but nonspecific. All patients withankylosing spondylitis in one populationsurvey reported symptom onset beforeage 40 years, making this history highlysensitive but nonspecific (SnNout, Ta¬ble l).23

Reduced spinal mobility results from"fusion" of adjacent vertebrae in thiscondition. The Schober Test, which mea¬sures distraction between two marks onthe skin during forward flexion, is a com¬

monly described method for quantify¬ing reduced flexion. Although it is mod¬erately reproducible,20·26 reduced spineflexion is not specific for inflammatoryspondylopathies, being equally commonin patients with chronic back pain or

spine tumors.27 Reduced chest expan¬sion (using a strict criterion for abnor¬mality, such as expansion <2.5 cm) ishighly specific (0.99, SpPin) but insen¬sitive in early ankylosing spondylitis(0.09),23·28 so that predictive values are

poor.Tests for sacroiliac joint tenderness

(to discriminate ankylosing spondylitisfrom mechanical spine conditions) in¬clude a hip extension test, anteroposte-rior pelvic pressure, lateral pelvic com¬

pression, and direct pressure on the sac¬roiliac joints. Unfortunately, these testsare poorly reproducible20·29 and inaccu¬rate in distinguishing ankylosingspondylitis from mechanical spine com¬

plaints.30·31 Early ankylosing spondylitisis most often suspected from roentgen-ograms obtained in the face of persis¬tent pain.

Although spine flexion is of limiteddiagnostic value, it may be useful in plan¬ning or monitoring physical therapy inpatients with low back pain of anycause.32 Range of motion in multiple di¬rections can be assessed with two incli¬nometers (used in the construction in¬dustry) with good precision.20·32 The tech¬nique is detailed elsewhere.32



IS THERE EVIDENCE OFNEUROLOGIC COMPROMISE?

The spinal cord, cauda equina, andnerve roots are vulnerable to several dis¬orders that cause back pain and sciatica.The most common of these is a herniatedintervertebral disk, but other causes in¬clude nerve root entrapment in the rootcanals by bony and ligamentous hyper¬trophy, spinal stenosis, spinal or paraspi-nal infections, and neoplasms. Irritationof neurological structures is manifestedas motor, reflex, or sensory dysfunctionin the lower extremities and (rarely) asbowel or bladder dysfunction.

The first clue to nerve root irritationis usually sciatica, a sharp or burningpain radiating down the posterior or lat¬eral aspect of the leg (usually to the footor ankle), often associated with numb¬ness or paresthesia. The pain is some¬times aggravated by coughing, sneez¬

ing, or the Valsalva maneuver. Amongpatients with low back pain alone (nosciatica or neurological symptoms), theprevalence of neurological impairmentsis so low that extensive neurological eval¬uation is usually unnecessary.

Lumbar Disk HerniationsSciatica has such a high sensitivity

(0.95) that its absence makes a clinicallyimportant lumbar disk herniation un¬

likely (SnNout).33·34 Using the accuracyof sciatica in Table 1 and a prevalence ofsurgically important disk herniations of2%, we would estimate the likelihood ofdisk herniation in a patient without sci¬atica to be one in 1000. Most patientshave a long history of recurrent backpain prior to the onset of sciatica, butwhen a frank disk herniation occurs, legpain usually overshadows the back pain.The peak incidence of herniated lumbardisks is in adults between the ages of 30and 55 years.33

A symptomatic disk herniation teth¬ers the affected nerve root, so pain re¬sults from stretching the nerve bystraight leg raising from the supine po¬sition. This is performed by cupping theheel in one hand and keeping the kneefully extended with the other. Thestraight leg is slowly raised from theexamining table until pain occurs. Ten¬sion is transmitted to the nerve rootsonce the leg is raised beyond 30°, butafter 70°, further movement of the nerveis negligible.36 A typical positive straightleg raising sign is one that reproducesthe patient's sciatica between 30° and60° of leg elevation.33·36·37

A related test is the "crossed straightleg raising sign." This occurs whenstraight leg raising is performed on thepatient's well leg and is found to elicitpain in the leg with sciatica. The preci-

Table 2.—Reproduclbility of Physical Examination Findings

Category Test

InterobserverUnit of Agreement

Measurement (Statistic) Source, yTenderness Bone tenderness Yes/no 0.40 ( ) McCombe et al,201S

Soft-tissuetenderness

Yes/no 0.24 ( ) McCombe et al,z0 1989

Muscle spasm Yes/no "Discarded—toounreliable"

Waddell et al,33 1982

SLR* Ipsilateral SLR,inclinometer

Degrees 0.78 to 0.97 (r) Hoehler and Tobis,39 1982;Hsiehetal,401983

Ipsilateral SLR,goniometer

Degrees 0.69 (r) McCombe et al,20 1989

SLR causes

leg painYes/no 0.66 ( ) McCombe et al,201989

Ipsilateral SLR<75° by visualestimation

Yes/no 0.56 ( ) Waddell et al,381982

Crossed SLR,causes pain

Yes/no 0.74 ( ) McCombe et al,201989

Neurologicexamination

Ankle dorsiflexlonweak

Yes/no 1.00 ( ) McCombe et al,20 1989

Great toeextensors weak


Ankle reflexesnormal

Yes/no 0.39-0.50 ( ) McCombe et al,201989;Schwartz et al,481990

Any sensory deficit Yes/no 0.68 ( ) McCombe et al,201989Calf wasting Yes/no 0.80 ( ) McCombe et al,201989

Inappropriatesigns

Superficialtenderness


Simulated rotationor axial loadingcauses pain


SLR withdistractioncauses pain


Inexplicablepattern,

¡ogicexamination


Overreaction Yes/no 0.29 ( ) McCombe et al,201989

*SLR indicates straight leg raising.

sion of tests for straight leg raising isshown in Table 2.20·38"40 Visual estima¬tion is reasonably accurate but a goni¬ometer or inclinometer improves inter-observer agreement.

Limited ipsilateral straight leg rais¬ing at 60° is moderately sensitive forherniated lumbar disks but nonspecific,since limitation is often observed in theabsence ofdisk herniations (Table 3).41"43Crossed straight leg raising is less sen¬sitive but highly specific.33·42"44 Thus, a

positive crossed straight leg raising testsubstantially increases the likelihood ofa disk herniation (SpPin), while a neg¬ative result is of limited value. The lowerthe angle of a positive straight leg rais¬ing test, the more specific the test be¬comes and the larger the disk protru¬sion found at surgery.45·46

Straight leg raising is most appropri¬ate for testing the lower lumbar nerveroots (L5 and SI), where the vast ma¬

jority of herniated disks occur. Irrita¬tion of higher lumbar roots is testedwith the femoral nerve stretch test (flex¬ing the knee with patient prone), butthe precision and accuracy of this testare unknown.

Assessment of Motor, Reflex, andSensory Function

Ninety-eight percent of clinically im¬portant lumbar disk herniations occurat either the L4-5 or L5-S1 interverte¬bral level,33·42·43·40 causing neurologic im¬pairments in the motor and sensory ter¬ritories of the L5 and SI nerve roots.Thus, the most common neurologic im¬pairments are weakness of the ankleand great toe dorsiflexors (L5), dimin¬ished ankle reflexes (SI), and sensoryloss in the feet (L5 and SI).33·42·43·46 In a

patient with sciatica, the neurologicalexamination can be concentrated onthese functions.

Ankle dorsiflexor strength is testedby having the supine patient dorsiflexthe ankle against the examiner's resis¬tance. Inability to maintain dorsiflexionagainst the examiner should be consid¬ered weakness, and the well side shouldbe checked for comparison. This methodshows excellent precision (Table 2) andis more reproducible than the patient'sability to heel stand.20 Ankle dorsiflexorweakness rarely occurs in isolation andis nearly always associated with weak



Table 3.—Estimated Accuracy of Physical Examination for Lumbar Disk Herniation Among Patients WithSciatica

Text Source, y Sensitivity* Specificity* Comments

Ipsllateralstraightleg raising

Kosteljanetz et al,411984; Hakellus andHindmarsh,421972

0.80 Positive test result:leg pain at <60°

Crossedstraightleg raising

Spangfort,33 1972;Hakelius andHindmarsh,42·431972

0.25 0.90 Positive test result;reproduction ofcontralateral pain

Ankledorsiflexionweakness

Spangfort,33 1972;Hakellus andHindmarsh,42 1972

0.35 HNPt usuallyat L4-5 (80%)

Great toeextensorweakness

Hakelius and Hindmarsh,421972; Kortelalnen et al,461985

0.50 0.70 HNP usuallyat L5-S1(60%)or L4-5 (30%)

Impairedanklereflex

Spangfort,33 1972;Hakelius andHindmarsh,421972

0.60 HNP usually atL5-S1; absentreflex increasesspecificity

Sensory loss Kosteljanetz et al,41 1984;Kortelalnen et al,46 1985

0.50 0.50 Area of losspoor predictorof HNP level

Patella reflex Aronson andDunsmore,50 1963

For upper lumbarHNP only

Ankle plantarflexionweakness

Hakellus andHindmarsh,42 1972

0.06

Quadricepsweakness

Hakelius andHindmarsh,42 1972

<0.01 0.99

*Sensitlvity and specificity were calculated by the authors of the present report. Values represent roundedaverages where multiple references were available. All results are from surgical case series.

tHNP indicates herniated nucleus pulposus.

toe dorsiflexion, sensory deficits, or im¬paired reflexes.47 For toe strength, thesupine patient is instructed to maximallydorsiflex the great toe ("point your bigtoe at your nose" seems to work well)and resist the examiner's effort to flexthe toe with two fingers.

Ankle reflexes are more difficult to re¬

produce, and patient positioning may beimportant. The side-lying, prone, andkneeling positions are probably best(rather than the sitting position), but weare unaware of comparative data. Thefoot is gently rocked until relaxation is ob¬tained, and the calfmuscles should be heldunder slight tension by dorsiflexing thefoot. Estimated values for the precisionof ankle reflexes range from 0.39 to0.50.20·48 Schwartz and colleagues48 foundthat a plantar tap is as good as an Achillestendon tap (estimated , 0.55). In thistechnique, the patient lies supine and theball of the foot is tapped with the reflexhammer. The plantar tap was preferredby patients and could be elicited in 91% ofpatients under age 65 years but in only71% of patients over age 65 years.

Ankle plantar flexion is an SI func¬tion, but only severe impairments canbe clinically detected, and sensitivity fordisk herniation is low (Table 3). Toe walk¬ing appears to be an unreliable methodof assessing plantar flexion strength( =0.00).20 Hamstring and hip extensorstrength have been used to evaluate SIroot injuries, but their precision and ac¬

curacy are unknown. Muscle wasting in¬dicates long-standing denervation or dis¬ease and may be detected visually. Good

precision was noted for observations ofanterior compartment and hamstringwasting in one study (Table 2).20

Sensory examination of the lower ex¬tremities can be time-consuming and ag¬gravating. Patients distinguish differ¬ences in pain intensity by pinprick more

accurately than differences in touch or

temperature, and sensory impairmentfrom nerve root compression is mostfrequent in the distal extremes of thedermatomes.40 Therefore, an efficientstrategy is to check for symmetry ofpain elicited by pinprick in the extremesof the L4, L5, and SI dermatomes (themedial aspect, dorsum, and lateral as¬

pect of the feet) (Figure).Higher lumbar nerve roots account

for only about 2% of lumbar disk her¬niations. They are suspected whennumbness or pain involves the anteriorthigh more prominently than the calf(Figure). Testing includes knee reflexes,quadriceps strength, and psoasstrength.33·47·50 Quadriceps weakness isvirtually always associated with impair¬ment in the patella reflex.47

The accuracy of neurologic findingsfor the diagnosis of a herniated disk isonly moderate (Table 3). Consideringcombinations is helpful, however, sincea finding of impaired ankle reflexes orweak foot dorsiflexion would have a sen¬

sitivity of almost 90% for patients withsurgically proven disk herniations.33 Mul¬tiple findings related to straight leg rais¬ing or neurologic examination increasethe probability that a herniated disk willbe found at surgery.51

S4L2

lL3 S2

\L4

L3 \ -S

L5

\L4

fsi;

Lower-extremity dermatomes.

Spinal StenosisThe mean age of patients at the time

ofsurgery for spinal stenosis is 55 years,with an average symptom duration of 4years.52 The characteristic history is thatof neurogenic claudication: pain in thelegs and occasionally neurologic deficitsthat occur after walking. In contrast toarterial ischemie claudication, neuro¬

genic claudication is more likely to occuron standing alone (without ambulation),may increase with cough or sneeze, andis associated with normal arterialpulses.53 The sensitivity of neurogenicclaudication is modest (about 0.60),52 butit is probably quite specific.

Few data are available concerning theaccuracy of physical examination, be¬cause stenosis has only been widely rec¬

ognized in recent years. Diagnostic cri¬teria, indications for surgery, and thenatural history are still being elucidated.Increased pain on spine extension is typ¬ical of stenosis (whereas flexion is usu-



ally most painful with herniated disks),but accuracy data are unavailable. Thesensitivity of leg pain is about 85%, neu¬

rologic abnormalities about 60%, and ab¬normal straight leg raising about 50%.52·53Cauda Equina Syndrome

A massive midline disk herniation maycause spinal cord or cauda equina com¬

pression, requiring immediate surgicalreferral. Fortunately, the cauda equinasyndrome occurs in only 1% to 2% of alllumbar disk herniations that come tosurgery,33 so its prevalence among allpatients with low back pain is about0.0004. The most consistent finding isurinary retention, with a sensitivity of0.90.54"56 Assuming a specificity of about95%, the predictive value of a negativetest (no urinary retention) would be al¬most 0.9999. Unilateral or bilateral sci¬atica, sensory and motor deficits, andabnormal straight leg raising are all com¬

mon, with sensitivities of over 0.80.5456The most common sensory deficit oc¬curs over the buttocks, posterior-supe¬rior thighs, and permeai regions ("sad¬dle anesthesia"), with a sensitivity ofabout 0.75.54"56 Anal sphincter tone isdiminished in 60% to 80% of cases.54·56Indications for Imaging Tests

There is a growing consensus thatplain roentgenograms are not necessaryfor every patient with low back painbecause of a low yield of useful findings,potentially misleading results, substan¬tial gonadal irradiation, and common in¬terpretive disagreements. The QuebecTask Force on Spinal Disorders sug¬gested that early roentgenography was

necessary only in the face of neurologicdeficits, age over 50 or under 20 years,fever, trauma, or signs of neoplasm.57Table 1 indicates "screening" questionsthat can virtually exclude neoplasm onthe basis of patient history alone.10

Magnetic resonance imaging and com¬

puted tomography can be used evenmore selectively, usually for surgicalplanning. The finding of herniated disksand spinal stenosis in many asymptom¬atic persons6·7 indicates that imaging re¬sults alone can be misleading, and validdecision making requires correlationwith the history and physical examina¬tion.58IS THERE EVIDENCE OF SOCIALOR PSYCHOLOGICAL DISTRESSTHAT MAY AMPLIFY ORPROLONG PAIN?

Some features of patient history in¬fluence management regardless of theexact spinal pathology. Chronic pain or

depression may be indications for theuse ofantidepressant medication ratherthan opiates. Alcohol or drug abuse in-

fluences the choice of medications andrequires specific intervention. Disabil¬ity compensation claims or litigation mayaffect initial evaluation and prognosis,and patients seeking compensation of¬ten respond poorly to a variety of treat¬ments.59

Patients with chronic low back pain(s3 months) present complex problems,and often a pathoanatomic cause is notapparent.60 Unlike acute pain, chronicpain is often not associated with ongo¬ing tissue injury, serves no biologicalusefulness, and is not accompanied bythe autonomie response of sympatheticoveractivity. Vegetative signs, such as

sleep disturbance, appetite disturbance,and irritability, appear, and pain is oftenreinforced or perpetuated by social andpsychological factors. Back pain can af¬fect employment, income, family, andsocial roles, producing psychological dis¬tress.60·61 Resulting somatic amplifica¬tion can serve the patient's needs foreconomic survival and maintenance ofself-esteem.61

In patients with chronic low back pain,the absence of systemic disease andtreatable anatomic abnormalities shouldbe confirmed by history, physical ex¬

amination, and review ofdiagnostic tests.Neurological abnormalities often proveto be long-standing and may persist af¬ter surgical interventions. Evidence ofpsychological distress should be sought,because this may respond to direct in¬tervention and improve the likelihood ofresponse to other treatments. The Min¬nesota Multiphasic Personality Inven¬tory (MMPI) is impractical in most pri¬mary care settings, and shorter depres¬sion scales are useful for screening.62·63

Waddell and colleagues64 proposed fivecategories of inappropriate or "nonor-ganic" signs that correlated with otherindicators of psychological distress: (1)inappropriate tenderness that is super¬ficial or widespread, (2) pain on simu¬lated axial loading by pressing on thetop of the head, or simulated spine ro¬tation (performed by holding the pa¬tient's arms to the side while rotatingthe hips, assuring that the shouldersand hips rotate together), (3) "distrac¬tion" signs, such as inconsistent perfor¬mance between straight leg raising inthe seated position vs the supine posi¬tion, (4) regional disturbances in strengthand sensation that do not correspondwith nerve root innervation patterns,and (5) overreaction during the physicalexamination. The occurrence of any one

sign was of limited value, but positivefindings in three of the five categoriessuggested psychological distress. Theprecision of nonorganic signs was re¬

ported by Waddell et al to be high, butsubsequent evaluation found poor pre-

cisión in the regional disturbance cate¬gory (Table 2).20

SUMMARY ANDRECOMMENDATIONSHistory

1. A few key questions can raise orlower the probability ofunderlying sys¬temic disease. The most useful itemsare age, history of cancer, unexplainedweight loss, duration of pain, and re¬

sponsiveness to previous therapy.2. Intravenous drug use or urinary

infection raises the suspicion of spinalinfection.

3. Ankylosing spondylitis is suggestedby the patient's age and sex (most com¬mon in young men), but most clinicalfindings have limited accuracy.

4. Failure of bed rest to relieve thepain is a sensitive finding for all these sys¬temic conditions, although not specific.

5. Neurologic involvement is sug¬gested by symptoms of sciatica or

pseudoclaudication. Pain radiating dis-tally (below the knee) is more likely torepresent a true radiculopathy than painradiating only to the posterior thigh. Ahistory of numbness or weakness in thelegs further increases the likelihood ofneurologic involvement.

6. Inquiry should be made concern¬

ing symptoms of the cauda equina syn¬drome: bladder dysfunction (especiallyurinary retention) and saddle anesthe¬sia in addition to sciatica and weakness.

7. The psychosocial history helps toestimate prognosis and plan therapy.The most useful items are a history offailed previous treatments, substanceabuse, and disability compensation. Briefscreening questionnaires for depressionmay suggest important therapeutic op¬portunities.Physical Examination

1. Fever suggests the possibility ofspinal infection. Vertebral tendernessis a sensitive finding for infection butnot specific.

2. The search for soft-tissue tender¬ness is unlikely to provide reproducibledata or demonstrably valid pathophys-iologic inferences.20·38

3. Limited lumbar flexion is not highlysensitive or specific for anklyosingspondylitis or other diagnoses. However,limited spinal motion may be useful inplanning physical therapy and monitor¬ing response.

4. In a patient with sciatica or possi¬ble neurogenic claudication, straight legraising should be assessed bilaterally,preferably using an inclinometer or go¬niometer.

5. Neurologic examination empha¬sizes ankle dorsiflexion strength, great



toe dorsiflexion strength, ankle reflexes,and the sensory examination. A rapidscreening sensory examination wouldtest pinprick sensation in the medial,dorsal, and lateral aspects of the foot.

6. For the patient with chronic pain,all of the evaluations described hereinshould be completed. Anatomically "in¬appropriate" signs may be helpful in

identifying psychological distress as aresult of or as an amplifier of low backsymptoms. The most reproducible ofthese signs are superficial tenderness,distracted straight leg raising, and theobservation ofpatient overreaction dur¬ing the physical examination.

This report was supported in part by the North¬west Health Services Research and Development

Field Prog. .. , Seattle Veterans Affairs MedicalCenter; by grant HS-06344 from the Agency forHealth Care Policy and Research (the Back PainOutcome Assessment Team); and by grant88298-2H from the John A. Hartford Foundationthrough the Multi-Institutional Technology As¬sessment Consortium.

Darlena Bursell assisted in preparation of themanuscript, and Monica Hayes helped to coordi¬nate study activities.

References1. Cypress BK. Characteristics of physician visitsfor back symptoms: a national perspective. Am JPublic Health. 1983;73:389-395.2. Deyo RA, Tsui-Wu JY. Descriptive epidemiol-ogy of low-back pain and its related medical care inthe United States. Spine. 1987;12:264-268.3. Deyo RA, Loeser JD, Bigos SF. Herniated lum-bar intervertebral disk. Ann Intern Med. 1990;112:598-603.4. White AA, Gordon SL. Synopsis: workshop on

idiopathic low-back pain. Spine. 1982;7:141-149.5. Nachemson A. The lumbar spine: an orthopedicchallenge. Spine. 1976;1:59-71.6. Weisel SE, Tsourmas N, Feffer H, Citrin CM,Patronas N. A study of computer-assisted tomog-raphy, I: the incidence of positive CAT scans in an

asymptomatic group of patients. Spine. 1984;9:549\x=req-\551.7. Boden SD, Davis DO, Dina TS, Patronas NJ,Wiesel SW. Abnormal magnetic resonance scans ofthe lumbar spine in asymptomatic subjects. J BoneJoint Surg Am. 1990;72:403-408.8. Liang M, Komaroff AL. Roentgenograms in pri-mary care patients with acute low back pain: acost-effectiveness analysis. Arch Intern Med. 1982;142:1108-1112.9. Deyo RA. Early diagnostic evaluation of lowback pain. J Gen Intern Med. 1986;1:328-338.10. Deyo RA, Diehl AK. Cancer as a cause of backpain: frequency, clinical presentation, and diagnos-tic strategies. J Gen Intern Med. 1988;3:230-238.11. Scavone JG, Latshaw RF, Weidner WA. An-teroposterior and lateral radiographs: an adequatelumbar spine examination. Am J Radiol. 1981;136:715-717.12. GehweilerJA, Daffner RH. Low back pain: thecontroversy of radiologic evaluation. Am J Radiol.1983;140:109-112.13. Brekkan A. Radiographic examination of thelumbosacral spine: an 'age-stratified' study. ClinRadiol. 1983;34:321-324.14. Carter ET, McKenna CH, Brian DD, KurlandLT. Epidemiology ofankylosingspondylitisin Roch-ester, Minnesota, 1935-1973. Arthritis Rheum. 1979;22:365-370.15. Hawkins BR, Dawkins RL, Christiansen FT,Zilko PJ. Use of the B27 test in the diagnosis ofankylosing spondylitis: a statistical evaluation. Ar-thritis Rheum. 1981;24:743-746.16. Waldvogel FA, Vasey H. Osteomyelitis: thepast decade. N Engl J Med. 1980;303:360-370.17. Sapico FL, Montgomerie JZ. Pyogenic verte-bral osteomyelitis: report of nine cases and reviewof the literature. Rev Infect Dis. 1979;1:754-776.18. Baker AS, Ojemann RG, Swartz MN, et al.Spinal epidural abscess. N Engl J Med. 1975;293:463-468.19. Chandrasekar PH. Low back pain and intra-venous drug abusers. Arch Intern Med. 1990;150:1125-1128.20. McCombe PF, Fairbank JCT, Cockersole BC,Pynsent PB. Reproducibility of physical signs inlow-back pain. Spine. 1989;14:908-918.21. Bauer RL, Deyo RA. Low risk of vertebralfracture in Mexican American women. Arch InternMed. 1987;147:1437-1439.22. Calin A, Porta J, Fries JF, Schurman DJ. Clin-ical history as a screening test for ankylosingspondylitis. JAMA. 1977;237:2613-2614.23. Gran JT. An epidemiological survey of the signsand symptoms ofankylosing spondylitis. Clin Rheu-

matol. 1985;4:161-169.24. Van der Linden S, Valkenburg HA, Cats A.Evaluation of diagnostic criteria for ankylosingspondylitis: a proposal for modification of the NewYork criteria. Arthritis Rheum. 1984;27:361-368.25. Calin A, Kaye B, Sternberg M, Antell B, ChanM. The prevalence and nature of back pain in anindustrial complex: questionnaire and radiographicand HLA analysis. Spine. 1980;5:201-205.26. Reynolds PMG. Measurement of spinal mobil-ity: a comparison of three methods. Rheumatol Re-habil. 1975;14:180-185.27. Rae PS, Waddell G, Venner RM. A simple tech-nique for measuring lumbar spinal flexion. JRCollSurg Edin. 1984;29:281-284.28. MollJMH, Wright V. An objective clinical studyof chest expansion. Ann Rheum Dis. 1972;31:1-8.29. Potter NA, Rothstein JM. Intertester reliabil-ity for selected clinical tests of the sacroiliac joint.Phys Ther. 1985;65:1671-1675.30. Russell AS, Maksymowych W, LeClercq S. Clin-ical examination of sacroiliac joints: a prospectivestudy. Arthritis Rheum. 1981;24:1575-1577.31. Blower PW, GriffinAJ. Clinical sacroiliac testsin ankylosing spondylitis and other causes of lowback pain. Ann Rheum Dis. 1984;43:192-195.32. Mayer TG, Tencer AF, Kristoferson S, MooneyV. Use ofnon invasive techniques for quantificationof spinal range of motion in normal subjects andchronic low-back pain dysfunction patients. Spine.1984;9:588-595.33. Spangfort EV. Lumbar disc herniation: a com-

puter aided analysis of 2504 operations. Acta Or-thop Scand. 1972;(suppl 142):1-93.34. Alpers BJ. The neurological aspects of sciatica.Med Clin North Am. 1953;37:503-510.35. Brieg A, Troup JDG. Biomechanical consider-ations in the straight-leg-raising test: cadaveric andclinical studies of medial hip rotation. Spine. 1979;4:242-250.36. CharnleyJ. Orthopedic signs in the diagnosis ofdisc protusion with special reference to the straightleg raising test. Lancet. 1951;1:186-192.37. Kosteljanetz M, Bang F, Schmidt-Olsen S. Theclinical significance of straight-leg-raising (Laseg-ue's sign) in the diagnosis of prolapsed lumbar disc.Spine. 1988;13:393-395.38. Waddell G, Main CS, Morris EW, et al. Nor-mality and reliability in the clinical assessment ofbackache. BMJ. 1982;284:1519-1523.39. Hoehler FK, Tobis JS. Low back pain and itstreatment by spinal manipulation: measures offlex-ibility and asymmetry. Rheumatol Rehabil. 1982;21:21-26.40. Hsieh CY, Walker JM, Gillis K. Straight legraising test: comparison ofthree instruments. PhysTher. 1983;63:1429-1432.41. Kosteljanetz M, Espersen JO, Halaburt H, Mi-letic T. Predictive value of clinical and surgicalfindings in patients with lumbago-sciatica: a pro-spective study (part 1). Acta Neurochir. 1984;73:67-76.42. Hakelius A, Hindmarsh J. The comparative re-

liability ofpreoperative diagnostic methods in lum-bar disc surgery. Acta Orthop Scand. 1972;43:234\x=req-\238.43. Hakelius A, Hindmarsh J. The significance ofneurological signs and myelographic findings in thediagnosis of lumbar root compression. Acta OrthopScand. 1972;43:239-246.44. Hudgins RW. The crossed straight leg raising

test: a diagnostic sign of herniated disc. J OccupMed. 1979;21:407-408.45. Shiqing X, Quanzhi Z, Dehao F. Significance ofstraight-leg-raising test in the diagnosis and clin-ical evaluation of lower lumbar intervertebral discprotrusion. JBone Joint Surg Am. 1987;69:517-522.46. Kortelainen P, PuranenJ, Koivisto E, Lahde S.Symptoms and signs ofsciatica and their relation tothe localization ofthe lumbar disc herniation. Spine.1985;10:88-92.47. Blower PW. Neurologic patterns in unilateralsciatica. Spine. 1981;6:175-179.48. Schwartz RS, Morris JGL, Crimmins D, et al.A comparison of two methods of eliciting the anklejerk. Aust N Z J Med. 1990;20:116-119.49. Keegan JJ. Dermatome hypalgesia associatedwith herniation of intervertebral disk. Arch NeurolPsychiatry. 1943;50:67-83.50. Aronson HA, Dunsmore RH. Herniated upperlumbar discs. J Bone Joint Surg Am. 1963;45:311-317.51. Morris EW, DiPaola M, Vallance R, Waddell G.Diagnosis and decision making in lumbar disc pro-lapse and nerve entrapment. Spine. 1986;11:436\x=req-\439.52. Turner JA, Ersek M, Herron L, Deyo R. Sur-gery for lumbar spinal stenosis: attempted meta-analysis of the literature. Spine. 1992;17:1-8.53. Hawkes CH, Roberts GM. Neurogenic and vas-cular claudication. J Neurol Sci. 1978;38:337-345.54. Kostuik JP, Harrington I, Alexander D, RandW, Evans D. Cauda equina syndrome and lumbardisc herniation. J Bone Joint Surg Am. 1986;68:386-391.55. O'Laoire SA, Crockard HA, Thomas DG. Prog-nosis for sphincter recovery after operation for caudaequina compression owing to lumbar disc prolapse.BMJ. 1981;282:1852-1854.56. Tay ECK, Chacha PB. Midline prolapse of alumbar intervertebral disc with compression of thecauda equina. J Bone Joint Surg Br. 1979;61:43-46.57. Spitzer WO, LeBlanc FE, Dupuis M, et al. Sci-entific approach to the assessment and manage-ment of activity-related spinal disorders: a mono-

graph for clinicians: report ofthe Quebec Task Forceon Spinal Disorders. Spine. 1987;12(suppl 7):S16-S21.58. Deyo RA, Bigos SJ, Maravilla KR. Diagnosticimaging procedures for the lumbar spine. Ann In-tern Med. 1989;111:865-867.59. Walsh NE, Dumitru D. The influence of com-

pensation on recovery from low back pain. OccupMed. 1988;3:109-121.60. Gatchel RJ, Mayer TG, Capra P, Diamond P,Barnett J. Quantification of lumbar function, VI:the use of psychological measures in guiding phys-ical functional restoration. Spine. 1986;11:36-42.61. Korbon GA, DeGood DE, Schroeder ME,Schwartz DP, Shutty MS. The development of asomatic amplification rating scale for low-back pain.Spine. 1987;12:787-791.62. Burnam AM, Wells KB, Leake B, Landsverk J.Development of a brief screening instrument fordetecting depressive disorders. Med Care. 1988;26:775-789.63. Rucker L, Frye EB, Cygan RW. Feasibilityand usefulness of depression screening in medicaloutpatients. Arch Intern Med. 1986;146:729-731.64. Waddell G, McCullochJA, Kummel E, VernnerRM. Nonorganic physical signs in low back pain.Spine. 1980;5:117-125.



what can history physical examination tell us about low back...what can the history and physical...

Documents