Evidence in practice: patient examples are designed to illustrate how evidence is gathered and used to guide clinical decision making.

? Clinical question: What signs and symptoms can be used to differentiate low back pain of a musculoskeletal origin from a potentially more serious non-musculoskeletal condition in a 12-year-old girl?

A 12-year-old girl with no previous history of back pain was referred to our outpatient physical therapy department after developing an insidious onset of back pain, bilateral thigh pain, and difficulty walking. Her symptoms began approximately 3 months prior to her referral to therapy. The initial symptoms were pain and tightness in the right posterior thigh, which progressed to pain in the lumbar region and the left posterior thigh and buttock. She had radiographs of the right hip that were reported by the radiologist to be negative for a fracture or any developmental abnormality. She reported her pain to be mild (2/10 on a pain rating scale) and mostly unaffected by activity. Her chief complaint was the inability to play soccer due to changes in her walking and running. She had received physical therapy 6 months before. At that time, she reported only pain and tightness in the right posterior thigh, and her primary care physician referred her to our department with a diagnosis of "hamstring strain." These symptoms resolved with 2 weeks of therapy that focused on stretching exercises. She returned to the same physician for the current episode, and the physician again made the diagnosis of "hamstring strain." She had not received any further imaging studies since the initial episode.

The initial examination of the patient revealed an abnormal gait pattern with reduced stride length bilaterally, which was worse on the right side than on the left. Observation showed a scoliotic curve in the lumbar region with the convexity on the right and what appeared to be visible muscle spasm in the lumbar paraspinal musculature. Active range of motion in the lumbar spine revealed 10 degrees of lumbar extension, 85 degrees of flexion, 5 degrees of right side bending, and 20 degrees of left side bending. Extension caused a mild increase in low back pain, whereas flexion did not increase pain; however, the patient deviated toward the right during the movement. Lower-extremity reflexes, muscle strength, and sensation to light touch and sharp/dull discrimination were intact bilaterally. The straight-leg-raise test was positive bilaterally, reproducing leg symptoms at approximately 20 degrees on the right side and 45 degrees on the left. Palpation of the lumbar spine, buttock, and posterior thigh regions did not reveal any areas of tenderness.

The signs and symptoms in this patient could have a musculoskeletal origin, such as a lumbar disk herniation, spondylolithesis, or other musculoskeletal impairment. However, the presentation of this patient at initial examination raised some concerns about the possibility of a non-musculoskeletal cause of her symptoms. We were particularly suspicious because she had been seen 6 months ago and had no gait abnormalities or visible scoliosis at that time. Because of the rather dramatic change in her presentation, we were concerned that this patient may require referral for further diagnostic evaluation to rule out a non-musculoskeletal cause, such as a tumor somewhere in her spinal column. We decided to search the literature to determine which signs and symptoms in the clinical examination may be most useful in distinguishing a musculoskeletal from a non-musculoskeletal cause of low back pain in a child.

* Database used for search: MEDLINE

MEDLINE is the National Library of Medicine's computerized database. We decided to search using MEDLINE because the answer to the question could come from one of many medical fields--including pediatrics, oncology, or orthopedics---or from the nursing or allied health literature. MEDLINE encompasses these and many other disciplines and, therefore, provided the scope that might be necessary to find the most useful information. We searched MEDLINE through Ovid Online (www.ovid.com). * This service requires a fee but is available free of charge at many academic institutions. As an alternative, MEDLINE can be accessed through PubMed, which is available to the public for free at www.ncbi.nlm.nih.gov/PubMed. Searches on PubMed, however, may produce different results. This search was performed on December 12, 2001.

* Initial keywords: child, non-musculoskeletal

The first keyword we entered was child in order to narrow the search to only articles dealing with younger patients. We did not use the term "pediatrics" because we were concerned that this term may be too specific, and we wanted to begin with a broad search. Using the keyword child resulted in 808,515 citations. The next keyword we entered was non-musculoskeletal. Because we wanted to keep the search as broad as possible, we clicked on the Map Term to Subject Heading checkbox, which is a default option, to deselect it. When the Map Term feature is selected, MEDLINE suggests controlled subject heading terms that match the query. This keyword, without the mapping function, resulted in only 11 references, none of which dealt with low back pain in children. We decided that this keyword was not helpful and selected a different keyword.

The primary non-musculoskeletal condition that we thought might present as low back pain was a neoplasm; therefore, we decided to try using the keyword neoplasm. According to Ovid's Help feature, selecting the Map Term checkbox "usually results in higher-quality retrieval." Therefore, we checked this box throughout the remainder of our search. When we typed in this keyword, MEDLINE suggested several more specific subject headings. In this list was the subject heading spinal neoplasms. We decided this term would be more useful for our patient, and we, therefore, used spinal neoplasms as a keyword. After entering this keyword, MEDLINE suggested several subheadings, including diagnosis, radiography, embryology, and many others. We clicked on the subheading diagnosis because we were concerned about the differential diagnosis of a spinal neoplasm in our patient. This search resulted in 1,763 citations. Next, we combined the keywords child and spinal neoplasms/diagnosis by clicking on the icon labeled Combine. We used an "AND" operator to combine the 2 keywords. This combination narrowed the citation list to 309.

A list of 309 was still too large, and we wanted to make the list more specific to this patient with low back pain. We, therefore, entered the keyword low back pain. MEDLINE suggested a number of subject headings, but, to avoid eliminating citations that might be useful by restricting the search too much, we chose the heading low back pain and then selected all subheadings. These actions resulted in 3,956 citations. We combined these with the list of 309 citations using an "AND" operator. This narrowed the list of citations to only 3 references. We felt this was too narrow, and we decided to return to our list of 309 citations. Because we were interested in the diagnostic accuracy of clinical findings for a diagnosis of spinal neoplasm, the most useful articles would be those reporting sensitivity and specificity values. We chose to use the keyword sensitivity. When we typed in this keyword, MEDLINE suggested the subject heading sensitivity and specificity. We chose to use this subject heading, which resulted in 78,279 citations. We combined this list with the 309 citations previously selected using an "AND" operator. This combination yielded 5 articles. We decided to view the titles of these 5 articles. All of the articles dealt with diagnostic imaging procedures such as magnetic resonance imaging (MRI). Because we were interested in the value of clinical findings, we did not believe these articles would be useful for our patient. We, therefore, returned to our list of 309 citations and decided that we would attempt to further narrow our search using MEDLINE's Limits feature.

* Limits: English, human, last 5 years

When using Ovid, MEDLINE permits the user to use several limits to the search. The most frequently used limits are listed under the heading Limit To:. We first decided to limit our search to articles written in English and those involving human subjects. We did this by clicking on Human and English Language in the Limit To: options. This narrowed the citation list from 309 to 207 references. This list was still too large to examine each reference individually; therefore, we decided to first examine the most recent references, then consider older references if we were unable to find useful information. We limited the search to the past 5 years by entering the dates 1996 and 2001 in the Publication Year field. This resulted in a list of 64 references. At this point, we felt that we could look over the titles from this list and select the most potentially relevant articles. We wanted to find articles that described the clinical presentation of pediatric patients with low back pain due to musculoskeletal or non-musculoskeletal causes. We were particularly interested in finding articles that compared the clinical findings in 2 groups of patients: one with musculoskeletal low back pain and another with low back pain of a non-musculoskeletal origin. Many of the articles reported single case studies or focused on diagnostic imaging techniques. Nine articles appeared to be of potential use in answering our clinical question because they appeared to describe clinical characteristics of patients. The citations to these 9 articles are listed in the box on the previous page.

* Selection of articles for review: After narrowing our search to these 9 articles, we reviewed each title more carefully in an attempt to find the most useful references for further consideration. The strongest evidence on the value of diagnostic findings comes from articles that compare patients with the condition of interest (in our case, low back pain of a non-musculoskeletal origin) with patients with similar conditions that are often confused with the condition of interest (ie, low back pain of a musculoskeletal origin). Ideally, an article would report sensitivity and specificity values for particular clinical findings, indicating their usefulness in diagnosing spinal neoplasms in pediatric patients with low back pain. We were not interested in review articles or single case studies because these articles will not provide data comparing different types of patients.

Unfortunately, most of the articles did not appear to be exactly what we were seeking. Only one article (Martinez-Lage et al) mentioned a comparison between children with low back pain of a non-musculoskeletal origin (spinal neoplasms) and those with back pain of a musculoskeletal origin (lumbar disk herniation). None of the articles mentioned anything about sensitivity or specificity in the title. We decided to select 4 articles and read their abstracts to see if we should obtain the full text version of these articles. We selected the article by Martinez-Lage et al because the title mentioned a comparison group. We selected 3 other articles that we hoped would at least provide descriptive information about children with low back pain due to spinal neoplasms and their clinical presentation. The abstracts are listed and the selected articles are discussed below. We were able to access the abstract for each article by clicking on the Abstract icon.

   Grattan-Smith PJ, Ryan MM, Procoopis PG. Persistent or severe back pain and
   stiffness are ominous symptoms requiring prompt attention. J Paediatr Child
   Health. 2000;36:208-12.

   BACKGROUND: Children with severe or persistent back pain and stiffness
   often have an underlying organic cause but there is a large differential
   diagnosis, examination may be difficult and the problem is relatively rare
   in general paediatric practice. These difficulties appeared to lead to
   delays in diagnosis and management of children with this problem.
   OBJECTIVES: To provide an approach to the diagnosis and management children
   with severe or persistent back pain or stiffness based on our clinical
   experience and the literature. METHODOLOGY: The case histories of 10
   children with severe back pain seen by the authors over a 5-year period
   were reviewed. They were chosen as illustrative examples of the diagnostic
   and management problems and did not represent a systematic review of all
   cases seen by the authors over that time. RESULTS: Underlying causes
   included infection, inflammation, neoplasm, trauma and vascular
   malformation. Four of the children had spinal cord compression which
   required urgent decompression. There was one child with a conversion
   disorder but three children with organic disease were initially felt to
   have a conversion disorder. Investigations generally proceeded relatively
   slowly and the problem was not regarded as a semiurgent situation carrying
   the risk of permanent paraplegia. Magnetic resonance imaging (MRI) scan of
   the spine was the investigation of choice. CONCLUSION: Children with severe
   or persistent back pain and stiffness have a wide variety of underlying
   causes. The possibility of underlying spinal cord compression should always
   be considered in children with this presentation. If the diagnosis is not
   obvious, MRI scan of the spine should be arranged without delay.

   [Abstract reprinted with permission of Blackwell Science Asia Pty Ltd.]

The abstract of this article indicated that the study included rely 10 children with back pain and not all had spinal neoplasms. The 10 cases were apparently selected by the authors; herefore, this paper is not a systematic or consecutive review of cases and, as a result, may be a biased sample of cases. There is no comparison group of children with back pain of a musculoskeletal origin. Because of the small number of cases and the lack of a comparison group, we did not believe that his article would be helpful in answering our clinical question, so we did not seek out the full text.

   Mehlman CT, Crawford AH, McMath JA. Pediatric vertebral and spinal cord
   tumors: a retrospective study of musculoskeletal aspects of presentation,
   treatment, and complications. Orthopedics. 1999;22:49-55.

   Fifty-two pediatric patients with documented vertebral and spinal cord
   tumors were reviewed to evaluate the musculoskeletal manifestations of
   presentation, treatment, and management of complications. Diagnoses
   included 16 malignant and 36 nonmalignant vertebral and spinal cord tumors.
   Mean age at diagnosis was 8 years 1 month, and the male-to-female ratio was
   1.7:1. Overall, there was an average delay in diagnosis of 12 months for
   this group of patients. Mean clinical follow-up was 5 years. The two most
   common reasons for presentation were pain (67%) and spinal deformity (46%).
   Initial plain radiographs were positive in 82% of patients. Postlaminectomy
   spinal deformity occurred at a rate of 45% in patients undergoing
   laminectomies, and 60% of those with deformity required spinal fusion.
   Overall, 67% of the 52 patients underwent spinal fusion, with a 9%
   asymptomatic pseudarthrosis rate. Complications were common, with 1.9
   complications per patient; 59% of complications required surgical
   management. There were no perioperative deaths. At follow-up, 19% of
   patients had neurologic deficits ranging from neurogenic bladder to lower
   extremity weakness and difficulty walking to quadriplegic. Eighty-six
   percent of patients were ambulating without difficulty at follow-up. The
   survival rate was 94% and the local recurrence rate was 27% at most recent
   follow-up.

   [Abstract reprinted with permission of Slack Inc.]

This study was a retrospective review of 52 children, all of whom had spinal neoplasms. The abstract indicated that patient characteristics as well as signs and symptoms are given in the article. Because this article involved a larger group of patients, all with spinal neoplasms, we thought that the article might contain some useful information about our patient, so we decided to examine the article more fully. We went to the medical library and copied the full text of the article.

The mean age of the 52 patients with spinal neoplasms reviewed in this article was about 8 years old, and many of the patients were about the age of our patient. The article contained useful tables of the percentage of patients with certain clinical signs and symptoms at the initial presentation. Some interesting descriptive information could be taken from this article. First, of the 52 patients reviewed, the majority (63%) were male. Most patients (67%) reported back pain, and the most common clinical findings were spinal deformity (46%), motor weakness (29%), local tenderness (23%), abnormal reflexes (21%), and abnormal gait (17%). No comparison group of pediatric patients with musculoskeletal low back pain was presented, so it was difficult to determine which of these clinical features would distinguish a spinal neoplasm from musculoskeletal low back pain.

We considered our patient in light of this information. She reported pain and also had a spinal deformity and gait abnormalities on examination. Our patient, however, was female and did not have any local tenderness or motor or reflex abnormalities. Because this article lacked a comparison group, it was interesting, but ultimately of limited usefulness.

   Beer SJ, Menezes AH. Primary tumors of the spine in children. Natural
   history, management and long-term follow-up. Spine. 1997;22:64%58.

   DESIGN: Patients 16 years of age and younger with primary vertebral
   neoplasms diagnosed between 1951 and 1996 at The University of Iowa were
   reviewed retrospectively with specific consideration given to follow-up.
   OBJECTIVE: Follow-up extending beyond the growth of the axial skeleton was
   used to establish the natural history, management outcome, and current
   approach to treatment. SUMMARY OF BACKGROUND DATA: Primary tumors of the
   spine are uncommon. The treatment of such lesions generally has been based
   on small series of tumors, or extrapolated from the treatment of tumors in
   other regions and tumors in adults. Given the unique developing anatomy and
   dynamic nature of the growing spine in children, delineating appropriate
   modalities of treatment for these tumors beyond the growth of the axial
   skeleton is essential. METHODS: Clinical history, radiographs, radiographic
   reports, and interviews were used to establish this database. Outcome with
   respect to the various approaches to treatment was then compared in detail.
   RESULTS: Forty-five patients were identified in which follow-up greater
   than 10 years was available for 58% of patients. There were 29
   histologically benign and 16 malignant tumors. Patients presented most
   frequently with pain (79%) and neurologic deficits (74%). The duration of
   symptoms was significantly shorter with malignant tumors (11 weeks)
   compared with benign tumors (26 weeks). Radiographic abnormalities were
   demonstrated on plain radiographs in 98% of cases. Tumor excision was
   achieved in 80%. Follow-up studies were available beyond the growth of the
   axial skeleton in these patients, with an average follow-up of 14 years.
   Recurrence was seen in 6 of 45 children, and the overall mortality rate was
   6.7%, occurring only with malignant tumors. CONCLUSION: Our current
   approach to treatment of spinal neoplasms varies from that used in the
   earlier part of the series and reflects the need for single stage complete
   resection and stabilization.

   [Abstract reprinted with permission from Lippincott Williams & Wilkins.]

This article was another retrospective review of children who had spinal neoplasms. This review included 45 patients. The abstract notes that the most common clinical findings were pain (79%) and neurological deficits (74%). The abstract of this article seems to indicate that the authors were most interested in comparing benign and malignant tumors and in the long-term outcome of these children. Because these issues were not our primary concern, we did not believe that this article would provide much useful information, so we did not seek out the full text.

   Martinez-Lage JF, Martinez-Robledo A, Loez F, et al. Disc protrusion in the
   child. Particular features and comparison with neoplasms. Childs Nerv Sys.
   1997;13:201-7.

   Lumbar intervertebral disc herniation, although common in adults, is
   infrequent in the young, and especially in patients under 17 years old. In
   this work we review clinical data pertaining to two pediatric groups of
   patients whose main complaint was low back pain and/or sciatica, trying to
   identify factors that might contribute to their earlier referral and to the
   differential diagnosis of protruded disc and spinal neoplasm in this
   population. Group A comprises 17 youngsters diagnosed as having lumbar
   herniated nucleus pulposus and group B, 16 children with neoplasms of the
   lower thoracic and lumbosacral regions. Both groups were similar in sex
   distribution and symptoms of pain and numbness. However, there was a
   striking difference in age at presentation. No patient in group A was
   younger than 11 years, while most of those in group B were in their first
   decade of life (P = 0.018). The classic clinical onset in the children with
   herniated discs started with low back pain and sciatica, as in the children
   with neoplasms, although in subgroup B leg pain tended to be bilateral. The
   usual examination findings in both groups were spinal rigidity and sensory
   loss, but motor weakness and impaired reflexes were found to be more
   frequent in the group with spinal growths (P = 0.02). Children with
   lumbosacral neoplasms also tended to present with atypical symptoms (acute
   onset, intracranial hypertension, subarachnoid hemorrhage and abdominal
   pain), while this was the exception in the group with herniated discs.
   Plain radiographs of the pediatric spine showed that X-ray examination is
   still a good tool for diagnosing spinal growths compared with their scant
   utility in disc herniations (P = 0.001). During the survey we were
   impressed by the children's apparent good tolerance to pain, which is
   probably due to the lack of the emotional component of pain in adults and
   explains their delayed referral for neurosurgical consultation. However,
   all modalities of treatment seemed to be effective in children,
   chemonucleolysis and surgery being extraordinarily effective in this age
   group. Accordingly, we see no reason for long-term conservative therapy in
   children with lumbar and sciatic pain; on the contrary, we believe these
   patients should be offered earlier neurosurgical treatment.

   [Abstract reprinted with permission of Springer-Verlag GmbH.]

Based on both the title and the abstract, it appeared that this article could be quite useful in answering our question because 2 groups of patients were included: one group of children with low back pain of a non-musculoskeletal origin (spinal neoplasms), and another group of children with low back pain of a musculoskeletal origin (disk herniation). Unfortunately, other types of musculoskeletal low back pain were not included, but at least a comparison group was used. As noted in the abstract, the authors found that children with musculoskeletal back pain were older than children with neoplasms (average age = 14.5 years and 6.8 years, respectively). No child with musculoskeletal back pain was under the age of 11, but most of the children with neoplasms were. Our patient is 12 years of age, which would appear to reduce the likelihood of a spinal neoplasm. The abstract also states that abnormal reflexes and motor weakness were more common in the children with neoplasms. Our patient had neither of these findings. We were interested to know exactly how useful these findings were in distinguishing children with spinal neoplasms from those with musculoskeletal back pain, and we wanted to know if other clinical findings were discussed in the paper but not in the abstract. We decided to retrieve the full text of the article from the library to determine whether any other useful information could be found.

The information in the article proved to be very useful. A table of the number of children with specific clinical findings in each group was presented (Tab. 1). After examining this information, we believed that the most important clinical findings in the diagnosis of a spinal neoplasm appear to be (1) having bilateral versus unilateral leg pain, (2) having motor and reflex impairments, and (3) not having a positive straight-leg-raise test. Of the 17 patients with musculoskeletal back pain, 12 had unilateral leg pain, and only 3 had bilateral leg pain. In the group of children with spinal neoplasms, however, 9 had bilateral leg pain, and no child had unilateral leg pain. Fifteen of the 17 children with musculoskeletal back pain had a positive straight-leg-raise test, but only 1 had a motor deficit and only 4 had impaired reflexes. In contrast, the children with spinal neoplasms were less likely to have a positive straight-leg-raise test (7 of 16) and were more likely to have impaired motor strength (8 of 16) or reflexes (11 of 16).

These ratios were helpful, but we wanted to calculate sensitivity and specificity values from the percentages to get a better idea of the relative importance of each of these findings. According to Sackett et al, (1) tests that have high sensitivity are useful for ruling out a diagnosis when the test is negative, and tests that have high specificity are useful for ruling in a diagnosis when the test is positive. The results of our calculations are in Table 2. Looking first at the tests for a spinal neoplasm, the most important finding that would help to rule in this diagnosis would be the presence of motor weakness, because this finding had the largest specificity value (0.94). The presence of bilateral leg pain also appears to be important (specificity = 0.82). Impaired reflexes had the largest sensitivity value (0.69), indicating that the absence of impaired reflexes would help to rule out a diagnosis of spinal neoplasm; however, this sensitivity value is not very high, and this finding may not be that important. For a diagnosis of musculoskeletal back pain (ie, disk herniation), the most important finding to help rule in the diagnosis would be the presence of unilateral leg pain (specificity=1.0). The straight-leg-raise test had the highest sensitivity (0.88), indicating that, if the straight-leg-raise test is negative, the chance of a disk herniation decreases.

* Clinical decision: When we summarized the information that we gathered from these articles, we found that our patient has some findings that might indicate a non-musculoskeletal source of her low back pain (ie, a spinal neoplasm), and others that would help rule out this diagnosis. Overall, we found the evidence to be rather sparse and of poor quality because only one study actually compared children who had a musculoskeletal source of low back pain with a group of children who had non-musculoskeletal low back pain. We could locate only one article that compared 2 groups of children, and we had to reexamine the data presented in order to extract the most useful information. There was, however, some helpful information that could be taken from these articles. Based on the article by Martinez-Lage et al, the presence of bilateral leg symptoms in our patient is an important finding that might indicate a spinal neoplasm. The study by Mehlman et al also noted that spinal deformity was common among patients with spinal neoplasms. Our patient clearly had a scoliotic curve. Unfortunately, none of the studies mentioned the relatively rapid onset of the deformity that our patient had experienced.

Several factors about our patient would appear to help rule out a diagnosis of a spinal neoplasm. Perhaps the most important factor was her age, because most patients with spinal neoplasms seem to be under the age of 10. According to Martinez-Lage et al, our patient's lack of neurological deficits, particularly impaired reflexes, would make a diagnosis of spinal neoplasm less likely. The fact that she has a positive straight-leg-raise test might make a diagnosis of disk herniation more likely.

Because our patient had several signs that were associated with spinal neoplasms and because of the high risk associated with a missed or delayed diagnosis in that case, we decided that the most prudent action to take would be to refer the patient to a pediatric neurology specialist for further diagnostic work-up. Our patient had an MRI performed on her spine. Fortunately, no spinal neoplasm was discovered; however, a grade II spondylolisthesis was found that required surgical management. Although the diagnosis we were most concerned about was not present, the decision to refer the patient to a neurology specialist did appear to be justified in the end.

Selected Citations Retrieved by Search Using the Keywords "Child" and "Spinal Neoplams/Diagnosis" With Limits Added *

1. Venkateswaran L, Rodriguez-Galindo C, Merchant TE, Poquette CA, Rao BN, Pappo AS. Primary Ewing tumor of the vertebrae: clinical characteristics, prognostic factors, and outcome. Med Pediatr Oncoh 2001 Jul;37(1):30-5.

2. Laredo JD, el Quessar A, Bossard P, Vuillemin-Bodaghi V. Vertebral tumors and pseudotumors. Radiol Clin North Am. 2001 Jan;39(1):137-63, vi. Review.

3. Grattan-Smith PJ, Ryan MM, Procoopis PG. Persistent or severe back pain and stiffness are ominous symptoms requiring prompt attention. J Paediatr Child Health. 2000 Jun;36(3):208-12.

4. King HA. Back pain in children. Orthop Clin North Am. 1999 Jul;30(3):467-74, ix. Review.

5. Mehlman CT, Crawford AH, McMath JA. Pediatric vertebral and spinal cord tumors: a retrospective study of musculoskeletal aspects of presentation, treatment, and complications. Orthopedics. 1999 Jan;22(1):49-55; discussion 55-6.

6. Martinez-Lage JF, Martinez Robledo A, Lopez F, Poza M. Disc protrusion in the child. Particular features and comparison with neoplasms. Childs Nerv Sys. 1997 Apr;13(14):201-7.

7. Beer SJ, Menezes AH. Primary tumors of the spine in children. Natural history, management and long-term follow-up. Spine. 1997 Mar 15;22(6):649-58, discussion 658-9.

8. Payne WK 3rd, Ogilvie JW. Back pain in children and adolescents. Pediatr Clin North Am. 1996 Aug;43(4):899-917. Review.

9. Sinha AK, Seki JT, Moreau G, Ventureyra E, Letts RMI. The management of spinal metastasis in children. Can J Surg. 1997 Jun;40(3):218-26.

* English language, Human studies, Publication years between 1996 and 2001.

Table 1.

Clinical findings in children with back pain due to a
spinal neoplasm or due to a lumbar disk
herniation. (a)

                                                    Musculoskeletal
                                     Spinal         LBP (b) (disk
                                     Neoplasm       herniation,
Characteristic                       (n = 16)       n = 17)

Male                                 11              8
Female                                5              9
Mean age at                           6.8           14.5
  diagnosis (y)
Symptoms
  LBP                                15             17
  Bilateral leg pain                  9              3
  Unilateral leg pain                 0             12
  Motor weakness                      4              3
  Paresthesia                         3              7
Signs
  Positive straight-                  7             15
    leg test
  Motor weakness                      8              1
  Impaired reflexes                  11              4
  Sensory loss                        6              3

(a) Adapted with permission from Table 2 in: Martinez-Lage JF,
Martinez Robledo A, Lopez F, Poza M. Disc protrusion in chil-
dren: particular features and comparison with neoplasms. Childs
Nerv Syst. 1997;13:203. [c] 1997, Springer-Verlag GmbH.

(b) LBP=low back pain.
Table 2.

Sensitivity and specificity values for particular clinical
findings for a diagnosis of spinal neoplasm or disk
herniation. (a)

                                      Sensitivity    Specificity

Characteristic favoring the
diagnosis of spinal neoplasm
  Bilateral leg pain                  0.56           0.82
  Motor deficits                      0.50           0.94
  Impaired reflexes                   0.69           0.76

Characteristic favoring the
diagnosis of disk herniation
  Positive straight-leg-raise test    0.88           0.56
  Unilateral leg pain                 0.71           1.0

(a) Values are based on the data presented by Martinez-Lage et al.

* Ovid Technologies, 333 Seventh Ave, 4th Floor, New York, NY 10001.

References

(1) Sackett DL, Straus SE, Richardson WS, et al. Evidence-Based Medicine: How to Practice and Teach EBM. 2nd ed. New York, NY: Churchill Livingstone Inc; 2000.

Julie M Fritz, FI, PhD, ATC, is Assistant Professor, Department of Physical Therapy, University of Pittsburgh, Pittsburgh, Pa.

M Kathleen Kelly, PT, MS, is Assistant Professor, Department of Physical Therapy, University of Pittsburgh.