Melatonin supplementation for quality of life in older patients with advanced cancer: a randomized controlled trial Authors: Angeline Ginzac, Emmanuelle Bourbouloux, Romain Rivoirard, Christelle Jouannaud, Marie-Odile Hager, Sophie Dubois, Fabrice Kwiatkowski, Ioana Molnar, Emilie Thivat, Xavier Durando Published: 2025 ABSTRACT BackgroundMaintaining quality of life (QoL) for older cancer patients (≥ 70 years) undergoing systemic treatment is challenging. Melatonin supplementation could be a potential strategy.AimsThe MEQAPAG trial investigated the effects of melatonin supplementation on this specific population’s quality of life.MethodsA multi-centre, double-blind, randomized placebo-controlled trial was implemented. Patients in the interventional arm received a daily dose of melatonin (2 mg/j for 90 days, 1–2 h before bedtime and after a meal) for the first three months of a new systemic treatment line. The primary endpoint was QoL assessment using the EORTC Quality of Life Questionnaire Core 30 (QLQ-C30) before and after supplementation. Among the secondary objectives, we evaluated the impact of supplementation on sleep and on urinary melatonin concentrations. We also studied safety and survival.ResultsOne hundred and twenty-three patients were included in the study between July 2015 and January 2021. The trial was prematurely terminated due to futility. Only 61% (n = 59) of the patients were assessable and considered as compliant (i.e., at least two months of supplementation), 30 and 29 were allocated to the experimental and control groups respectively. Median age was 76 years (range [70, 88]). No difference in the QLQ-C30 questionnaire global score was shown. After the supplementation, the aMT6s concentration is 8,4 ng/mg creatinine (range [0.15, 201], IQI [0.99, 105]) in the experimental group.ConclusionIn an older cancer population, supplementation with 2 mg melatonin was safe but did not lead to any significant improvement in the global quality of life score on the QLQ-C30 for compliant patients. MEQAPAG trial provides novelty data on urinary aMT6s concentration for older cancer patients.Trial registration numberNCT02454855, register on 05/12/2015.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12877-025-06899-1. Introduction In France, in 2018, around 120 000 patients aged over 75 were diagnosed with cancer. The therapeutic management of these patients requires age-related specificities, such as comorbidities, poly-medication and frailty to be taken into account [1]. Older patients combine the biological effects of ageing and those resulting from the disease itself and its treatments [2–5]. Fatigue and depression are the most common symptoms affecting older cancer patients’ quality of life (QoL) [6–8]. Preserving QoL must also be considered in the medical care of these patients. However, the multidimensional nature of QoL means that it is influenced by several factors. Several studies have shown that treatment-induced circadian disruption had a negative impact on patient QoL [9–12]. Indeed, altered circadian rhythms may be associated with symptoms related to cancer and/or treatments, such as fatigue, and may also lead to the onset of severe fatigue [13–15]. The high majority of older cancer patients experienced cancer-related fatigue (CRF) during their treatments [16]. CRF is defined by the National Comprehensive Cancer Network as a “distressing persistent, subjective sense of tiredness or exhaustion related to cancer or cancer treatment that is not proportional to recent activity and interferes with usual functioning”. By the way, CRF negatively impacts the quality of life and functioning of cancer patients [16–18]. According to the ESMO Clinical Practice Guidelines, the objective for management of CRF in older patients is to maintain the functional independence and to tailor the intervention on fatigue to each patient [16]. Furthermore, older people’s familial and/or social context can be complex, which worsens these effects and the repercussions on QoL. A multidisciplinary approach is necessary in order to maintain this specific population’s QoL, one of the major challenges in palliative care [19, 20]. Melatonin, a hormone regulating circadian rhythm, immunity, and cellular protection, decreases with age and is further reduced in cancer patients [21–27]. Melatonin could also promote apoptosis in cancer cells [28, 29]. Exogenous melatonin is used to treat circadian disruption in healthy populations. Melatonin could provide an interesting complementary treatment to standard systemic treatment in advanced cancer, particularly among older patients receiving chemotherapy for metastatic cancer, to maintain QoL [20, 30]. This population, because of the decline in melatonin observed with age, should benefit the most from this supplementation. However, older patients are underrepresented in clinical trials and there is a lack of data concerning the potential impact of an exogenous melatonin intake on QoL during treatment of cancer. The MEQAPAG trial is the first national, multi-center RCT evaluating melatonin’s effect on QoL in patients with advanced cancer over 70. Our hypothesis is that a 3-months supplementation in oral melatonin during cancer treatment could improve the mean global QoL score (QLQ-C30) of 5 points. We report here the results of this French study. Methods Study population Eligible patients were male or female (≥ 70 years old) with a solid tumour, either metastatic or advanced stage, and with an indication for systemic treatment (oral or intravenous chemotherapy, hormone therapy, targeted therapy or immunotherapy). Patients were to have a performance status ≤ 2, a mini-mental state score (MMS) ≥ 7 and at least 3 months of life expectancy. They were to be able to swallow a tablet, and to complete questionnaires in French. Patients undergoing treatment with melatonin or who had discontinued less than three months previously, and those with haematological cancer or autoimmune disease, were excluded. Patients with a known or suspected history of alcohol abuse were also excluded. Patients were eligible for inclusion at the start of a new line of treatment. All participants provided written informed consent. A total of 500 patients were expected to be enrolled: 250 in the experimental group and 250 in the control group. Four French comprehensive cancer centres were involved in the enrolment. Study design This was a national, prospective, multi-centre, randomized, double-blind, placebo-controlled phase-III study. The MEQAPAG trial was approved by the local ethics committee (Comité de Protection des Personnes Sud-Est VI) and the competent national authority (Agence Nationale des Médicaments et des produits de santé). It was registered on Clinicaltrials.gov (https://clinicaltrials.gov/study/NCT02454855, 05/12/2015). The study adheres to Consolidated Standards of Reporting Trials (CONSORT) guidelines [31]. Randomization and blinding Patients were randomly assigned to either the experimental or the control group (ratio 1:1). Randomization was stratified on performance status (0/1/2), type of cancer (breast, prostate, colorectal, lung, other) and the participating centre. The allocation sequence was generated by computer, and block sizes remained unknown to all investigators. The treatment allocations were centralised and were concealed from the patients, the investigators, and the study coordinators enrolling the patients. Treatment/intervention The choice of the reference anticancer treatment was at the investigator’s discretion. For the first three months of the new treatment line, patients received either oral melatonin (Circadin®) or placebo supplementation. Tablets for the two groups were identical (colour, size, packaging). Supplementation intake began after an initial urine collection on J1C1 or at the latest on J1C2. Circadin® was in the form of a 2 mg sustained-release tablet. The daily dose for patients was 2 mg/j for 90 days, 1–2 h before bedtime and after a meal. No changes in dose were planned. Treatment was discontinued in case of grade 3–4 toxicity related to the supplementation. Study objectives and endpoints The primary objective of the study was to study the effects of melatonin supplementation on the quality of life of older patients with advanced cancer. QoL was assessed by the score obtained on the EORTC QLQ-C30 questionnaire for global quality of life. For the secondary objectives, the safety of and compliance with melatonin supplementation were investigated. The impact of melatonin supplementation on the different QoL scores, on changes in quality of life over one year, on sleep disorders, asthenia, pain, depressive syndrome, anorexia, evolution of autonomy, cognitive functions, and global and disease-free survival were also studied. Toxicities were assessed using NCI-CTCAE v4.0. A visual analogue scale (VAS) was used to assess asthenia and pain. Compliance with supplementation was monitored by counting the remaining tablets returned to the pharmacy after the three months and/or by checking the patients’ compliance booklets. The other endpoints were based on the interpretation of questionnaires, and are presented in the next section. Geriatric assessment A geriatric assessment was conducted for each subject at inclusion, at the beginning and the end of melatonin supplementation (J1 and M3), and at follow-up visits (M6 and M12). It was composed of several questionnaire-based assessments. Comorbidities were assessed using the Cumulative Illness Rating Score - Geriatric (CIRS-G). The scale provides information on the degree of severity of potential physical impairments, ranging from “none” (score 0) to “extremely severe” (score 4) [32]. The Folstein Mini Mental State (MMS) is composed of 30 questions divided into six categories: orientation in time and space, ability to learn and transcribe information, attention and calculation skills, ability to recall information and retain memory, language and identification skills, constructive praxis [33]. A correct answer scores 1 and a wrong or approximate answer scores 0. The final score is out of 30 points. An unauthorized version of the French MMSE was initially used, but permission was subsequently obtained from PAR [https://www.parinc.com/]. The Instrumental Activities of Daily Living (IADL) was used to assess patients’ ability to carry out daily-life activities, such as house cleaning, managing money and paying bills, or shopping for groceries [34]. The higher the overall score, the less autonomous is the patient. The maximum score is 31 points and the minimum score is 8. The Geriatric depression scale is composed of 30 questions about patients’ feelings at the time of questioning or over the past week. Response options are “yes” or “no”, which score 0 or 1 point according to the question. If the total score is lower than 9 points, the result is normal, a score falling between 10 and 19 points corresponds to mild-to-moderate depression, and a score higher than 20 points corresponds to severe depression [35]. The patients’ state of malnutrition was assessed using the Mini Nutritional Assessment [36]. This scale score has a maximum score of 30 points. A score lower than 17 points corresponds to a poorer nutritional state, and a score between 17 and 23.5 reflects a risk of undernourishment. Pain, asthenia and appetite were evaluated with a visual analogue scale ranging from 0 (no pain/asthenia/no food intake) to 10 (maximum pain/asthenia/normal food intake). Self-administered questionnaires Patients were asked to complete three self-administered questionnaires four times during the study: at the beginning and the end of melatonin supplementation, and at follow-up visits M6 and M12. QLQ-C30 v3.0, developed by the European Organization for Research and Treatment of Cancer (EORTC), is composed of 30 questions assessing five functions, nine symptoms, and patients’ overall state of health [37]. The EORTC QLQ-ELD14 is specifically designed for assessment of the health-related quality of life of older patients with cancer [38]. It is composed of 14 items assessing 5 QoL scales and 2 single items. The responses are scored from 1 (not at all) to 4 (a lot). The Leeds questionnaire is composed of 10 visual analogue scales representing several aspects of sleep efficacy: ease of getting to sleep, quality of sleep, ease of awakening from sleep, integrity of behaviour following wakefulness [39]. Scoring for each question is on a scale from 0 to 100. Blood sample Several blood parameters were monitored during the study (J1, M3, M6, M12): creatinine, total protein, albumin, pre-albumin, CRP, AST, ALT, alkaline phosphatase, ferritin, iron. Urine sample Urine samples were collected twice during the study: before (J1C1 or J1C2) and at the end of supplementation (M3). Urine samples were collected over 12 h (from 7 p.m. to 7 a.m.) in provided containers, stored at 4 °C. The total volume collected was noted and then it was aliquoted in two 5mL tubes and stored at −20 °C until analyses. The samples were transported in dry ice to the Hospices Civils de Lyon and analysed for aMT6s using radioimmunoassay [40]. aMT6s data was corrected for creatinine (ng aMT6s/mg creatinine). Others measures At each visit (inclusion, J1, M3, M6, M12), a clinical examination (vital signs, weight, WHO performance index) was conducted and any concomitant treatment or toxicity was collected. Tumour assessment using RECIST 1.1 was carried out according to standard recommendations (every three months). Statistical analysis The study was intended to randomise 250 patients per arm, and was originally powered at 95% (alpha = 0.05, two-sided test) to detect a clinically meaningful between-group difference of 5 points in the change in global QoL score from baseline to M3, assuming a standard deviation (SD) of 15 and an attrition rate of 5%. Using the observed pooled SD in the per-protocol population (≈ 23.6), while keeping the pre-specified effect size of 5 points, the partially post-hoc power [41] for the actual per-protocol sample size (≈ 30 patients per arm) is approximately 13%. Statistical analyses were performed using R software (version 4.2). Group comparisons were conducted using Wilcoxon-Mann-Whitney tests, with effect size reported as the probabilistic index (PI), which represents the probability (P(E > C)) of the outcome of a random subject in the experimental (E) group exceeding that of a random subject in the control (C) group. Linear mixed models were applied for longitudinal data. Correlation between aMT6s and QoL changes was assessed using Spearman’s rank correlation due to non-normality of both variables. Progression-free survival (PFS), defined as the time from randomization to progression or death (with censoring at the last follow-up), and overall survival (OS), defined as the time from randomization to death (with censoring at the last follow-up), were estimated using the Kaplan-Meier method. Statistical significance was set at a two-sided p-value < 0.05, with an interpretation of the results focusing on clinical relevance. For scores derived from the same questionnaire, adjustments for multiple comparisons were made using the false discovery rate. All analyses were conducted on the per-protocol population (n = 59), except for safety outcomes, which were analysed on the modified intent-to-treat (mITT) population (n = 97), defined as all randomized patients who had received at least one treatment dose. An interim analysis was planned after the hundredth patient has been treated, without recruitment interruption. The aim of this analysis was to provide the independent data monitoring committee with the necessary information on two aspects: (i) the primary endpoint: the difference between quality of life at baseline and at 3 months (futility analysis), and (ii) safety. Results The trial was prematurely discontinued following a decision made by the independent data monitoring committee and the sponsor after the interim analysis, conducted on the first hundred treated patients. Indeed, the analysis estimated that, even with 500 treated patients, the trial was unlikely to show that there was a clinically significant difference in QoL. Furthermore, the recruitment rate was lower than expected. The results presented here cover all patients included up to the decision to stop the study prematurely. Patient characteristics From July 2015 to February 2021, 123 patients were enrolled in the study. Of these, 97 were randomized (51 in the Melatonin group and 46 in the Placebo group). A per-protocol population was defined and was composed of 59 patients (30 in the melatonin arm and 29 in the placebo arm) considered as compliant, i.e., having complied with at least two months of treatment out of the three planned (i.e., 60 tablets out of 90) (Fig. 1). The baseline characteristics of the per-protocol population are presented in Table 1. The median age was 76 years (range [70, 88]). The population was composed of 36 (61%) women and 23 (39%) men. Breast cancer was the most frequent (29%) followed by lung cancer (20%). The median daily intake of tablets per patient for the treatment of comorbidities was 8. Fig. 1CONSORT flow diagram CONSORT flow diagram Table 1Baseline characteristics of the per-protocol population (n= 59)Total (n=59)Melatonin group (n=30)Placebo group (n=29)Gender, n (%) Men23 (39)11 (36.7)12 (41.4) Women36 (61)19 (63.3)17 (58.6)Age (years), median [range]76 [70, 88]78,5 [70, 88]75 [70, 85]Initial performance status, n (%) 019 (32.2)8 (26.7)11 (37.9) 135 (59.3)20 (66.7)15 (51.7) 25 (8.5)2 (6.7)3 (10.3)Tumor localization, n (%) Colorectal4 (6,8)3 (10)1 (3,4) Lung12 (20,3)5 (16,7)7 (24,1) Prostate6 (10,2)4 (13,3)2 (6,9) Breast17 (28,8)7 (23,3)10 (34,5) Other20 (33,9)11 (36,7)9 (31)Stage, n (%) III1 (1.8)0 (0)1 (3.7) IV54 (98.2)28 (100)26 (96.3)T, n (%) 06 (14.3)1 (4.5)5 (25) 11 (2.4)1 (4.5)0 (0) 21 (2.4)1 (4.5)0 (0) 34 (9.5)3 (13.6)1 (5) 47 (16.7)4 (18.2)3 (15) X23 (54.8)12 (54.5)11 (55)N, n (%) 07 (16.3)2 (9.1)5 (23.8) 15 (11.6)4 (18,2)1 (4.8) 22 (4.7)1 (4,5)1 (4.8) 34 (9.3)2 (9.1)2 (9.5) X 25 (58.1)13 (59.1)12 (57.1)M, n (%) 03 (5.2)1 (3.4)2 (6.9) 152 (89.7)27 (93.1)25 (86.2) X3 (5.2)1 (3.4)2 (6.9)Smoker, n (%) Yes6 (10.9)2 (7.1)4 (14.8) No 49 (89.1)26 (92.9)23 (85.2) Table 1Baseline characteristics of the per-protocol population (n= 59)Total (n=59)Melatonin group (n=30)Placebo group (n=29)Gender, n (%) Men23 (39)11 (36.7)12 (41.4) Women36 (61)19 (63.3)17 (58.6)Age (years), median [range]76 [70, 88]78,5 [70, 88]75 [70, 85]Initial performance status, n (%) 019 (32.2)8 (26.7)11 (37.9) 135 (59.3)20 (66.7)15 (51.7) 25 (8.5)2 (6.7)3 (10.3)Tumor localization, n (%) Colorectal4 (6,8)3 (10)1 (3,4) Lung12 (20,3)5 (16,7)7 (24,1) Prostate6 (10,2)4 (13,3)2 (6,9) Breast17 (28,8)7 (23,3)10 (34,5) Other20 (33,9)11 (36,7)9 (31)Stage, n (%) III1 (1.8)0 (0)1 (3.7) IV54 (98.2)28 (100)26 (96.3)T, n (%) 06 (14.3)1 (4.5)5 (25) 11 (2.4)1 (4.5)0 (0) 21 (2.4)1 (4.5)0 (0) 34 (9.5)3 (13.6)1 (5) 47 (16.7)4 (18.2)3 (15) X23 (54.8)12 (54.5)11 (55)N, n (%) 07 (16.3)2 (9.1)5 (23.8) 15 (11.6)4 (18,2)1 (4.8) 22 (4.7)1 (4,5)1 (4.8) 34 (9.3)2 (9.1)2 (9.5) X 25 (58.1)13 (59.1)12 (57.1)M, n (%) 03 (5.2)1 (3.4)2 (6.9) 152 (89.7)27 (93.1)25 (86.2) X3 (5.2)1 (3.4)2 (6.9)Smoker, n (%) Yes6 (10.9)2 (7.1)4 (14.8) No 49 (89.1)26 (92.9)23 (85.2) Baseline characteristics of the per-protocol population (n= 59) Treatment compliance Sixty-one per cent (N = 59) of the entire per protocol population took at least two months of supplementation. Among them, 33% (N = 19) completed the three months. According to the tablet count at the end of supplementation, the mean intake was 85 (SD 7.5) tablets, reflecting good compliance. Seven treatment interruptions (melatonin group: N = 1; control group: N = 6) and three treatment discontinuations were observed (melatonin group: N = 2; control group: N = 1). The reasons for treatment discontinuations were the patient’s decision (N = 1 in the control group) and toxicities (N = 2 in the melatonin group); among the patients in the melatonin group. One patient presented headache considered to be attributable to Circadin®. The median daily oral concomitant tablets per group was 7 and 9 for melatonin and control group respectively. Efficacy outcomes for QoL and other parameters Among the 59 assessable patients, 55 completed both the inclusion and the M3 QLQ-C30 questionnaires. At inclusion, the global QoL score was well balanced between the experimental and control groups (PI = P(E > C) = 0.56, 95% CI (0.41, 0.69), p = 0.4). The variation in global scores between the beginning and the end of supplementation did not show any difference between the groups (PI = P(E > C) = 0.52, 95% CI (0.38, 0.67), p = 0.8) (Fig. 2). No clinically meaningful differences between the groups were observed in the QLQ-C30 sub-scores throughout the study (from inclusion to M12) (Table 2). Similarly, no differences were observed in the quality-of-life QLQ-ELD14 scores. Fig. 2Median QLQ-C30 global score from inclusion to 12-months supplementation Median QLQ-C30 global score from inclusion to 12-months supplementation Table 2Median QLQ-C30 global score from inclusion to 12-months supplementationMelatonin groupPlacebo group n Median (IQI) n Median (IQI)Inclusion3066.60 (50–75)2858.30 (50–66.6)M32966.60 (50–75)2658.30 (50–66.6)M62650.00 (41.6–66.6)2366.60 (50–66.6)M122366.60 (45.8–70.8)1854.15 (41.6–72.9)Abbreviations: M3 Three-months supplementation, M6 Six-months supplementation, M12 Twelve-months supplementation, IQI Interquartile interval Table 2Median QLQ-C30 global score from inclusion to 12-months supplementationMelatonin groupPlacebo group n Median (IQI) n Median (IQI)Inclusion3066.60 (50–75)2858.30 (50–66.6)M32966.60 (50–75)2658.30 (50–66.6)M62650.00 (41.6–66.6)2366.60 (50–66.6)M122366.60 (45.8–70.8)1854.15 (41.6–72.9)Abbreviations: M3 Three-months supplementation, M6 Six-months supplementation, M12 Twelve-months supplementation, IQI Interquartile interval Median QLQ-C30 global score from inclusion to 12-months supplementation Abbreviations: M3 Three-months supplementation, M6 Six-months supplementation, M12 Twelve-months supplementation, IQI Interquartile interval The analysis of the Leeds questionnaire evolution between arms showed a small, though not clinically significant, difference, with a mean score increase of about 0.62 points in the melatonin group compared to the control group (unadjusted p = 0.014, p = 0.07 after adjustment for multiple comparisons). Concerning the VAS used to assess pain, asthenia and appetite, no differences were found between groups. Similarly, no differences were observed for the depressive syndrome score (GDS questionnaire), nutritional status (MNA), autonomy score (IADL), and cognitive function (MMS). Urinary 6-sulfatoxymelatonin concentrations Median initial aMT6s concentrations in the experimental and control groups were: 0.43 ng/mg creatinine (range [0.07, 163], IQI [0.25, 0.84]) and 0.9 ng/mg creatinine (range [0.06, 5.7], IQI [0.47, 2.5]) respectively. After three months of supplementation, the median aMT6s was 8.4 ng/mg creatinine (range [0.15, 201], IQI [0.99, 105]) in the experimental group and 0.6 ng/mg creatinine (range [0.09, 7], IQI [0.35, 1.5]) in the control group (Table 3). A statistically significant difference in variation between 3 months and inclusion was observed in favour of the experimental group (PI = P(E > C) = 0.8, 95% CI (0.65, 0.89), p < 0.001). No correlation was found between the QLQ-C30 global score variation and urinary aMT6s variation between the beginning and the end of supplementation (Spearman’s correlation coefficient ρ = − 0.1, p = 0.5) (Fig. 3). Fig. 3Correlation between aMT6s and QoL changes Correlation between aMT6s and QoL changes Table 3Urinary aMT6s concentrations n MedianMinMaxIQIInclusionMelatonin group280.430.07163(0.25–0.84)Placebo group240.900.065.7(0.47–2.5)M3Melatonin group278.40.15201(0.99–105)Placebo group240.600.097(0.35–1.5)Abbreviation: M3 Three-months supplementation, IQI Interquartile intervalData presented as ng aMT6s/mg creatinine Table 3Urinary aMT6s concentrations n MedianMinMaxIQIInclusionMelatonin group280.430.07163(0.25–0.84)Placebo group240.900.065.7(0.47–2.5)M3Melatonin group278.40.15201(0.99–105)Placebo group240.600.097(0.35–1.5)Abbreviation: M3 Three-months supplementation, IQI Interquartile intervalData presented as ng aMT6s/mg creatinine Urinary aMT6s concentrations Abbreviation: M3 Three-months supplementation, IQI Interquartile interval Data presented as ng aMT6s/mg creatinine Safety outcomes Safety outcomes were analysed in the mITT population (N = 97). Twelve adverse events attributable to supplementation were observed (3 in the control group and 9 in the melatonin group). Two grade-3 and one grade-2 adverse events occurred in the melatonin group (nausea, drowsiness and asthenia, respectively) (Table 4). Table 4Adverse eventsNoDescriptionGradeImputability linked toAssociated treatmentArmCancer treatmentTreatment of the study1Nausea3YesYesYesMelatonin2DrowsinessUnkNoYesNoMelatonin3DizzinessUnkNoYesNoMelatonin4NightmareUnkNoYesYesPlacebo5Vomiting1NoYesNoPlacebo6HeadachesUnkNoYesNoMelatonin7Asthenia2NoYesNoMelatonin8Drowsiness3NoYesNoMelatonin9Nausea1YesYesYesMelatonin10Mucositis1YesYesYesMelatonin11Oesophageal burns1NoYesNoPlacebo12Cholestasis1NoYesNoMelatoninUnk Unknown Table 4Adverse eventsNoDescriptionGradeImputability linked toAssociated treatmentArmCancer treatmentTreatment of the study1Nausea3YesYesYesMelatonin2DrowsinessUnkNoYesNoMelatonin3DizzinessUnkNoYesNoMelatonin4NightmareUnkNoYesYesPlacebo5Vomiting1NoYesNoPlacebo6HeadachesUnkNoYesNoMelatonin7Asthenia2NoYesNoMelatonin8Drowsiness3NoYesNoMelatonin9Nausea1YesYesYesMelatonin10Mucositis1YesYesYesMelatonin11Oesophageal burns1NoYesNoPlacebo12Cholestasis1NoYesNoMelatoninUnk Unknown Adverse events Unk Unknown Thirty-three initial reports of serious adverse events (SAEs) were received in total (15 in the melatonin group and 18 in the control group), along with 5 follow-up reports (2 in the melatonin group and 3 in the control group). The majority were grade-3 events (N = 15, 40%). Hospitalization accounted for 74% of the declarations (N = 28) and concerned 71% of the melatonin group (N = 12) and 76% of the placebo group (N = 16). The remaining 26% were deaths (N = 10), with five occurrences in each group (29% in the melatonin group, and 24% in the control group). No SAEs reported during the study period were attributed to melatonin supplementation. Survival In the overall per-protocol population (N = 59) for progression-free survival (PFS), 28 cases of progression occurred (13 in the melatonin group and 15 in the control group). Median PFS was not reached in the melatonin group, and was 13 months in the control group. Concerning overall survival (OS), 9 events were observed (4 in the melatonin group and 5 in the control group). Median OS was not reached in either group. Discussion To the best of our knowledge, the MEQAPAG trial is the first national multi-centre randomized placebo-controlled trial evaluating the effects of melatonin supplementation on the QoL of older patients under systemic treatment for advanced cancer. Considering the comorbidities of this particular population, often associated with several oral treatments the rate of compliance (number of tablet intakes) in the study population was good. It can be recalled that, in the overall population, 61% of the patients are fully compliant, i.e., had at least two months of supplementation (melatonin or placebo) out of the planned three months. As expected, we observed that patients who had received supplementation in the form of melatonin presented significantly higher urinary aMT6s concentrations. Despite these observations, melatonin supplementation did not seem to improve the patients’ QoL compared to those treated with placebo. The global score on the QLQ-C30 was maintained during the three months of supplementation and over the 12-month follow-up in the overall population, and no deterioration was observed. Knowing that older patients have more risk factors for non-adherence in oral treatment, improving the adherence rate in this population is a real topic [42]. In our study, even if 100% of the per protocol population adhere to the supplementation after two-month, only 32,8% is at three-month. Previous works identified that potential barriers to adherence are related to therapy (e.g., adverse event, complexity of treatment), patient (e.g., age, polypharmacy, geriatric syndromes), or healthcare provider [43–47]. Furthermore, the compliance depends on the patient’s perception of the expected impact of the treatment on their symptoms and conversely, doubts about the effectiveness undermine compliance [48]. This loss of adherence across the study may be explain by all these parameters both patients’ perception and others barriers as change of therapeutic management or general condition of the patient. If the patient does not feel any positive effects from the treatment or, on the contrary, experiences adverse effects, they may become discouraged. The suggest strategies for these patients could relied on a geriatric evaluation to identify potential impairment and then personalised education program to sensibilize and support the patients and family members in the importance of adherence to benefit from the effects of treatment. The impact of melatonin on QoL and fatigue during cancer treatment is unclear. Even if some studies have concluded to an improvement in QoL, others have not observed any effect. For example, a prospective phase-II study conducted on metastatic breast cancer patients (N = 32) showed that oral intake of melatonin at 5 mg/day for two months improved global QoL [7]. A randomized, placebo-controlled study conducted on 151 patients treated with chemotherapy for non-small-cell lung cancer concluded that patients receiving melatonin supplementation (10 to 20 mg/day) for the six first months of chemotherapy tended to have a better QoL [30]. However, a recent meta-analysis (not focused on older patients) conducted on 6 clinical trials with cancer patients all taking a 20-mg melatonin supplementation, concluded it did not affect QoL or sleep quality [49]. Sedighi Pashaki’s team studied the short- and long-term effects of a melatonin-supplementation in breast cancer women during adjuvant treatment (chemotherapy and radiotherapy) [50, 51]. The patients were randomized in either experimental or control group. Experimental group received oral melatonin (18 mg/day) from 1 week before until 1 month after (for short-term study) and until 2 years after the adjuvant radiotherapy (for the long-term study) whereas control group received placebo. Brief Fatigue Inventory (BFI) was used to asses level of fatigue. The results showed similar BFI score at baseline in both studies. However, after the intervention, a significant lower score is observed in the experimental group, supporting that melatonin as a role to manage fatigue in breast cancer patients. In their long-term study, in addition to an improvement of fatigue level, the authors observed that the severity of fatigue is significantly lower in the experimental group. Furthermore, they showed that over time, patients taking oral melatonin have a greater decrease in fatigue score. In the MEQAPAG trial, we chose to use Circadin® 2 mg/day as recommended in the indication. Circadin® is indicated for the treatment of sleep disorders, which is one of the numerous aspects influencing QoL. This choice has been made notably taking into account that older patients often intake several tablets daily and here the supplementation add only one more tablet. Our team have already shown a significant correlation between nonadherence to treatment and the number of comedications [52].However, this dose may be insufficient for broader QoL benefits compared to higher doses (5–20 mg/day) used in prior studies [7, 30]. On the other hand, regarding fatigue, despite the small sample size and lack of power calculation, we found a slight improvement of quality of sleep in the experimental group on Leeds questionnaire. Although not clinically significant, this modest improvement may benefit patients with specific sleep complaints, warranting further investigation. This point suggests that the dose of 2 mg/day could lead to a positive effect on this QoL subdomain. We could hypothesize that if the dosage had been higher, it might have had a more significant impact on the result of the Leeds questionnaire score and possibly on the QoL global score. However, we cannot rule out the possibility that the placebo effect plays a role in improving symptoms and therefore dilutes the results. These various findings could be explained by the multidimensional aspects implicated in QoL. Improvements in symptoms are not sufficient to improve the global score, especially in an older advanced cancer population. Another explanation for the absence of any effect of melatonin supplementation on QoL could be that adjuvant treatment by chemotherapy and radiotherapy for solid tumours had no negative impact on older patients’ QoL (65 years and over) [53]. QoL in older cancer patients is influenced by multiple factors, including comorbidities, polypharmacy, and psychological resilience, which may explain the lack of melatonin’s effect [54–56]. This statement is still available for survivors as showed by a randomized clinical study conducted on breast cancer survivors (N = 20). The patients, not undergoing treatment, received either a 10-weeks Melatonin supplementation (6 mg/day) or placebo associated with an adapted physical activity. QoL was evaluated by World Health Organization Quality of Life Survey brief version (WHOQOL-BREF) and did not show any statistically significant difference between groups. So, with a similar length of supplementation, even if patients are younger than in our population (mean age around 66 years) and the Melatonin dosage higher, the conclusion drowned by the authors in terms of QoL are in line with ours [57]. Nevertheless, our findings on safety are consistent with the literature data regarding the fact that melatonin is well tolerated [58]. Regarding literature data related to melatonin and cancer, the questions about the modalities of the supplementation (duration, dose, moment) still raised and need to be investigate in further studies in order to identify how to enhance the QoL during cancer treatment. For the last four years, studies related to the use of crocin, derived from saffron, in cancer patients for improving anxiety and depression as emerged [59, 60]. The crocin daily supplementation (30 mg/day) was assessed during neoadjuvant chemoradiotherapy for oesophageal squamous cell carcinoma and during chemotherapy for breast cancer patients. A significant reduction of anxiety and depression level was observed. These papers open up perspectives on the potential of new molecules that could influence various aspects of quality of life. This work presents several limitations, which preclude concluding to significant results. The main limitation is the fact that the study was interrupted prematurely for futility after the interim analysis on the first hundred patients treated. Indeed, the accrual rate, below target, and the estimate of the probability of reaching a positive conclusion at the end of the trial were too low. Our population was heterogeneous in terms of primitive tumour localization and in terms of systemic treatment received (molecule and line); we also had difficulties in enrolling older patients, which led us to widen our selection criteria. The inclusion of multiple cancer types and treatment lines may have diluted melatonin’s effects in specific subgroups, such as breast or lung cancer patients. With a larger sample size, subgroup analyses would have been interesting. For example, assess the impact of treatment on specific groups of patients suffering from depression and/or insomnia or according to the type of cancer. Furthermore, many patients could not be assessed because they did not comply with the three-months of supplementation. This last point highlights the difficulties in implementing supplementation over several months for an older population. One of our hypotheses is that the majority of patients were already taking several daily oral treatments for concomitant pathologies (8 tablets/day in our study). In a study focused on secondary prevention of osteoporotic fractures in an acute rehabilitation setting for the older, the authors showed that the most common reason for non-participation was reluctance to take another medication (once weekly oral bisphosphonate) [61]. Moreover, survival analyses were underpowered due to the low number of events, precluding definitive conclusions. The MEQAPAG trial focused on a population of patients aged 70 years and older, which provides new insights into this population, often under-represented in clinical trials. A recent meta-analysis listed the reference intervals for urinary aMT6s concentrations according to age group and gender [62]. Very few articles are available on the 70–80-year-old population. There is only one for females [63] and two for males [64, 65] and none for both genders. Thus, our results provide new insights on urinary aMT6s concentrations for this category. In the MEQAPAG study, baseline concentrations in both the experimental and control groups were below 1 ng/mg creatinine, and supplementation enabled patients to achieve an average level of 8,4 ng/mg creatinine. Regarding the three study with data for the age group 70–80 years, Braam et al., reported higher urinary aMT6s concentrations than ours (15,2 ± 128,6 ng/mg creatinine) [62]. However, our results are not really comparable to those of the three studies cited above [63–65] because they focused on healthy older patients and concentrations were based on first morning urine samples while we measured aMT6s on 12-hour overnight samples in cancer patients. The peak secretion of melatonin occurs between midnight and 3 a.m., the urine collection over the night had been shown to be a tool in the studies of circadian rhythm disturbance [66]. Measuring hormone levels in 24-hour urine samples has the advantage of measuring hormone production throughout the entire nychthemeral cycle. Conclusion The present national, randomized, placebo-controlled trial focused on older cancer patients (70 years and older). Despite the good melatonin oral supplementation tolerance, QoL did not improve, but remained stable in both the experimental and placebo groups. Compliance rates were low, suggesting difficulties in implementing daily oral supplementation for this population. However, the MEQAPAG trial provides new insights for older cancer patients notably concerning urinary aMT6s concentrations. Supplementary Information Supplementary Material 1 Supplementary Material 1