How Do I Find Out Which Foods Spike Me Using My CGM Data

You find your trigger foods by logging the meal, reading the post-meal rise and peak, and comparing the same food across enough days to see a pattern. A single spike is not a food verdict. It is one data point in a repeated comparison.

Why a CGM "spike" is not just the highest number Your sensor reads glucose in the fluid under your skin, not blood, so it trails a finger-prick by roughly 5 to 15 minutes. After a mixed meal, glucose typically rises from your pre-meal baseline, peaks later in the post-meal window, and then returns toward baseline. The first day of a new sensor is the least reliable, so skip it when you are judging a food.

What you are looking for is the rise from your pre-meal baseline, not an absolute number. A meal that starts at 95 mg/dL and peaks at 130 mg/dL shows a clear rise. A meal that starts at 120 mg/dL and peaks at 130 mg/dL shows almost no rise. Both end at the same number, but the first is a bigger glucose response.

Look for a clear rise from your pre-meal baseline and a return toward that baseline. A larger rise or a slower return is what separates a notable response from a small one. There is no universal cutoff, and a rise alone is not harm.

The common mistake is to judge a food by the biggest number on the graph. Judge it by how far it moved you and how long it took to come back.

How do I find out which foods spike me using my CGM data? Run the same test three or more times. That sounds slow, but repeating the test helps separate the food from context.

• Log the meal before you eat. Include the specific food, portion size, and time.
• Write down your pre-meal glucose from the CGM, even if the app does it for you.
• Read the curve for the next few hours.
• Find the peak reading within that window.
• Calculate the rise: peak minus pre-meal.
• Record the peak time and whether the reading returns toward your pre-meal baseline.
• Repeat the same food on three or more separate days, in conditions as similar as possible.

In HealthOS, glucose is placed beside meals and workouts so patterns can be compared. The app is pre-launch with a waitlist at https://healthosx.com/start. HealthOS reads third-party sensor data from an export file or from Apple Health where the maker's app writes data. It does not connect directly to a Libre or other sensor. If you already have a sensor export, the free browser report shows your biggest rises, and the meal-spike checker can calculate a single meal's rise from your export.

The comparison that actually works: same food, same context A single test of a food can be noisy. Your response changes with meal size, time of day, recent exercise, stress, illness, where you are in a menstrual cycle, and sleep quality if you note it. A food that spikes you before breakfast may not spike you after a workout. A food eaten alone may look worse than the same food eaten with fat and protein, which slows the glucose rise.

That does not make the data useless. It tells you exactly what a glucose sensor can and cannot answer. It can answer what happened to you on that day, under those conditions. It cannot answer what the food always does until you repeat it.

One spike is a data point. A pattern is a finding.

To compare fairly, keep the portion the same, the time of day roughly the same, and note anything unusual: a bad night, a run, a stressful meeting, an illness. If you note those, you may find that your biggest glucose days are not about food at all.

What if a food spikes me once? Is that dangerous? No, not on its own. Rises after meals are normal. The 70 to 180 mg/dL time-in-range band was set for people with diabetes (Battelino et al., 2019). In people without diabetes, one study reported about 3 hours per day above 140 mg/dL (Spartano et al., 2024), while another reported about 96% of time between 70 and 140 mg/dL (Shah et al., 2019). These are context, not pass marks. The point is not to chase a flat line. A flat line is not normal either.

A single high reading can come from a sensor's first day, an infection, a larger portion than usual, or a stressful day. Pressure on a sensor while you sleep is more likely to fake a low than a high. Treat one dramatic trace as a cue to retest, not as a reason to ban a food.

Pay attention when the same food repeatedly produces a large rise, a high late peak, or a slow return. If you have diabetes, use the targets your care team gave you. Without diabetes, there is no settled dangerous threshold for a single post-meal excursion. Persistent high readings or readings that worry you are worth discussing with a doctor.

Turn it into a trigger-food list You do not need a perfect list. You need a short list of repeat offenders.

• Rank foods by average rise across your repeated tests.
• Mark foods that stay above your baseline for a prolonged period, not just foods that peak high.
• Retest the top offenders with a smaller portion, a different time of day, or paired with protein, fat or a walk.
• Keep the foods that produce a small rise and a quick return.

This is a personal list. It is not a ranking of good and bad foods for everyone. Your job is to find which ones move you.

If you already have a sensor export from LibreView, Dexcom Clarity, Stelo, Lingo or SiBionics, upload it to the free CGM report to see your AGP, time in range and biggest rises in the browser, no account. If you do not have a sensor yet, you can learn about HealthOS Sensor, a wellness product, not a medical device.