Amy Sandford, Alison Haywood, Kirsty Rickett, Phillip Good, Sohil Khan, Karyn Foster, Janet R Hardy

Fatigue is the most commonly reported symptom in people with advanced cancer. Cancer-related fatigue (CRF) is pervasive and debilitating, and can greatly impact quality of life (QoL). CRF has a highly variable clinical presentation, likely due to a complex interaction of multiple factors. Corticosteroids are commonly used to improve CRF, but the benefits are unclear and there are significant adverse effects associated with long-term use. With the increasing survival of people with metastatic cancer, the long-term effects of medications are becoming increasingly relevant. Since the impact of CRF can be immensely debilitating and can negatively affect QoL, its treatment warrants further review. To determine the benefits and harms of corticosteroids compared with placebo or an active comparator in adults with advanced cancer and CRF. We searched CENTRAL, MEDLINE, Embase, CINAHL, Science Citation Index (ISI Web of Science), LILACS, and two clinical trial registries from inception to 18 July 2022.  SELECTION CRITERIA: We included randomised controlled trials in adults aged ≥18 years. We included participants with advanced cancer who were suffering from CRF. We included trials that randomised participants to corticosteroids at any dose, by any route, administered for the relief of CRF; compared to placebo or an active comparator, including supportive care or non-pharmacological treatments. Three review authors independently assessed titles identified by the search strategy; two review authors assessed risk of bias; and two extracted data. We extracted the primary outcome of participant-reported fatigue relief using validated scales and secondary outcomes of adverse events, serious adverse events and QoL. We calculated the risk ratio with 95% confidence intervals (CIs) between groups for dichotomous outcomes. We measured arithmetic mean and standard deviation, and reported the mean difference (MD) with 95% CI between groups for continuous outcomes. We used standardised mean difference (SMD) with 95% CIs when an outcome was measured with different instruments measuring the same construct. We used a random-effects model to meta-analyse the outcome data. We rated the certainty of the evidence using GRADE and created two summary of findings tables.  MAIN RESULTS: We included four studies with 297 enroled participants; data were available for only 239 participants. Three studies compared corticosteroid (equivalent ≤ 8 mg dexamethasone) to placebo. One study compared corticosteroid (dexamethasone 4 mg) to an active comparator (modafinil 100 mg). There were insufficient data to evaluate subgroups, such as dose and duration of treatment. One study had a high risk of performance and detection bias due to lack of blinding, and one study had a high risk of attrition bias. Otherwise, we assessed risks of bias as low or unclear. Comparison 1: corticosteroids compared with placebo Participant-reported fatigue relief The was no clear difference between corticosteroids and placebo (SMD -0.46, 95% CI -1.07 to 0.14; 3 RCTs, 165 participants, very low-certainty evidence) for relief of fatigue at one week of the intervention. We downgraded the certainty of the evidence three times for study limitations due to unclear risk of bias, imprecision, and inconsistency. Adverse events There was no clear difference in the occurrence of adverse events between groups, but the evidence is very uncertain (3 RCTs, 165 participants; very low-certainty evidence). Serious adverse events There was no clear difference in the occurrence of serious adverse events between groups, but the evidence is very uncertain (2 RCTs, 118 participants; very low-certainty evidence). Quality of lIfe One study reported QoL at one week using the Edmonton Symptom Assessment System (ESAS) well-being, and found no clear difference in QoL between groups (MD -0.58, 95% CI -1.93 to 0.77). Another study measured QoL using the Quality of Life Questionnaire for Cancer Patients Treated with Anticancer Drugs (QoL-ACD), and found no clear difference between groups. There was no clear difference between groups for either study, but the evidence is very uncertain (2 RCTs, 118 participants; very low-certainty evidence).  Comparison 2: corticosteroids compared with active comparator (modafinil) Participant-reported fatigue relief There was improvement in fatigue from baseline to two weeks in both groups (modafinil MD 10.15, 95% CI 7.43 to 12.87; dexamethasone MD 9.21, 95% CI 6.73 to 11.69), however no clear difference between the two groups (MD -0.94, 95% CI -4.49 to 2.61; 1 RCT, 73 participants, very low-certainty evidence). We downgraded the certainty of the evidence three times for very serious study limitations and imprecision. Adverse events There was no clear difference in the occurrence of adverse events between groups, but the evidence is very uncertain (1 RCT, 73 participants; very low-certainty evidence).  Serious adverse events There were no serious adverse events reported in either group (1 RCT, 73 participants; very low-certainty evidence).  Quality of lIfe One study measured QoL at two weeks, using the ESAS-well-being. There was marked improvement in QoL from baseline in both groups (modafinil MD -2.43, 95% CI -2.88 to -1.98; dexamethasone MD -2.16, 95% CI -2.68 to -1.64), however no clear difference between the two groups (MD 0.27, 95% CI -0.39 to 0.93; 1 RCT, 73 participants, very low-certainty evidence). There is insufficient evidence to support or refute the use of systemic corticosteroids in adults with cancer and CRF. We included four small studies that provided very low-certainty of evidence for the efficacy of corticosteroids in the management of CRF. Further high-quality randomised controlled trials with larger sample sizes are required to determine the effectiveness of corticosteroids in this setting.