Diabetes and Sleep Apnoea: The Link That Often Goes Unexamined

How common the overlap is

Reported prevalence varies widely with the population studied and the testing method. A review in European Endocrinology summarising epidemiological studies found moderate-to-severe obstructive sleep apnoea in 23.8 to 70 per cent of people with type 2 diabetes, against general-population estimates of roughly 9 to 14 per cent of men and 2 to 7 per cent of women. Lower figures came from primary care samples, higher ones from secondary care.

In the Sleep AHEAD cohort, 305 overweight and obese adults with type 2 diabetes underwent home sleep studies, and 86 per cent had an apnoea-hypopnoea index of 5 or more. That is a selected group and not a general estimate, but it indicates how high the figure can go where obesity and diabetes coexist.

Looking the other way, a review in Chest estimated that 15 to 30 per cent of people with obstructive sleep apnoea have type 2 diabetes, with higher rates in severe apnoea. Much of that association is attenuated when BMI and other confounders are accounted for, which is a reminder that obesity contributes substantially to both.

The bidirectional mechanism

The proposed pathways from apnoea to disordered glucose metabolism are reasonably well characterised:

• Repeated cycles of oxygen desaturation and reoxygenation generate oxidative stress. In healthy volunteers, a few hours of intermittent hypoxia has been shown to reduce insulin sensitivity without a compensatory rise in insulin secretion.
• Recurrent arousals and hypoxia increase sympathetic nervous activity, which affects both insulin secretion and hepatic glucose output.
• Sleep fragmentation and loss of slow-wave sleep are themselves associated with lower insulin sensitivity.
• Apnoea alters adipokine and inflammatory signalling, lowering adiponectin and raising inflammatory cytokines.

A meta-analysis found that moderate-to-severe obstructive sleep apnoea was associated with incident type 2 diabetes with a relative risk of 1.63; for mild apnoea the association was not statistically significant.

The reverse direction is less established but plausible. Diabetic autonomic and peripheral neuropathy may impair central respiratory control and upper airway reflexes, and in people with type 1 diabetes those with apnoea had higher rates of autonomic and peripheral neuropathy. This remains an association rather than a demonstrated causal mechanism.

Why it is frequently missed

Snoring is normalised. Daytime tiredness in someone with obesity and diabetes is attributed to the diabetes, to low mood, or to being busy. Nocturia is attributed to glucose. Morning headache is attributed to sleep quality in a circular way. And the classic picture of a sleepy middle-aged man is a poor guide: women, people who are not markedly sleepy, and people of lower body weight all have apnoea and are diagnosed later.

NICE lists type 2 diabetes among the risk factors that should prompt assessment for obstructive sleep apnoea-hypopnoea syndrome, alongside obesity, treatment-resistant hypertension, atrial fibrillation, previous stroke or transient ischaemic attack, heart failure, polycystic ovary syndrome, hypothyroidism and acromegaly. The features it lists as suggesting the condition include snoring, witnessed apnoeas, unrefreshing sleep, waking headaches, excessive sleepiness or fatigue, nocturia, choking during sleep, sleep fragmentation and cognitive or memory difficulty.

What assessment involves

Assessment begins with the history, including an account from a bed partner where possible. NICE advises using the Epworth Sleepiness Scale but not on its own to determine referral, and suggests considering the STOP-Bang questionnaire alongside the sleep history. Questionnaires identify who should be tested; they do not diagnose.

Diagnosis requires a sleep study. NICE recommends offering home respiratory polygraphy first, with home oximetry considered where access to polygraphy is limited. If oximetry is negative but symptoms are significant, respiratory polygraphy or polysomnography can be considered, and polysomnography is reserved for cases where polygraphy is negative but symptoms persist. Severity is graded from the sleep study result.

What treating it does, and does not do

For symptoms, treatment can be transformative, and NICE recommends continuous positive airway pressure for moderate or severe disease and for mild disease where symptoms affect quality of life and daytime activities. Mandibular advancement splints, positional modifiers and surgical assessment have defined roles where CPAP is declined or not tolerated.

For glucose control, the evidence is weaker than the mechanism predicts, and this is where overstatement is common. A review in Chest summarised seven controlled studies in type 2 diabetes with 498 participants. Of the four reporting HbA1c at three to six months, three found no significant reduction, including the largest with 256 completers. One found a reduction of around 0.4 per cent, and another a similar reduction only after excluding dropouts and medication changes. An earlier meta-analysis found no significant effect on HbA1c. In the Sleep AHEAD trial of intensive lifestyle modification, weight loss of 10.8 kg reduced the apnoea-hypopnoea index and lowered HbA1c, but the change in HbA1c was not related to the change in apnoea severity.

The honest summary is that CPAP has not been shown to produce a reliable improvement in HbA1c in people with type 2 diabetes. Adherence appears to matter, and benefit may be greater in those with poorer baseline control, but it should be regarded as adjunctive. Treat apnoea because it causes sleepiness, impairs quality of life and carries cardiovascular and driving risk, not on the promise of a better HbA1c.

That is still a strong reason to look for it. If you snore, wake unrefreshed, or your partner has noticed you stop breathing, raise it at your next diabetes review.