Introduction
Hodgkin’s lymphoma (HL) survivors frequently experience unexplained chronic fatigue, which can negatively impact long-term physical and psychosocial wellbeing [1]. Mantle radiation therapy (RT) revolutionized HL treatment and was the first-line curative treatment for almost all HL patients from the 1970s to the late 1990s; however, it is also an important cause of morbidity and mortality in long-term survivors [2]. RT for head and neck cancers is associated with both fatigue and obstructive sleep apnea (OSA), likely due to RT-associated anatomic and/or neural changes [3]. As mantle RT for HL extends from the chest into the neck region, we hypothesized that OSA in this population may be an under-recognized, potential contributor to fatigue.
Methods
All HL survivors treated with mantle RT at three teaching hospitals, ≥ 15 years old at diagnosis, > 5 years post-treatment, asymptomatic for cardiovascular disease, and who participated in a cardiac screening protocol between March 2004 and July 2008 were included. Full cohort details were previously published [4]. The Dana-Farber/Harvard Cancer Center Institutional Review Board approved this study.
Part I: prospective analysis
During a prospective clinical cardiology screening visit, patients were interviewed about (1) fatigue and (2) sleep-related symptoms. All patients completed the Functional Assessment of Chronic Illness Therapy—Fatigue (FACIT-F) questionnaire. Lower FACIT-F score indicates greater fatigue. Since fatigue may be caused by sleep issues, patients were also assessed for (1) daytime sleepiness, (2) sleep issues, including trouble falling/staying asleep, (3) snoring, and (4) observed/reported apneas. Patients with ≥ 2 sleep symptoms were all referred for clinical diagnostic overnight polysomnography (PSG) at accredited sleep centers. FACIT-F scores were not used as referral criteria. Clinical diagnoses of OSA positivity were made by sleep physicians based on PSG apnea–hypopnea index (AHI) > 5 events/h. Neck circumferences were recorded.
Part II: retrospective analysis
Approximately 20 years after the original prospective protocol, systematic retrospective chart review of all participants was undertaken. Subsequent new clinical OSA diagnoses (conferred outside the prospective protocol) and home/in-lab sleep study results were recorded. Clinical history of dropped head syndrome (severe weakness of the neck extensor muscles causing the chin to drop toward the chest) and history of atrial fibrillation/flutter, heart failure, and all-cause mortality were extracted from medical records. Relationships between OSA diagnosis and events were explored using Poisson regression and presented as rate ratios (number of events divided by years follow-up in OSA versus non-OSA groups). Statistical analyses were performed using SAS 9.4 or Stata v.17.
Results
One hundred eighty-two HL survivors with prior mantle RT were included (Table 1). During the prospective study (part I), 45 patients had ≥ 2 sleep symptoms and were all referred for clinical PSG. These patients had lower median FACIT-F scores than the rest of the cohort (44 versus 48, p < 0.001). Twenty-five of the 45 patients referred to in-lab PSG chose to attend testing. Chart review (part II), performed ~ 20 years after the prospective protocol, found an additional 17 patients from the cohort who underwent sleep testing. Therefore, a total of 42 patients underwent sleep testing in this cohort. Results are expressed as count (%) for categorical data or median (range) for continuous data. Categorical data was compared between groups using Fisher’s exact test while continuous data was compared using the Wilcoxon rank sum test
Table 1.
Demographics/treatment, sleep study data, and cardiac events: overall and by obstructive sleep apnea diagnosis
| Total (n = 182) | OSA-positive patients (n = 36) | Part I* OSA-positive patients (n = 21) | Part II* OSA-positive patients (n = 15) | p value (part I vs. II) | ||
|---|---|---|---|---|---|---|
| Baseline characteristics | ||||||
| Male sex | 73 (40%) | 22 (61%) | 17 (81%) | 5 (33%) | 0.01 | |
| Age at completed RT (years) | 28 (13–55) | 27 (14–42) | 28 (14–42) | 26 (17–41) | 0.57 | |
| Age at study enrollment (years) | 43 (21–65) | 44 (31–62) | 47 (31–62) | 44 (40–49) | 0.63 | |
| Follow up time since RT to study (years) | 15 (5–36) | 16 (6–36) | 13 (5–35) | 17 (5–26) | 0.75 | |
| Total RT dose (Gray) (n = 181, 36, 21, 15) | 3960 (2550–5325) | 4000 (3240–4564) | 3860 (3320–4410) | 4060 (3240–4564) | 0.04 | |
| BMI at enrollment (kg/m2) | 25.7 (17.9–45.5) | 26.8 (19.6–45.5) | 26.6 (22.5–45.5) | 26.9 (19.6–39.4) | 0.52 | |
| Overweight (BMI ≥ 25 and < 30) | 66 (36%) | 19 (53%) | 13 (62%) | 6 (40%) | 0.31 | |
| Obese (BMI ≥ 30) | 37 (20%) | 7 (19%) | 4 (19%) | 3 (20%) | 1.0 | |
| Family history of CVD (n = 181, 36, 21, 15) | 55 (31%) | 14 (39%) | 8 (38%) | 6 (40%) | 1.0 | |
| Taking HTN medication | 24 (13%) | 9 (25%) | 6 (29%) | 3 (20%) | 0.71 | |
| Elevated hs-CRP (n = 177, 36, 21, 14) | 66 (37%) | 10 (28%) | 6 (29%) | 4 (29%) | 1.0 | |
| Elevated glucose or DM medication | 8 (4%) | 2 (6%) | 1 (5%) | 1 (7%) | 1.0 | |
| Prior smoking | 63 (35%) | 13 (36%) | 8 (38%) | 5 (33%) | 1.0 | |
| Recent smoking | 11 (6%) | 4 (11%) | 3 (14%) | 1 (7%) | 0.63 | |
| Exercise (days/week) | 3 (0–7) | 2 (0–7) | 3 (0–7) | 2 (0–5) | 0.30 | |
| FACIT-F (n = 176, 34, 20, 14) | 47 (17–56) | 45 (18–51) | 43 (18–51) | 48 (27–51) | 0.07 | |
| Sleep study characteristics | ||||||
| Age at sleep study | 55 (28–66) | 50 (28–62) | 60 (53–66) | 0.001 | ||
| BMI at sleep study (kg/m2) | 27.5 (21.0–44.8) | 27.5 (21.6–44.8) | 24.6 (21.0–41.0) | 0.45 | ||
| Epworth Sleepiness Scale, out of 24 (n = 26, 18, 8) | 6 (1–15) | 6 (1–15) | 6 (2–12) | 0.82 | ||
| Snoring (n = 23, 17, 6) | 18 (78%) | 14 (82.4%) | 4 (66.7%) | 0.58 | ||
| Total AHI (events/h) (n = 28, 18, 10) | 19.5 (5.8–49.4) | 21.0 (7.0–49.4) | 14.2 (5.8–42.7) | 0.38 | ||
| Total hypopneas (n = 24, 16, 8) | 67 (6–432) | 67 (20–432) | 71 (6–173) | 0.56 | ||
| Obstructive apneas (n = 21, 13, 8) | 1 (0–48) | 1 (0–23) | 2 (0–48) | 0.28 | ||
| Central apneas (n = 21, 13, 8) | 1 (0–7) | 1 (0–7) | 1 (0–5) | 0.97 | ||
| Mixed apneas (n = 21, 13, 8) | 0 (0–4) | 0 (0–4) | 0 (0–1) | 0.71 | ||
| Classification of OSA (n = 34, 21, 13) | 0.71 | |||||
| Mild | 17 (50%) | 9 (43%) | 8 (62%) | |||
| Moderate | 11 (32%) | 7 (33%) | 4 (31%) | |||
| Severe | 6 (18%) | 5 (24%) | 1 (8%) | |||
| Supine AHI (events/h) (n = 23, 15, 8) | 19.6 (2.7–164.4) | 26.7 (2.7–164.4) | 17.1 (5.9–90.0) | 0.46 | ||
| Non-supine AHI (events/h) (n = 21, 13, 8) | 9.3 (0–48.0) | 9.3 (0.0–48.0) | 7.9 (0.7–22.5) | 0.94 | ||
| Positional (supine-dependent) OSA (n = 22, 14, 8) | 14 (64%) | 9 (64%) | 5 (63%) | 1.0 | ||
| Severe supine OSA (AHI > 30 events/h) (n = 14, 9, 5) | 8 (57%) | 7 (78%) | 1 (20%) | 0.21 | ||
| Non-REM AHI (n = 15, 12, 3) | 11.1 (1.0–42.8) | 12.1 (1.5–33.4) | 9.8 (1.0–42.8) | N/Aa | ||
| REM AHI (n = 17, 14, 3) | 17.7 (0.0–51.7) | 23.4 (0.0–51.7) | 5.4 (0–41.5) | N/Aa | ||
| Baseline SaO2 (n = 23, 15, 8) | 94% (90–99%) | 94% (92–99%) | 95% (90–98%) | 0.45 | ||
| Min SaO2 (n = 28, 17, 11) | 86% (64–96%) | 88% (64–96%) | 84% (78–89%) | 0.029 | ||
| Cardiac events and all-cause mortalityb | ||||||
| OSA-positive patients (n = 36) | No documented OSA (n = 146) | Adjusted Rate Ratio | P value | |||
| Heart failureb | 17 (47%) | 31 (21%) | 1.94 | 0.007 | ||
| Atrial fibrillation or atrial flutter | 12 (33%) | 24 (16%) | 1.51 | 0.19 | ||
| All-cause death | 9 (25%) | 20 (14%) | 1.22 | 0.62 | ||
AHI apnea–hypopnea index, BMI body mass index, DM diabetes mellitus, FACIT-F Functional Assessment of Chronic Illness Therapy—Fatigue, hs-CRP high sensitivity C-reactive protein, HTN hypertension, N/A not available, OSA obstructive sleep apnea, PSG polysomnography, REM rapid eye movement, RT radiation therapy, SaO2 oxygen saturation
*Of the original 45 patients with ≥ 2 sleep symptoms, all of whom were referred to sleep testing, 30 underwent testing in either part I or II: 25 were OSA-positive and 5 were negative. Fifteen declined testing
aAHI during REM/NREM was not compared due to limited number of participants in part II with these data (n = 3), since most patients were diagnosed on home sleep studies
bEvent rate ratios are adjusted for sex and BMI. Patients could have more than 1 event. Of the 25 unique OSA patients with recorded cardiac events, 20 experienced the first event after OSA diagnosis (median 7.6 years) and 5 experienced events before date of diagnosis (median 3.3 years). Heart failure clinical events included both heart failure with systolic dysfunction and also preserved left ventricular ejection fraction
In part I, 21/25 tested patients were OSA-positive. In part II, 15/17 patients were OSA-positive, totaling 36 OSA-positive individuals out of 42 who underwent testing. Of the 6 OSA-negative patients, 3 had elevated respiratory disturbance index (31.2, 14.7, and 14.4 events/h). Overall, OSA-positive patients in parts I and II did not differ in demographic/treatment characteristics, although part I patients were younger at diagnosis. OSA patients were not obese (median BMI at sleep study = 27.5 kg/m2). All neck circumferences (median = 30 cm [women]; 38 cm [men]) were smaller than those usually associated with OSA risk in the general population (median = 38 cm [women]; 43 cm [men]) [5] (Fig. 1). Patients had far more hypopneas than apneas. Most patients with recorded supine/non-supine AHI (14/22, 64%) had positional OSA (supine/non-supine AHI-ratio > 2).
Fig. 1.
Characterization of obstructive sleep apnea in Hodgkin lymphoma survivors. A Lateral view of the thin neck phenotype in a representative patient. Note the neck muscle atrophy and small size of the neck (circumference: 31 cm) as compared with the patient’s head and shoulders. B Sagittal posterior view of the neck with prominent atrophy of the neck muscles. Overnight polysomnography showed severe OSA with a total of 289 hypopneas and 5 apneas. Written patient consent was obtained for photography
Finally, retrospective chart review (median 17.5 years follow-up) found that OSA diagnosis (n = 36) as compared to no documented diagnosis (n = 146) was significantly associated with dropped head syndrome, even after adjusting for sex and BMI (rate ratio = 1.97, p = 0.012). Regarding cardiac events, heart failure had a higher rate ratio in OSA-positive patients, adjusting for sex and BMI (Table 1). Atrial fibrillation/flutter rate ratio was elevated only on unadjusted analysis. All-cause mortality rate ratio was greater in the OSA group than in the rest of the cohort but not statistically significant.
Discussion
This is the first study to characterize OSA in long-term HL survivors who received mantle (neck and chest) radiation. The demographics and phenotype of OSA-positive HL patients in our cohort deviated from those typically considered high-risk [5]: they were not obese, were younger in age, and had normal-sized/slender necks. Without these typical risk factors, at least 3 patients referred to PSG could not obtain insurance coverage and could not undergo testing. Given that HL survivors have a higher risk of cardiovascular death relative to the general population, and that OSA-positive HL survivors had a higher heart failure rate ratio compared to non-OSA survivors in this study, OSA diagnosis and treatment may be important for long-term survivorship care [6]. With greater awareness of this atypical OSA phenotype in HL survivors, physicians may sooner suspect OSA in this population.
OSA is classically more common in patients with large necks due to prolapse of redundant tissue into the upper airway. This may be less common in HL patients treated with RT, since they had slender/normal-sized necks in our study. Neck radiation has been associated with decreased snoring in HL survivors with a dose-dependent relationship, indicating RT might reduce airway collapsibility [7]. This may be consistent with our observed preponderance of hypopneas compared with apneas and positional OSA [8]. We conjecture that potential OSA mechanisms in this population may involve RT-related fibrosis, neuromuscular injury, and altered ventilatory control (loop gain); however, this remains untested.
Future studies may explore OSA screening methodologies for HL survivors. STOP-BANG’s anatomic/demographic criteria—BMI > 35 kg/m2, age > 50 years, neck circumference > 40 cm, and male gender—may be insufficient to assess OSA risk in this cohort due to these patients’ atypical anatomical features [9]. FACIT-F score, which was lower in OSA patients, may contribute an additional clue to any occult OSA, beyond traditional variables.
This study is limited in its generalizability, as data may not reflect findings associated with newer, radiation-sparing RT protocols. Since patients underwent PSG at multiple institutions/time-points, data accessibility was incomplete for some patients. Selection bias is also present, since (a) only symptomatic individuals were referred for clinical PSG, and (b) some referred symptomatic patients did not attend their PSG; therefore, the true number of OSA patients may be underestimated. OSA may be under-diagnosed in the comparator group, and ascertainment and detection bias may inflate observed associations. Underestimation of AHI may be present in part II patients diagnosed on home sleep studies (n = 12) [10]. Larger, prospective, and multicenter investigations can further characterize sleep demographics in this population and estimate prevalence.
In conclusion, unsuspected OSA may be present in HL survivors treated with high-dose chest and neck radiation, who often present with an atypical OSA phenotype. These survivors are typically non-obese and relatively young, with many having slender necks and positional OSA with predominant hypopneas. HL survivors with OSA had an increased heart failure rate ratio compared to those with no known OSA. Therefore, urgently increasing awareness of this OSA phenotype among patients and their internists, pulmonologists, oncologists, sleep specialists, and cardiologists may improve long-term outcomes in HL survivors treated with RT.
Acknowledgements
Ming Hui Chen, MD, is grateful to the participation of all our Hodgkin lymphoma survivors, and especially the late Dr. Martha Praught, who shared her own story to increase awareness of OSA and the difficulty in getting a diagnosis secondary to a previously undescribed OSA phenotype.
Author contribution
Conceptualization: Ming Hui Chen; methodology: Anita Li, Andrea K. Ng, Kimberlee Gauvreau, Ming Hui Chen; formal analysis and investigation: Anita Li, Laura C. Petishnok, Sonia F. Epstein, Michelle L. Walsh, Kimberlee Gauvreau, Ming Hui Chen; writing—original draft preparation: Anita Li, Laura C. Petishnok, Sonia F. Epstein, Michelle L. Walsh, Ming Hui Chen; writing—review and editing: Anita Li, Laura C. Petishnok, Sonia F. Epstein, Kimberlee Gauvreau, Elizabeth B. Klerman, Ming Hui Chen; funding acquisition, resources, and supervision: Ming Hui Chen.
Funding
This work was supported by the Boston Children's Hospital Translational Fund in Cardiology and Oncology.
Data availability
The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request. Data are under controlled access at Boston Children’s Hospital.
Declarations
Ethics approval and consent to participate
This study was approved by the Dana-Farber/Harvard Cancer Center Institutional Review Board and performed in line with the principles of the Declaration of Helsinki. Informed consent was obtained from all individual participants included in the prospective study.
Consent for publication
The authors affirm a research participant provided informed consent for publication of the images in Fig. 1.
Competing interests
EBK—(a) Travel: Sleep Research Society, Lighten Up/EPFL Pavilions, World Sleep Society, Sante Fe Institutes, Society for Research in Biological Rhythms, Lorenz Center. (b) Consulting: National Sleep Foundation, Circadian Therapeutics, Buck Institute for Research on Aging, Guidepoint. (c) Honoraria: Sleep Research Society. (d) Scientific Advisory Board: Chronsulting (unpaid). Partner is founder and CSO of Chronsulting. All other authors have no competing interests.
Footnotes
Anita Li and Laura C. Petishnok are first authors, contributed equally, and listed alphabetically.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Daniëls LA, Oerlemans S, Krol AD, van de Poll-Franse LV, Creutzberg CL. Persisting fatigue in Hodgkin lymphoma survivors: a systematic review. Ann Hematol. 2013;92(8):1023–32. 10.1007/s00277-013-1793-2. [DOI] [PubMed] [Google Scholar]
- 2.Adams MJ, Constine LS, Lipshultz SE. Late effects of therapy for Hodgkin’s lymphoma. Curr Hematol Malig Rep. 2007;2(3):143–50. 10.1007/s11899-007-0020-4. [DOI] [PubMed] [Google Scholar]
- 3.Zhou J, Jolly S. Obstructive sleep apnea and fatigue in head and neck cancer patients. Am J Clin Oncol. 2015;38(4):411–4. 10.1097/01.coc.0000436086.61460.cb. [DOI] [PubMed] [Google Scholar]
- 4.Chen MH, Blackington LH, Zhou J, et al. Blood pressure is associated with occult cardiovascular disease in prospectively studied Hodgkin lymphoma survivors after chest radiation. Leuk Lymphoma. 2014;55(11):2477–83. 10.3109/10428194.2013.879716. [DOI] [PubMed] [Google Scholar]
- 5.Yayan J, Rasche K. A systematic review of risk factors for sleep apnea. Prev Med Rep. 2024;42:102750. 10.1016/j.pmedr.2024.102750. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.de Vries S, Schaapveld M, Janus CPM, et al. Long-term cause-specific mortality in Hodgkin lymphoma patients. J Natl Cancer Inst. 2021;113(6):760–9. 10.1093/jnci/djaa194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Daniel LC, Wang H, Brinkman TM, et al. Mechanisms of sleep disturbances in long-term cancer survivors: a childhood cancer survivor study report. JNCI Cancer Spectr. 2024;8(2):pkae010. 10.1093/jncics/pkae010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Wang X, Zhou T, Huang W, Zhang J, Zou J, Guan J, et al. Differences in physiologic endotypes between nonpositional and positional OSA: results from the Shanghai Sleep Health Study cohort. Chest. 2024;166(1):212–25. 10.1016/j.chest.2024.01.021. [DOI] [PubMed] [Google Scholar]
- 9.Chung F, Yegneswaran B, Liao P, et al. STOP questionnaire: a tool to screen patients for obstructive sleep apnea. Anesthesiology. 2008;108(5):812–21. 10.1097/ALN.0b013e31816d83e4. [DOI] [PubMed] [Google Scholar]
- 10.Borsini E, Nogueira F, Nigro C. Apnea-hypopnea index in sleep studies and the risk of over-simplification. Sleep Sci. 2018;11(1):45–8. 10.5935/1984-0063.20180010. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are not openly available due to reasons of sensitivity and are available from the corresponding author upon reasonable request. Data are under controlled access at Boston Children’s Hospital.