Failure to seroconvert (primary vaccine failure) is believed to be the principal reason (approx. > 95%) why some vaccinees remain susceptible to measles and is often attributed to the persistence of maternal antibodies in children vaccinated at a young age. Avidity testing is able to separate primary from secondary vaccine failures (waning and/or incomplete immunity), but has not been utilized in measles epidemiology. Low-avidity (LA) and high-avidity (HA) virus-specific IgG antibodies indicate primary and secondary failure, respectively. Measles vaccine failures (n = 142; mean age 10.1 years, range 2-22 years) from an outbreak in 1988-9 in Finland were tested for measles-virus IgG avidity using a protein denaturating EIA. Severity of measles was recorded in 89 failures and 169 non-vaccinees (mean age 16.2 years, range 2-22 years). The patients with HA antibodies (n = 28) tended to have clinically mild measles and rapid IgG response. Among failures vaccinated at < 12, 12-15 and > 15 months of age with single doses of Schwarz-strain vaccine in the 1970s, 50 (95% CI 1-99), 36 (CI 16-56) and 25% (CI 8-42) had HA antibodies, respectively. When a single measles, mumps and rubella (MMR) vaccine had been given after 1982 at 15 months of age, only 7% (CI 0-14) showed HA antibodies. Omitting re-vaccinees and those vaccinated at < 15 months, Schwarz-strain recipients had 3.6 (CI 1.1-11.5) higher occurrence of HA responses compared to MMR recipients. Apart from one municipality, where even re-vaccinees had high risk of primary infection, 89% (CI 69 to approximately 100) of the infected re-vaccinees had an HA response. Secondary measles-vaccine failures are more common than was more previously thought, particularly among individuals vaccinated in early life, long ago, and among re-vaccinees. Waning immunity even among individuals vaccinated after 15 months of age, without the boosting effect of natural infections should be considered a relevant possibility in future planning of vaccination against measles.