Patient demographic and preoperative characteristics: median, interquartile range, (mean ± standard deviation) or percent (count) (Reprinted with permission from Gornet MF, Burkus JK, Shaffrey ME, Argires PJ, Nian H, Harrell FE: Cervical disc arthroplasty with PRESTIGE LP disc versus anterior cervical discectomy and fusion: a prospective, multicenter investigational device exemption study. J Neurosurg Spine. 2015 Jul 31:1-16. [Epub ahead of print] PMID: 26230424.22).
| Investigational (N = 280) | Control (N = 265) | p Value before PS Adjustment* | P value after PS Adjustment | ||||
|---|---|---|---|---|---|---|---|
| PS as Continuous Covariate† | PS Stratificationt‡ | ||||||
| Age (yrs) | 39 44 49 | (44.5±8.8) | 38 44 49 | (43.9±8.8) | 0.369 | 0.997 | 0.873 |
| Height (in) | 64 67 71 | (67.7±4.1) | 64 67 71 | (67.5±4.2) | 0.622 | 0.998 | 0.946 |
| Weight (lbs) | 154 180 218 | (186.9±45.0) | 155 181 210 | (184.7±41.5) | 0.565 | 0.998 | 0.962 |
| NDI score | 44 54 66 | (55.5±14.7) | 44 58 68 | (56.4±15.9) | 0.499 | 0.997 | 0.952 |
| SF36 PCS score | 27.4 32.4 36.7 | (32.2±7.4) | 27.1 31.5 36.6 | (32.0±7.5) | 0.775 | 0.999 | 0.898 |
| SF36 MCS score | 34.9 46.5 53.6 | (44.5±11.5) | 33.1 42.1 53.0 | (42.7±12.4) | 0.078 | 0.993 | 0.924 |
| Neck pain score | 50 70 81 | (60.7±20.8) | 56 72 81 | (69.3±21.5) | 0.190 | 0.995 | 0.987 |
| Arm pain score | 40 64 80 | (59.6±26.3) | 42 67 83 | (62.4±28.5) | 0.237 | 0.995 | 0.960 |
| Female | 53.9% | (151) | 54.0% | (143) | 1.000 | 1.000 | 0.881 |
| Race | 0.075 | 0.997 | 0.708 | ||||
| Caucasian | 96.8% | (271) | 91.7% | (243) | |||
| Black | 2.5% | (7) | 4.9% | (13) | |||
| Asian | 0.0% | (0) | 0.8% | (2) | |||
| Hispanic | 0.4% | (1) | 2.3% | (6) | |||
| Other | 0.4% | (1) | 0.4% | (1) | |||
| Marital status | 0.109 | 0.990 | 0.749 | ||||
| Single | 14.3% | (40) | 12.1% | (32) | |||
| Married | 67.5% | (189) | 77.0% | (204) | |||
| Divorced | 15.0% | (42) | 9.1% | (24) | |||
| Separated | 2.5% | (7) | 1.1% | (3) | |||
| Widowed | 0.7% | (2) | 0.8% | (2) | |||
| Education level | 0.063 | 0.991 | 0.877 | ||||
| < High school | 5.4% | (15) | 5.3% | (14) | |||
| High school | 20.5% | (57) | 29.2% | (77) | |||
| > High school | 74.1% | (206) | 65.5% | (173) | |||
| Workers’ compensation case | 11.4% | (32) | 13.2% | (35) | 0.616 | 0.998 | 0.864 |
| Unresolved spinal litigation case | 12.1% | (34) | 12.1% | (32) | 1.000 | 1.000 | 0.928 |
| Tobacco used | 26.4% | (74) | 34.7% | (92) | 0.045 | 0.991 | 0.893 |
| Alcohol used | 53.6% | (150) | 53.2% | (141) | 1.000 | 1.000 | 0.900 |
| Working before operation | 67.1% | (188) | 62.6% | (166) | 0.312 | 0.996 | 0.968 |
| Non-narcotic relaxant medication use | 74.3% | (208) | 71.1% | (187) | 0.462 | 0.997 | 0.836 |
| Weak narcotic medication use | 47.7% | (133) | 48.3% | (127) | 0.954 | 1.000 | 0.843 |
| Strong narcotic medication use | 22.2% | (62) | 22.0% | (58) | 1.000 | 1.000 | 0.995 |
| Muscle relaxant medication use | 35.8% | (100) | 43.2% | (114) | 0.097 | 0.993 | 0.977 |
| Time to start having symptoms | 0.488 | 0.999 | 0.992 | ||||
| < 6 weeks | 7.9% | (22) | 5.7% | (15) | |||
| 6 weeks to 6 months | 30.4% | (85) | 33.6% | (89) | |||
| > 6 months | 61.8% | (173) | 60.8% | (161) | |||
| Normal motor functions | 38.2% | (107) | 59.6% | (158) | <0.001 | 0.979 | 0.751 |
| Normal sensory functions | 41.8% | (117) | 50.9% | (135) | 0.040 | 0.991 | 0.890 |
| Normal reflexes | 66.4% | (186) | 61.1% | (162) | 0.231 | 0.995 | 0.888 |
| Normal gait score | 93.6% | (262) | 77.0% | (204) | <0.001 | 0.881 | 0.467 |
| Positive foraminal compression test | 42.9% | (120) | 54.3% | (144) | 0.009 | 0.989 | 0.875 |
| Treatment level | 0.201 | 0.989 | 0.981 | ||||
| C3-C4 | 1.4% | (4) | 3.8% | (10) | |||
| C4-C5 | 7.5% | (21) | 5.7% | (15) | |||
| C5-C6 | 52.5% | (147) | 56.2% | (149) | |||
| C6-C7 | 38.6% | (108) | 34.3% | (91) | |||
NDI = Neck Disability Index, MCS = Mental Component Score, PCS = Physical Component Score, PS = propensity score.
↵* P values are from ANOVA for continuous variables and from Chi-square test for categorical variables.
↵† For continuous variables, p-values are from ANCOVA and for categorical variables, from logistic regression; propensity score as a continuous covariate for both.
‡ For continuous variables, p-values are from ANCOVA and for categorical variables, from the CMH test.
†,‡For categorical variables with multiple categories, they were dichotomized (except for treatment level) in the logistic regression models for calculating propensity scores to increase model stability and to check the covariate balance between treatment groups.