Disturbed sleep after lung transplantation is associated with worse patient-reported outcomes and chronic lung allograft dysfunction

Introduction

Lung transplantation aims to extend survival, relieve disability, and improve health-related quality of life (HRQL).^1,2 Although many do well, perioperative complications are increasing, 20-30% of survivors do not report substantive improvements in physical functioning or HRQL, and one in three die within the first three post-operative years,.^1,3–8 In response to these frequent and increasing barriers to transplant success, identifying modifiable factors that could support transplant recovery and improve HRQL is needed. Sleep may be one potential behavioral pathway, although its impacts on outcomes after lung transplantation are not well characterized.

Sleep plays a fundamental role in physical health, well-being, and recovery from illness. The potential for sleep disturbances after lung transplant is high, particularly during the dynamic peri- and early post-operative period that is increasingly complicated by Primary Graft Dysfunction (PGD; a form of acute post-transplant lung injury), delirium, prolonged hospital stays, post-traumatic stress, readmissions, and intense immunosuppression.^6,9–13 Disturbed sleep has the potential to significantly impair transplant success. Outside of the lung transplant context, disturbed sleep is associated with impaired cognition, mood dysregulation (particularly depression), risk of frailty, and immune dysfunction.^14–22 In addition to being important outcomes to lung transplant recipients themselves, depression, frailty and immune dysfunction have previously been shown to also predict disability, chronic lung graft dysfunction (CLAD), and mortality after transplant.^23–27

To date, a relatively limited number of studies have examined sleep after lung transplantation and even fewer have examined the impact of disturbed sleep on post-transplant outcomes.²⁸ Depending on the sleep measure used, between 30-74% of lung transplant recipients reported disturbed sleep; when present, disturbed sleep was associated with depression and anxiety.²⁸

To address these gaps, in this single-center cohort study we sought to examine potential perioperative predictors of disturbed sleep after lung transplant and whether disturbed sleep is associated with disability, HRQL, and frailty. We additionally sought to test whether disturbed sleep was associated with risk of CLAD onset and mortality.

Methods

Results

During the Breathe Again study period, of the 392 participants enrolled, 259 underwent lung transplant. Of these, 141 completed the pilot survey that included MOS-Sleep and formed our study cohort (Table 1 and Figure 1). These 141 participants were 43% female with a mean age of 58 years (standard deviation [SD] ± 13); 75% were non-Hispanic white. Most participants underwent transplant for pulmonary fibrosis (72%) followed by non-suppurative obstructive lung diseases (15%). Participants completed the study battery used in these analyses at a median of 1.5 years after transplant (Interquartile range [IQR]: 0.6, 2.4). Over the study period, 52 (37%) developed CLAD and 20 (14%) died.

• Download figure
• Open in new tab

Figure 1.

Study Flow

Of the predictors tested, prolonged post-transplant hospital length of stay (LOS) was associated with disturbed sleep and insomnia, after adjusting for age, sex, and lung function. Considering SPI and insomnia as continuous measures, prolonged LOS was associated with 12-points worse SPI scores (12.18; 95%CI: 3.38, 20.99; p = 0.01; SPI MID = 8) and 15-points worse insomnia scores (14.56; 95%CI: 2.08, 27.04; p = 0.02; insomnia MID = 12). Prolonged LOS was associated with a 7-fold higher odds of being in the worst quartile of SPI disturbed sleep (adjusted Odds Ratio [aOR] 7.09; 95%CI: 2.10, 23.95; p <0.01). Prolonged LOS, however, was not significantly associated with being in the worst quartile of insomnia (aOR: 1.69; 95%CI: 0.47, 6.04; p = 0.42). Neither PGD nor delirium were significantly associated with disturbed sleep or insomnia (all p >0.05) (Appendix Table 1).

We also found consistent associations between disturbed sleep, either as a continuous variable or dichotomized as worst versus the top three quartiles, and all patient-reported outcomes of interest, after adjusting for age, sex, transplant indication, and allograft function (Table 2). For example, for each MID worsening in the Sleep Problems Index (SPI) participants reported worse depressive symptoms (GDS: 0.63; 95%CI: 0.32, 0.94; p < 0.01; MID = 1.65), generic mental HRQL (SF36 MCS: −2.04; 95%CI: −2.75, −1.33; p <0.01; MID=5), and health utilities (EQ5D: −0.03; 95%CI: −0.04, −0.02; p<0.01; MID=0.06). Similarly, participants in worst quartile of SPI reported substantially worse respiratory-specific HRQL (AQ20-R: 3.53; 95%CI: 1.79, 5.26; p<0.01; MID=1.75), cognitive functioning (LTQOL-Cog: 0.63; 95%CI: 0.31, 0.96; p <0.01; MID=0.47), and generic physical HRQL (SF36-PCS: −6.99; 95%CI: −10.67, −3.30; p <0.01; MID=5). Insomnia symptoms were similarly strongly associated with our PROs of interest after adjusting for prespecified confounders (Table 3).

Poorer sleep and insomnia symptoms were also generally associated with worse physical functioning and frailty by FFP (Tables 2 and 3). For example, each MID worsening in SPI was associated with disability (LT-VLA: 0.05; 95%CI: 0.02, 0.09; p <0.01; MID = 0.3) and frailty (FFP: 0.13; 95%CI: 0.04, 0.21; p <0.01). Each MID worsening in insomnia was also associated with disability (LT-VLA: 0.05; 95%CI: 0.02, 0.09; p <0.01; MID = 0.3) and frailty (FFP: 0.12; 95%CI: 0.03, 0.21; p <0.01). Sleep did not appear to be significantly associated with frailty by SPPB.

Finally, poorer sleep and insomnia also appeared to be associated with the development of CLAD and death after lung transplantation. For example, adjusting for age, sex, and transplant indication, each MID worsening in sleep by SPI was associated with a 1.14-fold increased risk of CLAD (HR 1.14; 95%CI: 1.00, 1.30; p=0.04) and a 1.29-fold increased risk of death (HR: 1.29; 95%CI: 1.05, 1.58; p=0.01). The direction and magnitude of the associations were relatively consistent across analyses, although the strength of associations between binary SPI and insomnia with death after transplant did not reach statistical significance in tests. Results are detailed in Table 4 and in Figures 2 and 3.

• Download figure
• Open in new tab

Figure 2.

Kaplan Meier estimate of the association of disturbed sleep by MOS-Sleep Problems Index (panel A) and insomnia-specific subscale (panel B) defined as categorical variables on time to CLAD. Survival Area Plot illustrating the association of disturbed sleep by MOS-Sleep Problems Index (panel C) and insomnia-specific subscale (panel D) defined as continuous variables on time to CLAD.

• Download figure
• Open in new tab

Figure 3.

Kaplan Meier estimate of the association of disturbed sleep by MOS-Sleep Problems Index (panel A) and insomnia-specific subscale (panel B) defined as categorical variables on time to death. Survival Area Plot illustrating the association of disturbed sleep by MOS-Sleep Problems Index (panel C) and insomnia-specific subscale (panel D) defined as continuous variables on time to death.

Discussion

In this single-center study, we found that prolonged hospital stay after lung transplantation was associated with disturbed sleep and insomnia which, in turn, were associated with worse cognitive functioning, depressive symptoms and poorer HRQL. Disturbed sleep and insomnia were also associated with disability and frailty as well as increased risk of CLAD and mortality.

Our findings are consistent with evidence that emphasizes the critical role sleep plays in physical and mental health outcomes. They also add important insights into the relatively little that is currently known about the impact sleep has on patient-reported outcomes after lung transplantation. Among a small number of modest sized studies, poorer sleep after lung transplant has been associated with worse depressive symptoms and poorer scores on the general mental health component of the SF-36..^40–43 An understanding of poor sleep on other outcomes, however, remains largely unknown. Outside of lung transplantation, poorer sleep is associated with cognitive dysfunction and risk of frailty.^{16,18,44–46} For example, experimental and epidemiologic data supports links between poor sleep and impaired cognition, including difficulties in executive functioning, memory, and attention processes.¹⁷ Notably, cognitive dysfunction following major cardiothoracic surgery is common and, in lung transplant, is associated with increased risk of mortality.^26,47–49 Sleep disturbance during this important recovery period may contribute to future cognitive outcomes. Further, a pooled analysis supports short and excessively long sleep duration, as well as a longer sleep onset latency as risk factors for frailty.¹⁸

Our study has several notable strengths. First, it is one of the largest studies of sleep and patient reported outcomes in lung transplant. A recent scoping literature review of sleep quality following transplant identified just nine research studies with sample sizes ranging from 20 to 219 participants. Our study is the second largest to date and the first to link measures of sleep quality to outcomes beyond mental health. We newly identified associations between sleep and physical disability, frailty, CLAD, and even death after lung transplant, raising the possibility that sleep disturbance may be an important and modifiable behavioral pathway to improve quality life and clinical outcomes among lung transplant recipients.

Despite these strengths, it is important to highlight that our study design precludes causal inference testing or understanding mechanisms of association. For example, our measures of sleep were collected concurrent to other PROs. Thus, it remains unclear whether disturbed sleep caused depressive symptoms, poorer HRQL, and other PROs or whether the reverse may be true. Understanding the directionality of these relationships are key to further efforts including designing interventions aimed at improving patient-centered outcomes. Further, while the associations between disturbed sleep and cognitive, psychological, and HRQL outcomes may be more direct, the associations with physical impairments including frailty and CLAD may be less-so. Behaviorally-driven factors such as impaired motivation to exercise regularly and to consume a balanced diet could drive disability and frailty whereas depression and cognitive impairment could impact medication adherence, thereby increasing risk of CLAD. It is also plausible that sleep’s essential role in immune system dysregulation from disturbed sleep via enhanced systemic inflammation^20,50 and accelerated biological aging^19,51 could identify biological links between sleep and disability, frailty and CLAD. While plausible, these potential explanations should be considered speculative until more definitive research is performed.

There are additional limitations to bear in mind. This cohort was not originally designed to study the impact of sleep disturbances on outcomes after lung transplantation. Our convenience sampling strategy yielded a group of participants whose sleep was queried over a range of early post-operative years potentially limiting our ability to identify important peri- and post-operative predictors of poor sleep after transplant. We also lack information on how sleep disturbances change over the early post-operative period. In prior work by our team and by others, other factors after transplant improve for many in the first post-operative year. Whether persistently poor sleep is differentially associated with clinical outcomes than sleep that worsens or improves over time is unknown. Answers to this unknown are important when considering when to screen for sleep after transplant and when designing potential interventions. We also did not perform objective tests of sleep, which precluded us from determining which type of sleep impairment drove our findings. This has critical treatment implications as cognitive behavioral and pharmacologic interventions used to treat problems like insomnia are less effective for sleep disordered breathing conditions. While our insomnia subscale has face validity, we did not use a validated insomnia-specific measure. Thus, our findings related to insomnia should be interpreted with caution. Finally, although this cohort represents one of the largest studies of sleep after lung transplantation, it, nevertheless, is relatively modest in size.

Future efforts to study sleep longitudinally in the peri- and early post-operative period may help shed light on the causal role of sleep in lung transplant recovery. If our findings are confirmed, treatments for disturbed sleep attributable to disorders like insomnia and obstructive sleep apnea are effective, relatively easy to implement, and could be incorporated into post-transplant care both in person and remotely. Further, if our findings linking disturbed sleep with CLAD and mortality are confirmed, efforts to disambiguate their behavioral or immunological underpinnings could shed new insights into important outcomes in lung transplantation.