The 0.70 Cutoff in COPD Diagnosis: Fixed Ratio Versus Lower Limit of Normal

The short version#

A person has spirometric COPD, in the eyes of the GOLD program, when their post-bronchodilator FEV1/FVC sits below 0.70. That single fixed number is the anchor because it is reproducible, easy to teach, and does not change when a laboratory swaps in a different reference equation. Its rival, the lower limit of normal (LLN), adjusts the threshold for each patient's age, height, and sex, which is biologically tidier but leaves the cutoff dependent on whichever equation was chosen. The tension is real, and the 2025 GOLD report settles it not by picking a winner but by keeping 0.70 as the diagnostic anchor and asking clinicians to cross-check the LLN in the two groups the fixed ratio most often gets wrong.

Key points#

• The fixed ratio flags obstruction whenever post-bronchodilator FEV1/FVC is below 0.70, the same line for everyone.
• The LLN flags a ratio below the fifth percentile of a healthy reference population matched on age, height, and sex.
• Because the healthy ratio falls with age, 0.70 tends to overcall COPD in older adults and undercall early disease in younger ones.
• Large outcome data show 0.70 predicts hospitalization and death at least as well as the LLN, which is GOLD's main reason for keeping it.
• The practical rule for 2025 is to use 0.70 as the anchor and read the LLN or z-score alongside it in borderline younger and older patients.

What the two thresholds are actually counting#

Spirometry after a bronchodilator produces two volumes that define airflow obstruction: how much air you can force out in the first second (FEV1) and how much you can exhale in total (FVC). Divide the first by the second and a low result means air is leaving too slowly, the signature of obstruction. Everything in this debate turns on where to draw the low line.

The fixed approach puts that line at 0.70 for every adult, no adjustment. The LLN approach compares each person against a reference distribution of healthy never-smokers of the same age, height, and sex, and calls the result abnormal when it falls below the fifth percentile, which corresponds to a z-score of roughly -1.645. Those reference values come from pooled datasets, most prominently the Global Lung Function Initiative 2012 equations, which Quanjer and colleagues built in the European Respiratory Journal from more than 70,000 healthy nonsmokers spanning ages 3 to 95.

Why a fixed line slips at both ends of adult life#

Lungs lose elastic recoil as they age. FVC holds up somewhat better than FEV1 over the years, so the healthy ratio drifts steadily downward across a normal lifespan. By the seventies and eighties, the true fifth percentile in people who never smoked can fall below 0.70 on its own. Hold the line at a fixed 0.70 and you end up counting ordinary aging as disease, which is how the fixed ratio overdiagnoses older adults.

Run the same physiology forward and the opposite error appears in the young. A healthy thirty-year-old often sits near 0.80 or higher. Someone that age with genuine early obstruction can post a ratio of 0.72, comfortably above the fixed line yet below their own age-matched LLN, and walk away reassured that their lungs are fine. Both mistakes trace back to a single cause: a fixed threshold cannot follow a moving physiological target. GOLD names both errors in the 2025 report and adds guidance for younger patients precisely because a missed early diagnosis is the harder one to catch later.

The evidence GOLD leans on to keep 0.70#

If the fixed ratio drifts wrong at the edges, why keep it? The defense comes down to three arguments, and the strongest is about outcomes rather than tidiness.

The first argument is operational. One number applies in any clinic on the planet without loading population-specific software or picking a reference set. The second, and the one that carries the most weight, is predictive validity. In a pooled analysis of more than 24,000 US adults published in JAMA in 2019, Bhatt and colleagues followed participants for a median of about 15 years and found that defining obstruction as FEV1/FVC below 0.70 discriminated COPD-related hospitalization and death as well as or better than the LLN. When they let the data choose the statistically optimal threshold, it landed at 0.71, close enough to 0.70 to be indistinguishable in practice.

The third argument is that spirometry never stands alone. Writing in the European Respiratory Journal in 2025, the GOLD Science Committee stressed that a COPD diagnosis requires symptoms plus a relevant risk factor such as tobacco or biomass smoke, and that a confirmatory post-bronchodilator test further trims false positives. The same review noted that overdiagnosis estimates, which range from roughly 11 to 35 percent across cohorts when spirometry is read in isolation, shrink once symptoms and risk factors are required before the label is applied.

The catch the LLN cannot shake#

The LLN looks like the objective fix, but it is only as stable as the equation sitting behind it. Change the reference population and the threshold moves with it, even though the patient's lungs have not changed at all. This stopped being hypothetical in 2023, when the American Thoracic Society recommended race-neutral interpretation using the GLI Global equation, which averages across ancestry groups instead of applying separate coefficients. That shift reset predicted values and LLN thresholds for many patients, reclassifying some from normal to obstructed and others the other way. It is the underappreciated weakness of any individualized cutoff: it carries every assumption baked into its reference model, and those assumptions are still being revised.

Putting both numbers to work#

The 2025 report's answer is to use the two tools for what each does best rather than crown one. The fixed 0.70 stays the diagnostic anchor for consistency across clinics and for its track record against hard outcomes. The LLN or z-score becomes a targeted cross-check in the patients where the fixed ratio is least trustworthy. In an adult under 50 whose ratio sits at or just above 0.70 but who has symptoms and a risk factor, comparing against the predicted LLN can catch an early diagnosis the fixed line would miss. In an older adult, the LLN can temper an overcall driven by age alone. In both cases the number on the report is where your clinical reasoning begins, not where it ends.