Lumbar disc herniation is one of the few causes of spinal pain that can be successfully treated surgically. Perhaps because of the insufficient knowledge of most other causes of spinal pain, it has become a popular diagnosis. The overall prognosis is favorable, with a strong tendency toward spontaneous healing: herniations shrink and may be resorbed.23 In most cases conservative treatment, pain control, and awaiting the natural course are enough.1,5,29 However, incidence of disc surgery varies tenfold among industrialized countries, differences too conspicuous to be explained by biologic factors alone.13 Rather, secondary gain, healthcare organization, social insurance, availability of medical resources, and local traditions seem to be important. The benefit of disc surgery is difficult to assess because evaluation protocols vary: Howe and Frymoyer14 using different evaluation systems reassessed one group of patients, and found the fraction with good or excellent results ranged between 60% and 97%.
Despite advancements in diagnostic imaging and refinement of surgical technique, the results after lumbar disc surgery do not seem to have improved during recent decades1,22,25: 15% to 20% of the patients do not recover, and a few have severe chronic pain syndromes. The bad results more often are attributed to error in diagnosis (a protruded disc without a ruptured anulus was not the main cause of the patient's leg pain) than to technical errors or complications.24 Disc protrusions or herniations are common in symptom free adults, as shown by autopsy21 and diagnostic imaging.3,4,18 In many cases, the pain is relieved after conservative treatment, but the herniation persists unchanged.5 Even at surgery there is no sharp demarcation between negative exploration and bulging discs. The diagnostic uncertainty and bad surgical results have prompted spine surgeons to seek other ways to improve preoperative assessment. Psychologic tests have proved efficient in predicting surgical outcome, often having a higher impact than pathologic anatomy.25 Distress, abnormal illness behavior, and other psychosocial disincentives to recovery may overshadow the organic lesion.28
After clinical history, physical examination is the clinician's most helpful tool. However, many commonly used physical signs and tests have a surprisingly low reliability and reproducibility.20 After more than 60 years' experience with lumbar disc surgery, the unsolved questions concerning lumbar disc herniation mostly are centered around the basic clinical evaluation of patients. The purpose of this study was to analyze the diagnostic and prognostic value of clinical history and physical examination.
MATERIALS AND METHODS
One hundred sixty consecutive patients with clinical and radiographic signs of lumbar disc herniation, 85 men and 75 women, with a mean age of 43 ± 10 years (mean ± standard deviation), underwent primary surgery between 1981 and 1984 because of sciatica attributable to suspected lumbar disc herniation. Before undergoing surgery the patients completed a form regarding social background factors and medical and psychiatric history, and all previous medical records were reviewed. Only patients with a definite diagnosis who were under treatment were considered to have comorbidity. Education was classified as high (academics and skilled workers) or low (no additional education or vocational training after elementary school). Four patients were retired, and eight were unemployed.
Pain was graded qualitatively: (1) incapacitating, the patient was mostly bedridden because of pain and in constant need of analgesics; (2) severe, the patient had constant pain and was in constant need of analgesics, had difficulty walking, and was intermittently bedridden; (3) moderate, the patient had intermittent pain that allowed activities of daily life and had intermittent use of analgesics; and (4) no pain at all. For computation purposes, pain was dichotomized into incapacitating and nonincapacitating pain before surgery and into pain or no pain after surgery. The patients also made pain drawings, as previously described.31
Lumbar sagittal range of motion (ROM) was measured with a kyphometer according to Debrunner8 and Vucetic and Svensson.32 Lasègue and crossed Lasègue tests were performed in two maneuvers, as originally described by Forst12: first, passive flexion of the hip with the knee extended; second, passive hip flexion with the knee flexed. The test was considered positive if the first, but not the second, maneuver elicited or aggravated pain radiating down to the foot. Tendon reflexes were tested with the patient supine and the hips and knees flexed and were considered pathologic if absent. The power of the foot extensors was tested without resistance, with the patient in a supine position with hips and knees flexed. A difference in the range of extension was classed as pathologic in the absence of local disease. In addition, all patients underwent preoperative myelography. The films were reassessed in a blinded manner for dislocation of the root or dura. In 39 cases the films had been discarded, and the information was obtained from the patients' records.
At surgery, disc pathology was graded as intact anulus (negative exploration [n = 13] and protruding disc [n = 49]) or ruptured anulus (subligamentary perforation [n = 39] and complete perforation [n = 59]). There were 98 patients with ruptured anulus and 62 patients with intact anulus. One patient had a postoperative discitis. The patients were examined clinically after 2 years, as described.
Continuous variables such as duration of leg pain, patient age, and lumbar sagittal ROM were dichotomized around the medians of 7 months, 41 years, and 33°, respectively. Chi square analysis or Fisher's exact test were used at a rejection level of 5% to test the significance of individual factors for herniation. To discover the factors that were jointly the best predictors for herniation, stepwise logistic regression was used, again at a rejection level of 5%.7
The clinical results after 2 years are shown in Table 1. For predicting anatomic diagnosis, pain related factors dominated. From Table 2 it can be seen that a patient with all favorable factors would have odds of a ruptured anulus of 1600, obtained by multiplication of the individual factors (= 0.008 * 30.5 * 6.1 * 5.5 * 4.9 * 3.6 * 3.5* 3.1); a patient with all negative factors would obtain have odds of 0.001 (R2 = 0.495). (The R2 measure for models with dichotomous response variable generally is smaller than for models with continuous response. The reason for this is that unless all of the probabilities p given by the model are 0 or 1, no single observation will ever show a perfect fit: after all, the actual response will be either 0 or 1, whereas the prediction of the model will be a probability p, which is somewhere in between.)
The diagnosis was the most important predictor for pain relief. In Table 3 it can be seen that a patient with all positive factors has an odds of 4.7 of having no leg pain (R2 = 0.106). The corresponding R2 value for back pain was 0.153 (Table 4). For predicting anatomic diagnosis, pain and pain related factors dominated in the current study, accounting for not less than 70% of the R2 value in the logistic regression analysis. Social and behavioral factors accounted for approximately 22%. Thus, the structural factors of dislocation of root or dura on the myelogram added only 8%. However, for return to work anatomic diagnosis had no direct impact (Table 5). A patient with all positive factors has an odds for returning to work of 40. The corresponding figure for a patient with all unfavorable factors is 0.02 (R2=0.304).
When the factors were entered in the same sequence as clinically presented, history and pain related factors became the most important (Fig 1).
In 1951 Charnley6 remarked that clinically there appears to be two kinds of disc herniations: one type with severe pain, restricted lumbar sagittal mobility, and a favorable postoperative prognosis; and another one with less pain, less restricted mobility, and a less favorable postoperative prognosis. This essentially was confirmed in this study, where the anatomic continuity of the anulus was a key factor. The pathologic anatomy also appeared to carry statistical information about the patient's education, gender, comorbidity, and previous nonspinal surgery. Of course it does not seem to be likely, from a biologic point of view, that education would weaken the anulus. Rather, it is more likely that the patients were selected for surgery partly because of their disability caused by comorbidity and behavioral and psychosocial factors. In this, as in many other disc herniation studies, women fared less well than did men. In general, women also appear to be more prone than men to have chronic pain syndrome: cultural, social, and biologic factors act in concert to make middle aged women particularly vulnerable.17
In addition, the influence of somatic disorders on the outcome of disc surgery has been described.10,11,17,22 In this study the increased prevalence of previous nonspinal operations in patients with intact anulus may suggest that these patients already had established abnormal illness behaviors.30 Thus, a simple clinical history seems to be a good alternative to intricate psychologic tests in the preoperative assessment. In addition, there may be a danger in an uncritical use of psychologic tests: it does not seem reasonable a priori to deny patients adequate therapy just because they have a weak psyche in addition to a weak anulus.
The distinction between excellent or good outcome in lumbar disc surgery often is complete pain relief, and return to work often constitutes the difference between good or fair. In contrast to pain, return to work is easy to assess, but interpretation is another matter: the ability to work certainly is not synonymous with return to work. A blue collar worker has less possibility of changing his working conditions than does a white collar worker. In this and other studies, return to work largely was governed by extraspinal factors. Return to work does not seem to be a valid outcome factor in lumbar disc surgery. In industrialized countries, chronic disability is linked closely to the legal framework of social insurance.28 In musculoskeletal conditions, the relationships between disease, illness, and suffering are particularly difficult to analyze because the symptom of pain often allows collection of Social Security benefits.
Apart from the rare cases with massive neurologic deficits, pain is the main indication for lumbar disc surgery. However, pain is subjective by definition. Thus, many clinicians prefer to rely on criteria that seem to be more objective, and traditionally, much emphasis has been attached to neurologic signs. However, such signs often are absent and have no strong association with preoperative and postoperative pain, impairment, or disability. In this model, ankle areflexia was slightly more prevalent in patients with ruptured anulus but became redundant when crossed Lasègue sign and lumbar ROM were entered in the computation. These pain related, orthopaedic,6 or mechanical15,16 signs seem to be more important from a surgical point of view. The neglected physical sign, lumbar ROM, is measurable on a ratio scale with a reasonable reliability.32
Kortelainen et al19 showed that pain pattern is a better criterion for localization of disc herniation than are neurologic signs. Pain localization on pain drawing aids in determining the grade and the level of disc herniation.31 For determining the grade of disc herniation, pain localization became redundant when a simple qualitative assessment of preoperative pain severity was entered in the computations. Hurme and Alaranta17 also found that the severity of preoperative pain was the strongest factor predicting outcome. Likewise, in a prospective multivariate analysis on the outcome of lumbar disc surgery, Abramovitz and Neff1 found that pain related factors were the most important.
In clinical assessment it is reasonable to start with the history and pain analysis (Fig 1) that in this study explained most of the variability. However, in the current study the patients had been selected for discectomy. An investigation conducted in primary care or at an orthopaedic outpatient department might provide different results. Perhaps the impact of physical examination would increase, but it seems unlikely that the overall pattern would be radically different because it is pain that prompts the patient to consult a physician.
The current study was performed when myelography was the gold standard. In a comparative study, Szypryt et al26 showed the magnetic resonance (MR) imaging is only slightly better than myelography in diagnosing lumbar disc protrusion. In that study the authors focused on dislocation of root and dura, and it is possible to translate this concept to modern diagnostic imaging. In an MR imaging study, dislocation of the neural structures was the only imaging parameter that attained statistical significance. Even if modern diagnostic imaging could have improved the preoperative diagnosis to a certain degree, it seems unlikely that it would beat the pathologic diagnosis at surgery, and actually most diagnostic and prognostic information available before surgery could be obtained by history.
Weber29 in a prospective randomized long term trial compared surgical and conservative treatment: after 1 year, 36% of the conservatively treated patients had improved, as compared with 65% of the surgically treated group. But after 4 and 10 years, the two groups ended up with a similar result. However, selection bias makes it hard to interpret the results. Before randomization, 24% of the patients were selected for surgery and 31% for conservative treatment, and during the study, 26% of the patients randomized for conservative treatment had surgery because of pain. (These figures show the difficulties in running randomized spinal surgery trials.) In contrast, better results after surgery than after conservative treatment were seen in a meta-analysis13 and in a prospective cohort study.2
The data from this study indicate that it often is possible by simple means to determine which patients would benefit from early surgery: prolonged waiting in patients with sequestrated herniations is meaningless and only causes unnecessary suffering.25 Equally important is that it often is feasible to pinpoint patients who are unlikely to benefit from surgery.9-11 The fact that only approximately half of patients with disc protrusion have improvement in condition after surgery indicates that many protruding discs removed at surgery are not the sole culprit and that some of these patients probably had referred pain and clinical and radiologic signs mimicking disc herniation.23 In four of five patients who had more severe pain after surgery, the anulus was intact. Four of them later had several reoperations with no improvement. The pathogenesis in failed back surgery syndrome is not known; perhaps the cause is surgical trauma per se, but perhaps it also is the chemical irritation from the incised nucleus pulposus. In the authors' opinion, patients with bulging discs only rarely benefit from surgery more than what can be expected from the natural regression toward the mean and the placebo effects of surgery. Even after inefficacious surgical procedures ⅓ of the patients subjectively feel improvement.27 Such effects probably are achieved better by more harmless means than through surgery.
The authors thank Nils Hövmöller for linguistic revision, Gudrun Brattström for statistical analysis, and Kristina Jönsson for data processing.
1. Abramovitz J, Neff S: Lumbar disc surgery: Results of the prospective lumbar discectomy study of the joint section on disorders of the spine and peripheral nerves of the American Association of Neurological Surgeons and the Congress of Neurological Surgeons. Neurosurgery 29:301-308, 1991.
2. Atlas SJ, Deyo RA, Keller RB, et al: The Maine lumbar spine study, Part II. One year outcomes of surgical and nonsurgical management of sciatica. Spine 21:1777-1786, 1996.
3. Boden SD, Davis OD, Dina TS, et al: Abnormal magnetic-resonance scans of the lumbar spine in asymptomatic subjects. J Bone Joint Surg 72A:403-407, 1990.
4. Boos N, Rieder, R, Schade V, et al: The diagnostic accuracy of magnetic resonance imaging, work perception, and psychosocial factors in identifying symptomatic disc herniations. Spine 20:13-25, 1995.
5. Bush K, Cowan N, Katz DE, et al: The natural history of sciatica associated with disc pathology. A prospective study with clinical and independent radiological follow-up. Spine 17:1205-1212, 1992.
6. Charnley J: Orthopaedic signs in the diagnosis of disc protrusion With special reference to the straightleg-raising test. Lancet 260:186-192, 1951.
7. Christensen R: Log-Linear Models. Berlin, Springer-Verlag 345-360, 1990.
8. Debrunner HU: Das Kyphometer. Z Orthop Grenzgeb 110:389-392, 1972.
9. Fager CA, Freidberg SR: Analysis of failures and poor results fo lumbar spine surgery. Spine 5:87-94, 1980.
10. Finneson BE: A lumbar disc surgery predictive score card. Spine 3:186-188, 1978.
11. Finneson BE, Cooper VR: A lumbar disc surgery predictive score card. A retrospective evaluation. Spine 4:141-144, 1979.
12. Forst JJ: Contribution a l'étude clinique de la sciatique. Thèse pour le doctorat en medicine. Paris A Parant 1881.
13. Hoffman RM, Wheeler KJ, Deyo RA: Surgery for herniated lumbar discs: A literature synthesis. J Gen Intern Med 8:487-496, 1993.
14. Howe J, Frymoyer JW: Effects of questionnaire design on the determination of end results in lumbar spinal surgery. Spine 10:804-805, 1985.
15. Hudgins WR: The predictive value of myelography in the diagnosis of ruptured lumbar discs. J Neurosurg 32:152-162, 1970.
16. Hudgins WR: The crossed straight leg raising test: A diagnostic sign of herniated discs. J Occup Med 21:407-408, 1979.
17. Hurme M, Alaranta H: Factors predicting the result of surgery for lumbar intervertebral disc herniation. Spine 12:933-938, 1987.
18. Jensen MC, Brant-Zawadzki MN, Obuchowski N, et al: Magnetic resonance imaging of the lumbar spine in people without back pain. N Engl J Med 331:369-373, 1994.
19. Korterlainen P, Puranen J, Koivisto E, et al: Symptoms and signs of sciatica and their relation to the localization of the lumbar disc herniation. Spine 10:88-92, 1985.
20. McCombe PF, Fairbank JCT, Cockersole BC, et al: Reproducibility of physical signs in low-back pain. Spine 14:908-918, 1989.
21. McRae DL: Asymptomatic intervertebral disc protrusion. Acta Radiol 46:9-27, 1956.
22. Rish BL: A critique of the surgical management of lumbar disc disease in private neurosurgical practice. Spine 9:500-504, 1984.
23. Saal JA, Saal JS, Herzog RJ: The natural history of lumbar intervertebral disc extrusions treated nonoperatively. Spine 15:683-686, 1990.
24. Spangfort EV: The lumbar disc herniation. A computer-aided analysis of 2504 operations. Acta Orthop Scand 142 (Suppl):69-70, 1972.
25. Spengler DM, Ouellette E, Battiè M, et al: Elective discectomy for herniation of lumbar disc. Additional experience with an objective method. J Bone Joint Surg 72A:230-237, 1990.
26. Szpryt EP, Tvining P, Wilde GP, et al: Diagnosis of lumbar disc protrusion. A comparison between magnetic resonance imaging and radiculography. J Bone Joint Surg 70B:717-722, 1988.
27. Turner JA, Deyo RA, Loeser JD, et al: The importance of placebo effects in pain treatment and research. JAMA 271:1609-1614, 1994.
28. Waddell G, Morris EW, Di Paola MP, et al: A concept of illness tested as an improved basis for surgical decision in low-back disorders. Spine 11:712-719, 1986.
29. Weber H: Lumbar disc herniation. A controlled, prospective study with ten years of observation. Spine 8:131-140, 1983.
30. Vucetic N, De Bri E, Svensson O: Clinical history in lumbar disc herniation. A prospective study in 160 patients. Acta Orthop Scand 68:116-120, 1997.
31. Vucetic N, Mááttánen H, Svensson O: Pain and pathology in lumbar disc herniation. Clin Orthop 320:65-72, 1995.
© 1999 Lippincott Williams & Wilkins, Inc.
32. Vucetic N, Svensson O: Physical signs in lumbar disc herniation. Clin Orthop 335:192-201, 1996.