Letter to the Editor Regarding “Development and Internal Evaluation of a Nomogram for Clinically Significant Depressive Symptoms in Obstructive Sleep Apnea-Hypopnea Syndrome” [Letter]

Dear editor

We read with great interest the article by Yi et al1 recently published in Nature and Science of Sleep, titled “Development and Internal Evaluation of a Nomogram for Clinically Significant Depressive Symptoms in Obstructive Sleep Apnea-Hypopnea Syndrome.” In this single-center retrospective study of 820 hospitalized patients with PSG-confirmed OSAHS, the authors employed LASSO regression to select five predictors—ODI, PSQI, ESS, TSH, and FT3—from 36 candidate variables and constructed a nomogram for identifying clinically significant depressive symptoms as defined by an SDS standardized score ≥53. The model demonstrated favorable discrimination in the internal validation cohort (AUC = 0.915). As an exploratory study aimed at developing a screening tool for depressive symptoms in the OSAHS population, this work is of considerable clinical relevance. Nevertheless, while fully acknowledging the value of this study, we would like to highlight several neurological and methodological issues that merit further discussion.

First, the model predicts “depressive symptom burden” rather than “depressive disorder.” The authors defined clinically significant depressive symptoms using an SDS standardized score ≥53, but there is a fundamental distinction between a self-report scale score and a structured psychiatric interview-based diagnosis of depressive disorder—the latter requires a duration of at least two weeks, clinically significant distress, and functional impairment. Previous studies have indicated that somatic symptom items on the SDS (such as palpitations, weight loss, constipation, sleep disturbance, and fatigue) may be more closely related to the experience of chronic disease itself rather than specifically reflecting depressive disorder.2 In neurological practice, patients with OSAHS often attribute fatigue, sleepiness, and cognitive symptoms to sleep apnea itself rather than to mood disturbances, which may lead to either underestimation or overestimation of depressive symptoms. Therefore, this model is better suited as a screening tool for further psychological assessment rather than as a decision-making aid for initiating antidepressant treatment. This distinction directly affects the model’s practical applicability in clinical pathways.

Second, there is a substantial neurobiological overlap between OSAHS-related sleepiness, fatigue, and depressive symptoms. ESS, PSQI, and SDS overlap at both the conceptual and neurobiological levels—they collectively reflect intermittent hypoxia-induced alterations in prefrontal-limbic functional connectivity, neuroinflammation, and neurotransmitter dysregulation.3,4 Thus, the strong association between ESS/PSQI and SDS may reflect the manifestation of dysfunction within the same neural circuits across multiple behavioral dimensions, rather than identifying independently actionable clinical risk signals. Although the authors’ sensitivity analysis, which excluded ESS and PSQI, demonstrated that objective predictors retained predictive information, this finding was not sufficiently emphasized in the main text. A clearer presentation of these results—for example, by showing the discrimination of the objective-only model alongside the full model in a forest plot or supplementary table—would help clarify the model’s robustness in the absence of subjective questionnaire data and enhance its practical utility in neurological clinical pathways.

Third, ODI as a hypoxemia metric has informational limitations. The authors selected ODI over AHI as the final predictor, which is biologically plausible—intermittent hypoxia is a core mechanism of brain injury in OSAHS. However, ODI primarily reflects the frequency of desaturation events and does not integrate information on the depth and duration of oxygen desaturation. The recently proposed “hypoxic burden” metric, which integrates the depth, duration, and frequency of desaturation events, has been shown to provide additional information beyond AHI in predicting cardiovascular outcomes and functional impairment.5 If more comprehensive hypoxemia metrics such as hypoxic burden were incorporated into the model, its explanatory power for depressive symptoms might be enhanced, and the cumulative effects of intermittent hypoxia on emotion-regulating neural circuits could be more accurately captured.

Fourth, information on treatment status is lacking. OSAHS treatment—particularly CPAP and upper airway surgery—is known to influence depressive symptom trajectories. Longitudinal studies have demonstrated that CPAP therapy significantly improves sleepiness, depressive symptoms, and anxiety in patients with OSA.6 A three-year prospective study found that patients who adhered to CPAP treatment showed a significant reduction in depressive symptoms (P = 0.002), as well as greater improvements in sleep quality (P < 0.001) and sleepiness (P = 0.031) compared with non-users.7 The present cross-sectional study did not report patients’ current treatment status or track symptom changes following treatment initiation. In clinical practice, if some patients have already received treatment, their baseline PSG parameters may no longer reflect current disease severity. More importantly, treatment response information is of core value to neurological clinical decision-making: after identifying a patient with OSAHS and comorbid depressive symptoms, should CPAP therapy be prioritized over antidepressant medication? This question requires a longitudinal design to answer. Incorporating information on whether patients are currently receiving or have previously received CPAP treatment, or performing subgroup analyses on treatment responders, would help distinguish “hypoxic sequelae” from “primary depressive disorder,” thereby guiding individualized treatment strategies.

Fifth, the clinical operability and causal direction of thyroid-axis markers remain unclear. The authors included FT3 and TSH as predictors; however, thyroid function tests are not part of the routine assessment for OSAHS patients in standard clinical practice. Widespread adoption of this model would require additional blood tests, increasing both cost and implementation barriers. More importantly, Liang et al, in a study of 73 male patients with OSAHS, found that serum FT3 levels in the severe hypoxemia group were significantly higher than in the control group [(5.4 ± 0.7) ng/L vs (4.5 ± 0.6) ng/L, P < 0.05]. However, when the severe hypoxemia group was further stratified, patients with comorbid depressive symptoms had lower serum FT3 [(5.0 ± 0.5) ng/L] and FT4 [(16.2 ± 1.9) ng/L] levels compared with those without depressive symptoms [(5.5 ± 0.7) ng/L and (18.2 ± 2.3) ng/L, respectively; both P < 0.05].8 This finding suggests that the relationship between FT3 and depressive symptoms may not be simply linear, and whether FT3 reduction is a downstream consequence of OSAHS-related chronic stress or an independent predictor of depressive symptoms remains unclear. This uncertainty regarding causal direction limits confidence in incorporating this marker into clinical decision-making pathways.

Sixth, the study lacks neurocognitive assessment and a temporal dimension. Cognitive dysfunction in patients with OSAHS—including impairments in attention, executive function, and working memory—is closely associated with depressive symptoms. Hong et al, in a study of 102 patients who underwent PSG, found that SDS scores in the moderate-to-severe OSA group were significantly higher than those in the no/mild OSA group, and that AHI was independently associated with SDS scores.9 Furthermore, neuroimaging studies have confirmed that patients with OSA exhibit functional connectivity impairments in the posterior default mode network, which may constitute the neural basis for cognitive and depressive symptoms.3,4 The present study did not include any neurocognitive assessment tools, and therefore cannot distinguish whether depressive symptoms represent a primary mood disorder or are secondary to cognitive dysfunction. Moreover, as a cross-sectional study with only a single SDS assessment, it cannot differentiate between “state” depressive symptoms (OSAHS-related, potentially improving with treatment) and “trait” depressive symptoms (chronic depressive tendencies independent of OSAHS). Longitudinal follow-up comparing depressive symptom trajectories before and after treatment would be required to distinguish between these two types and thereby guide individualized treatment strategies.

In summary, Yi et al have provided an exploratory study of considerable clinical significance, offering preliminary evidence supporting the potential value of the combined ODI, PSQI, ESS, TSH, and FT3 model in identifying depressive symptoms in patients with OSAHS. We look forward to future studies that adopt prospective longitudinal designs, incorporate neurocognitive assessments, clarify the model’s clinical boundaries (screening vs diagnosis), collect treatment status information to evaluate the impact of CPAP intervention on depressive symptoms, and explore the value of more refined hypoxemia metrics—such as hypoxic burden—in predicting mood disorders. Such studies would contribute to a more precise understanding of the neurobiological basis of OSAHS-related depressive symptoms and provide a more robust evidence base for the comprehensive neuropsychiatric management of patients with OSAHS.

Data Sharing Statement

No datasets were generated or analysed during the current study.

Author Contributions

Rijuan Jin: Conceptualization, Writing -review & editing, Supervision. All authors have read and approved the final manuscript.

Disclosure

The authors report no conflicts of interest in this communication.