Table 2.

Multivariable analysis for diabetes risk by treatment among TC survivors (case-cohort study with detailed treatment data)

	Adjusted for age			Adjusted for age and BMI at TC diagnosis (continuously)
	n/N	HR	95% CI	HR	95% CI
Radiotherapya
No	27/196	1	ref	1	ref
Yes	134/379	1.66	1.05–2.62	1.54	0.96–2.45
P-heterogeneity			0.030		0.074
Radiotherapy field a
No PAO-radiotherapy	27/196	1	ref	1	ref
PAO-radiotherapy, left(incl. bilateral)	58/172	1.62	0.99–2.68	1.46	0.87–2.46
PAO-radiotherapy, right b	74/199	1.71	1.05–2.77	1.61	0.98–2.64
Other infradiaphragmatic fields	2/8	1.26	0.25–6.45	1.16	0.24–5.67
P -heterogeneity			0.156		0.204
Radiotherapy dosea
No para-aortic radiotherapy	27/196	1	ref	1	ref
≤26 Gray	80/281	1.45	0.91–2.32	1.29	0.79–2.11
27–32 Gray	35/46	2.39	1.31–4.35	2.26	1.22–4.21
≥33 Gray	19/52	1.88c	0.92–3.86	1.94c	0.95–3.97
P -heterogeneity			0.024		0.028
P -trend			0.013		0.011
BMI at TC diagnosis
<30 kg/m2	129/531	1	ref	—	—
≥30 kg/m2	32/44	2.52	1.31–4.87	—	—
P -heterogeneity			P<0.001
Smoking status at TC diagnosis
No	94/330	1	ref	—	—
Yes	67/245	0.94	0.64-1.37	—	—
P -heterogeneity			0.625

n/N Median number of cases/median number of non-cases in the 20 imputed and pooled data sets

PAO para-aortic radiotherapy, BMI body mass index (kg/m²)

^aPrimary and/or follow-up treatment

^bDiabetes risk was not significantly different between left and right sided PAO irradiation, P-heterogeneity, adjusted for age: 0.084; P-heterogeneity, adjusted for age and BMI: 0.489

^cPara-aortic RT dose 27-32 Gy and >32 Gy versus ≤26 Gy, P-hetero ≤0.016 and P-hetero ≤0.093, respectively (median dose: 26 Gy in category ≤26Gy, 30 Gy in category 27–32 Gy and 40 Gy in category ≥33 Gy)