Table 2.

HRs and 95% CIs for the association between diet quality^a and disease progression (Gleason score upgrading) in localised prostate cancer patients on active surveillance

	Range	N	Events	Base Modelb			Base + Clinical Characteristics Modelc
				HR	95% CI	P value	HR	95% CI	P value
Baseline diet quality
Low	34.81–63.30	137	29	1.00	Ref.	Ref.	1.00	Ref.	Ref.
Med	63.32–72.73	137	28	0.90	0.53–1.53	0.70	0.90	0.52–1.53	0.69
High	72.87–95.14	137	19	0.62	0.34–1.12	0.11	0.59	0.32–1.08	0.09
P trend d						0.11			0.06
6-Month clinical follow-up diet quality
Low	38.97–65.08	88	19	1.00	Ref.	Ref.	1.00	Ref.	Ref.
Med	65.17–74.45	88	13	0.66	0.32–1.33	0.24	0.77	0.37–1.60	0.49
High	74.66–93.84	87	19	0.99	0.52–1.89	0.98	1.05	0.54–2.04	0.90
P trend d						0.94			0.67
Proportional change in diet quality from baseline to 6-month follow-up
Decline	−0.15% to −25.4%	116	21	1.00	Ref.	Ref.	1.00	Ref.	Ref.
Improve	0.02–29.6%	147	30	1.10	0.59–2.05	0.76	1.18	0.62–2.25	0.62

BMI body mass index, HR hazard ratio, CI confidence interval, ANOVA analysis of variance, PSA prostate-specific antigen, HEI Healthy Eating Index

^aBaseline diet quality was defined by baseline HEI-2015 assessed in all 411 patients; 6-month clinical follow-up diet quality was defined by post-diagnostic HEI-2015 in a subset of 263 patients; and proportional change in diet quality from baseline to 6-month follow-up was calculated as the HEI-2015 score change from baseline to 6 months (baseline–6 months)

^bBase Model adjusted for age and total energy intake

^cBase + Clinical Characteristics model additionally includes PSA and composite tumour length

^dP_trend was calculated by using the median of each tertile as a continuous variable in the Cox proportional hazard model