Table 2. Frailty related characteristics, contents of intervention, and results (n=12).

First author	Conceptual/operational definitions of frailty	Instruments of frailty		Frailty related variable		Contents of intervention/Results related to frailty
Mohile (2007) [34]	Vulnerable older people by Saliba et al [41]	· VES-13		N/A		- The reliability of the VES-13 was 0.92 (Pearson correlation coefficient).
		· CGA
			- ADL, IADL			- The cut-off score of 3 on the VES-13 had 72.7% sensitivity and 85.7% specificity for CGA deficits and was highly predictive for identifying impairment.
			- SPPB
			- Comorbidity score
			- Number of medications
			- MOS Social Support Scale
			- SPMSQ
Bylow (2011) [35]	The frailty phenotype by Fried et al & Modified “obese” frailty based on the definition by Fried et al [4]	· Fried’s 5 criteria		· General & Clinical characteristics		- Men with BCR on ADT are frailer with lower performance status and more falls (p=0.02).
			- Unintentional weight loss: self-reported weight loss>10 pounds over the last year		- ADT
					- Self-reported comorbidities
			- Exhaustion: self-reported exhaustion based on two items from CES-D			- Comorbidity significantly increased the likelihood of “obese” frailty (p=0.01) and falls (p=0.01).
			- Weakness/grip strength: using hand dynamometer	· Physiological index
					- Hemoglobin
			- Walking speed: using a timed 15-foot walk,	· Physical function		- Mild anemia, which has been associated with both androgen depletion and frailty, is more prevalent in men on ADT (p=0.01).
					- Falls
			- Physical activity: self-reported physical activity based on weighted score of kilocalories expended per week
						- When controlling for age, clinical characteristics, and comorbidities, the men with BCR on ADT group showed significant trends in “obese” frailty (p=0.01) and falls (p=0.01).
		· Modified “obese” frailty criteria (substituting obesity [BMI>30.0 kg/m2] for the ‘weight loss’ criteria of Fried’s 5 criteria)
Cheung (2016) [20]	The frailty phenotype by Fried et al [4]	· Fried’s 5 criteria		· General & Clinical characteristics		Within 12 months of commencement, frailty increased with ADT (p<0.001), which was related to decreased testosterone (p=0.028), and less to fat mass (p=0.056) or lean mass (p=0.79).
			- Unintentional weight loss: self-reported weight loss>10 pounds in the last year
					- ADT
			- Exhaustion: self-reported exhaustion based on two items from CES-D	· Physiological index
					- Total testosterone
			- Weakness/grip strength: three times in the dominant and non-dominant hand with hand dynamometer
			- Walking speed: time to walk 4 m
			- Physical activity: Minnesota Leisure Time Physical Activity Questionnaire
Mareschal (2017) [33]	A cancer-specific definition of frailty by Hurria et al [42]	Old (age≥75 years) and/or presenting with at least one out of the following functional or physiological frailty criteria		N/A		(1) Contents of intervention: nutritional coaching, supervised group physical training, psychological counseling.
			- Cardiovascular/pulmonary comorbidities with CCI≥3
						(2) Follow-up period: start of ADT and 3, 6, 9, 12, 18, and 24 months or at discharge, thereafter (patients were discharged from the study when total testosterone blood levels turned back to normal. The last follow-up was planned 12 months after study discharge).
			- VES-13≥3
			- Balance unipedal stance test<5 seconds
						(3) Results: no significant change over the two years, including post-study follow-up. Means of QOL, nutritional, physical, and psychological variables remained stable over more than 2 years.
Winters-Stone (2017) [7]	Positive by the FRAIL scale which was developed by a European, Canadian, and American GAP & Modified “obese” frailty based on the definition by the GAP	· Self-reported FRAIL scale		· General & Clinical characteristics		- Current (43%) or past (40%) ADT users were more likely to be classified as prefrail or frail than never users (15%) (p<0.001), and the prevalence of combined obese frailty and prefrailty was even greater in current (59%) or past (62%) ADT users than never users (25%) (p<0.001).
		· Modified “obese” frailty criteria (substituting obesity (BMI>30.0 kg/m2) for the ‘shrinking’ criteria of the FRAIL scale)
					- ADT
Cheung (2018) [36]	The frailty phenotype by Fried et al [4]	· Fried’s 5 criteria		N/A		- Two years after cessation of ADT, participants had no difference in frailty score with those who did not (p=0.51).
			Specific measurement methods are not mentioned.
Navarro-Martínez (2019) [39]	The frailty phenotype by Fried et al [4]	· Fried’s 5 criteria		· Physiological index		- The severity of frailty syndrome and lymphocyte count were significantly negatively correlated (p<0.01). The concentration of IL-6 (p<0.05), CRP (p<0.05), and fibrinogen (p<0.01) were significantly associated with frailty syndrome.
			- Unintentional weight loss: self-reported weight loss which was 5% or 4.5 kg or more in the last year		- Lymphocyte count
					- IL-6
					- CRP
			- Exhaustion: self-reported exhaustion based on one item from CES-D		- Fibrinogen
					- Creatinine
			- Weakness/grip strength: three times in each hand alternately with a hand dynamometer
						- Creatinine had a significantly higher mean serum concentration in robust individuals compared to prefrail and frail patients (p=0.037)
			- Walking speed: time to walk 4.6 m
			- Physical activity: IPAQ, the total amount of energy spent on activities for 1 week
Buigues (2020) [38]	The frailty phenotype by Fried et al [4]	· Fried’s 5 criteria		· General & Clinical characteristics		- At baseline, older patients were more frailer than younger patients (p=0.009) and there were significant differences in the CCI between robust, prefrail and frail individuals at follow-up (p=0.002).
			- Unintentional weight loss: self-reported weight loss of 4.5 kg or more in the last year		- Age
					- ADT
					- CCI adjusted for age
			- Exhaustion: self-reported exhaustion based on one item from CES-D	· Physiological index
					- Lymphocyte count
			- Weakness/grip strength: three times in each hand alternately with a hand dynamometer		- IL-6	- At the baseline, a high IL-6 was associated with a significantly increased OR of being frail (p=0.013), as was a high IL-8 (p=0.014) compared to being non-frail. A higher lymphocyte count associated with a significant OR of being frailty (p=0.036).
					- IL-8
					- CRP
					- Monocyte count
			- Walking speed: time to walk 4.6 m		- Lymphocyte count
			- Physical activity: IPAQ, the total amount of energy spent on activities for 1 week
						- At baseline, there was a significant difference in CRP concentration in blood between robust, prefrail, and frail individuals (p=0.04).
						- At the follow up, there were significant differences between robust, prefrail and frail individuals for IL-6 (p<0.001), and the monocyte count (p=0.04). A higher IL-6 was associated with a significantly increased OR of being frail (p=0.011).
						- A higher log (IL-6) was associated with a significantly increased risk of frailty progression at 1 year follow-up (p<0.05), as was a lower lymphocytes count (p<0.05).
Couderc (2020) [27]	The frailty phenotype by Fried et al [4]	· CGA		N/A		Compared with older PC patients scheduled to receive RT or ADT without sarcopenia, those with probable sarcopenia had more geriatric frailties, although the differences did not reach significance.
			- G8 score
			- ECOG-PS
			- ADL, IADL
Momota (2020) [37]	Geriatric frailties by G8 score & The frailty phenotype by Fried et al [4]	· Fried’s 5 criteria		· General & Clinical characteristics		- The number of patients with frailty was significantly lower among the patients treated with RARP than those treated with RT (p<0.001) and ADT alone (p<0.001), and those with M1 disease (p<0.001).
			Specific measurement methods are not mentioned.		- Metastases
					- Treatment modalities
		· G8 score
						- The G8 score was significantly higher in the M0 group than in the M1 group (p<0.001).
						- The G8 score was significantly higher in the RARP group compared with the RT and ADT-alone groups (p<0.001)
Inglis (2021) [32]	A phase angle value <5.7 in older men	Phase angle value from BIA		· Body composition		(1) Contents of intervention: high-dose vitamin D supplementation
					- Phase angle value
							- Experimental group: high dose vitamin D
							- Control group: low dose vitamin D
						(2) Follow-up period: week 12 and week 24
							- Subjects took the supplements for 24 weeks.
						(3) Results
							- The experimental group had significantly wider phase angle values at week 12 (p=0.014) and at week 24 (p=0.018)
Hamaya (2021) [40]	Geriatric frailties by G8 score	G8 score		· General & Clinical characteristics		- The G8 scores of the patients with ADT alone were significantly lower than the patients with RARP and RT in the M0 group (p<0.001).
					- ADT
					- Metastases
						- A significant difference was observed between patients with M0 and mCSPC groups in the G8 score (p=0.002).

VES: vulnerable elders survey, CGA: comprehensive geriatric assessment, ADL: activities daily living, IADL: instrumental activities daily living, SPPB: short physical performance battery, MOS: medical outcomes study, SPMSQ: short portable mental status questionnaire, N/A: not applicable, CES-D: center for epidemiological studies-depression, ADT: androgen deprivation therapy, BMI: body mass index, BCR: biochemical recurrence, CCI: charlson comorbidity index, QOL: quality of life, FRAIL: fatigue, resistance, ambulation, illnesses, and loss of weight, GAP: geriatric advisory panel, IPAQ: international physical activity questionnaire, IL: Interleukin, CRP: C-reactive protein, OR: odds ratio, ECOG-PS: European cooperative oncology group-performance status, PC: prostate cancer, G8: geriatric 8 screening tool, RT: radiotherapy, RARP: robot assisted radical prostatectomy, M1: metastatic prostate cancer, M0: localized prostate cancer, mCSPC: metastatic castration-sensitive prostate cancer, BIA: bioelectrical impedance analysis.