We conducted and reported this systematic review according to the PRISMA statement (23). This systematic review has been registered (Reg. No. CRD42023455278) in PROSPERO.

(1) Nonclinical trial publications such as animal studies and review papers; (2) review papers or meta-analysis papers; (3) repeat publications (the first published was selected); (4) papers that did not clearly state the selection and method of use of aromatherapy; (5) papers with nonstandardized experimental designs or inappropriate use of statistical methods; (6) papers without standards for efficacy evaluation or with unscientific study results.

The English databases PubMed and Web of Science and the Chinese databases CNKI Database, Wanfang Database, and VIP Database were searched. The literature search was limited to Chinese and English. The Chinese search terms used were “insomnia (不寐)”, “insomnia (失眠)”, “sleep disorders (睡眠障碍)” “sleep quality (睡眠质量)”, “sachet (香囊)”, “aromatherapy (芳香疗法)”, “plant essential oil (植物精油)”, and “aromatherapy (香薰疗法)”. The English search terms used were “insomnia”, “sleep disorder”, “sleep quality”, “aromatherapy”, and “essential oil”. The retrieval time frame ranged from database inception to August 30, 2023.

(1) Psychological state evaluation: The subjects selected in this study all had underlying disease. Underlying disease creates a negative psychological state which is closely related to sleep quality. Therefore, it was necessary to evaluate patients’ psychological state. The SAS and SDS were used to evaluate the negative psychology of the patients. (2) Sleep quality: The PSQI was used to evaluate the sleep quality of the patients. A total of 18 items formed seven dimensions, each dimension having 0-3 points, for a total score of 21 points. The total score of the PSQI ranged from 0-21 points.

The GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach was applied to evaluate the level of evidence in the included literature (24, 25). The GRADE process includes all important and critical outcomes explicitly. The main domains used to assess the certainty of the evidence are risk of bias, inconsistency, indirectness of evidence, imprecision, and publication bias. The factors that can increase the certainty of the evidence are dose-response gradient, large magnitude of an effect, and effect of plausible residual confounding. Finally, we could summarize the methodology in four grades (high, moderate, low, very low).

Two reviewers (YQ and HYJ) independently screened all literature and extracted data using EndNote X9 software. The form was utilized to gather data on the included studies, including the first author, year of publication, region where the study was conducted, number of participants in each group, sex of the patients, interventions (category of essential oils and duration of treatment), and scores on the SAS, SDS, and PSQI. Differences of opinion were mediated by a third reviewer (HBC). The identification content mainly included whether aromatherapy was inhalation therapy, whether the experimental design was RCT, and whether insomnia was comorbidity.

Two reviewers (YQ and HYJ) categorized the risk of bias for each trial as high, low, or unclear based on the Cochrane Collaboration Risk of Bias Assessment Tool 1.0.15. The following risk of bias domains were evaluated: (1) random sequence generation (selection bias); (2) allocation concealment (selection bias); (3) blinding of participant and personnel (performance bias); (4) blinding of outcome assessment (detection bias); (5) incomplete outcome data (attrition bias), (6) selective reporting (reporting bias); and (7) any other bias. Differences of opinion were resolved by a third reviewer (HBC). All studies were assessed and categorized as high risk, low risk, or unclear risk according to their respective domains. If the number of evaluated trials exceeded five, funnel plots were employed to assess the presence of publication bias.

Two reviewers independently used RevMan 5.3 software to synthesize and statistically analyze the efficacy data. Mean differences (MDs) with 95% confidence intervals (CIs) were presented for continuous data, and data from studies were pooled using the inverse variance method. Dichotomous data were presented as odds ratios with 95% CIs and pooled using the Mantel-Haenszel method. The heterogeneity was assessed using the Q-test and I² statistic. The fixed-effects model was used in instances of low heterogeneity (I²<50%), while the random effects model was used in situations of moderate heterogeneity (I²>50%). Sensitivity analysis was used to observe the effect of a single article on heterogeneity. Subgroup analysis was performed according to the type of disease, duration of intervention, agents used, and control methods. P-values less than 0.05 were defined as statistically significant.

According to the retrieval methods, a total of 1092 publications that met the inclusion criteria were found, including 699 publications in Chinese and 393 publications in English. After reading the titles and abstracts, the publications that did not meet the inclusion criteria were excluded, and a total of 27 publications (26–52) met the inclusion criteria ( Figure 1 ). Of these, six papers (34, 36, 39, 41, 50, 51) were master’s theses, and the other 21 papers were journal papers.

The 27 studies included a total of 2072 patients, including 1039 patients in the experimental group and 1033 patients in the control group. The basic information of the included studies is listed in Table 1 .

All 27 studies were RCTs. The studies all used correct randomization methods, including random number tables and permutation block randomization. The statistical methods of all the studies were correct, and the distribution between groups was balanced. Two studies (26, 28) reported blinding during the experiment, and six (27, 29, 33, 34, 50, 51) reported data loss, including loss to follow-up ( Figures 2 , 3 ).

Twenty-seven articles were included in this study. Our meta-analysis showed that aromatherapy well relieved anxiety and depression in the patients and improve their sleep quality during their illness. The funnel plot showed a scattered distribution of data, without obvious symmetry, suggesting that there is a certain possibility of publication bias. The pooled results were highly heterogeneous (I²>50%), indicating that some factors of the included studies affected the accuracy of the intragroup heterogeneity. Sensitivity analysis did not find that a single study significantly affected the heterogeneity. This indicated some stability in this result. Subgroup analysis of the sleep quality index score, which was performed in many of the included studies, confirmed that the type of disease and the duration of intervention were not factors affecting the heterogeneity, and the category of essential oils and the control method could not be excluded as not underlying factors of the heterogeneity. The results of sensitivity and subgroup analyses suggested that the heterogeneity of this study was not methodologically or statistically induced. We suspect that it may be related to the different diagnostic criteria for inclusion in the study. A meta-analysis of the six dimensions of the PSQI showed that aromatherapy can improve sleep quality (MD -0.54, 95% CI: -0.77 to -0.32, p<0.00001), shorten the time to falling asleep (MD -0.58, 95% CI: -0.87 to -0.29, p=0.0001), improve sleep efficiency (MD -0.37, 95% CI: -0.64 to -0.11, p=0.006), relieve sleep disorder (MD -0.57, 95% CI: -0.84 to -0.30, p<0.0001), enhance daytime function (MD -0.44, 95% CI: -0.69 to -0.20, p=0.0004) and lengthen sleep duration (MD -0.51, 95% CI: -0.94 to -0.08, p=0.02).

Sleep problems are common in the modern world, classified from mild to severe into sleep disorders, sleep difficulties, and insomnia (53). Primary insomnia patients showing isolated symptoms account for only 10-20% of the total number of insomnia patients. Comorbid insomnia (i.e., insomnia combined with other diseases) makes up the bulk of insomnia cases. Comorbid insomnia has many causes, which involve complex diseases, neurological function, and other factors. Therefore, modern medicine has difficulty improving this symptom based on its etiology (54). Modern studies have shown that essential oils such as lavender, rose, sweet orange, and valerian, which have sleep-promoting, sedative, and anti-anxiety effects, all contain one or both of linalool and limonene (55). Inhalation of linalool and limonene essential oils can relieve insomnia and anxiety (56). This type of treatment has long been used in traditional medicine in China, most often involving the use of vectors such as sachets and medicated pillows for disease prevention and treatment. Aromatherapy is also widely used internationally (57), showing therapeutic effects on different types of patients in pain, nausea, anxiety, depression, insomnia, and other symptoms and also benefiting patients with preoperative anxiety, tumors, palliative care, hospice care, and end of life (58–67).

Future research should aim to determine the optimal formulation of the essential oil preparation, which may include the composition and the extraction method of essential oils, the amount of essential oils to be used, and the duration of oil use. In order to further validate the long-term efficacy of inhalation aromatherapy for the treatment of comorbid insomnia, future research needs to be conducted in the following areas. First, laboratory techniques need to used to clarify the effective ingredients of aromatic traditional Chinese medicines, innovate and improve the dosage form and method of use. At the same time, large-scale clinical trials of aromatic Chinese medicines should be carried out to optimize these safe and efficient Chinese medicines. Second, syndrome analysis of traditional Chinese medicine should be conducted on secondary insomnia patients with different diseases, syndrome types and dialectical prescription should be summarized, and the clinical application potential of Chinese medicine essential oil compounds should be developed to promote the right, good, and active use of aromatherapy in first-line clinical practice. Third is to carry out more rigorous and detailed randomized controlled trial studies in China, focusing on the development of placebo oil to avoid data bias caused by breaking the blinding of patients in the control group. At the same time, patients can be guided to develop the habit of long-term use of aromatherapy, and long-term follow-up can be performed to determine its long-term efficacy.

To the best of our knowledge, this is the first meta-analysis to investigate inhalation aromatherapy versus conventional treatment for comorbid insomnia. To reduce heterogeneity, we purposely excluded other treatment methods using aromatic essential oils. We performed subgroup analyses to identify potential sources of heterogeneity. While these analyses reduced heterogeneity between studies, our review was not without potential limitations (1): The methodological quality of the included trials was found to be low, and detailed descriptions of blinded conduct, allocation concealment, attrition, and lost cases were lacking, especially the lack of blinded design may cause a large bias in this study; (2) Most of the included trials were of short duration and did not examine the long-term efficacy of inhalation aromatherapy. Therefore, high-quality methods, further studies with large samples, and long-term interventions are still needed to evaluate the safety and effectiveness of inhalation aromatherapy in improving comorbid insomnia; (3) The level of certainty of the evidence was considered to be average; (4) The diagnostic criteria for comorbid insomnia had not been standardized at this stage. There are limitations in the studies themselves, which may have an impact on the results. Therefore, the results of this meta-analysis should be treated with caution.