BAC Levels
ES

How Accurate Are Breathalyzers?

Updated May 2026 · 12 min read

Breathalyzers are science, but not infallible science. The devices measure alcohol in your breath and estimate blood alcohol content using an assumed conversion ratio — and that assumption fails for millions of people depending on their physiology, medications, and health conditions. A breathalyzer can be off by 20%, 30%, or more without anyone being the wiser.

This guide covers what breathalyzers actually measure, why they're fallible, which medical conditions and substances throw off readings, and when test results hold up in court and when they crack under scrutiny.

How Breathalyzers Actually Work

A breathalyzer doesn't directly measure blood alcohol. Instead, it measures ethanol vapor in your breath and converts that measurement to a blood alcohol estimate using a fixed formula.

Alcohol is absorbed into the bloodstream through your stomach and small intestine. Your lungs then exhale a portion of that alcohol as vapor — roughly 1 unit of alcohol in your blood equilibrates with 2100 units in your breath (the "breath-to-blood ratio"). The device assumes this 2100:1 ratio holds true for everyone, which is the core flaw in the system.

Several methods exist to measure breath alcohol:

All roadside breathalyzers require calibration against known-alcohol solutions. If a device isn't calibrated regularly (often required every 3-6 months), accuracy degrades quickly.

The Fatal Flaw: Individual Breath-to-Blood Ratios

The 2100:1 breath-to-blood ratio is an average derived from population-level studies. Real people vary dramatically:

What determines your ratio? Lung capacity, breathing patterns, hematocrit (red blood cell percentage), metabolism, body composition, and genetics all play roles. A forensic toxicologist can estimate your ratio from blood tests and prior arrests, but roadside breathalyzers can't account for individual variation — they just assume 2100:1.

In criminal defense, this is known as the "partition ratio defense." If a defendant's actual ratio is significantly lower than 2100:1, the breathalyzer overstates their BAC. A breathalyzer reading of 0.10% might represent a true BAC of 0.075% — under the legal limit.

Mouth Alcohol: The Easiest Mistake

Mouth alcohol is the single most common source of false positive readings. If you burped, vomited, or had alcohol residue in your mouth within 15-20 minutes of testing, the breathalyzer detects that oral alcohol and inflates your reading.

Common sources of mouth alcohol:

Proper police procedure requires a 15-20 minute observation period before administering a breath test. The officer watches to ensure the subject doesn't eat, drink, smoke, or vomit during this window. If a police report omits this observation period, it's a strong defense — the breathalyzer reading becomes questionable.

Medical Conditions That Spike Breathalyzer Readings

GERD (Gastroesophageal Reflux Disease)

GERD affects roughly 20% of the US population. In people with GERD, stomach acid and contents, including alcohol, backflow into the esophagus and mouth. This means a person with GERD can blow 0.05% or higher on a breathalyzer from a single drink hours after consumption, even if their BAC is near zero. Medications like Nexium or Prilosec can reduce (but not eliminate) this effect.

Diabetes (Uncontrolled)

Poorly controlled diabetes causes ketoacidosis — a buildup of ketones in the blood and breath. Some infrared breathalyzers misidentify ketones as ethanol because both molecules absorb light at similar wavelengths. A person in diabetic ketoacidosis can blow 0.03-0.08% falsely on a breathalyzer despite being sober. This is a legitimate medical defense in court if documented via blood glucose and ketone testing.

Ketosis (Low-Carb Diets)

Strict keto or low-carb diets trigger ketosis — elevated ketone bodies in the breath. While not as severe as diabetic ketoacidosis, ketosis can add 0.01-0.02% to breathalyzer readings. Athletes and people on Atkins/keto diets have documented cases of breathalyzer overstatement.

Anemia (Low Hematocrit)

Red blood cells (hematocrit) affect the breath-to-blood ratio. People with anemia have lower hematocrit — they read 5-10% higher on breathalyzers compared to people with normal red blood cell counts. Women tend to have lower hematocrit than men, which is one reason women's BAC can spike higher from the same drinks.

Liver Disease

Cirrhosis, hepatitis, and fatty liver disease impair alcohol metabolism. People with liver disease metabolize alcohol more slowly, keeping BAC elevated longer. Additionally, portal hypertension (high blood pressure in liver veins) can cause varices and bleeding that ingests blood — fresh blood alcohol levels differ from steady-state levels. The dynamics are complex and often favor a DUI defense.

Medications & Substances That Interfere

Several medications and substances produce false positives or artificially inflate readings:

If you were taking any medication before a breathalyzer test, disclose it immediately to your attorney. It's a potential defense.

Device Calibration & Maintenance

Breathalyzer accuracy depends entirely on calibration. An uncalibrated or drifting device can read 20-40% off. Many jurisdictions require calibration every 3 months; some only annually. If a device hasn't been calibrated recently or if calibration records are missing, the reading is challengeable in court.

Police departments must maintain calibration logs. Defense attorneys routinely request these records. If they're incomplete or show late calibration, it's grounds for excluding the breathalyzer result as unreliable.

Breathing Patterns & Technique

How you blow into the device affects the reading:

Proper procedure requires the subject to blow steadily for at least 6 seconds. If the officer didn't demonstrate proper technique or give clear instructions, the result is questionable.

Challenging Breathalyzer Results in Court

If you've been charged with DUI based on a breathalyzer test, several lines of defense exist:

Critical: Many states allow you to request a blood test to verify a breathalyzer result. If you blew above the limit and immediately requested a blood test, you preserved your defense. A blood test is more accurate and harder to challenge than a breath test.

What Breathalyzers Do Well

Despite their limitations, properly calibrated breathalyzers are useful tools for screening. They're quick, non-invasive, and reasonably reliable for BAC estimates in the 0.08-0.15% range in people without complicating health conditions. In laboratory conditions with careful calibration, they're accurate to within ±5%.

The real-world accuracy gap — 10-20% or more — emerges from device calibration drift, operator error, mouth alcohol, medical conditions, and breathing technique. These variables explain why the same person can blow 0.08% on one device and 0.095% on another within minutes.

The Bottom Line

Breathalyzers are not infallible. They're tools that measure breath alcohol and estimate blood alcohol using population-average assumptions — assumptions that fail for millions of people due to medical conditions, medications, device calibration errors, and physiological variation. A breathalyzer reading is evidence, not proof. If you've been charged with DUI based on a breath test, challenge it with expert testimony, medical records, and procedural scrutiny. A skilled DUI attorney can often raise reasonable doubt about the reliability of the result.

And if you're trying to estimate your own BAC to decide if you should drive, remember: breathalyzers are approximations at best. Our BAC Calculator uses the same Widmark formula law enforcement relies on, and it's free. When in doubt, don't drive — call a cab or rideshare.

Frequently Asked Questions

How accurate are police breathalyzers?
The most reliable roadside breathalyzers (like the Alco-Sensor III) claim ±10% accuracy under lab conditions, meaning a true BAC of 0.08% might register anywhere from 0.072% to 0.088%. Real-world accuracy is often worse due to improper calibration, operator error, and environmental factors. Breath tests measure breath alcohol, not blood alcohol directly — the conversion relies on an average 2100:1 breath-to-blood ratio that varies by individual.
Can mouth alcohol affect the test?
Yes. If you burped, regurgitated, or had alcohol residue in your mouth within 15-20 minutes of the test, the breathalyzer reads the mouth alcohol in your breath, not the BAC in your lungs. A mouthwash sip, cough syrup, or beer-flavored candy can spike readings. Proper police procedure includes a 15-20 minute observation period before testing to prevent this — but not all officers follow it.
Do medical conditions affect breathalyzer readings?
Absolutely. GERD (acid reflux) pushes stomach alcohol into the esophagus and mouth, inflating readings. Uncontrolled diabetes causes ketones in the breath that some breathalyzers confuse with alcohol, producing false positives. High blood sugar and low insulin also trigger ketosis. Yeast infections in the mouth can ferment residual sugars into ethanol, falsely elevating readings. These conditions can make a sober person blow 0.05% or higher on a roadside breathalyzer.
What is the legal breath-to-blood ratio?
The assumption is 2100:1 — for every 1 unit of alcohol in the blood, there are 2100 units in the breath. This ratio varies by individual from roughly 1500:1 to 2400:1 depending on lung capacity, breathing patterns, metabolism, and body composition. A person with a 1500:1 ratio will blow 40% higher than someone with a 2100:1 ratio at the same BAC. The law uses the average ratio, which disadvantages some people.
What factors influence breathalyzer accuracy?
Device calibration (the biggest factor — out-of-calibration units can be off by 20%+), operator training, ambient temperature, subject's breathing depth, mouth temperature, hematocrit (red blood cell count), and recent food/drink. A person with low hematocrit (anemia) reads 5-10% higher. High hematocrit reads lower. Breathing shallowly underestimates BAC; deep breathing overestimates it. These variables explain why identical BACs produce different readings across devices and people.