Association of cohort-corrected fasting peripheral blood concentration of IGFBP-1 (ng/mL) with risk of multiple myeloma, in the pooled study population and by year from blood draw to multiple myeloma diagnosis
| Fasting IGFBP-1 concentration, ng/mL* . | Case . | Control . | OR (95% CI)† . | P‡ . |
|---|---|---|---|---|
| Per SD increase (SD = 32.29 ng/mL)§ | ||||
| Pooled population | 354 | 709 | 1.2 (1.0-1.5) | .08 |
| By years, blood draw to multiple myeloma diagnosis | ||||
| ≤ 3 | 76 | 141 | 2.3 (1.4-3.8) | .001 |
| 4- ≤ 6 | 104 | 213 | 1.1 (0.8-1.6) | .50 |
| > 6 | 170 | 329 | 1.0 (0.7-1.4) | .83 |
| Per quartile increase‖ | ||||
| Pooled population | 354 | 709 | ||
| Q1 | 80 | 176 | 1.0 (ref) | |
| Q2 | 72 | 178 | 0.9 (0.6-1.3) | |
| Q3 | 94 | 176 | 1.2 (0.8-1.8) | |
| Q4 | 107 | 171 | 1.4 (0.9-2.1) | .04 |
| By years, blood draw to multiple myeloma diagnosis | ||||
| ≤ 3 | 76 | 141 | ||
| Q1 | 14 | 44 | 1.0 (ref) | |
| Q2 | 14 | 37 | 1.0 (0.4-2.6) | |
| Q3 | 18 | 34 | 1.7 (0.7-4.2) | |
| Q4 | 30 | 26 | 3.6 (1.4-9.5) | .002 |
| 4- ≤ 6 | 104 | 213 | ||
| Q1 | 24 | 51 | 1.0 (ref) | |
| Q2 | 16 | 54 | 0.6 (0.3-1.3) | |
| Q3 | 31 | 49 | 1.3 (0.6-2.7) | |
| Q4 | 33 | 59 | 1.2 (0.6-2.6) | .28 |
| > 6 | 170 | 329 | ||
| Q1 | 42 | 74 | 1.0 (ref) | |
| Q2 | 42 | 84 | 0.8 (0.5-1.5) | |
| Q3 | 45 | 89 | 0.9 (0.5-1.7) | |
| Q4 | 41 | 82 | 1.0 (0.5-1.8) | .83 |
| Fasting IGFBP-1 concentration, ng/mL* . | Case . | Control . | OR (95% CI)† . | P‡ . |
|---|---|---|---|---|
| Per SD increase (SD = 32.29 ng/mL)§ | ||||
| Pooled population | 354 | 709 | 1.2 (1.0-1.5) | .08 |
| By years, blood draw to multiple myeloma diagnosis | ||||
| ≤ 3 | 76 | 141 | 2.3 (1.4-3.8) | .001 |
| 4- ≤ 6 | 104 | 213 | 1.1 (0.8-1.6) | .50 |
| > 6 | 170 | 329 | 1.0 (0.7-1.4) | .83 |
| Per quartile increase‖ | ||||
| Pooled population | 354 | 709 | ||
| Q1 | 80 | 176 | 1.0 (ref) | |
| Q2 | 72 | 178 | 0.9 (0.6-1.3) | |
| Q3 | 94 | 176 | 1.2 (0.8-1.8) | |
| Q4 | 107 | 171 | 1.4 (0.9-2.1) | .04 |
| By years, blood draw to multiple myeloma diagnosis | ||||
| ≤ 3 | 76 | 141 | ||
| Q1 | 14 | 44 | 1.0 (ref) | |
| Q2 | 14 | 37 | 1.0 (0.4-2.6) | |
| Q3 | 18 | 34 | 1.7 (0.7-4.2) | |
| Q4 | 30 | 26 | 3.6 (1.4-9.5) | .002 |
| 4- ≤ 6 | 104 | 213 | ||
| Q1 | 24 | 51 | 1.0 (ref) | |
| Q2 | 16 | 54 | 0.6 (0.3-1.3) | |
| Q3 | 31 | 49 | 1.3 (0.6-2.7) | |
| Q4 | 33 | 59 | 1.2 (0.6-2.6) | .28 |
| > 6 | 170 | 329 | ||
| Q1 | 42 | 74 | 1.0 (ref) | |
| Q2 | 42 | 84 | 0.8 (0.5-1.5) | |
| Q3 | 45 | 89 | 0.9 (0.5-1.7) | |
| Q4 | 41 | 82 | 1.0 (0.5-1.8) | .83 |
Measurement of IGFBP-1 concentration was restricted to blood samples known to be collected ≥ 8 hours after the last meal.
ORs and 95% CIs were calculated in conditional logistic regression models that were stratified on matched set and further adjusted for BMI at blood draw (kg/m2). Persons with outlier IGFBP-1 values (ie, cohort-corrected levels > 121.68 ng/mL) and those with missing BMI data were excluded.
The P values associated with the per-SD ORs and CIs are Wald P values estimated in the corresponding conditional logistical regression models. The P values associated with the quartile-based results are from trend tests performed in conditional logistic regression models that were identical to those run for the quartile variables (ie, stratified on matched set and further adjusted for BMI at blood draw; kg/m2). Persons with outlier values and those with missing BMI data were excluded.
The SD was obtained from the distribution of cohort-corrected values in the pooled fasting controls.
Quartile of IGFBP-1 level was determined from the distribution of cohort-corrected values among the pooled controls.