The Dawn Phenomenon: Why Glucose Rises Before Breakfast and When It Matters

The body does not maintain constant insulin sensitivity around the clock. The circadian clock, located in the hypothalamus (the suprachiasmatic nucleus), regulates the daily rhythm of liver glucose production and hepatic insulin sensitivity. Towards morning — in the second part of the night and just before waking — cortisol secretion rises, and during deep sleep pulses of growth hormone appear. Both hormones act opposite to insulin: they boost gluconeogenesis and glycogenolysis in the liver, and growth hormone additionally reduces glucose uptake in the muscles. The review by Carroll and Schade (2005) indicates that the most likely mechanism of the dawn phenomenon is precisely growth hormone-mediated impairment of insulin sensitivity in the liver and muscles, although the exact biochemical pathways remain unexplained.

In a healthy person, this morning rise in insulin resistance is offset by the pancreas: the beta cells secrete slightly more insulin and glucose stays in a narrow range, often with no noticeable change. The dawn phenomenon becomes visible in the results when the compensatory reserve is limited — in type 2 diabetes, in earlier insulin resistance, and in people treated with insulin, in whom the morning rise in insulin requirement is not automatically covered. As Rybicka et al. (2011) put it, the dawn phenomenon occurs when endogenous insulin secretion falls or the action of insulin given the previous day wears off, while at the same time the level of insulin-antagonistic hormones physiologically rises.

The review by Peng et al. (2022) adds an important thread: disruption of the circadian clock is associated with an extended dawn phenomenon in type 2 diabetes. This means that irregular sleep, shift work and disorganization of the daily rhythm may be factors that amplify what in a healthy person is a barely noticeable fluctuation. These data are review-level and mechanistic, so they do not prove that improving sleep regularity will reduce dawn glucose rise in a given person, but they explain why changes to daily rhythm are sometimes part of practical advice.

This result organizes intuition. The dawn phenomenon is not 'yes or no' — it is a continuum that intensifies as glycemic regulation worsens. Only about 9% of people with normal glucose tolerance exceeded the 20 mg/dl threshold adopted in this study, while in type 2 diabetes more than half of those studied did. This does not mean that 9% of healthy people have a problem: the authors themselves note that the threshold definition is conventional and used mainly to describe the intensity of the phenomenon, not to diagnose disease. A smaller rise (the secondary threshold of 0.56 mmol/l, i.e. about 10 mg/dl) occurs much more often, including in people without diabetes.

The classic review by Carroll and Schade, drawing on older studies in patients, proposes a clinical definition: a glucose rise of more than 10 mg/dl from the nocturnal nadir or an increase in insulin requirement of at least 20%. By this definition the dawn phenomenon affects about half of patients with type 1 and type 2 diabetes (54% and 55% in the quoted data). The authors, however, point out that controversy exists about its frequency, reproducibility and pathogenesis — the size of the dawn rise can vary from night to night in the same person. A single night is therefore not enough to draw conclusions.

Definitely yes, on a small scale. With the popularization of CGM sensors among healthy people, more and more people see nocturnal 'steps' of glucose or a morning rise before breakfast on the chart and interpret it as a sign of disease. A multicenter reference study of 153 healthy people without diabetes (Shah et al., 2019) showed that mean glucose was about 98–99 mg/dl (104 mg/dl in people over 60), and that glucose stayed within 70–140 mg/dl 96% of the time; the median time below 70 mg/dl was 1.1% of the day (about 15 minutes). This means that even in healthy people glucose fluctuates, and short episodes outside the 'ideal' range are normal.

For understanding the dawn phenomenon it also matters that CGM sensors measure glucose in interstitial fluid, not in blood, and have their own measurement error and lag. Known artefacts, such as false 'nocturnal lows' after pressure on the sensor while sleeping on your side, can distort the picture of the night hours. A single rise of 10–20 mg/dl visible on the chart on the way to waking should therefore not be interpreted as proof of a disorder. A broader discussion of CGM limitations in people without diabetes can be found in our article on continuous glucose monitoring for healthy people.

The Somogyi effect is a hypothesis according to which nocturnal hypoglycemia triggers a rebound response of counter-regulatory hormones (adrenaline, glucagon, cortisol), resulting in high glucose in the morning. It concerns only people using drugs capable of causing hypoglycemia — insulin or sulfonylureas — and in theory would call for reducing, not increasing, the evening drug dose. In people not using such drugs it has no practical significance: a healthy pancreas and liver do not cause a 'rebound' after a nocturnal glucose dip.

The practical conclusion from this table for people without diabetes is simple: the Somogyi effect does not apply here, and the morning change in glucose usually results from the dawn phenomenon, from the influence of dinner, or from ordinary variability. For people with diabetes, distinguishing the dawn phenomenon from the Somogyi effect and from an insufficient evening dose requires glucose measurements at night (e.g. around 3:00–5:00 a.m. for several nights) or CGM, as Rybicka et al. (2011) note — and therapeutic decisions belong to the treating doctor.

In type 2 diabetes, the dawn phenomenon has measurable consequences. Monnier et al. (2013) analyzed CGM data from 248 people with type 2 diabetes not treated with insulin (diet, insulin sensitizers or secretagogues). The median glucose rise from the nocturnal nadir to the pre-breakfast value was 16 mg/dl (interquartile range 0–31.5 mg/dl) and did not differ significantly between groups selected by treatment or HbA1c. The average impact of the dawn phenomenon on HbA1c was estimated at about 0.4 percentage points (0.39%), and on the 24-hour mean glucose at about 12 mg/dl; the authors note that this effect is not eliminated by any of the available oral antidiabetic drugs.

Epidemiological data suggest that a very large dawn phenomenon may have prognostic significance. In a prospective study of 5,542 hospitalized adults with type 2 diabetes (Cai et al., 2024), after a median follow-up of 9.4 years, a glucose rise from the nocturnal nadir to the pre-breakfast value of more than 5.55 mmol/l (about 100 mg/dl) was associated with a 30% higher risk of all-cause death compared with the group with a rise of 1.12–3.33 mmol/l (about 20–60 mg/dl), after accounting for HbA1c. This is an observational study conducted in patients requiring hospitalization, so it does not prove causation and does not carry over to healthy people, but it suggests that a severe dawn phenomenon is worth treating as an element of the assessment of glycemic control.

Not every morning rise in glucose is the dawn phenomenon. The most common coexisting or mimicking causes include: a late, large dinner rich in high-glycemic-index carbohydrates (we discuss this in a separate article), short or interrupted sleep, alcohol in the evening, intense stress or illness (infection, inflammation), taking drugs that raise glucose (e.g. glucocorticoids), and inappropriate timing of antidiabetic drugs in people who are ill. Conversely, if fasting glucose stays in the 100–125 mg/dl range, it is worth having HbA1c done and considering an OGTT, rather than focusing solely on the shape of the nocturnal CGM curve.

It is also worth remembering nocturnal hypoglycemia in people using insulin or secretagogues: low glucose at night (below 70 mg/dl) can look like 'high glucose in the morning' on the next reading if the person ate something sweet after waking. This is the best reason not to change doses in people treated with insulin on the basis of single morning results and not to guess whether we are dealing with the dawn phenomenon or the Somogyi effect.

What can realistically be done for a person without diabetes if a dawn rise is visible but laboratory results are normal? The answer is unspectacular but honest: most often nothing beyond caring for the general factors that support insulin sensitivity — regular sleep, physical activity, a sensible composition of the last meal and avoiding very late, large dinners. Evidence that specific interventions 'extinguish' a physiological dawn rise in healthy people is limited, and the dawn phenomenon itself is not a treatment target.

The dawn phenomenon is one of those topics that, thanks to the popularization of CGM, has started to worry healthy people, even though in their case it simply describes the normal hormonal rhythm of the day. A sensible approach separates two questions: is the morning rise physiological, that is small and reproducible, and do the absolute glucose values, HbA1c and any OGTT fall within the normal range. If so, the CGM chart is a curiosity, not an indication for action. If not, a conversation with a doctor matters more than another day with the sensor.