Treatment of Hot Flashes in Men With Prostate Cancer Undergoing Androgen Deprivation Therapy

ABSTRACT

Background

Hot flashes are common and often debilitating side effects of castration therapy; a cornerstone in the treatment of metastatic prostate cancer (PCa) as well as in localized or locally advanced PCa when combined with radiotherapy. The evidence for relieving vasomotor symptoms is limited. This systematic review presents both pharmacological and non‐pharmacological interventions for treating hot flashes in men with PCa undergoing castration therapy, primarily androgen deprivation therapy (ADT). Methods: a systematic literature search was conducted in PubMed using (“hot flash*” OR “hot flush*” OR “vasomotor*”) AND (“prostate”) as keywords. Studies with intervention for hot flashes due to castration therapy, estimation of treatment response (e.g., reduction in frequency or impact on quality of life) and patients with PCa (any stage) were eligible.

Results

Of 469 papers, 35 were included in the review. The included studies evaluated cyproterone acetate, estrogen or estrogen derivatives, progesterone derivatives, selective serotonin reuptake inhibitors (SSRIs), gabapentin, oxybutynin, and clonidine, as well as non‐pharmacological interventions such as acupuncture, cognitive behavioral therapy (CBT), and dietary supplements (e.g., Dong Quai/Angelica Sinensis, Serelys Homme, soy protein, and Salvia officinalis). Across all blinded pharmacological interventions, the relative reduction in hot flash frequency ranged from –21% to –84%, and the placebo effect was between –19% and –30%.

Conclusion

Hormonal agents such as cyproterone acetate and estrogen appear to be the most effective treatments, although they are associated with side effects. Non‐pharmacological options like acupuncture and CBT may offer some benefit, while dietary supplements seem to be ineffective. Future studies are needed also to evaluate newer treatments, such as fezolinetant, in this patient population.

1. Introduction

Pharmacological or surgical castration therapy is a cornerstone in the treatment of metastatic prostate cancer (PCa) and as (neo)adjuvant therapy combined with curative‐intent radiotherapy for patients with localized or locally advanced PCa [1, 2]. Vasomotor symptoms, such as hot flashes, are common side effects, affecting between 44% and 80% of patients receiving pharmacological castration therapy [3] and the frequency and severity of hot flashes tend to continue through out the treatment period (follow‐up between 1 and 5 years) [4, 5, 6]. Patients undergoing surgical castration usually start experiencing hot flashes after 3 months post‐surgery lasting about 30 months [7]. The symptoms can significantly impact quality of life and sleep quality [3, 8, 9] and may lead to early discontinuation of treatment [10]. Younger age and low or increased body mass index (BMI) are associated with an increased frequency of hot flashes [4, 11]. Symptoms range from a sensation of warmth to flushing, palpitations and severe sweating. The pathogenesis is unknown, but the most widely accepted theory involves the drop in testosterone levels, altering neurotransmitter function and ultimately leading to changes in thermoregulation within the hypothalamus [12, 13]. Treatment of hot flashes in men is not well‐studied, and most treatment strategies are based on research in women, despite uncertainties regarding whether these treatments are equally effective in men [14].

Pharmacological castration therapy/androgen deprivation therapy (ADT) involves subcutaneous or intramuscular administration of luteinizing hormone‐releasing hormone (LHRH) agonists or antagonists every 4–24 weeks. Recently, daily oral administration of LHRH antagonist has also become available [15]. Depending on the duration of ADT and type of treatment testosterone levels may return to normal 6–22 months after discontinuation of LHRH agonists, although approximately 20% of patients never recover testosterone production [16, 17]. For LHRH antagonists, the testosterone recovery is between 1 and 6 months after discontinuation [15, 18]. Further, hot flashes are more common with LHRH antagonists than with agonists [19].

The aim of this systematic review was to evaluate the currently available evidence on pharmacological and non‐pharmacological interventions for treating of hot flashes in men with PCa receiving ADT, and to provide evidence‐based treatment recommendations as well as to identifying current gaps in the existing knowledge.

2. Methods

This review was conducted to address the following research question: “What are the effects on frequency and severity of hot flashes in men with PCa receiving ADT following different pharmacological and non‐pharmacological therapies.” The review was conducted and reported in accordance with the Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) guideline from 2020 and checklist [20].

2.1. Eligibility Criteria and Sources of Information

Eligible criteria included systematic reviews, placebo‐controlled studies, clinical intervention studies, and case reports involving men with PCa (all stages) who had undergone castration therapy. Narrative reviews, trial protocols, letter/comments, and conference abstracts were excluded. Additional eligibility criteria were studies that mentioned pharmacological or non‐pharmacological (nutritional supplements and acupuncture were included, but hypnosis and thermal devices were excluded) interventions for treating hot flashes as well as presenting an estimate of the frequency and/or severity of hot flashes.

2.2. Search Strategy

The literature search consisted of two searches that both was conducted in PubMed using the keywords (“hot flashes” OR “hot flushes”) AND (“prostate”) for the first search and (“hot flash*” OR “hot flush*” OR “vasomotor*”) AND (“prostate”) for the second search, supplemented by reference searches from eligible systematic reviews. Only literature written in English or Danish was included.

2.3. Selection and Data Collection Process

Articles were reviewed by a single investigator (ES) using PubMed's article tagging system and Endnote reference software. First a screening of title and abstracts was conducted and then a full‐text screening of studies identified during the first screening was performed.

2.4. Data Collection and Data Items

Data collection was performed by a single reviewer (ES). Data items collected included the following: first author, year of publication, study design, number of patients, time of follow‐up, treatment type, treatment administration (oral of transdermal), daily dose, measure of hot flashes (frequency, severity or in combination presented as a hot flash score (frequency x severity) if this was the only available measure, baseline data, follow‐up data for frequency and severity, and adverse events.

If only abstracts were available for original literature, the abstracts were included if they provided information on number of patients, treatment characteristics and results. We did not include abstracts on systematic reviews.

2.5. Risk‐of‐Bias Assessment

Placebo‐controlled double‐blind parallel or crossover trials were assessed for risk‐of‐bias (ROB) by using a visual tool (ROBvis tool) available as an open‐source Shiny web application [21] by a single reviewer (ES). Within the ROBvis template, the Cochrane ROB 2 tool [22] was used, applying the correct risk‐of‐bias domain names to the figures, and label the judgment levels appropriately (e.g., “Low,” “Some Concerns,” and “High”). Further, a narrative summary of the findings and describing the cohort studies is provided in the main text.

2.6. Synthesis Methods

We presented the selected data items in tables and addressed any missing data as not applicable (N/A).

The studies were categorized according to treatment type (i.e., pharmacological or non‐pharmacological as well as by specific treatment type), and by level of evidence. We did not make a meta‐analysis, but effectiveness on frequency and impact on quality of life was standardized by calculating the relative reduction in frequency and severity in percentage compared to baseline. Stata for Windows (version 11.2; IMB Corp., Armonk, NY) was used for data analysis.

In addition, the most common and serious side effects of the treatments were described. Finally, other medical interventions not yet studied in men with PCa but with prior or anecdotal evidence, were presented in the text but not reported in the tables.

3. Results

The initial search was conducted in October 2024 identified a total of 414 citations published between 1983 and 2023. To ensure an updated overview of literature, an additional search identifying studies published between 2023 and January 2026 was performed in January 2026. After title and abstract screening 55 potentially relevant studies were identified and reviewed in full text. Of these, 34 articles were eligible (Flowcharts of the two searches can be found in Supporting Information: Figures S1 and S2). Also four systematic reviews [23, 24, 25, 26] were identified and by looking through the reference lists from the eligible full text papers and the reviews additional 1 studies were identified; thus, 35 studies were included for analysis, 23 evaluated pharmacological interventions, 11 evaluated non‐pharmacological treatments, and 1 evaluated both pharmacological and non‐pharmacological interventions. Table 1 summarizes the primary studies by type of medical intervention. The published studies evaluated cyproterone acetate [27, 28], estrogen or estrogen derivatives [29, 30, 31, 32], progesterone derivatives [27, 33, 34, 35, 36, 37], selective serotonin reuptake inhibitors (SSRIs) [10, 14, 27, 38, 39, 40, 41], gabapentin [42, 43, 44], oxybutynin [45, 46], and clonidine [47, 48, 49, 50]. Non‐pharmacological interventions (Table 2) included acupuncture [51, 52, 53, 54, 55, 56], cognitive behavioral therapy (CBT) [57], and nutritional supplements (Dong Quai/Angelica Sinensis [58], Serelys Homme [59], soy protein [14, 60], and Salvia Officinalis [61]). Reasons for study exclusion were wrong patient population, wrong design, non‐English or non‐Danish language (one study in Japanese), not mentioning treatment of hot flashes, and wrong publication type.

Table 1.

Pharmacological interventions in treatment of hot flashes in men with prostate cancer receiving androgen deprivation therapy.

First author, year, design,		Patients, N	Follow‐up, weeks	Daily dose	HF measures	Frequency, mean ± SD (or otherwise stated)			Severity, mean ± SD (or otherwise stated)			Adverse events
						Baseline, HF/day	Follow‐up	Relative reduction (%)	Baseline, test	Follow‐up	Relative reduction (%)
Cyproterone acetate
Irani et al., 2010, double‐blind, randomized, no control, 3 different treatments		100	8	100 mg	Frequency + severity (hot‐flash score)		Included in hot‐flash score		Median: 6.5	Median: 0.0, p < 0.001	–100%	Gastrointestinal, nausea, edema, venous thrombosis, head ache, weight gain, and tiredness.
Eaton et al., 1983, placebo, crossover, double‐blind		12	3, 1 week wash out	100 mg × 3	Frequency	No baseline registrations.	Cyproterone: 2.3 ± SEM: 0.99, Placebo: 9.4 ± SEM: 1.6, p < 0.001	N/A	N/A			Tiredness.
Estrogen or estrogen derivatives
Gerber et al., 2000, randomized, crossover, high or low dose estrogen		12	4, 4 weeks wash out	Transdermal, 0.05 or 0.1 mg twice a week	Frequency & Severity, VAS 1–10	6.9	Low dose: 5.3, vs baseline p = 0.09 High dose 4.4, vs baseline p = 0.02	Low dose: –23% High dose: –36%	VAS: 6.5	Low dose: 4.8, vs baseline p = 0.02 High dose 4.5, vs baseline p = 0.02	Low dose: –26% High dose: 31%	Breast tenderness, gynecomastia.
Russel et al., 2022, placebo, randomized, double‐blind, study		39/39	24	Estradiol gel 0.9 mg 0.1%	Frequency & Frequency + severity (hot‐flash score)	Median (IQR) Estradiol gel: 4.1 (1.4; 6.7) Placebo: 3.4 (1.2; 7.1)	Median (IQR) Estradiol gel: 0.7 (0.0; 2.7) Placebo: 2.7 (0.0; 6.8) Mean adjusted difference: –1.6 (95% CI: –2.7; –0.5), p = 0.04.	Estradiol gel vs baseline: –85% Placebo vs baseline: –21%	Median (IQR) Estradiol gel: 33 (11.2; 68.8) Placebo: 28 (10; 79.5)	Median (IQR) Estradiol gel: 5.0 (0.0; 19) Placebo: 21.5 (0.5; 69.2) Mean adjusted difference: –19.6 (95% CI: –35.5; –3.8), p = 0.11.	Estradiol gel vs baseline: –85% Placebo vs baseline: –23%	Breast tenderness, gynecomastia
Zimmerman et al., 2022, placebo, double‐blind, parallel study		41/21	24	Estetrol 40 mg Prophylactic, i.e., treatment initiated at start of ADT	Frequency & Frequency + severity (hot‐flash score)	Estetrol: 0 HF/week Placebo: 0 HF/week	Estetrol: 0.6 ± 1.61 Placebo: 22.5 ± 46.22 (no p‐value or similar was presented)	N/A	Estetrol: 0 Placebo: 0	Estetrol: 0.1. ± 0.3 Placebo: 3.7 ± 7.3 (no p‐value or similar was presented)	N/A	Breast tenderness, gynecomastia, no treatment‐related cardiovascular events
Progesterone og progesterone derivatives
Sartor et al., 1999, case		1	72	Megestrol acetate 20 mg twice	Frequency & Severity	N/A	Effect: yes, but only descriptive	N/A	N/A	N/A	N/A	PSA increase
Langenstroer et al., 2005, cohort.		48	N/A	Intra‐muscular injection of medroxyprogesterone acetate 400 mg (n = 40) or 150 mg (n = 8) every 10th month	Frequency & Severity: mild: 1, moderate: 2, Severe: 3	Low dose: 5.25 High dose: 4.63	Low dose: 2.88 High dose: 1.40No statistically significant differences btw. doses, but vs baseline p < 0.05	Low dose: –45% High dose: –69%	Low dose: 2.75 High dose: 2.08	Low dose: 1.75 High dose: 0.56 No statistically significant differences btw. doses, but vs baseline p < 0.05	Low dose: –36 High dose: –73	N/A
Irani et al., 2010, double‐blind, randomized, no control, 3 different treatments		108	8	Medroxyprogesterone acetate 20 mg	Frequency + severity (hot‐flash score)		Included in hot‐flash score		Median: 9.6	Median: 0.3, p < 0.001	–97%	Gastrointestinal, nausea, edema, venous thrombosis, headache, weight gain, and tiredness.
Koike et al., 2013, cohort		32	8	Chlormadinone acetate start dose 100 mg, if effective, dose was decreased to 50 mg and then to 25 mg	Presence	32 patients with HF	15 patients with HF	–84%	N/A	N/A	N/A	Hyperhidrosis, ALAT and ASAT increase.
Loprinzi et al., 1994, placebo, crossover, double‐blind		30/30	4, no washout	Megestrol acetate 20 mg × 2	Frequency & Severity	Megestrol acetate4–9: 39% ≥ 10: 61% Placebo: 4–9: 36% ≥ 10: 58%	First treatment period only due to significant carryover effects: % of baseline daily average Megestrol acetate: 20% Placebo 81% p< 0.001	Megestrol acetate: –80% Placebo: –19%	N/A	First treatment period: % of baseline daily average Megestrol acetate: 13% Placebo 84 % p< 0.001	Megestrol acetate: –87% Placebo: –16%	N/A
Ronzoni et al., 1998, cohort		37	N/A	Cyproterone acetate or medroxyprogesterone acetate. Only abstract available	“Therapeutic efficacy”	N/A	80% cyproterone acetate70% medroxyprogesterone acetate	N/A	N/A	N/A	N/A	N/A
Selective serotonin reuptake inhibitors
Roth et al., 1998, case		5 cases	4–6	Sertraline 25 – 125 mg	Frequency	N/A	Reduced frequency, e.g., from 2 to 1 min pr HF	N/A	N/A	N/A	N/A	Nausea, gastrointestinal symptoms
Quella et al., 1999, pilot, cohort		16	4	Venlafaxine 12.5 mg twice daily	Frequency & Frequency + severity (hot‐flash score)	10 (range: 0.7; 33.9)	6 (range: 0; 15)	–46%	N/A	11 patients had < 60% of baseline hot‐flash score	N/A	Sleep disturbances, dryness of mouth, tiredness, dizziness, nausea,
Loprinzi et al., 2004, pilot, cohort		24	5	Paroxetine start dose: 12.5 mg increased to 37.5 mg	Frequency & Frequency + severity (hot‐flash score)	Median: 6.2	Median: 2.5	–60%	Median: 10.6	Median: 3.0	–72%	Well tolerated
Naoe et al., 2006, cohort		10	4	Paroxetine 10 mg	Frequency & Severity (severity grade; 1: not at all, 5: intermediate, 10: extremelysevere)	3.5	2.0 p = 0.009	–43%	4.6	2.0 p = 0.033	–57%	Tiredness, dryness of mouth,
Irani et al., 2010, double–blind, randomized, no control, 3 different treatments		102	8	Venlafaxine 75 mg	Frequency + severity (hot‐flash score)		Included in hot‐flash score		Median: 6.9	Median: 3.6, p < 0.001	–57%	Gastrointestinal, nausea, edema, venous thrombosis, headache, weight gain, and tiredness.
Vitolins et al., 2013, placebo controlled, double–blind, parallel study		30/30/30	12	Venlafaxine 75 mg 20 g Soy protein	Frequency & Severity (1, mild; 2, moderate; 3, severe)	Venlafaxine + soy protein: 9.8 ± 8.0 Venlafaxine + milk protein: 8.9 ± 5.2 Placebo: 8.6 ± 4.4	Venlafaxine + soy protein: 5.5 ± 5.0Venlafaxine + milk protein: 4.8 ± 2.7 Placebo: 5.9 ± 5.1 p > 0.05 (all, ANCOVA)	Venlafaxine + soy protein: –44% Venlafaxine + milk protein: –36% Placebo: –30%,	Venlafaxine + soy protein: 2.1 ± 0.4 Venlafaxine + milk protein: 2.3 ± 0.5 Placebo: 2.4 ± 0.5	Venlafaxine + soy protein: 1.6 ± 1.0 Venlafaxine + milk protein: 1.7 ± 0.7 Placebo: 1.8 ± 0.8 p > 0.05 (all, ANCOVA)	Venlafaxine + soy protein: –24% Venlafaxine + milk protein: –26% Placebo: –25%	Few side effects, not specified
Nishiyama et al., 2004, cohort		12/20	2	Fluvoxamine maleate 25 mg in patient who wishes to be treated for HF (n = 12)	Frequency & Frequency + severity	Chose treatment: 9.3 ± 5.6	Chose treatment: 1.5 ± 2.2, p = 0.001	–84%	N/A	Median (IQR) Chose treatment: 8 (2;21) No treatment: 3 (1; 16), p = 0.007	N/A	None
Gabapentin
Jeffery et al., 2002, case		1	4	600 mg	Frequency	Daily HF	1 HF/week	N/A	N/A	N/A	N/A	N/A
Loprinzi et al., 2009, double–blind, placebo–controlled		54/53/54/54	4	300, 600, or 900 mg	Frequency & Frequency + severity	N/A	Median % change (95% CI): 300 mg: 22.8 (12.1; 33.0) 600 mg: 31.8 (16.5; 40.5) 900 mg: 45.5 (31.1; 50.6) Placebo: 21.5 (11.3; 30.9) p = 0.19 for gabapentintotal vs placebo	See results column	N/A	Median % change (95% CI): 300 mg: 29.7 (13.1; 36.9) 600 mg: 33.8 (22.2;47.1) 900 mg: 44.4 (35.2; 56.3) Placebo: 27.0 (12.1; 36.1) p = 0.46 for gabapentintotal vs placebo, p = 0.02 for 900 mg vs placebo	See result column	Nausea, obstipation
Moraska et al., 2010, cohort		117	8	300 – 900 mg	Frequency & Frequency + severity (hot–flash score)	N/A	Sustained reduction of frequency	N/A	N/A	Sustained reduction of hot–flash score	N/A	None, well tolerated
Other
Oxybutynin
Smith et al., 2018, case		1	6	1.25 mg increasing to 2.5 mg × 2	Frequency	N/A	Subjective improvement	N/A	N/A	N/A	N/A	Sleep disturbances, dryness of mouth, restless legs syndrome,
Stish et al., 2025, double‐blind, randomized		88	6	2.5 or 5 mg × 2	Frequency & Frequency + severity (hot‐flash score)	Placebo: 9.9 ± 7.2 2.5 mg: 10.3 ± 5.5 5 mg: 10.3 ± 3.9	Mean reduction HF/day from baseline: Placebo: 2.15 2.5 mg: 4.77 (p = 0.02 compared to placebo) 5 mg: 6.89 (p < 0.001 compared to placebo)	N/A	Placebo: 17.5 ± 16.1 2.5 mg: 17.5 ± 11.6 19.7 ± 13.0	Mean reduction hot‐flash score from baseline: Placebo: 4.85 2.5 mg: 9.94 (p = 0.07 compared to placebo) 5 mg: 13.95 (p < 0.02 compared to placebo)	N/A	Dry mouth
Clonidine
Parra et al., 1990, cohort		7	N/A	Transdermal 0.1 mg/week	Frequency	N/A	HF ablated in 3 patients, considerably reduced in 4 patients	N/A	N/A	N/A	N/A	No significant side effects
Bressler et al., 1993, pilot, case report		4	min 4	Transdermal or oral 0.1–0.2 mg in increasing dosage until effect or adverse events (0.1–0.3 mg)	Frequency & Severity	N/A	Symptomatic improvement, e.g., from 1 HF per hour to 1 pr 2 h	N/A	N/A	N/A	N/A	Adverse effects in one patient (sedation, dry mouth local irritation), but continued treatment
Loprinzi et al., 1994 Double‐blind, placebo, randomized, crossover, only abstract		70	N/A	N/A	Frequency & Severity	N/A	No effect	N/A	N/A	No effect	N/A	N/A
Smith et al., 1994, cohort, descriptive study		68 (different treatments)	N/A	Unclear number of patients treated with clonidin	Frequency & Severity	N/A	Clonidin: No effect	N/A	N/A	Clonidin: No effect	N/A	N/A

Abbreviations: ALAT, alanine aminotransferase; ANCOVA, analysis of covariance; ASAT, aspartate aminotransferase; Daily hot‐flash score: Number of hot flashes × average severity (mild = 1, moderate = 2, severe = 3, very severe = 4); HF: hot flashes; MAD, mean adjusted difference; N/A: not applicable; SEM: standard error of the mean; SSRI: selective serotonin reuptake inhibitor; VAS: visual analog scale (1 = least severe to 10 = most severe).

Table 2.

Non‐pharmacological interventions and nutritional supplements in treatment of hot flashes in men with prostate cancer receiving androgen deprivation therapy.

First author, year, design	Patients, N	Follow‐up, weeks	Daily dose	HF measures	Frequency, mean ± SD (or otherwise stated)			Severity, mean ± SD (or otherwise stated)			Adverse events
					Baseline, HF/day	Follow‐up	Relative reduction (%)	Baseline	Follow‐up	Relative reduction (%)
Acupuncture
Hammar et al., 2009, cohort	7	24	30 min, 1–2 times/week	Frequency	7.9	3.6 HF/ day, p = 0.004 (ANOVA)	–54 %	N/A	N/A	N/A	N/A
Frisk et al., 2009, cohort	31	4–52 weeks	Electrostimulated (n = 15) and traditional acupuncture (n = 16) 30 min 1–2 times/week in 12 weeks	Frequency & Severity	Median (IQR) Electrostimulated: 7.4 (5.5; 12.9) Traditional: 6.4 (5.2; 9.4)	Median (IQR) Electrostimulated: 12 weeks: 4.1 (2.0; 6.5), p = 0.002 52 weeks: 6.2 (4.2; 6.5), p = 0.27 Traditional 12 weeks: 3.4 (1.8; 6.3), p = 0.001 52 weeks: 4.1 (2.7; 5.2), p = 0.009 No difference between groups, p = 0.65.	Electrostimulated: 12 weeks: –45 % 52 weeks: –16 % Traditional: 12 weeks: –47 % 52 weeks: –36 %	Median (IQR) Electrostimulated: 8.2 (6.5; 10.7) Traditional: 7.6 (4.7; 8.3)	Median (IQR) Electrostimulated: 12 weeks: 3.3 (0.3; 8.1), p = 0.003 52 weeks: 5.5 (3.8; 6.9), p = 0.016 Traditional: 12 weeks: 3.4 (2.0; 5.6) p = 0.001 52 weeks: 4.3 (1.3; 7.7) p = 0.036	Electrostimulated: 12 weeks: –60 % 52 weeks: –33 % Traditional: 12 weeks: –55 % 52 weeks: –43 %	No serious side effects recorded
Harding et al., 2009, pilot, cohort	60	10	Auricular acupuncture, weekly	Frequency & Severity (0 no, 6 max intense)	Daytime 7.2 ± 4.9 Night–time 6.3 ± 3.9	Daytime 2.2 ± 2.1 p < 0.05, ANOVA Night–time 1.9 ± 1.4 p < 0.05, ANOVA	Daytime: –69 % Night–time: –70 %	Daytime 3.2 ± 0.8 Night–time 4.3 ± 0.9	Daytime 1.6 ±1.4 p < 0.05, ANOVA Night–time 1.6 ± 1.3 p < 0.05, ANOVA	Daytime: –50 % Night–time: –63 %	Minimal adverse events
Beer et al., 2010, cohort	22	8	1–2 times/week	Frequency + severity (hot–flash score)	N/A	N/A	N/A	16.2 ± 12.3/day	52 % of baseline	–48 %	No clinically significant adverse events
Ashamalla et al., 2011, cohort	14	24	2 times/week	Frequency + severity (hot–flash score)	N/A	N/A	N/A	28.3 ± 29.3	7.0 ± 8.4 p = 0.001	–75 %	None
Hirsch et al., 2015, cohort	7	6	2 times/week	Frequency & Severity (1 mild, 2 moderate, 3 severe)	N/A (at least 10)	> 50 % reduction from baseline	N/A	N/A (at least moderate)	Grade 2 or grade 1 (not specified)	N/A	N/A
Cognitive behavioral therapy
Stefanopoulou et al., 2015, randomized, parallel study	33/35	6 + 32	Self–help CBT, 4 weeks vs treatment as usual	Frequency +Severity (0–10, 10 worst)	CBT: 56.09 ± 30.16 HF/week Control: 52.95 ± 50.07	6 weeks: CBT: 36.06 ± 25.03/week Control: 44.58 ± 38.7/week, p = 0.05 (adjusted) 32 weeks: CBT: 32.11 ± 24.5/week Control: 42.08 ± 54.9/week p > 0.05 (adjusted)	6 weeks: CBT: –36 % Control: –13 % 32 weeks: CBT: –43 % Control: –21 %	CBT: 4.5 ± 1.98 Control: 4.8 ± 2.4	6 weeks: CBT: 2.76 ± 1.53 Control: 4.19 ± 2.2 p = 0.001 (adjusted) 32 weeks: CBT: 2.66 ± 1.67 Control: 3.33 ± 1.84 p > 0.05 (adjusted)	6 weeks: CBT: –38 % Control: –13 % 32 weeks: CBT: –40 % Control: –31 %	None
Nutritional supplements
Dong Quai (Angelica Sinensis)
Al–Bareeq et al., 2020, placebo controlled, randomized, double–blind, parallel study	11/11	12	500 mg	Frequency & Severity (0–10)	Dong Quai: 6.80 ± 4.33 Placebo: 3.42 ± 1.42 p = 0.02	Dong Quai: 6.9 ± 4.1 Placebo: 2.84 ± 1.9 Between group: p = 0.02 Either group: p > 0.05	Dong Quai: 1 % Placebo: –18 %	Dong Quai: 3.76 ± 1.67 Placebo: 3.64 ± 1.37 p = 0.79	Dong Quai: 3.71 ± N/A Placebo: 2.47 ± N/A Between group: p> 0.05	Dong Quai: –1 % Placebo: –32	None
Serelys Homme
Belkacemi et al., 2023, cohort	46	24	2 tablets of 160 mg	Frequency & Severity (0–4)	≥ 7: 21.7 %	≥ 7: 8 %, p < 0.05	–63 %	Moderate, severe or very severe (2–4): 47.9 %	Moderate, severe or very severe: 24 % p = 0.05	–50 %	None
Soy protein
Sharma et al., 2009, randomized, placebo controlled, double–blind	17/16	12	20 g (170 mg isoflavones)	Severity, Blatt–Kupperman scale (vasomotor score)	N/A	N/A	N/A	Soy protein: 7 ± 1 Placebo: 4.5 ± 0.6 p = 0.05	Soy protein: 7 ± 0.9 Placebo: 4.5 ± 0.7 Between groups: p = 0.04 Within soy protein group: p = 0.90	Soy protein: 0 % Placebo: 0 %	None
Vitolins et al., 2013, randomized, placebo controlled, double–blind	30/30	12	20g (160 mg isoflavones)	Frequency & Severity.	Soy protein: 10.0 ± 4.2 Placebo: 8.6 ± 4.4	Soy protein: 5.9 ± 5.1 Placebo: 4.1 ± 3.1 Between groups: p > 0.05 Within groups: p < 0.001	Soy protein: –41 % Placebo: –52 %	Soy protein: 2.3 ± 0.5 Placebo: 2.4 ± 0.5	Soy protein: 1.8 ± 0.9 Placebo: 1.8 ± 0.8 Between groups: p > 0.05 Within groups: p < 0.001	Soy protein: –27 % Placebo: –25 %	None definitely related to treatment
Salvia officinalis
Vandecasteele et al., 2012, cohort	10	4	150 mg 3 times	Frequency & Frequency + severity (hot–flash score).	58 ± 40 HF/week	33 ± 23 HF/week, p = 0.0001	–42 %	112 ± 71	58 ± 50 p = 0.018	‐48 %	Skin eruption (n = 1).

Abbreviations: ANOVA, analysis of variance; Blatt–Kupperman scale, Used for rating menopausal symptoms, both self‐reported and physician‐assessed, including hot flashes; CBT, Cognitive behavioral therapy; Daily hot‐flash score, Number of hot flashes × average severity (mild = 1, moderate = 2, severe = 3, very severe = 4); HF, hot flashes; N/A, not applicable.

Study designs ranged from case reports to randomized placebo‐controlled trials, with follow‐up periods ranging from three to 72 weeks and sample sizes between one and 117 patients. Methods for assessing hot flashes varied across studies, including symptom diaries, hot flash scales, and quality‐of‐life questionnaires. Three studies included only patients treated with orchiectomy [28, 62, 63], while the remaining studies included patients primarily treated with ADT. Treatment was initiated after the onset of hot flashes symptoms except in one study, where treatment was given with prophylactic intent [32].

3.1. Risk of Bias

As presented in Figure 1, the risk of bias varied across the randomized studies, where some had overall low risk of bias whereas others had high risk, mainly due to side‐effect unmasking the blinding. The risk of selection bias in the cohort studies was high. Bias due to missing data was low, but in some of the studies there was a high risk of detection bias due to variations of reporting frequency of hot flashes. The case reports had high risk of bias.

Figure 1.

Visualization of risk of bias assessment of the randomized controlled trials. [Color figure can be viewed at wileyonlinelibrary.com]

3.2. Pharmacological Treatment

Statistically significant reductions in both frequency and impact on daily life compared to placebo were observed for cyproterone acetate, estrogen, progesterone derivatives, and oxybutin [28, 30, 32, 36, 46]. In contrast, SSRIs, clonidine, oxybutin, and gabapentin did not demonstrate significant differences compared to placebo [14, 43, 46, 49]. Placebo effects ranged from –19% to –30% [14, 30, 36, 64]. Across all pharmacological interventions, including open studies, the relative reduction in hot flash frequency ranged from –21% to –84%.

3.3. Non‐Pharmacological Treatment

For all non‐pharmacological interventions, the relative reduction in frequency ranged from ‐16% to ‐70%. Acupuncture may be effective, although controlled studies are lacking [51, 52, 53, 54]. One controlled study of CBT showed short‐term benefits after 6 weeks but no difference after 32 weeks compared to usual care [57]. Dietary supplements and soy protein had limited to no effect, with no differences between intervention and placebo [14, 58, 59, 60, 61].

3.4. Diethylstilbestrol

Studies on diethylstilbestrol are not included because the drug has been withdrawn [62, 63].

3.5. New Drugs Not Studied in Men

Fezolinetant is a non‐hormonal selective neurokinin‐3 receptor antagonist that reduces the frequency and severity of hot flashes by blocking specific receptors in the thermoregulatory center of the hypothalamus. In postmenopausal women, it has been observed to cause a transient decrease in LH without affecting follicle‐stimulating hormone, testosterone, estrogen, or dehydroepiandrosterone. The Institute for Rational Pharmacotherapy in Denmark (IRF) does not recommend fezolinetant as a first‐line treatment for vasomotor symptoms in women due to limited experience with the drug and its high cost compared to hormonal alternatives. Few publications have suggested that fezolinetant may be used to treat hot flashes in prostate cancer and calls for further studies [65, 66, 67]. To date, two studies have been listed on clinicaltrials. gov, and one case study have demonstrated positive findings of fezolinetant in reducing vasomotor symptoms in a man with PCa receiving ADT [68].

4. Discussion

The treatment of hot flashes in men with PCa is not well‐studied. Most treatment approaches are based on research on hot flashes in women or based on standard clinical practice within individual departments. Additionally, there is likely a significant publication bias, as only two studies on medical treatment for hot flashes have reported negative outcomes [14, 43]. The placebo effect ranges between 19% and 30%, suggesting that a substantial proportion of the observed treatment benefit in uncontrolled studies may not be attributable to the intervention itself. As hot flashes are patient‐reported symptoms that are susceptible to expectation and reporting biases, treatment efficacy may easily be overestimated in the absence of a placebo control group underlining the need for placebo‐controlled studies.

Estrogen preparations appear to be the most effective, although only three studies have been published on men, with the remaining studies based on studies of women or mixed populations [13, 69]. Also, a recently published randomized trial investigating treatment of men with locally advanced prostate cancer with transdermal estradiol or LHRH agonists, estradiol was just as effective as ADT and was associated with a lower risk of hot flashes, thus underlining the possible untapped potential of estrogen [70].

In placebo‐controlled studies on SSRIs and gabapentin, no significant differences from placebo were found. In a study comparing cyproterone acetate, progesterone, and venlafaxine, venlafaxine was found to be less effective than the other two drugs [27].

Overall, nutritional supplements showed no effect, while cognitive behavioral therapy may have some benefit, particularly in the early phase. Also taking the large placebo effect into consideration, we suggest that psychological interventions such as cognitive behavioral therapy may have great potential as ameliorating hot flashes, although further studies are needed. Acupuncture may be a promising non‐pharmacological treatment, although a systematic review concluded it to be ineffective [71]. There is a need for controlled studies, and despite methodological challenges using sham acupuncture, one such study is currently underway [72].

Other reviews investigating pharmacological treatment of hot flashes concludes that more studies are needed and that hormonal treatments (estrogens, megestrol acetate, medroxyprogesterone, diethylstilbestrol, cyproterone acetate) are considered the most effective treatments [23, 25, 73, 74, 75, 76, 77, 78]. Further, the reviews focus on side effects associated with the interventions.

Regarding non‐pharmacological treatments, a systematic review on acupuncture concluded that acupuncture is ineffective; however, the quality of the studies included in the analysis were low, leaving the potential effect uncertain [24], another review concluded that acupuncture is popular but the effect is inconclusive [71] and finally, Qan'ir et al., suggested acupuncture as a potential tool in reducing hot flashes without side effects [78].

Treatment decisions should involve thorough patient counseling and shared decision making processes, focusing on balancing between efficacy and potential side effects aiming for the lowest effective dose.

4.1. Side Effects

In general, side effects such as weight gain, tiredness, edema, gastrointestinal side effects like constipation, diarrhea and nausea are frequent and thus experienced in between 1% and 10% of patients [79]. For estrogen derivatives, breast tenderness and gynecomastia are also relevant side effects. Hyperhidrosis, dizziness and an increase of alanine aminotransferase and aspartate aminotransferase are also observed for progesterone derivatives. Below is a discussion of selected specific side effects categorized by drug.

4.2. Cyproterone Acetate: Risk of Meningioma

In 2020, the European Medicines Agency (EMA) published a referral highlighting the increased risk of meningioma with use of cyproterone acetate, although they did not change the recommendations for PCa treatment (as an anti‐androgen indication or flare prevention) [80]. A recent review from The Danish Pharmaceutical Information reported that the absolute risk of meningioma is low (0.01%–0.1%), but the relative risk increases with the dose and duration of cyproterone acetate [81]. For example, 6 months of treatment of cyproterone acetate is estimated to increase the relative risk by 6, while the absolute risk is approximately 1 per 1,000. The dose should be reduced to the lowest effective level, and cyproterone acetate is contraindicated in patients with a history of meningioma.

4.3. Estrogen: Thromboembolic Events

No thromboembolic events related to estrogen treatment were reported in the current studies. Oral estrogen therapy has been associated with an increased risk of venous thromboembolism. Routine thromboprophylaxis is not recommended in patients receiving oral estrogen therapy; however, management should be individualized in patients with additional major risk factors for venous thromboembolism [82].

4.4. Progesterone Derivative (Medroxyprogesterone Acetate): Risk of Meningioma

Recently, Pfizer, the Danish Medicines Agency, and the EMA have published information on the increased risk of meningioma with high doses of medroxyprogesterone acetate (all injectable and oral formulations ≥ 100 mg), particularly after several years of use [83]. In current studies on hot flashes in PCa, doses ranged from 20 to 100 mg/day administered either intramuscularly or orally. It is recommended that the dose of medroxyprogesterone acetate remain below 100 mg/day.

4.5. Progesterone Derivative (Megestrol Acetate)

Pulmonary embolism and thrombophlebitis are listed as very common side effects [79]. Additionally, a case report observed an increase in PSA levels following megestrol acetate treatment, with a subsequent decrease after discontinuation [33]. However, this finding was later deemed insignificant [84], and studies involving larger patient groups have shown decreases in PSA levels [85, 86]. Given the potential risks, the possible benefits of megestrol acetate do not outweigh its side effects.

5. Limitations

Since the literature search was conducted to be used in the writing of a local treatment guideline we did not make a PROSPERO registration prior to the search. Only having one reviewer, as well as the use of simple search terms and only including one database are also limitations increasing the risk of missing relevant studies. Based on the different measures of hot‐flashes (e.g., frequency reported as median, mean or as a combined hot‐flash‐score), we decided not to include any quantitative synthesis of the results.

6. Conclusion

Hot flashes due to castration therapy for PCa are common and influence quality of life. There is limited evidence supporting interventions reducing vasomotor symptoms. Placebo effects are substantial, emphasizing the need for placebo‐controlled trials. Hormonal agents such as cyproterone acetate and estrogen appear to be the most effective and results on studies on neurokinin‐3 receptor antagonists are awaiting. Non‐pharmacological options like acupuncture and CBT may offer some benefit, while nutritional supplements are ineffective.

Ethics Statement

This study synthesizes findings from existing literature and does not involve direct interaction with patients or collection of primary data. As such, ethical approval was not required for this study.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Flowchart of literature search performed in October 2024.

Figure S2: Flowchart of literature search performed in January 2026.

Data Availability Statement

The datasets generated during and/or analyzed during this study are available from the first author upon reasonable request.