CONTINUOUS GLUCOSE MONITORING (CGM) 2026: Must know & What is new

 

CONTINUOUS GLUCOSE MONITORING (CGM)

(2026, Must know & What is new)


Dr. Sree Vani Gayathri 

MD (General Medicine), Associate Professor

Dr. Sidhartha Mullapudi 

DrNB Cardiology, MD (General Medicine), PGDHSc Diabetology,


meritmedscript.blogspot.com

ADA Standards of Care 2026  ·  ATTD Consensus  ·  Endocrine Society  ·  Braunwald 13th Ed

One-line viva summary:

rtCGM with alerts is the standard.

Fix lows before highs.

Need 14 days and ≥70% wear.

TIR >70%, TBR <4% / <1% under 54, %CV ≤36%.

One or two changes only.

Arrows plus IOB — never stack insulin.


Part 1. What Is New in CGM (2025–2026)

1. Indication expanded beyond insulin dependence

Guidelines now support CGM from diabetes onset and anytime thereafter in:

  • Adults with diabetes on insulin (any regimen).

  • Adults on non-insulin drugs that trigger hypoglycemia (sulfonylureas, glinides).

  • Any patient in whom a real-time glucose profile will change titration or lifestyle advice.

2. Intermittently scanned CGM (isCGM) is no longer the standard

Current tech tables have dropped isCGM as the default. Real-time CGM (rtCGM) with automated high/low alerts and predictive alarms is the definitive standard of care.

3. No arbitrary barriers to pumps or AID

Initiation of CSII or automated insulin delivery must not require C-peptide cut-offs, islet autoantibody confirmation, or a mandatory period of prior MDI.

4. OTC / non-prescription sensors

Biosensors cleared for type 2 diabetes not on insulin and for prediabetes / wellness use have widened access. The clinician still needs a structured interpretation framework (AGP + DATUM steps below). An OTC trace is data — it is not automatically a complete diabetes plan.

5. Inpatient and critical-care use

CGM is validated for ward and step-down monitoring using hybrid protocols: believe the trend, confirm with capillary glucose when values are out of range or the trajectory is changing rapidly. ICU / vasopressor / shock states remain confirmation-heavy.

6. Interference is reduced, not abolished

Newer sensors have far less electrochemical interference from hydroxyurea, high-dose acetaminophen, ascorbic acid and maltose. High-dose IV vitamin C can still falsely elevate some enzymatic platforms. When the number and the patient disagree, check capillary glucose.


Part 2. Core Metrics and AGP Targets

A clinically actionable Ambulatory Glucose Profile needs a standardized 14-day collection with at least 70% active sensor wear. Less than that: do not make major regimen changes on a thin dataset.

Metric

Definition

Target

TIR

70–180 mg/dL (3.9–10.0 mmol/L)

T1D / T2D: >70%. Older / high-risk / unaware: >50%. Pregnancy T1D (63–140): >70%

TBR Level 1

54–69 mg/dL

Standard <4% (<1 h/day). Frail / high-risk <1% (<15 min/day)

TBR Level 2

<54 mg/dL (<3.0 mmol/L)

All cohorts: <1% (<15 min/day) — non-negotiable

TAR Level 1

181–250 mg/dL

Standard <25%

TAR Level 2

>250 mg/dL (>13.9 mmol/L)

Standard <5%

%CV

Coefficient of variation

≤36%. Above 36% = unstable control, high neuroglycopenia risk

GMI

14-day mean sensor glucose → estimated A1c

GMI vs lab A1c mismatch → altered RBC turnover (CKD, iron deficiency, hemolysis, hemoglobinopathy)

Exam memory: TIR + TBR + %CV beat a lone A1c. Pregnancy uses a tighter range (63–140 mg/dL), not 70–180.


Part 3. DATUM — Systematic AGP Interpretation

Use this five-step sequence in clinic and in viva. Do not start with the pretty median curve.

Step

What to do

1. Data quality

Confirm ≥70% active time over consecutive 14 days. If not, repeat the profile before major changes.

2. Safety first (TBR)

Read TBR <70 and <54 first. Time the lows: nocturnal vs post-meal vs post-exercise. Cut basal insulin or secretagogues before you chase highs.

3. Variability

%CV >36% or a wide 10th–90th percentile band means unstable days. Fix meal pattern, ICR, and rebound treatment of lows — not only the basal rate.

4. Hyperglycemia

Place the 50th-percentile (median) line against 70–180. High overnight/fasting median → basal or evening gluconeogenesis. Post-meal spikes → bolus 15–20 min before eating, ICR too weak, or high-GI meals.

5. One or two actions

Change basal or one post-meal problem — not both plus ICR plus ISF in the same visit. Recheck in 2–4 weeks.

DATUM mnemonic for viva: Data → Avoid lows → Turbulence (%CV) → Upward (TAR/median) → Minimal changes.


Part 4. Rate-of-Change Arrows and Correction

Arrows estimate interstitial velocity over the next 15–20 minutes. Always combine with insulin-on-board. Units below are teaching defaults — individualize by ISF.

Arrow

Velocity / 30-min projection

Correction strategy

↑↑  Two up

>3 mg/dL/min  ·  rise >90 mg/dL

Anticipate IOB before any extra bolus. High stacking risk.

↑  One up

2–3 mg/dL/min  ·  rise 60–90

Add ~1.5–2.0 U (by ISF).

↗  Angled up

1–2 mg/dL/min  ·  rise 30–60

Add ~1.0 U.

→  Flat

<1 mg/dL/min  ·  stable

Standard correction if above target.

↘  Angled down

1–2 mg/dL/min  ·  drop 30–60

Subtract ~1.0 U, or suspend if already near target.

↓  One down

2–3 mg/dL/min  ·  drop 60–90

Subtract 1.5–2.0 U. If glucose <100, take 10–15 g fast carbohydrate now.

↓↓  Two down

>3 mg/dL/min  ·  drop >90

Treat immediately: 15–20 g rapid carbohydrate. Do not wait for the low alarm.


Part 5. High-Yield Pitfalls

  • Physiological lag (5–15 min). Interstitial glucose trails capillary blood during exercise or a fast carbohydrate load. Symptoms ≠ sensor → fingerstick.

  • Compression lows. Sleeping on the sensor arm starves local perfusion. Overnight “severe hypo” that recovers as a vertical line when the patient turns is mechanical, not insulin. Check the raw overnight trace before cutting nocturnal basal.

  • Drug interference. Newer membranes block most paracetamol effect. High-dose IV ascorbic acid can still falsely raise some enzymatic sensors.

  • Lag-driven insulin stacking. Rising arrows tempt repeat corrections before the last subcutaneous dose has peaked. Delayed severe hypo 2–3 hours later. Teach IOB every visit.

  • Treating the A1c while ignoring TBR. A “good” A1c with TBR Level 2 >1% is a failed AGP.

  • Changing five settings at once. You will not know what worked. One or two actions, then 2–4 weeks.

  • Trusting <70% wear. Incomplete AGP is not a license for large titration.

  • OTC sensor without a plan. Wellness traces still need DATUM if they will change drugs.


Part 6. Quick Viva Cards

Question

Model answer

Minimum AGP quality?

14 consecutive days and ≥70% active wear.

Standard TIR / TBR / TAR?

TIR >70% (70–180). TBR <4% and <1% below 54. TAR <25% / Level 2 <5%.

Frail or unaware target?

TIR >50%. TBR Level 1 <1%. Never trade safety for a prettier A1c.

Pregnancy T1D range?

63–140 mg/dL with TIR >70%.

Unstable control number?

%CV >36%.

GMI ≠ lab A1c?

Think RBC lifespan: CKD, iron deficiency, hemolysis, hemoglobinopathy — not a “broken sensor” first.

First AGP move?

Eliminate hypoglycemia. Then variability. Then hyperglycemia.

Two arrows down?

Treat now with 15–20 g rapid carbohydrate. Do not wait for the alarm.

Need antibodies or C-peptide for AID?

No. Arbitrary C-peptide, autoantibody, or mandatory MDI-duration barriers are removed.

isCGM vs rtCGM?

rtCGM with predictive alerts is the 2026 standard. isCGM is no longer the default listed technology.


Sources

ADA Standards of Care in Diabetes (2026); ATTD consensus on CGM metrics and AGP; Endocrine Society diabetes technology statements; Braunwald 13th edition context for inpatient and cardiometabolic comorbidity. Class language is summarised for teaching — confirm exact COR/LOE in the parent document before quoting in a protocol.

meritmedscript.blogspot.com  ·  Clinical exam notes  ·  September 2026


Comments