Figures
- FIGURE 1
Study flow chart. SDB: sleep disordered breathing.
- FIGURE 2
Cumulative adjusted hazard for incident malignancy in 1239 patients with diabetes mellitus type 2 and a) apnoea–hypopnoea index (AHI) <30 events·h−1 versus ≥30 events·h−1; b) oxygen desaturation index (ODI) <30 events·h−1 versus ≥30 events·h−1; and c) percentage of night-time spent with oxygen saturation <90% (tsat90%) <10.4% versus ≥10.4%.
- FIGURE 3
Cumulative adjusted hazard for incident malignancy in i) 511 females and ii) 728 males with a) apnoea–hypopnoea index (AHI) <30 events·h−1 versus ≥30 events·h−1; b) oxygen desaturation index (ODI) <30 events·h−1 versus ≥30 events·h−1; and c) percentage of night-time spent with oxygen saturation <90% (tsat90%) <10.4% versus ≥10.4%.
Tables
- TABLE 1
Baseline characteristics of the 1239 patients overall and according to severity of sleep disordered breathing
Entire cohort AHI <30 events·h−1 AHI ≥30 events·h−1 Patients 1239 1111 (89.7) 128 (10.3) Age years 67±9 66±9 68±8 Female 511 (41.2) 482 (43.4) 29 (22.7) BMI kg·m−2 30.9±5.3 30.5±5.2 33.6±5.9 Waist–hip ratio 0.95±0.08 0.95±0.08 1.00±0.08 Obesity 622 (50.2) 533 (48.0) 89 (69.5) HbA1c mmol·mol−1 51±11 51±12 50±9 Former or current smokers 701 (56.6) 624 (56.2) 77 (60.2) High alcohol intake# 353 (28.5) 309 (27.8) 44 (34.4) Socioeconomic status¶ 2 (1–2) 2 (1–2) 2 (2–2) Physical inactivity+ 699 (56.4) 620 (55.8) 79 (61.7) Follow-up time years 2.7 (2.2–4.5) 2.8 (2.2–4.5) 2.4 (2.1–4.5) Excessive daytime sleepiness§ 87 (7.1) 77 (7.0) 10 (7.8) AHI events·h−1 10 (5–19) 9 (4–15) 44 (36–52) ODI events·h−1 9 (5–19) 8 (5–15) 39 (33–49) tsat90% 10.4 (2.5–30.6) 8.8 (2.1–28.4) 23.5 (13.7–45.4) Data are presented as n, mean±sd, n (%) or median (interquartile range). AHI: apnoea–hypopnoea index; BMI: body mass index; ODI: oxygen-desaturation index; tsat90%: percentage of night-time spent with oxygen saturation <90%. #: ≥3 drinks per week; ¶: according to Robert Koch Institute (Germany) [37] subdivided into four groups ranging from 1 (lowest) to 4 (highest) and including educational level, professional qualification and income; +: light activity ≤2 times per week; §: Epworth Sleepiness Scale ≥11.
- TABLE 2
Tumour entities
Total sample 1239 Skin tumours# 20 (25.3) Prostate carcinoma 9 (11.5) Colorectal carcinoma 7 (8.9) Pancreatic carcinoma 7 (8.9) Breast cancer 5 (6.3) Pulmonary carcinoma 5 (6.3) Malignancy of the haematopoietic and lymphatic system 5 (6.3) Malignancy of the female genital organ 5 (6.3) Malignancy of the urinary organ 5 (6.3) Others¶ 11 (13.9) Total 79 (100) Males 728 Skin tumours# 15 (28.3) Prostate carcinoma 9 (17.0) Colorectal carcinoma 7 (13.2) Malignancy of the urinary organ 5 (9.4) Pulmonary carcinoma 4 (7.5) Pancreatic carcinoma 3 (5.7) Malignancy of the haematopoietic and lymphatic system 2 (3.8) Others¶ 8 (15.1) Total 53 (100) Females 511 Skin tumours# 5 (19.2) Breast cancer 5 (19.2) Malignancy of the female genital organ 5 (19.2) Pancreatic carcinoma 4 (15.4) Malignancy of the haematopoietic and lymphatic system 3 (11.6) Pulmonary carcinoma 1 (3.8) Others¶ 3 (11.6) Total 26 (100) Data are presented as n or n (%). #: malignant melanomas, squamous cell carcinomas and lentiginous melanomas (basal cell carcinomas were excluded); ¶: cancer types observed only in few or single persons, e.g. liver tumour, parotid tumour, meningioma.
- TABLE 3
Adjusted hazard ratios (HRs) for the incidence of malignancy in 1239 patients with type 2 diabetes during a median follow-up of 2.7 years (number of events 79)
Adjusted HR (95% CI)# p-value AHI events·h−1 AHI (continuous) 1.01 (0.99–1.03) 0.416 AHI ≥30 1.30 (0.64–2.62) 0.473 ODI events·h−1 ODI (continuous) 1.00 (0.98–1.02) 0.799 ODI ≥30 1.24 (0.57–2.68) 0.584 tsat90% tsat90% (continuous) 1.00 (1.00–1.01) 0.395 tsat90% ≥10.4% 1.59 (0.94–2.68) 0.085 HRs calculated using Cox proportional hazards regression analysis. The multivariable analyses were adjusted for sex, age, body mass index, smoking status, alcohol intake, socioeconomic status and HbA1c. AHI: apnoea–hypopnoea index; ODI: oxygen desaturation index; tsat90%: percentage of night-time spent with oxygen saturation <90%.
- TABLE 4
Adjusted hazard ratios (HRs) for the incidence of malignancy in 1239 patients with type 2 diabetes during a median follow-up of 2.7 years stratified by sex and age
Males Females ≥70 years <70 years Adjusted HR (95% CI) p-value Adjusted HR (95% CI) p-value Adjusted HR (95% CI) p-value Adjusted HR (95% CI) p-value Patients n 728 511 472 767 Events n 53 26 43 36 AHI events·h−1 AHI (continuous) 1.00 (0.97–1.02) 0.659 1.03 (1.00–1.06) 0.028 1.00 (0.98–1.03) 0.751 1.01 (0.98–1.03) 0.538 AHI ≥30 0.76 (0.31–1.87) 0.550 4.19 (1.38–12.77) 0.012 1.21 (0.48–3.09) 0.686 1.28 (0.44–3.78) 0.651 ODI events·h−1 ODI (continuous) 0.99 (0.96–1.01) 0.357 1.04 (1.00–1.07) 0.032 1.00 (0.97–1.03) 0.933 1.00 (0.97–1.03) 0.972 ODI ≥30 0.62 (0.22–1.79) 0.380 5.64 (1.78–17.90) 0.003 1.34 (0.53–3.36) 0.539 0.88 (0.20–3.83) 0.863 tsat90% tsat90% (continuous) 1.00 (0.99–1.01) 0.670 1.02 (1.00–1.03) 0.035 1.00 (0.99–1.01) 0.912 1.01 (1.00–1.02) 0.236 tsat90% ≥10.4% 1.28 (0.66–2.46) 0.467 2.57 (1.03–6.41) 0.043 0.91 (0.51–2.11) 0.913 2.43 (1.09–5.42) 0.030 HRs calculated by Cox proportional hazards regression analysis. The multivariable analyses were adjusted for sex, age, body mass index, smoking status, alcohol intake, socioeconomic status and HbA1c. AHI: apnoea–hypopnoea index; ODI: oxygen desaturation index; tsat90%: percentage of night-time spent with oxygen saturation <90%.
- TABLE 5
Association of sleep apnoea and incidence# of malignancy, comparison of existing studies
First author (year) [ref.] Population n (% female) T2D % Study design Follow-up years Sleep apnoea diagnosis Main outcomes Main findings Key limitations Sleep apnoea diagnosis with PSG or PG Prospective Driendl (2020; present study) 1239
(41)100 Prospective cohort study 2.7 PG Cancer incidence (n=79) Association between AHI ≥30 events·h−1 and cancer incidence in females Use of PG Marshall (2014) [39] 390
(26)3 Prospective cohort study 20 PG Cancer incidence (n=125) Association between elevated RDI (≥15 events·h−1) and cancer incidence Small population
Lack of control of some cancer risk factors
Lack of information on PAP therapyRetrospective Justeau (2020) [40] 8748
(36)15 Retrospective cohort study, multicentre 5.8 PSG, PG Cancer incidence (n=718) Association between nocturnal hypoxaemia (tsat90% >13%) and cancer incidence Lack of control of some cancer risk factors
Partial use of PGCampos- Rodriguez (2013) [9] 4910
(33)n/s Retrospective cohort study, multicentre 4.5 PSG (32%), PG (68%) Cancer incidence
(n=261)Association between severe OSA (tsat90% >12%) and cancer incidence, limited to male patients <65 years Lack of control of some cancer risk factors
Major use of PGBrenner (2019) [10] 5243
(26)n/s Retrospective cohort study 5.9 PSG Cancer incidence (n=265) Association between AHI >57 events·h−1 and cancer incidence for patients <45 years Lack of control of some cancer risk factors
Lack of information on PAP therapyKendzerska (2014) [41] 10 149
(38)14 Retrospective cohort study, multicentre 7.8 PSG Cancer incidence (n=627) No association between OSA and cancer incidence Cross-sectional Pataka (2019) [42] 19 556
(29)n/s Cross-sectional analysis, multicentre PSG, PG Cancer prevalence (n=388) Association between cancer prevalence and OSA and nocturnal hypoxaemia in females Lack of control of some cancer risk factors
Lack of information on PAP therapyMeta-analysis Shantha (2015) [21] 112 228
(26–100)4–22 Meta-analysis, five studies 4.5–20 PSG, PG Cancer incidence
(n=864)Patients with SDB had a nearly 50% greater overall cancer risk compared with patients without SDB Zhang (2017) [22] 86 460
(26–38)n/s Meta-analysis, six studies 4.5–20 PSG, PG Cancer incidence (n=965) OSA was not independently associated with cancer incidence OSA diagnosed according to ICD-9 or symptoms Prospective Christensen (2013) [43] 8783
(55)n/s Prospective cohort study 13 Symptoms of OSA Cancer incidence (n=1985) No association between symptoms of OSA and cancer incidence OSA diagnosis based on symptoms
Lack of information on PAP therapyRetrospective Gozal (2016) [11] 5.6 million
(50)14 in OSA-group Retrospective cohort study 3.2–3.9 According to ICD-9-CM Cancer incidence (n=167 022) Elevated risk for malignant melanoma and kidney and pancreatic cancer for patients with OSA
Lower risk for colorectal, breast and prostate cancer for patients with OSAPotential bias by use of administrative claims database
Lack of control of some cancer risk factorsSillah (2018) [12] 34 402
(43)n/s Retrospective cohort study 5.3 According to ICD-9-CM Cancer incidence (n=1575) Elevated risk for malignant melanoma and kidney, uterine and breast cancer for patients with OSA
Lower risk for colorectal and lung cancer for patients with OSALack of control of some cancer risk factors
Lack of information on PAP therapyT2D: type 2 diabetes mellitus; PSG: polysomnography; PG: polygraphy; OSA: obstructive sleep apnoea; ICD-9-CM: International Classification of Diseases, 9th Revision, Clinical Modification; AHI: apnoea–hypopnoea index; RDI: respiratory disturbance index; n/s: not specified; PAP: positive airway pressure; SDB: sleep disordered breathing; tsat90%: percentage of night-time spent with oxygen saturation <90%. #: prevalence shown for Pataka et al. [42].
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary tables 00036-2021.supplement
