The significance of correlation of radiographic variables and MOS short-form health survey for clinical decision in symptomatic low back pain patients.

Objectives: To assess any correlation between SF-36 scores and several lateral roentgenographic variables of the lumbar spine, both in low back pain patients and asymptomatic volunteers. The clinical relevance of the method proposed in this study to make a surgical strategy on the basis of distinct lateral roentgenographic parameters and conversely was assessed by independent radiologists and physicians.

Methods: One hundred asymptomatic male volunteers, used as controls and an equal number of age-, height- and weight- comparable consecutive patients with chronic low back pain were examined both roentgenographically and with the SF-36 questionnaire. The roentgenographic variables measured were: lumbar lordosis, sacral inclination, L1 to S1 vertebral inclination, L4-S1 distal lordosis, disc index, and L1 to L5 vertebral index. These variables were correlated with the eight SF-36 items both in patients and controls. Both sensitivity and specificity of the method were calculated to assess the impact of subjectivity on the clinical decision.

Results: Low back patients showed significantly lower scores than their asymptomatic counterparts, in physical role limitations (P<0.01), severity of bodily pain (P<0.01), interference of bodily pain (P<0.01) and mental health (P<0.001). In the controls, general health, physical functioning, social functioning, and role limitations displayed a statistically significant correlation with age, height, weight, lumbar lordosis, sacral inclination, inclination of L1,L3 & L5-vertebra, L1 to L5-vertebral index and L1-L2, L3-L4, L4-L5 and L5-S1 disc index. In LBP-patients previous general health, physical functioning, role limitations, social functioning, bodily pain were significantly correlated with age, height, L1-and L2-inclination, distal lordosis, L5-index, and disc index L4-L5, and L5-S1. Although the consensus between either radiologists suggesting a strategy on the basis of roentgenographic data only or between physicians based on only SF-36 data was significant (P<0.001), there was no consensus in clinical decision between physicians and radiologists. Clinical decision based on matched SF-36 and roentgenographic data made either by radiologists or surgeons showed significant correlations (P<0.001). However, both sensitivity and specificity of our method to make a clinical decision on the basis of radiology were low: 0.48 and 0.36 respectively.

Conclusions: SF-36 scores were correlated with distinct lateral roentgenographic variables of the lowermost lumbar spine (L4-S1) in low back patients, and of the whole lumbar spine in asymptomatic individuals. Clinical decision should not be taken on the basis of radiological evidence of pathology because clinical decision seems to be more accurate when is taken on the basis of combined SF-36 and roentgenographic data. However, clinical examination is mandatory to SF-36 questionnaire and radiographic analysis.