Depression in the rationale, insomnia as the target
- The European Sleep Research Society and the European Insomnia Network developed the updated guideline, published on 28 November 2023 in the Journal of Sleep Research, which Wiley issues on behalf of the ESRS (volume 32, issue 6, e14035; Riemann and 44 co-authors, the first of them at the University of Freiburg). It runs to 36 pages and can be read in full without registration or payment; no course, fee or certificate is attached to it.
- The licence is a separate fact from the free access: the article is published under CC BY 4.0, which permits reuse and adaptation with attribution. Table 2 gives the diagnostic criteria "according to ICSD-3 (AASM, 2014)", and the document does not settle whether the article's licence extends to that table.
- The argument that insomnia is more than a symptom rests on prediction: on page 6 the view of insomnia as primarily a symptom "was strongly challenged by evidence that chronic insomnia is an independent risk factor for mental health conditions such as depression or anxiety disorders", with two meta-analyses cited whose titles name insomnia as a predictor of depression (2011) and of mental disorders (2019). The grade A recommendation names insomnia itself as its object: CBT-I first-line for chronic insomnia in adults of any age, including those with comorbidities, in person or digitally. Grades follow a table of study types, and in some cases a consensus decision was decisive. Assessment by clinical interview, sleep questionnaires and sleep diaries is graded A; a diary kept for at least 7–14 days and a course of four to eight individual or group sessions are described without a grade; actigraphy is not recommended for routine evaluation (C).
- Depression and anxiety scales are among the endpoints of the meta-analyses in Tables 6–8, which list endpoints without dividing them into primary and secondary. Where a size is named, the effects on these scales are small to medium, from "small effects for depression" (47 studies, 4,317 patients) to small to medium effects on depression and anxiety (86 studies, 15,578 patients); a single row, 31 studies, finds "no significant effects on anxiety and mental health". The exception is Squires and colleagues (2022) in people with cancer, with "small to large" effects across nine endpoints, anxiety and depressive symptoms among them, without saying which endpoint the large effect belongs to. For insomnia disorder with depression (17 studies, 1,756 patients) Table 7 records significant effects on ISI and PSQI and "weaker effects for depressive symptoms".
Table 6, starting on page 9, is where depression and anxiety scales first appear as endpoints, in meta-analyses of CBT-I for insomnia without comorbidities. One of its rows splits down the middle. Xu and colleagues (2021), 31 studies, list significant effects on seven sleep measures, depression and fatigue, then "no significant effects on anxiety and mental health". In the next row, Yu and colleagues (2021), 14 studies, find significant effects on both anxiety and depression. Ballesio and colleagues (2018), 47 studies and 4,317 patients, record "small effects for depression"; Benz and colleagues (2020), 86 studies and 15,578 patients, small to medium effects on depression and anxiety. For these 25 meta-analyses the authors warn on page 11 that waiting-list controls and overlap between analyses "may overemphasise the effects of CBT-I".
Table 7, on page 12, covers insomnia with a second condition in 15 meta-analyses. Feng and colleagues (2020) pooled 17 studies of insomnia disorder with depression, 1,756 patients: "Significant ES for ISI/PSQI, weaker effects for depressive symptoms". Hertenstein and colleagues (2022), 22 studies and 1,083 patients with comorbid mental disorders, report medium to large effects on ISI and significant effects on symptoms of the comorbid disorder. Table 8, on self-help and digital formats, adds moderate effects on anxiety and depression (Ye and colleagues, 2015) and, in insomnia with depression, a smaller effect on depressive than on insomnia symptoms (Ho and colleagues, 2020). ISI and PSQI are questionnaires; for latency, wake time, total sleep time and efficiency most rows do not name the method, two in Table 6 name polysomnography and actigraphy, and one lists an undefined "AKT".
Two titles in the reference list carry the word predictor: "Insomnia is a predictor of depression", a 2011 meta-analysis of longitudinal epidemiological studies, and "Insomnia as a predictor of mental disorders" from 2019. On page 6 both stand behind one sentence. The view of insomnia as primarily a symptom "was strongly challenged by evidence that chronic insomnia is an independent risk factor for mental health conditions such as depression or anxiety disorders". The grade A recommendation takes insomnia itself as its object: CBT-I "as the first-line treatment for chronic insomnia in adults of any age (including patients with comorbidities), either applied in-person or digitally". No graded recommendation, in the summary or in Table 16, names a psychiatric outcome or the prevention of depression as its aim.
Table 9, a list of digital programs, is where prevention turns up: a trial by Leerssen and colleagues (2022) in insomnia with high risk of depression, "to prevent worsening of depressive symptoms", cited once as the evidence for the Dutch program i-Cycle. Its findings are not described.
Comorbid mental disorders on page 11
For insomnia alongside depression, posttraumatic stress disorder or alcohol dependency, the text reports medium to large effects on insomnia severity; for bipolar and psychotic disorders there "is not sufficient evidence". A search of the published text for nightmare and suicide found neither word. The supplementary material was not opened.
The risk argument names depression and anxiety, while the grade A recommendation names chronic insomnia.
This is a clinical guideline, not a trial. The effect descriptions quoted from Tables 6 to 8 are the guideline's verbal summaries of other authors' meta-analyses; the tables give no numeric effect sizes, and those meta-analyses were not opened. The trial by Leerssen and colleagues was not read, and nothing here describes its results. Statements about what the guideline does not contain rest on a reading of the summary and of Table 16 and a text search of the published 36-page version; the supplementary material was not opened. The literature search covered June 2016 to October 2022, with a further update till May 2023. The guideline's sections on medication and light therapy are outside this note.