CGM in Functional Practice: When Continuous Glucose Monitoring Adds Clinical Value

Clinical scenario: when to order CGM

CGM technology has moved rapidly from specialist diabetes clinics into direct-to-consumer wellness apps. That trajectory creates a real risk: practitioners either dismiss CGM entirely because it feels like a consumer fad, or they recommend it to every patient because the data look impressive. Neither position is defensible. The question is not whether CGM produces interesting data -- it does -- but whether that data will change your clinical decision for this particular patient.

Presentations where CGM adds clear value

Unexplained postprandial fatigue and energy crashes. A patient reports consistent afternoon energy collapse despite adequate sleep, reasonable macronutrient balance, and no anaemia on standard bloods. HbA1c is 36 mmol/mol (5.4%) -- normal. Fasting glucose is 4.8 mmol/L -- normal. Yet something is driving the fatigue pattern. A 14-day CGM trial can reveal whether postprandial glucose excursions are spiking to 9--10 mmol/L and then dropping sharply, creating a reactive hypoglycaemia pattern that would be invisible on any single-point blood test. Hall et al. (2018) demonstrated that even individuals classified as normoglycaemic by standard measures can exhibit distinct "glucotypes" with markedly different glycaemic variability profiles.

Suspected early insulin resistance with normal HbA1c. HbA1c reflects a 90-day average and can remain within reference range whilst postprandial glucose handling is already deteriorating. CGM captures the postprandial excursion that fasting glucose and HbA1c miss. This is particularly useful when waist circumference is increasing, triglycerides are creeping upward, or there is a strong family history of type 2 diabetes. The patient's standard bloods look reassuring; the CGM data may tell a different story.

Dietary intervention plateau. A patient has been following a well-designed nutrition protocol for 8--12 weeks but progress has stalled. CGM can identify specific foods or meal compositions that provoke disproportionate glycaemic responses. Zeevi et al. (2015) showed that individual postprandial glycaemic responses to identical foods vary enormously -- a finding confirmed by the PREDICT study (Berry et al., 2020). CGM converts "eat less sugar" into "this specific breakfast combination is producing a 4 mmol/L spike followed by a crash at 10:30."

Dawn phenomenon investigation. A patient reports waking unrefreshed despite 7--8 hours of sleep, or presents with elevated fasting glucose on morning blood draws despite good dietary control. CGM reveals whether a nocturnal glucose rise (the dawn phenomenon, driven by counter-regulatory cortisol and growth hormone secretion) is producing hyperglycaemia that resolves before the patient eats. This finding changes management: it shifts the focus from dinner-time carbohydrate restriction to cortisol-modulating strategies, evening exercise timing, or circadian hygiene.

Presentations where CGM is unlikely to add value

• Patients with well-controlled metabolic markers and no fatigue, energy, or weight complaints.
• Patients whose primary concern is a non-metabolic presentation (e.g. dermatological, musculoskeletal) unless there is a plausible glucose-mediated mechanism.
• Patients who are not willing to log food alongside the CGM data. Without a food diary, the glucose trace is clinically uninterpretable -- you see the response but cannot identify the trigger.
• Patients with high health anxiety who are likely to catastrophise normal physiological glucose fluctuations. A glucose spike to 8.5 mmol/L after a meal containing 60 g of carbohydrate is physiologically normal; a highly anxious patient may read it as evidence of disease.

The clinical decision rule is straightforward: will a 14-day CGM dataset change what I recommend? If the answer is no, save the patient the cost and complexity.

What CGM measures

A continuous glucose monitor is a subcutaneous sensor -- typically worn on the back of the upper arm -- that measures interstitial fluid glucose every 1--5 minutes, producing roughly 1,440 data points per day. It does not measure blood glucose directly; interstitial glucose lags venous blood glucose by approximately 5--15 minutes, which matters during rapid glucose changes but is clinically negligible for the pattern-level analysis that functional practice requires (Hoss & Budiman, 2017).

Core metrics

Time in range (TIR). The percentage of a 24-hour period spent within a defined glucose range. The international consensus (Battelino et al., 2019) defines the standard range as 3.9--10.0 mmol/L (70--180 mg/dL) for people with diabetes. For non-diabetic individuals, Shah et al. (2019) found that healthy adults spend approximately 96% of time between 3.9--7.8 mmol/L (70--140 mg/dL). In functional practice, this tighter range -- sometimes called "time in tight range" (TITR) -- is the more clinically useful metric. A non-diabetic patient spending less than 85% of time in the 3.9--7.8 mmol/L window warrants investigation.

Glucose coefficient of variation (CV). The standard deviation of glucose divided by the mean, expressed as a percentage. It quantifies overall glycaemic variability independent of mean glucose level. Rodbard (2009) established that a CV below 36% indicates stable glycaemic control in the diabetes context. For functional practice with non-diabetic patients, a CV below 20% represents well-regulated glucose handling; 20--30% suggests early dysregulation worth monitoring; above 30% in a non-diabetic individual is clinically significant.

Mean amplitude of glycaemic excursions (MAGE). Captures the average magnitude of glucose swings that exceed one standard deviation. Whilst CV tells you how variable glucose is overall, MAGE tells you how large the swings are. A patient can have moderate CV but large infrequent swings (high MAGE) or frequent small oscillations (low MAGE). Both patterns have clinical implications but require different interventions. Monnier et al. (2006) demonstrated that acute glucose fluctuations activate oxidative stress pathways more potently than sustained chronic hyperglycaemia -- making MAGE a marker of metabolic stress, not just glucose control.

Dawn phenomenon. A rise in glucose concentration occurring in the early morning hours (typically 04:00--08:00), driven by the circadian surge in cortisol, growth hormone, and catecholamines. In non-diabetic individuals, endogenous insulin secretion compensates and the rise is modest (0.5--1.0 mmol/L). In individuals with early insulin resistance or cortisol dysregulation, the rise can exceed 2.0 mmol/L and persist into the first meal, confounding fasting glucose readings. CGM is the only practical way to characterise dawn phenomenon outside a clinical research setting.

Nadir glucose and hypoglycaemic events. The lowest glucose recorded, and the number and duration of episodes below 3.9 mmol/L (or below 3.0 mmol/L for clinically significant hypoglycaemia). Reactive hypoglycaemia -- a postprandial glucose crash below 3.5 mmol/L occurring 2--4 hours after eating -- is a common finding in functional practice and is often the mechanism behind "afternoon slump" presentations.

Reference ranges vs optimal ranges

One of the most common errors in CGM interpretation is conflating diabetes care targets with optimal metabolic health targets. They serve different clinical purposes.

Conventional diabetes care targets (Battelino et al., 2019)

These targets were developed for people with diagnosed diabetes. They are designed to minimise acute risk (severe hypoglycaemia) and long-term complications (microvascular disease). They are not aspirational targets for metabolic optimisation.

Functional practice optimal targets (non-diabetic, metabolically healthy)

The important principle: functional targets are tighter than diabetes targets, but they must be communicated to patients as optimisation goals rather than disease thresholds. A patient whose TIR (3.9--7.8 mmol/L) is 82% is not ill -- they have room for improvement. Framing matters for compliance and psychological safety.

Where nutrient status screening fits

Glucose regulation does not operate in isolation. Chromium, magnesium, vitamin D, and B vitamins all modulate insulin sensitivity and glucose disposal. Before attributing a suboptimal CGM profile entirely to dietary composition, consider whether underlying micronutrient insufficiency is contributing. A structured nutrient deficiency screen -- such as the Stewart nutritional assessment framework, which systematically evaluates signs and symptoms of subclinical deficiency across key micronutrients -- can identify correctable factors that improve glucose handling independent of macronutrient manipulation. In practice, this means checking serum magnesium (or better, red blood cell magnesium), 25(OH)D, active B12, and ferritin alongside any CGM trial. Correcting a magnesium insufficiency, for example, may improve insulin sensitivity enough to normalise a borderline CGM profile without dietary restriction.

Case study: CGM results informing treatment decisions

Note: This case study is a composite illustration based on common clinical presentations. It does not represent a single identifiable patient.

Presentation

Sarah, 42, presents with a 2-year history of progressive afternoon fatigue, difficulty concentrating after lunch, and a 6 kg weight gain concentrated around the waist. She exercises three times per week (running and yoga), sleeps 7 hours per night, and eats a diet she describes as "healthy -- lots of whole grains, fruit, and salads." She has no diagnosis of diabetes. Her GP bloods 3 months ago were unremarkable: HbA1c 37 mmol/mol (5.5%), fasting glucose 5.1 mmol/L, total cholesterol 5.2 mmol/L, triglycerides 1.6 mmol/L.

On the Functional Health Matrix, her Transport node scores 4/10 (below optimal), driven by the elevated triglycerides relative to HDL, central adiposity, and symptom pattern suggestive of glucose dysregulation.

Clinical reasoning for CGM

Standard bloods are normal by conventional thresholds but the triglyceride-to-HDL ratio (1.6 / 1.4 = 1.14 in mmol/L terms) is above the optimal functional target of <0.8, and waist circumference at 88 cm places her at the metabolic syndrome threshold for women. The energy crash pattern is temporally linked to meals. A 14-day CGM trial is indicated to characterise postprandial glucose handling and identify whether specific meal compositions are driving disproportionate glycaemic responses.

CGM findings (14-day FreeStyle Libre 3 trial)

Pattern analysis

The CGM reveals two clinically actionable patterns:

Pattern 1: Postprandial spikes followed by reactive crashes. Sarah's lunch -- typically a large mixed salad with quinoa, chickpeas, and dried fruit -- produces glucose peaks of 9.2--10.1 mmol/L within 45 minutes, followed by crashes to 3.2--3.5 mmol/L by 14:30. This explains the afternoon fatigue and difficulty concentrating. The combination of high-glycaemic dried fruit with a large carbohydrate load from quinoa overwhelms her current insulin sensitivity. Notably, her dinners -- which contain protein, fat, and lower-glycaemic vegetables -- produce much smaller excursions (peak 6.8 mmol/L average).

Pattern 2: Significant dawn phenomenon. Glucose rises from 4.6 mmol/L at 03:00 to 6.4 mmol/L at 07:00, before any food intake. This pattern, repeated on 12 of 14 days, suggests cortisol dysregulation or early hepatic insulin resistance. It also means her fasting glucose at the GP (drawn at 08:30) reflects the dawn phenomenon rather than true fasting baseline.

Treatment decisions informed by CGM

1. Lunch restructuring. Replace dried fruit with fresh berries. Reduce quinoa portion by half. Add 15--20 g of protein (e.g. grilled chicken, tinned sardines) and a tablespoon of olive oil or avocado to blunt the glycaemic response. The target: postprandial peak below 7.8 mmol/L, elimination of reactive hypoglycaemia.
2. Dawn phenomenon investigation. Order a 4-point salivary cortisol (waking, noon, evening, bedtime) to characterise the cortisol awakening response. Introduce evening magnesium glycinate (300 mg elemental magnesium) and review sleep hygiene -- specifically blue-light exposure after 21:00 and consistency of wake time.
3. Nutrient status screen. Check RBC magnesium, 25(OH)D, fasting insulin (to calculate HOMA-IR), and chromium status. A structured Stewart nutritional deficiency assessment identified dry skin at the lateral upper arms (possible essential fatty acid insufficiency) and longitudinal nail ridging (possible zinc or iron insufficiency).
4. Follow-up CGM. Repeat 14-day trial after 8 weeks of intervention to quantify improvement. Target: TIR (3.9--7.8 mmol/L) >90%, CV <20%, elimination of reactive hypoglycaemic episodes, dawn rise <1.0 mmol/L.

This case illustrates the core value proposition of CGM in functional practice: it converts a vague symptom ("I crash after lunch") into a quantified physiological pattern with a measurable intervention target. Without CGM, the practitioner would have adjusted the diet based on general principles. With CGM, the adjustment is specific, targeted, and auditable.

Cost considerations: UK context

CGM is not inexpensive, and cost-effectiveness is a legitimate clinical consideration -- particularly in functional practice, where patients typically pay out of pocket.

NHS prescription criteria

NHS CGM prescription is primarily restricted to patients with type 1 diabetes who meet NICE NG17 criteria: those experiencing recurrent hypoglycaemia, impaired hypoglycaemia awareness, or inability to self-monitor with fingerstick testing. Some clinical commissioning groups extend access to type 2 diabetes patients on intensive insulin therapy with recurrent hypoglycaemia, but this is inconsistent across regions. For non-diabetic patients -- the population most relevant to functional practice -- NHS CGM prescription is not available.

Private CGM costs (as of May 2026)

Accuracy note. The FreeStyle Libre 3 has an overall mean absolute relative difference (MARD) of 7.8%, making it the most accurate factory-calibrated sensor currently available in the UK. The Libre 2 Plus has a MARD of approximately 9.2%. Lower MARD means greater accuracy relative to venous blood glucose. For functional practice purposes, either generation provides sufficient accuracy for pattern-level analysis (Hoss & Budiman, 2017).

Cost-effectiveness in functional practice

A single 14-day CGM trial costs the patient GBP 50--75. In the context of a functional nutrition consultation (typically GBP 120--200), the CGM adds 25--60% to the cost of that visit cycle. The question is whether the data it produces justifies the expense.

The strongest cost-effectiveness argument is specificity. Without CGM, a practitioner might recommend a generic "reduce refined carbohydrates" protocol that the patient partially follows for 6 weeks with modest results. With CGM, the practitioner identifies that three specific meal compositions (not all carbohydrates) are driving the problem, and the patient sees real-time feedback that reinforces compliance. The Zeevi et al. (2015) and Berry et al. (2020) data both demonstrate that individual glycaemic responses to identical foods vary enormously -- generic advice fails because it assumes homogeneity that does not exist.

The pragmatic recommendation: use CGM as a targeted diagnostic tool for a defined 14-day period, not as an ongoing monitoring subscription. One well-designed CGM trial with concurrent food logging, followed by a structured interpretation session, delivers more clinical value than months of ad hoc wear.

How to prepare patients

Patient preparation determines whether a CGM trial produces clinically useful data or becomes a source of confusion and anxiety. The following protocol maximises data quality.

Before the sensor

1. Set expectations. Explain that glucose is supposed to fluctuate -- it rises after meals and falls between them. The goal is not to achieve a flat line. A flat glucose trace would indicate either starvation or sensor malfunction. The goal is to understand the pattern and identify specific triggers worth modifying.
2. Define the clinical question. Tell the patient what you are looking for: "We are investigating whether your afternoon fatigue is related to postprandial glucose crashes" or "We want to see whether your overnight glucose suggests a dawn phenomenon." A defined question prevents the patient from catastrophising every minor fluctuation.
3. Prescribe a food and activity log. The CGM trace without a concurrent food diary is clinically useless. Every meal, snack, and drink must be logged with approximate quantities and timing. Activity (type, duration, intensity) should also be recorded. Many CGM apps allow photo-based food logging, which reduces compliance burden.
4. Advise on sensor placement and practical considerations. The sensor goes on the back of the upper arm. Advise the patient to avoid placing it on the dominant arm if they sleep on that side. Warn that the sensor is water-resistant but not waterproof under sustained pressure -- swimming is fine, but prolonged hot tub use can dislodge the adhesive. The first 24 hours of sensor data are less reliable due to the "bedding-in" period; some practitioners discard day one data from analysis.
5. Address psychological readiness. For patients with a history of disordered eating, orthorexia, or high health anxiety, CGM can be counterproductive. The continuous visibility of glucose data may reinforce obsessive food monitoring. Screen for these presentations before recommending CGM. If in doubt, a single fasting insulin plus 2-hour oral glucose tolerance test (OGTT) with insulin measured at 0, 60, and 120 minutes provides similar clinical information without the 14-day data exposure.

During the sensor period

• Instruct the patient to eat their normal diet for the initial 7 days. The purpose of the sensor is to characterise their current glucose handling, not their behaviour when they know they are being monitored. Changing diet during the observation period invalidates the data.
• In the final 7 days, introduce targeted experiments. Ask the patient to try specific meal modifications (e.g. adding protein to breakfast, replacing a high-glycaemic snack with a lower-glycaemic alternative) and observe the difference. This creates a within-subject comparison that is far more compelling than cross-sectional dietary advice.
• Check in at day 3--4. A brief message or call to address questions and ensure the food log is being maintained prevents data loss and reassures the patient.

After the sensor

• Schedule a dedicated interpretation session. Do not attempt CGM interpretation in the final 5 minutes of a general consultation. Allocate 30--45 minutes to walk through the data with the patient, identify the 2--3 most actionable patterns, and agree a specific intervention plan with measurable targets.
• Produce a written summary. Key metrics (TIR, CV, MAGE, dawn phenomenon), the identified patterns, the agreed interventions, and the criteria for a follow-up CGM trial. This document becomes part of the clinical record and gives the patient a reference they can review at home.

Functional Health Matrix: Transport node

CGM data feeds directly into the Transport node of the Functional Health Matrix. The Transport node encompasses cardiovascular function, haemodynamic regulation, and metabolic substrate delivery -- of which glucose regulation is a core component. A patient whose CGM reveals elevated glycaemic variability (CV >25%), significant postprandial spikes (>8.5 mmol/L in a non-diabetic individual), or reactive hypoglycaemia is exhibiting impaired metabolic transport that affects energy delivery to tissues, contributes to endothelial dysfunction via oxidative stress (Monnier et al., 2006), and correlates with cardiometabolic risk markers including elevated triglycerides, low HDL, and increased arterial stiffness (Hjort, Iggman & Rosqvist, 2024).

Scoring the Transport node with CGM data provides a quantified, longitudinal metric that is more sensitive to early metabolic change than HbA1c or fasting glucose alone. It allows the practitioner to track improvement across intervention cycles and to identify when glucose regulation has been optimised to a point where other Transport node components (e.g. lipid metabolism, blood pressure regulation, iron transport) become the priority.

Limitations and safety considerations

CGM is a powerful tool, but it has boundaries that practitioners must respect.

CGM does not diagnose diabetes. Diagnosis requires venous blood sampling (fasting glucose, OGTT, or HbA1c) per WHO and NICE criteria. CGM data may raise clinical suspicion, but the practitioner must refer for confirmatory testing rather than diagnosing from sensor data alone.

Sensor accuracy has limits. All CGM devices measure interstitial glucose, which lags venous glucose by 5--15 minutes. During rapid glucose changes (e.g. the first 30 minutes after a high-glycaemic meal, or during intense exercise), the sensor reading may underestimate the true peak or overestimate the nadir. Pattern-level interpretation over days is reliable; minute-to-minute readings should not be over-interpreted.

Compression lows are artefacts, not real hypoglycaemia. If the patient sleeps on the sensor arm, pressure can temporarily reduce interstitial fluid flow and produce a false low reading (sometimes below 3.0 mmol/L) that resolves when the pressure is removed. These are identifiable by their sudden onset and resolution and should be excluded from clinical analysis.

Scope of practice. Functional nutrition practitioners should use CGM data to guide dietary, lifestyle, and supplementation strategies within their scope of practice. If CGM data suggests undiagnosed diabetes (e.g. TIR 3.9--10.0 mmol/L below 70%, fasting glucose consistently above 7.0 mmol/L, or recurrent hypoglycaemia without obvious dietary trigger), the patient must be referred to their GP or an endocrinologist. If a patient reports symptoms of severe hypoglycaemia (confusion, loss of consciousness, seizure), this is a medical emergency -- call 999 immediately.

Conclusion

Continuous glucose monitoring adds genuine clinical value in functional practice when it is deployed with discipline: for the right patient, with a defined clinical question, alongside a concurrent food log, and interpreted within a structured consultation. It does not replace foundational metabolic blood work, and it is not a universal screening tool. But for the subset of patients whose glucose variability is driving symptoms that conventional static tests cannot explain, a well-designed 14-day CGM trial transforms clinical reasoning from pattern-guessing to pattern-quantifying.

The evidence base for CGM in non-diabetic populations is growing. Glycaemic variability is now established as an independent correlate of cardiometabolic risk markers, and the inter-individual variability in postprandial responses means that personalised nutrition advice based on CGM data is more defensible than generic dietary guidance. For functional practitioners, CGM is not the answer to every clinical question -- but when it is the right tool, it provides data that nothing else can match.