Abstract
Aim
A bidirectional relationship exists between obstructive sleep apnoea (OSA) and type 2 diabetes (T2D). We aimed to examine the cumulative impact of having both OSA and T2D on patient outcomes, relative to having either condition alone.
Materials and methods
Using TriNetX, a global federated research network (n = 128 million), we undertook two retrospective cohort studies, using time‐to‐event analysis. Analysis 1 compared OSA with T2D versus OSA alone; analysis 2 compared T2D with OSA versus T2D alone. Propensity score matching using greedy nearest neighbour (calliper 0.1) balanced the cohorts (1:1) for significant covariates. Primary outcomes were cardiovascular, liver, diabetes‐related (microvascular) and cancer events over 1–5 years.
Results
Analysis 1 (n = 179 688): A codiagnosis of T2D/OSA significantly increased risk of all‐cause mortality (hazard ratio [HR] 1.52; confidence interval [CI]: 1.48, 1.57), dementia (HR 1.19; CI: 1.12, 1.26), liver (HR 2.20; CI: 1.77, 2.73), pancreatic (HR 1.62; CI: 1.35, 1.93), colon, renal and endometrial cancers; all cardiovascular, microvascular and liver related outcomes versus OSA alone over 1–5 5 years following OSA diagnosis. Analysis 2 (n = 240 094): A codiagnosis of OSA/T2D significantly increased the risk of peripheral (HR 1.39; CI: 1.36, 1.43) and autonomic (HR 1.63; CI: 1.51, 1.75) neuropathy; retinopathy (HR 1.13; CI: 1.09, 1.18), CKD (HR 1.21; CI: 1.18, 1.23); all cardiovascular and liver outcomes; all‐cause mortality and several obesity related cancers versus T2D alone.
Conclusions
T2D significantly potentiates risk of cardiovascular, malignancy and liver‐related outcomes in individuals with OSA. OSA, in individuals with T2D, significantly potentiates risk of cardiovascular disease, malignancy, death and several microvascular complications (retinopathy, CKD, peripheral/autonomic neuropathy).
Keywords: cardiovascular disease, dementia, malignancy, microvascular complications, obstructive sleep apnoea, type 2 diabetes
1. INTRODUCTION
The obesity pandemic has driven an adiposopathy epidemic, with a rapid rise in incidence of adiposity‐related complications including cardiovascular disease (CVD), metabolic dysfunction–associated steatotic liver disease (MASLD), type 2 diabetes (T2D) and obstructive sleep apnoea (OSA). 1 This clustering of such complications is associated with significant morbidity, mortality and impaired quality of life. 2 The cooccurrence of T2D and OSA suggests a bidirectional relationship likely exists independent of BMI or waist circumference as surrogates for adiposity, 3 posing a distinct challenge. This is biologically plausible given features of OSA (intermittent hypoxaemia, autonomic nervous system [ANS] overactivity and sleep fragmentation) promote elevated systemic inflammation, insulin resistance, glucose intolerance and β‐cell dysfunction, whilst, conversely, features of T2D (diabetic neuropathy, insulin resistance, chronic inflammation and oxidative stress) promote sleep‐disordered breathing through ANS and ventilatory control dysfunction 4 ; however, as these studies did not directly measure adiposity or fat distribution, it remains possible associations are purely related to differences in these measures between individuals.
This bidirectional relationship is crucial given the individual risks inherent with both T2D and OSA. T2D is associated with major adverse cardiovascular events (MACE), heart failure, major adverse liver outcomes (MALO) and site‐specific cancers. 5 Previously, MACE was the leading cause of death in patients with T2D 6 ; however, over the last decade, cancer mortality, and specifically, colorectal, hepatocellular, gallbladder, breast, endometrial and pancreatic cancers, has surpassed MACE. 7 , 8 Similarly, OSA amplifies CVD risk, 9 with the American Heart Association advocating CVD risk reduction in patients with OSA. 10 Moreover, OSA is associated with more aggressive MASLD, 11 whilst also increasing risk of separate site‐specific cancers to T2D. 12
Clinical trials have shown effective cardiovascular risk reduction with lipid‐lowering therapy, antihypertensive, sodium‐glucose cotransporter‐2 inhibitors (SGLT2i) and novel incretin–based glucose‐lowering therapy has reduced mortality, incidence of CVD, including heart failure 13 , 14 and microvascular complications in patients with T2D. 15 However, given the global prevalence and increasing incidence of T2D, the burden of these complications remains challenging on a population level. Therefore, identifying specific at‐risk groups to target preventive strategies towards is crucial. Patients with concurrent T2D and OSA may represent one such high‐risk phenotype, although current literature has not adequately defined the extent to which the cooccurrence of T2D and OSA impacts on CVD, MASLD, diabetes‐related complications and cancer, beyond either pathology in isolation. It is plausible that multiple, distinct mechanistic pathways are implicated that may potentiate risk of these complications, and these patients warrant targeted screening.
The aims of this study were to assess whether a cumulative impact exists on cardiovascular, liver, diabetes‐related and cancer outcomes, by having a co‐diagnosis of T2D and OSA.
2. MATERIALS AND METHODS
2.1. Study populations
We performed a large retrospective cohort study using the TriNetX dataset. TriNetX is a global federated research network with access to inpatient and outpatient health‐related data, including diagnoses, procedures, medications and laboratory values. Using their Global Collaborative Network, we can access the anonymised health data of 128 million patients aged 18 or over, from 17 different countries, from predominantly US healthcare organizations (HCOs). To assimilate data from HCOs that join the network, data are mapped to a common data model to reflect individual institution, country and regional standards regarding electronic health record data. All data collection, processing and transmission are performed in compliance with all Data Protection laws applicable to the contributing HCOs, including the EU Data Protection Law Regulation 2016/679, the General Data Protection Regulation on the protection of natural persons regarding the processing of personal data and the Health Insurance Portability and Accountability Act, the US federal law which protects the privacy and security of healthcare data. The TriNetX Global Collaborative Network is a distributed network (with most HCOs located in the USA), and analytics are performed at the HCO with only aggregate results being surfaced and returned to the platform. Data usage and publication agreements are in place with all HCOs. Data for this paper were extracted from TriNetX on the 21st of June 2024. Further information on the network has been described in detail elsewhere. 16
2.2. Building cohorts in TriNetX
2.2.1. Analysis 1: Comparison of OSA with T2D versus OSA alone (OSA + T2D vs. OSA)
Adult patients (age ≥ 18 years) diagnosed with OSA were identified using the ICD‐10 CM: G47.33 code; additionally, diagnosis must have been at least 5 years before the date of data extraction. Patients were then split into two cohorts based on their T2D status, ICD‐10 CM: E11. In the OSA + T2D cohort, patients were included if they had their first coding of T2D within 12 months of OSA diagnosis. Patients with type 1 diabetes (ICD‐10 CM: E10) are excluded from this cohort. In the OSA only cohort, patients were excluded if they had a diagnosis of diabetes, ICD‐10 CM: E08‐E013. The two cohorts were propensity score matched (PSM) 1:1, using greedy nearest neighbour matching and a calliper of 0.1, for age at index event, sex, ethnicity (white, black or African American), prior diagnoses of hypertension, ischaemic heart disease (IHD), dyslipidaemia, chronic kidney disease (CKD) and body mass index (BMI). Covariates were deemed well matched if the standardized mean difference (SMD) was <0.1. The participants in this analysis were also matched on continuous positive airway pressure (CPAP) initiation. Importantly, not all patients had BMI data available at baseline, as such, a separate sensitivity analysis was also undertaken to ensure this had no significant impact on the study findings.
2.2.2. Analysis 2: Comparison of T2D with OSA versus T2D alone (T2D + OSA vs. T2D)
Adult patients (age ≥ 18 years) diagnosed with T2D were identified using the ICD‐10 CM: E11 code; additionally, the diagnosis must have been at least 5 years before the date of data extraction. Patients with type 1 diabetes were excluded using the ICD‐10 CM: E10 code. Patients were then split into two cohorts based on their OSA status, ICD‐10 CM: G47.33. In the T2D + OSA cohort, patients were included if their first coding of OSA was within 12 months of T2D diagnosis. In the T2D, only cohort patients were excluded if they had a diagnosis of OSA. PSM for this analysis was the same as analysis one, except CPAP is replaced with HbA1c. Like BMI, not all patients had baseline HbA1c data available, so a sensitivity analysis was undertaken where both cohorts had 100% BMI and HbA1c data available to ensure this absence of data had no significant impact on the study findings.
2.3. Outcomes
Time to event analysis, using Cox proportional hazards models, was performed over a 1‐to‐5‐year period from the index event. Outcomes were not collected in the first 12 months after the index event as this would introduce an immortal time bias in favour of the combined cohorts (analysis one: OSA + T2D and analysis two: T2D + OSA). Outcome data were included from 366 to 1825 days after the index event. The primary outcomes were split into three main groups: cardiovascular events (CVE) (IHD, ischaemic stroke, heart failure and atrial fibrillation/flutter); cancer (hepatocellular, pancreatic, breast, colon, cholangiocarcinoma, renal, oesophageal and endometrial); microvascular complications (peripheral neuropathy, macular oedema, retinopathy, amputation, autonomic neuropathy, CKD and foot ulcers). Additional outcomes include all‐cause mortality, dementia, MASLD and metabolic dysfunction‐associated steatohepatitis (MASH). All data were collected directly from electronic medical records; no data were collected from government sources or registered persons databases, so mortality data may be underreported, see Table S1 for a full list of codes used for each outcome.
2.4. Identification of the index event and the time window
The index event defines the point at which outcome data can start to be collected, day 1. In analysis 1, the index event was OSA diagnosis, and in analysis 2, T2D diagnosis. In both analyses, outcome data were included from days 366 to 1825 after the index event. The results of a sensitivity analysis using data from days 1 to 1825 can be found in Tables S2 and S3. The choice to only use data from days 366 to 1825 was done to prevent an immortal time bias in favour of the combined cohorts, OSA + T2D in analysis 1 and T2D + OSA in analysis 2.
2.5. Sensitivity analyses
Several sensitivity analyses were completed and are included as supplemental information. These sensitivity analyses explored the effect of including the first year in the analyses, the impact of the absent BMI and HbA1c data at baseline and the effect of excluding patients initiated on CPAP in analysis one and excluding patients initiated on SGLT2i, glucagon‐like peptide‐1 receptor agonists (GLP‐1RA) and other incretin‐based therapies in analysis 2 (Tables [Link], [Link]).
2.6. Statistical analysis
Cox regression analyses were performed using the TriNetX platform for each outcome. These analyses are unadjusted for other covariates as the platform does not allow for the generation of adjusted Cox regression models. Patients who had experienced an outcome prior to the index event were excluded from that specific outcome analysis. Data were collected from 12 months after the date of the index event till 5 years after and was used to generate a hazard ratio (HR) with 95% confidence interval (CI), log rank test and Kaplan–Meier's survival curve. Censoring was applied to patients when there was no more data available. A patient was removed (censored) from the analysis after the last event in their electronic record. TriNetX uses the R's Survival package v3.2‐3 for survival analyses.
3. RESULTS
3.1. Propensity score matching and follow‐up
Prior to PSM in analysis one (OSA + T2D vs. OSA), there were 180 129 patients in the combined cohort and 1 126 548 patients in the OSA only cohort. The greatest differences in covariates were seen in age (SMD 0.538), IHD (SMD 0.410), CKD (SMD 0.369), hypertension (SMD 0.476), lipid disorders (SMD 0.376) and BMI (SMD 0.516). Following PSM, all covariates had a SMD <0.1, and were deemed well matched; cohort sizes were now matched at 179 688 patients. Median follow‐up (days) after PSM was 1775 (IQR 1349) for the combined cohort and 1825 (IQR 1019) in the OSA only group (Table 1).
TABLE 1.
Propensity score matching.
| Analysis 1: OSA + T2D vs. OSA | |||||||
|---|---|---|---|---|---|---|---|
| Covariate | Prepropensity score matching | Postpropensity score matching | |||||
| OSA + T2D | OSA | SMD | OSA + T2D | OSA | SMD | ||
| Sample size a | 180 129 | 1 126 548 | 179 688 | 179 688 | |||
| Age at index event—Mean ± SD (years) | 60.0 ± 12.8 | 51.5 ± 18.2 | 0.538 | 60.0 ± 12.8 | 60.1 ± 13.2 | 0.007 | |
| Sex (%) | Female | 38.7 | 39.8 | 0.022 | 38.7 | 38.9 | 0.004 |
| Male | 56.7 | 56.4 | 0.005 | 56.7 | 56.3 | 0.008 | |
| Ethnicity—White (%) | 64.5 | 68.0 | 0.075 | 64.5 | 64.6 | 0.003 | |
| Ethnicity—Black or African American (%) | 13.2 | 10.1 | 0.095 | 13.2 | 12.9 | 0.009 | |
| Hypertension a | 113 908 (63.4%) | 452 387 (40.2%) | 0.476 | 113 907 (63.4%) | 119 950 (66.8%) | 0.071 | |
| IHD a | 49 646 (27.6%) | 131 237 (11.7%) | 0.410 | 49 645 (27.6%) | 48 241 (26.8%) | 0.018 | |
| CKD a | 27 167 (15.1%) | 48 925 (4.4%) | 0.369 | 27 166 (15.1%) | 24 471 (13.6%) | 0.043 | |
| Dyslipidaemia a | 91 681 (51.0%) | 368 680 (32.8%) | 0.376 | 91 680 (51.0%) | 92 121 (51.3%) | 0.005 | |
| BMI—Mean ± SD (kg/m2) | 38.2 ± 8.9 | 33.7 ± 8.5 | 0.516 | 38.2 ± 8.9 | 37.9 ± 8.6 | 0.033 | |
| CPAP a | 6392 (3.6%) | 15 674 (1.4%) | 0.140 | 6391 (3.6%) | 5772 (3.2%) | 0.019 | |
| Analysis 2: T2D + OSA versus T2D | |||||||
|---|---|---|---|---|---|---|---|
| Prepropensity score matching | Postpropensity score matching | ||||||
| Covariate | T2D + OSA | T2D | SMD | T2D + OSA | T2D | SMD | |
| Sample size a | 240 533 | 4 829 108 | 240 094 | 240 094 | |||
| Age at index event—Mean ± SD (yrs) | 60.3 ± 12.6 | 61.4 ± 14.6 | 0.084 | 60.3 ± 12.6 | 60.3 ± 12.6 | 0.001 | |
| Sex (%) | Female | 40.0 | 48.6 | 0.173 | 40.0 | 40.0 | <0.001 |
| Male | 55.3 | 48.7 | 0.133 | 55.3 | 55.3 | <0.001 | |
| Ethnicity—White | 64.1 | 54.8 | 0.190 | 64.1 | 64.1 | <0.001 | |
| Ethnicity—Black or African American | 14.4 | 14.2 | 0.005 | 14.4 | 14.4 | <0.001 | |
| Hypertension a | 153 939 (64.1%) | 2 558 622 (55.5%) | 0.177 | 153 939 (64.1%) | 154 005 (64.1%) | 0.001 | |
| IHD a | 61 921 (25.8%) | 839 897 (18.2%) | 0.184 | 61 921 (25.8%) | 61 849 (25.8%) | 0.001 | |
| CKD a | 34 839 (14.5%) | 439 795 (9.5%) | 0.153 | 34 839 (14.5%) | 34 898 (14.5%) | 0.001 | |
| Dyslipidaemia a | 120 486 (50.2%) | 1 820 215 (39.5%) | 0.217 | 153 939 (64.1%) | 154 005 (64.1%) | <0.001 | |
| BMI—Mean ± SD (kg/m2) | 37.9 ± 8.8 | 32.2 ± 8.1 | 0.674 | 37.9 ± 8.8 | 37.4 ± 8.3 | 0.063 | |
| HbA1c—Mean ± SD (%) | 7.2 ± 1.9 | 7.3 ± 2.0 | 0.022 | 7.2 ± 1.9 | 7.3 ± 2.0 | 0.035 | |
Abbreviations: BMI, body mass index; CKD, chronic kidney disease; CPAP, continuous positive airway pressure; IHD, ischaemic heart disease; OSA, obstructive sleep apnoea; SD, standard deviation; SMD, standardized mean difference; T2D, type 2 diabetes.
Number of participants.
Prior to PSM in analysis 2 (T2D + OSA vs. T2D), there were 240 533 patients in the T2D + OSA cohort and 4 829 108 patients in the T2D only cohort. The greatest difference in covariate was seen in BMI (SMD 0.674). Following PSM, all covariates had a SMD <0.1; cohort sizes were now matched at 240 094. Median follow‐up (days) after PSM was 1818 (IQR 1249) in the T2D + OSA cohort and 1825 (IQR 1296) in the T2D only cohort (Figure 1).
FIGURE 1.
Study overview. (A) A timeline showing the period of data collection and how it relates to diagnoses of T2D and OSA. (B) A schematic demonstrating how the cohorts were created for the two analyses. OSA, obstructive sleep apnoea; T2D, type 2 diabetes.
3.2. Log rank test
For analysis 1 (OSA + T2D vs. OSA), the log rank test demonstrated a significant difference (p < 0.05) in the survival curves for all cardiovascular, microvascular, dementia, all‐cause mortality and steatotic liver disease–related outcomes. In cancer outcomes, there was a significant difference in endometrial, renal, colon, pancreatic and hepatocellular cancers with no significant difference in breast or cholangiocarcinoma. In analyses 2 (T2D + OSA vs. T2D), there was a significant difference in the survival curves for all cardiovascular outcomes, dementia, all‐cause mortality and steatotic liver disease–related outcomes. In microvascular and cancer outcomes, there was a significant difference in the survival curves for peripheral neuropathy, retinopathy, autonomic neuropathy, CKD, foot ulcers, breast and renal cancers. There was no significant difference in the remaining microvascular and neoplastic outcomes. See Table 2A,B for the complete log rank test results for analyses 1 and 2, respectively.
TABLE 2.
Summary of outcome incidence, survival probability and log rank test for analysis 1 (OSA + T2D vs. OSA) and analysis 2 (T2D + OSA vs. T2D).
| Analysis 1 (OSA + T2D vs. OSA) | Analysis 2 (T2D + OSA vs. T2D) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value | |
| Peripheral neuropathy | OSA + T2D | 159 220 | 11 424 | 88.91% | 4756.417 | <0.001 | T2D + OSA | 206 555 | 17 582 | 87.25% | 851.726 | <0.001 |
| OSA | 173 915 | 4440 | 96.34% | T2D | 219 922 | 13 636 | 90.60% | |||||
| Macular oedema | OSA + T2D | 178 340 | 953 | 99.17% | 690.635 | <0.001 | T2D + OSA | 237 255 | 1936 | 98.77% | 0.336 | 0.562 |
| OSA | 179 408 | 141 | 99.89% | T2D | 237 011 | 1969 | 98.75% | |||||
| Retinopathy (excluding macular oedema) | OSA + T2D | 175 744 | 2786 | 97.52% | 1849.826 | <0.001 | T2D + OSA | 232 377 | 5262 | 96.57% | 38.752 | <0.001 |
| OSA | 178 768 | 516 | 99.59% | T2D | 233 305 | 4671 | 96.96% | |||||
| Amputations | OSA + T2D | 179 247 | 335 | 99.70% | 171.261 | <0.001 | T2D + OSA | 239 324 | 574 | 99.63% | 0.013 | 0.909 |
| OSA | 179 539 | 87 | 99.93% | T2D | 239 278 | 579 | 99.63% | |||||
| Autonomic neuropathy | OSA + T2D | 178 474 | 1174 | 98.96% | 1206.978 | <0.001 | T2D + OSA | 237 806 | 1956 | 98.74% | 179.902 | <0.001 |
| OSA | 179 650 | 32 | 99.97% | T2D | 238 852 | 1212 | 99.23% | |||||
| CKD | OSA + T2D | 143 756 | 10 792 | 88.85% | 1680.037 | <0.001 | T2D + OSA | 188 323 | 16 381 | 87.47% | 268.630 | <0.001 |
| OSA | 151 987 | 6693 | 93.87% | T2D | 196 462 | 14 188 | 89.41% | |||||
| Foot ulcers | OSA + T2D | 175 287 | 3215 | 97.13% | 2038.687 | <0.001 | T2D + OSA | 232 493 | 5264 | 96.58% | 74.880 | <0.001 |
| OSA | 178 894 | 648 | 99.48% | T2D | 233 635 | 4449 | 97.11% | |||||
| Cardiovascular outcomes | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ischaemic heart disease | OSA + T2D | 119 135 | 11 361 | 85.97% | 884.953 | <0.001 | T2D + OSA | 156 775 | 16 572 | 84.88% | 311.632 | <0.001 |
| OSA | 126 465 | 8684 | 90.51% | T2D | 167 389 | 14 526 | 87.39% | |||||
| Heart failure | OSA + T2D | 135 033 | 10 093 | 89.08% | 918.152 | <0.001 | T2D + OSA | 179 064 | 14 604 | 88.43% | 405.814 | <0.001 |
| OSA | 151 270 | 7815 | 92.90% | T2D | 203 909 | 12 906 | 90.76% | |||||
| Atrial fibrillation | OSA + T2D | 144 102 | 7169 | 92.57% | 186.632 | <0.001 | T2D + OSA | 193 088 | 10 409 | 92.15% | 155.220 | <0.001 |
| OSA | 148 903 | 6366 | 94.04% | T2D | 210 877 | 9437 | 93.35% | |||||
| Ischaemic stroke | OSA + T2D | 170 287 | 3700 | 96.59% | 203.670 | <0.001 | T2D + OSA | 226 610 | 5640 | 96.21% | 42.181 | <0.001 |
| OSA | 172 303 | 2900 | 97.59% | T2D | 229 170 | 5048 | 96.65% |
| Neoplastic outcomes | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Liver cancer | OSA + T2D | 179 126 | 248 | 99.78% | 54.115 | <0.001 | T2D + OSA | 239 254 | 401 | 99.74% | 0.070 | 0.791 |
| OSA | 179 391 | 124 | 99.90% | T2D | 239 190 | 394 | 99.75% | |||||
| Pancreatic cancer | OSA + T2D | 178 804 | 298 | 99.74% | 28.315 | <0.001 | T2D + OSA | 238 891 | 411 | 99.74% | 0.432 | 0.511 |
| OSA | 179 338 | 203 | 99.84% | T2D | 239 018 | 431 | 99.72% | |||||
| Breast cancer | OSA + T2D | 177 274 | 861 | 99.24% | 0.219 | 0.640 | T2D + OSA | 236 677 | 1294 | 99.16% | 7.082 | 0.008 |
| OSA | 176 739 | 922 | 99.25% | T2D | 236 899 | 1165 | 99.24% | |||||
| Colon cancer | OSA + T2D | 178 427 | 501 | 99.56% | 15.550 | <0.001 | T2D + OSA | 238 275 | 747 | 99.52% | 0.096 | 0.756 |
| OSA | 178 473 | 425 | 99.66% | T2D | 238 350 | 760 | 99.51% | |||||
| Cholangiocarcinoma | OSA + T2D | 179 650 | 25 | 99.98% | 1.853 | 0.173 | T2D + OSA | 240 034 | 30 | 99.98% | 0.544 | 0.461 |
| OSA | 179 643 | 18 | 99.99% | T2D | 240 028 | 36 | 99.98% | |||||
| Renal cancer | OSA + T2D | 178 197 | 510 | 99.55% | 5.025 | 0.025 | T2D + OSA | 237 928 | 768 | 99.51% | 8.075 | 0.004 |
| OSA | 178 168 | 486 | 99.61% | T2D | 238 498 | 663 | 99.58% | |||||
| Oesophageal cancer | OSA + T2D | 179 323 | 128 | 99.89% | 2.459 | 0.117 | T2D + OSA | 239 544 | 186 | 99.88% | 0.567 | 0.451 |
| OSA | 179 391 | 115 | 99.91% | T2D | 239 680 | 172 | 99.89% | |||||
| Endometrial cancer | OSA + T2D | 178 757 | 284 | 99.75% | 10.159 | 0.001 | T2D + OSA | 238 740 | 402 | 99.74% | 0.075 | 0.784 |
| OSA | 178 894 | 236 | 99.81% | T2D | 238 872 | 395 | 99.74% |
| All‐cause mortality, dementia and liver outcomes | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value | Cohorts | Sample size a | Outcome | 5‐year survival probability | Log rank test | p value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All‐cause mortality | OSA + T2D | 179 688 | 14 841 | 87.73% | 1013.070 | <0.001 | T2D + OSA | 240 094 | 21 751 | 86.92% | 296.774 | <0.001 |
| OSA | 179 688 | 10 977 | 91.57% | T2D | 240 094 | 18 376 | 88.84% | |||||
| Dementia | OSA + T2D | 175 351 | 2524 | 97.76% | 35.657 | <0.001 | T2D + OSA | 233 954 | 3854 | 97.50% | 45.036 | <0.001 |
| OSA | 175 912 | 2343 | 98.11% | T2D | 234 870 | 3312 | 97.85% | |||||
| Metabolic dysfunction‐associated steatotic liver disease | OSA + T2D | 169 014 | 6435 | 94.01% | 719.643 | <0.001 | T2D + OSA | 225 473 | 9122 | 93.83% | 467.791 | <0.001 |
| OSA | 172 631 | 4366 | 96.37% | T2D | 230 440 | 6667 | 95.56% | |||||
| Metabolic dysfunction‐associated steatohepatitis | OSA + T2D | 177 528 | 1492 | 98.68% | 488.367 | <0.001 | T2D + OSA | 236 912 | 2137 | 98.63% | 150.173 | <0.001 |
| OSA | 178 839 | 594 | 99.52% | T2D | 238 383 | 1422 | 99.09% |
Abbreviations: CKD, chronic kidney disease; OSA, obstructive sleep apnoea; T2D: type 2 diabetes.
Number of participants.
3.3. Analysis 1: The impact of T2D on outcomes for individuals with OSA (OSA + T2D vs. OSA) from 12 months to 5 years post‐OSA diagnosis
3.3.1. Cardiovascular events (Figure 2A)
FIGURE 2.
Hazard ratios of people with OSA and T2D versus OSA over a 1‐ to 5‐year period from OSA diagnosis: (A) cardiovascular events, (B) malignancy, (C) microvascular complications, and (D) all‐cause mortality, dementia and fatty liver disease. CKD, chronic kidney disease; MASH, metabolic dysfunction–associated steatohepatitis; MASLD, metabolic dysfunction–associated steatotic liver disease; OSA, obstructive sleep apnoea; T2D, type 2 diabetes.
The development of T2D within a year of OSA diagnosis was associated with a significant increased risk for all CVE when compared to a T2D diabetes naive cohort. The largest increase in risk was seen in the development of heart failure (HR 1.57 CI: 1.53, 1.62).
3.3.2. Cancer diagnoses (Figure 2B)
A codiagnosis of T2D in individuals with OSA was associated with a significant increased risk of endometrial, renal, colon, pancreatic and hepatocellular cancers with the greatest increase in risk seen with hepatocellular cancer (HR 2.20 CI: 1.77, 2.73). There was no significant increase in the rate of breast, cholangiocarcinoma or oesophageal cancers.
3.3.3. Microvascular complications (Figure 2C)
The presence of T2D in individuals with OSA significantly increased the risk of all microvascular outcomes with the greatest increase in risk seen with autonomic neuropathy (HR 40.73 CI: 28.67, 57.86).
3.3.4. All‐cause mortality, dementia and fatty liver disease (Figure 2D)
The development of T2D within a year of OSA diagnosis was associated with a significant increased risk of all‐cause mortality (HR 1.49 CI: 1.45, 1.53), dementia (HR 1.19 CI: 1.12, 1.26) and steatotic liver disease.
3.4. Analysis 2: The impact of OSA on outcomes for individuals with T2D (T2D + OSA vs. T2D) from 12 months to 5 years postdiabetes diagnosis
3.4.1. Cardiovascular events (Figure 3A)
FIGURE 3.
Hazard ratios of people with T2D and OSA versus T2D over a 1‐ to 5‐year period from T2D diagnosis: (A) cardiovascular events, (B) malignancy, (C) microvascular complications, and (D) All‐cause mortality, dementia, and fatty liver disease. CKD, chronic kidney disease; MASH, metabolic dysfunction–associated steatohepatitis; MASLD, metabolic dysfunction–associated steatotic liver disease; OSA, obstructive sleep apnoea; T2D, type 2 diabetes.
The development of OSA within a year of T2D diagnosis was associated with a significant increased risk for all CVE when compared to a OSA naive cohort. The largest increase in risk was seen in the development of heart failure (HR 1.28 CI: 1.25, 1.31).
3.4.2. Cancer diagnosis (Figure 3B)
A codiagnosis of OSA in individuals with T2D was associated with a significant increased risk of breast (HR 1.11 CI: 1.03, 1.21) and renal (HR 1.16 CI: 1.05, 1.29) cancer. There was no significant increase in the rate of endometrial, cholangiocarcinoma, colon, pancreatic, hepatocellular or oesophageal cancer.
3.4.3. Microvascular complications (Figure 3C)
The presence of OSA in individuals with T2D significantly increased the risk of peripheral neuropathy, retinopathy, autonomic neuropathy, CKD and foot ulcer formation. The greatest increase in risk was seen with autonomic neuropathy (HR 1.63. CI: 1.51, 1.75). There was no significant change in the risk of macular oedema or amputation.
3.4.4. All‐cause mortality, dementia and fatty liver disease (Figure 3D)
The development of OSA within a year of T2D diagnosis was associated with a significant increased risk of all‐cause mortality (HR 1.19 CI: 1.17, 1.21), dementia (HR 1.17 CI: 1.12, 1.23) and fatty liver disease.
4. DISCUSSION
In this large retrospective real‐world study, we explored the impact of a codiagnosis of T2D and OSA on incident rates of CVE, obesity‐related malignancy, microvascular complications, MASLD, dementia and all‐cause mortality. We robustly demonstrate the strong cumulative impact of these conditions in amplification of risk for several significant clinical outcomes: CVE (IHD, heart failure, ischaemic stroke and AF), MASLD, dementia, microvascular complications, certain obesity related cancers and thus all‐cause mortality over a 4‐year period, when compared to each condition independently. Tentatively, our data also suggest a greater impact on the development of T2D in individuals with OSA on cardiovascular, microvascular and liver related outcomes compared to those with T2D who develop OSA supporting the notion that aggressive weight loss management in individuals with OSA may provide a greater health economic benefit through preventing the development of T2D. However, this finding could be due to variation in visceral and ectopic fat distribution between the cohorts and is not something controlled for in this study 17 ; this finding would need confirmation in prospective studies.
The increased risk of CVE and all‐cause mortality with both diagnoses is consistent with current literature. 18 , 19 , 20 With respect to microvascular complications, unsurprisingly, our data support the well‐established paradigm that development of T2D increases the risk of these complications. However, in analysis 2 (T2D + OSA vs. T2D), we establish that addition of a diagnosis of OSA in individuals with T2D significantly impacts the rate these complications develop, above that seen in individuals with only T2D, specifically autonomic neuropathy, peripheral neuropathy, CKD, diabetic foot ulcers and retinopathy. There is generally a paucity of large‐scale studies assessing the impact of OSA on diabetes‐related microvascular complications, and this study represents the largest study trying to address this. Previously, Zhang et al. demonstrated the degree of nocturnal hypoxia in individuals with OSA and T2D directly impairs renal function with estimated glomerular filtration rate (eGFR) independently correlating with oxygen desaturation index but failed to demonstrate a significant relationship between severity of OSA and another microvascular complication, retinopathy. 21 This finding was further confirmed in a cohort study by Tahrani et al., who demonstrated development of OSA in individuals with T2D was independently associated with both the development of diabetic nephropathy and a faster rate of decline in renal function over an average follow up period of 2.5 years. 22 With respect to peripheral and autonomic neuropathy, the independent association of OSA as a contributing factor has been evidenced previously, 23 although data quality is generally poor. A systematic review by Abelleira et al. highlighted a lack of longitudinal data, heterogenous design and small sample sizes as limitations in the ability of researchers to draw significant conclusions; however, they conclude a relationship exists between OSA with T2D and occurrence of peripheral neuropathy. 24 The literature on the impact of OSA on retinopathy development in individuals with T2D is mixed, with several studies demonstrating no significant relationship. 21 , 25 , 26 Conversely, an extensive systematic review and meta‐analysis by Simonson et al. identified a strong association between OSA and retinopathy, but due to the cross‐sectional natures of most of the studies, they were unable to comment on causality. 27
A novel inclusion in this study was assessing the impact of OSA on the development of MASLD (previously termed non‐alcoholic fatty liver disease) and dementia in individuals with T2D. OSA is a known risk factor for the development of MASLD 11 , 28 and dementia 29 with previous studies focusing on exploring its effect in the general population. We believe this is the largest study demonstrating this same effect in a population of individuals with T2D. We also demonstrate an increased association in rates of MASH in individuals with T2D who develop OSA. Conversely, it is an expected finding individuals with OSA who develop T2D have an increased risk of MASLD, MASH and dementia 30 as T2D has been a well‐established, independent risk factor for these conditions. 31 , 32
Finally, we also looked at the impact a codiagnosis of OSA and T2D on the risk of obesity‐related cancers. 33 The greatest significant increase in risk was seen for both pancreatic and liver cancer in individuals with OSA who developed T2D with relatively small, but still significant, increases in risk of renal, colon and endometrial cancers. These results should be interpreted within the constraints of this style of analysis and should be seen as identifying significant association rather than causation. However, the findings from this study largely mirror the findings from previous research examining the impact of diabetes on cancer risk. 34 The impact of a diagnosis of OSA in individuals with T2D was more trivial, but significant associations were still seen for renal and breast cancer. The general relationship between OSA and cancer risk is still relatively unclear in the literature; although largely accepted to increase overall cancer risk of, 35 , 36 the effect on specific cancer types is less clear.
The strengths of this study are the size of the population being assessed and the comprehensive array of outcomes explored over a moderate time frame. The limitations reflect the nature of using real‐world data, with data neither randomized nor controlled. Specifically, a significant limitation is that a proportion of patients with T2D and/or OSA are currently undiagnosed. This potentially leads to patients being included in the incorrect cohort and is a limitation inherent to retrospective real‐world studies. Additionally, clinical practices at the HCOs may differ when it comes to investigating, diagnosing and coding the different conditions and outcomes in this study. As the study only work with aggregated results, it is not possible to account for variation in practice between HCOs. Separately, we were unable to account for the effect of smoking or social economic status when designing this study, as this information is poorly captured in the TriNetX network. Additionally, we are reliant on electronic medical records, so data may be incomplete or absent. Data related to mortality are collected only from the electronic medical records and not from government sources and as such may be incomplete. With regard to microvascular outcomes, there is no record of how these diagnoses were made or the criteria/tests used, as such we are unable to account for variation in clinician habits when examining and reporting these health outcomes. Specifically, as microvascular complications are not routinely screened for in an OSA population without concomitant T2D, there may be under reporting of these conditions in that cohort. Other limitations relate to the inability to account for OSA severity, absent BMI data at baseline, CPAP initiation or variation in prescribing habits in individuals with T2D after the index event. We have performed several sensitivity analyses excluding patients who commenced CPAP during the study period, commenced SGLT2i or GLP‐1RA, and one that only included patients with BMI data at baseline, there was no significant difference in the results for cardiovascular, microvascular, liver, all‐cause mortality or dementia outcomes (Tables [Link], [Link]). There are small changes in the cancer outcomes, with supplemental analyses that had a reduced cohort size having wider confidence intervals due to fewer reported outcomes (Tables [Link], [Link], [Link], [Link]), and those with a longer time frame having smaller confidence intervals due to an increased number of outcomes (Tables S2 and S3), no result changed direction.
In conclusion, individuals who codevelop OSA and T2D have a significant increased risk for all CVE, dementia and MASLD, above that seen with either condition in isolation. OSA, in individuals with T2D, increases the risk for autonomic and peripheral neuropathy, diabetic retinopathy and diabetic nephropathy than in individuals with T2D, but no OSA. Both conditions have a significant association with certain obesity‐related malignancies. With the growing armament of licensed and pipeline antiobesity medication, the cumulative risk of these two common obesity‐related comorbidities is relevant in accurately risk stratifying patients and targeting highest risk individuals accordingly. The findings in this study will need confirmation in prospective longitudinal studies.
AUTHOR CONTRIBUTIONS
DRR contributed to the generation of the results and analysis using the TriNetX platform and took lead on writing the manuscript. AH assisted with the writing, reviewing and editing of the manuscript. MA assisted with reviewing and editing of the manuscript. GH facilitated access to the TriNetX platform and assisted in generating the results, analysis and review of the final manuscript. SSZ provided statistical support and review of the manuscript. UA, SC and JPHW provided senior author review of the manuscript, and DJC oversaw the study development, manuscript writing and provided senior author review of the manuscript. DJC is also the guarantor of this work and had full access to all the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.
FUNDING INFORMATION
No funding was received for this work.
CONFLICT OF INTEREST STATEMENT
DRR reports no conflict of interest. GH is an employee of TriNetX LLC. SSZ reports no conflict of interest. AH reports no conflict of interest. MA reports no conflict of interest. SC reports no conflict of interest. DJC has received investigator‐initiated grants from Astra Zeneca and Novo Nordisk and support for education from Perspectum. JPHW reports consultancy/advisory board work for the pharmaceutical industry contracted via the University of Liverpool (no personal payment) for Altimmune, AstraZeneca, Boehringer Ingelheim, Cytoki, Lilly, Napp, Novo Nordisk, Menarini, Pfizer, Rhythm Pharmaceuticals, Sanofi, Saniona, Tern and Shionogi; research grants for clinical trials from AstraZeneca and Novo Nordisk and personal honoraria/lecture fees from AstraZeneca, Boehringer Ingelheim, Medscape, Napp, Novo Nordisk and Rhythm. He is past president of the World Obesity Federation, a member of the Association for the Study of Obesity, Diabetes UK, EASD, ADA, Society for Endocrinology and the Rank Prize Funds Nutrition Committee. From 2009 to 2024, he was national lead for the Metabolic and Endocrine Speciality Group of the UK NIHR Clinical Research Network. UA has received honoraria from Procter & Gamble, Viatris, Grunenthal, Eli Lilly and Sanofi for educational meetings. UA has also received investigator‐led funding by Procter & Gamble.
PEER REVIEW
The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.16059.
ETHICS STATEMENT
No ethical approval is required for this study. As a federated network, research studies using TriNetX do not require ethical approval. To comply with legal frameworks and ethical guidelines guarding against data reidentification, the identity of participating healthcare organizations and their individual contribution to each dataset are not disclosed. The TriNetX platform only uses aggregated counts and statistical summaries of deidentified information. No protected health information or personal data are made available to the users of the platform.
Supporting information
Supplemental Table S1. Codes used for outcome analyses.
Supplemental Table S2. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only addition that the time window is from 1 to 1825 days.
Supplemental Table S3. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only addition that the time window is from 1 to 1825 days.
Supplemental Table S4. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only addition that patients are required to have BMI data available at baseline.
Supplemental Table S5. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only addition that patients are required to have both BMI and HbA1c data available at baseline.
Supplemental Table S6. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only change that patients who initiated CPAP are excluded.
Supplemental Table S7. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only change is that patients prescribed SGLT2i or GLP‐1RA prior to or within the time window are excluded.
ACKNOWLEDGMENTS
We would like to thank TriNetX in allowing us to use their global federated research network to conduct this study.
Riley DR, Henney A, Anson M, et al. The cumulative impact of type 2 diabetes and obstructive sleep apnoea on cardiovascular, liver, diabetes‐related and cancer outcomes. Diabetes Obes Metab. 2025;27(2):663‐674. doi: 10.1111/dom.16059
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from TriNetX, LLC, https://trinetx.com/, but third‐party restrictions apply to the availability of these data. The data were used under licence for this study with restrictions that do not allow for the data to be redistributed or made publicly available. However, for accredited researchers, the TriNetX data are available for licensing at TriNetX, LLC. Data access may require a data sharing agreement and may incur data access fees. Data used in the generation of this paper was collected from the global TriNetX network, and local data at LUHFT were not used.
REFERENCES
- 1. Swinburn BA, Sacks G, Hall KD, et al. The global obesity pandemic: shaped by global drivers and local environments. The Lancet. 2011;378(9793):804‐814. [DOI] [PubMed] [Google Scholar]
- 2. Wang L, Ding H, Deng Y, Huang J, Lao X, Wong MCS. Associations of obesity indices change with cardiovascular outcomes: a dose‐response meta‐analysis. Int J Obes (Lond). 2024;48(5):635‐645. [DOI] [PubMed] [Google Scholar]
- 3. Huang T, Lin BM, Stampfer MJ, Tworoger SS, Hu FB, Redline S. A population‐based study of the bidirectional association between obstructive sleep apnea and type 2 diabetes in three prospective U.S. cohorts. Diabetes Care. 2018;41(10):2111‐2119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rajan P, Greenberg H. Obstructive sleep apnea as a risk factor for type 2 diabetes mellitus. Nat Sci Sleep. 2015;7:113‐125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Shang Y, Grip ET, Modica A, et al. Metabolic syndrome traits increase the risk of major adverse liver outcomes in type 2 diabetes. Diabetes Care. 2024;47(6):978‐985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Einarson TR, Acs A, Ludwig C, Panton UH. Prevalence of cardiovascular disease in type 2 diabetes: a systematic literature review of scientific evidence from across the world in 2007‐2017. Cardiovasc Diabetol. 2018;17(1):83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Pearson‐Stuttard J, Bennett J, Cheng YJ, et al. Trends in predominant causes of death in individuals with and without diabetes in England from 2001 to 2018: an epidemiological analysis of linked primary care records. Lancet Diabetes Endocrinol. 2021;9(3):165‐173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Pearson‐Stuttard J, Papadimitriou N, Markozannes G, et al. Type 2 diabetes and cancer: an umbrella review of observational and Mendelian randomization studies. Cancer Epidemiol Biomarkers Prev. 2021;30(6):1218‐1228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long‐term cardiovascular outcomes in men with obstructive sleep apnoea‐hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365(9464):1046‐1053. [DOI] [PubMed] [Google Scholar]
- 10. Yeghiazarians Y, Jneid H, Tietjens JR, et al. Obstructive sleep apnea and cardiovascular disease: a scientific statement from the American Heart Association. Circulation. 2021;144(3):e56‐e67. [DOI] [PubMed] [Google Scholar]
- 11. Musso G, Cassader M, Olivetti C, Rosina F, Carbone G, Gambino R. Association of obstructive sleep apnoea with the presence and severity of non‐alcoholic fatty liver disease. A systematic review and meta‐analysis. Obes Rev. 2013;14(5):417‐431. [DOI] [PubMed] [Google Scholar]
- 12. Cho J, Jo S. Association of obstructive sleep apnea with risk of lung cancer: a nationwide cohort study in Korea. Sci Rep. 2024;14(1):12394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kristensen SL, Rorth R, Jhund PS, et al. Cardiovascular, mortality, and kidney outcomes with GLP‐1 receptor agonists in patients with type 2 diabetes: a systematic review and meta‐analysis of cardiovascular outcome trials. Lancet Diabetes Endocrinol. 2019;7(10):776‐785. [DOI] [PubMed] [Google Scholar]
- 14. Zelniker TA, Wiviott SD, Raz I, et al. SGLT2 inhibitors for primary and secondary prevention of cardiovascular and renal outcomes in type 2 diabetes: a systematic review and meta‐analysis of cardiovascular outcome trials. Lancet. 2019;393(10166):31‐39. [DOI] [PubMed] [Google Scholar]
- 15. Kunutsor SK, Zaccardi F, Balasubramanian VG, et al. Glycaemic control and macrovascular and microvascular outcomes in type 2 diabetes: systematic review and meta‐analysis of cardiovascular outcome trials of novel glucose‐lowering agents. Diabetes Obes Metab. 2024;26(5):1837‐1849. [DOI] [PubMed] [Google Scholar]
- 16. Palchuk MB, London JW, Perez‐Rey D, et al. A global federated real‐world data and analytics platform for research. JAMIA Open. 2023;6(2):ooad035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Kang PS, Neeland IJ. Body fat distribution, diabetes mellitus, and cardiovascular disease: an update. Curr Cardiol Rep. 2023;25(11):1555‐1564. [DOI] [PubMed] [Google Scholar]
- 18. Bushi G, Padhi BK, Shabil M, et al. Cardiovascular disease outcomes associated with obstructive sleep apnea in diabetics: a systematic review and meta‐analysis. Diseases. 2023;11(3):103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Gottlieb DJ. Sleep apnea and cardiovascular disease. Curr Diab Rep. 2021;21(12):64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Viigimaa M, Sachinidis A, Toumpourleka M, Koutsampasopoulos K, Alliksoo S, Titma T. Macrovascular complications of type 2 diabetes mellitus. Curr Vasc Pharmacol. 2020;18(2):110‐116. [DOI] [PubMed] [Google Scholar]
- 21. Zhang R, Zhang P, Zhao F, Han X, Ji L. Association of Diabetic Microvascular Complications and Parameters of obstructive sleep apnea in patients with type 2 diabetes. Diabetes Technol Ther. 2016;18(7):415‐420. [DOI] [PubMed] [Google Scholar]
- 22. Tahrani AA, Ali A, Raymond NT, et al. Obstructive sleep apnea and diabetic nephropathy: a cohort study. Diabetes Care. 2013;36(11):3718‐3725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Tahrani AA, Ali A, Raymond NT, et al. Obstructive sleep apnea and diabetic neuropathy: a novel association in patients with type 2 diabetes. Am J Respir Crit Care Med. 2012;186(5):434‐441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Abelleira R, Zamarron C, Ruano A, et al. Impact of sleep apnoea‐hypopnoea syndrome on diabetic neuropathy. A systematic review. Med Clin (Barc). 2021;157(3):106‐113. [DOI] [PubMed] [Google Scholar]
- 25. Liu H, Li L, Zan X, Wei J. No bidirectional relationship between sleep phenotypes and risk of proliferative diabetic retinopathy: a two‐sample Mendelian randomization study. Sci Rep. 2024;14(1):9585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Storgaard H, Mortensen B, Almdal T, Laub M, Tarnow L. At least one in three people with type 2 diabetes mellitus referred to a diabetes centre has symptomatic obstructive sleep apnoea. Diabet Med. 2014;31(11):1460‐1467. [DOI] [PubMed] [Google Scholar]
- 27. Simonson M, Li Y, Zhu B, et al. Multidimensional sleep health and diabetic retinopathy: systematic review and meta‐analysis. Sleep Med Rev. 2024;74:101891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Bu LF, Xiong CY, Zhong JY, et al. Non‐alcoholic fatty liver disease and sleep disorders. World J Hepatol. 2024;16(3):304‐315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Guay‐Gagnon M, Vat S, Forget MF, et al. Sleep apnea and the risk of dementia: a systematic review and meta‐analysis. J Sleep Res. 2022;31(5):e13589. [DOI] [PubMed] [Google Scholar]
- 30. Xue M, Xu W, Ou YN, et al. Diabetes mellitus and risks of cognitive impairment and dementia: a systematic review and meta‐analysis of 144 prospective studies. Ageing Res Rev. 2019;55:100944. [DOI] [PubMed] [Google Scholar]
- 31. Xie J, Huang H, Liu Z, et al. The associations between modifiable risk factors and nonalcoholic fatty liver disease: a comprehensive Mendelian randomization study. Hepatology. 2023;77(3):949‐964. [DOI] [PubMed] [Google Scholar]
- 32. Jarvis H, Craig D, Barker R, et al. Metabolic risk factors and incident advanced liver disease in non‐alcoholic fatty liver disease (NAFLD): a systematic review and meta‐analysis of population‐based observational studies. PLoS Med. 2020;17(4):e1003100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Pati S, Irfan W, Jameel A, Ahmed S, Shahid RK. Obesity and cancer: a current overview of epidemiology, pathogenesis, outcomes, and management. Cancers (Basel). 2023;15(2):485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Pearson‐Stuttard J, Zhou B, Kontis V, Bentham J, Gunter MJ, Ezzati M. Worldwide burden of cancer attributable to diabetes and high body‐mass index: a comparative risk assessment. Lancet Diabetes Endocrinol. 2018;6(6):e6‐e15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Marrone O, Bonsignore MR. Obstructive sleep apnea and cancer: a complex relationship. Curr Opin Pulm Med. 2020;26(6):657‐667. [DOI] [PubMed] [Google Scholar]
- 36. Wu D, Zhao Z, Chen C, et al. Impact of obstructive sleep apnea on cancer risk: a systematic review and meta‐analysis. Sleep Breath. 2023;27(3):843‐852. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Table S1. Codes used for outcome analyses.
Supplemental Table S2. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only addition that the time window is from 1 to 1825 days.
Supplemental Table S3. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only addition that the time window is from 1 to 1825 days.
Supplemental Table S4. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only addition that patients are required to have BMI data available at baseline.
Supplemental Table S5. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only addition that patients are required to have both BMI and HbA1c data available at baseline.
Supplemental Table S6. Sensitivity analysis of analysis one (OSA+T2D vs. OSA)—This analysis uses the same methodology as described in the main manuscript with the only change that patients who initiated CPAP are excluded.
Supplemental Table S7. Sensitivity analysis of analysis two (T2D+OSA vs. T2D)—This analysis uses the same methodology as described in the main manuscript with the only change is that patients prescribed SGLT2i or GLP‐1RA prior to or within the time window are excluded.
Data Availability Statement
The data that support the findings of this study are available from TriNetX, LLC, https://trinetx.com/, but third‐party restrictions apply to the availability of these data. The data were used under licence for this study with restrictions that do not allow for the data to be redistributed or made publicly available. However, for accredited researchers, the TriNetX data are available for licensing at TriNetX, LLC. Data access may require a data sharing agreement and may incur data access fees. Data used in the generation of this paper was collected from the global TriNetX network, and local data at LUHFT were not used.