Skip to main content
. 2015 Mar 27;29(8):1713–1720. doi: 10.1038/leu.2015.65

Table 3. Logistic regression analysis for factors associated with ‘any CIg <2.8'.

Variable N Any CIg <2.8 Any CIg ≥2.8 OR (95% CI) P-value
Univariate
 B2M ⩾3.5 mg/l 137 44/82 (54%) 14/55 (25%) 3.39 (1.61, 7.15) 0.0013
 B2M >5.5 mg/l 137 28/43 (65%) 30/94 (32%) 3.98 (1.86, 8.54) 0.0004
 Hb <10 g/dl 139 27/43 (63%) 33/96 (34%) 3.22 (1.52, 6.81) 0.0022
 CRP ⩾8 mg/l 139 27/45 (60%) 33/94 (35%) 2.77 (1.33, 5.76) 0.0063
 LDH ⩾190 U/l 139 21/31 (68%) 39/108 (36%) 3.72 (1.59, 8.69) 0.0025
 BMPC% ⩾33% 134 49/93 (53%) 10/41 (24%) 3.45 (1.52, 7.85) 0.0031
 Cytogenetic abnormalities 136 30/57 (53%) 28/79 (35%) 2.02 (1.01, 4.05) 0.0468
 GEP70 high risk 139 20/32 (63%) 40/107 (37%) 2.79 (1.23, 6.31) 0.0136
 GEP80 high risk 139 11/16 (69%) 49/123 (40%) 3.32 (1.09, 10.15) 0.0352
 GEP MS subgroup 139 4/20 (20%) 56/119 (47%) 0.28 (0.09, 0.89) 0.0311
 GEP centrosome index ⩾3 139 41/69 (59%) 19/70 (27%) 3.93 (1.93, 8.02) 0.0002
 Number of stem lines >2 139 11/17 (65%) 49/122 (40%) 2.73 (0.95, 7.87) 0.0628
 Total LCR% >50% 139 21/28 (75%) 39/111 (35%) 5.54 (2.16, 14.18) 0.0004
 
Multivariate
 B2M >5.5 mg/l 132 28/42 (67%) 29/90 (32%) 3.04 (1.27, 7.25) 0.0121
 CRP ⩾8 mg/l 132 25/42 (60%) 32/90 (36%) 3.35 (1.40, 8.01) 0.0065
 GEP centrosome index ⩾3 132 38/65 (58%) 19/67 (28%) 2.56 (1.12, 5.87) 0.0263
 LCR% >50 132 20/26 (77%) 37/106 (35%) 4.97 (1.68, 14.72) 0.0038

Abbreviations: B2M, beta-2 microglobulin; BMPC%, bone marrow plasma cell percentage; CRP, C-reactive protein; CI, confidence interval; GEP, gene expression profile; Hb, hemoglobin; LDH, lactate dehydrogenase; MS, MMSET; LCR%, light chain-restricted percentage; OR, odds ratio.

P-value from Wald χ2-test in logistic regression. NS2 multivariate results not statistically significant at 0.05 level. Univariate P-values reported if <0.1. Multivariate model uses stepwise selection with entry level 0.1 and variable remains if meets the 0.05 level. A multivariate P-value >0.05 indicates variable forced into model with significant variables chosen using stepwise selection.