. 2017 May 24;47(9):832–839. doi: 10.1093/jjco/hyx075

Table 4.

Univariate and multivariate analysis of risk factors for developing ILD

Exploratory variables Criterion variable Evaluation variable	Total number of patients	Incidence of ILD, %	Univariate analysis^a			Multivariate analysis^b
	Total number of patients	Incidence of ILD, %	HR	95% CI	P-value	HR	95% CI	P-value
Overall	829	6.3
Gender			1.3	0.7–2.2	0.4033
Female	346	5.5
Male	483	6.8
Age, years
<75	722	5.7	1.8	0.9–3.6	0.0719
≥75	107	10.3	1.8	0.9–3.6	0.0719
Previous or concurrent lung disease
No	746	5.8	1.9	0.9–3.9	0.0849	2.2	1.0–4.5	0.0365
Yes	83	10.8	1.9	0.9–3.9	0.0849	2.2	1.0–4.5	0.0365
Smoking history
No	460	4.6	1.8	1.1–3.2	0.0298
Yes	369	8.4	1.8	1.1–3.2	0.0298
ECOG PS
0	572	6.1	1.5	0.9–2.7	0.1495
≥1	257	6.6	1.5	0.9–2.7	0.1495
Number of chemotherapy regimens for primary disease: two categories
≥1	162	2.5	2.6	0.9–7.2	0.0667
0	667	7.2	2.6	0.9–7.2	0.0667
Number of affected organs: two categories
0 or ≤2	748	5.3	3.9	2.1–7.5	<0.0001	4.2	2.2–8.2	<0.0001
≥3	81	14.8	3.9	2.1–7.5	<0.0001	4.2	2.2–8.2	<0.0001

CI, confidence interval; HR, hazard ratio.

^aFourteen patients for whom not all exploratory variable data were obtained were excluded from the univariate and multivariate analyses.

^bExploratory variables included: gender, age, number of organs affected by metastases (categorical variable <3 vs. ≥3), previous/concurrent lung disease, smoking history, ECOG PS and previous chemotherapy regimens. These baseline demographics were selected as ILD risk factors because they were previously reported as ILD risk factors in Japan (26) or were statistically different with a P-value of <0.05 in univariate analyses. 75 years was selected as the exploratory variable for age because it had a higher univariate HR compared with 65 years.