In this issue, Delnord et al. use gestational age (GA) data from 30 high-income countries to evaluate rates of very preterm birth (VPTB; < 32 weeks of gestation) in liveborn and stillborn singleton infants (Delnord et al. BJOG 2016; DOI: 10.1111/ 1471-0528.14273). They were especially interested in the contribution of periviable 22–23 weeks of gestation infants to VPTB and their influence on variability of VPTB among countries. The median birth rate per 1000 total births across all countries for 22–23 weeks of gestation was 0.9/ 1000; for 24–27 weeks was 2.8/1000; and for 28–31 weeks was 5.5/1000 for a total VPTB rate of 9.5/1000, ranging from 5.7 to 15.7. The proportion of stillbirths to all births was 58.3% at 22–23 weeks; 24.8% at 24–27 weeks and 10.6% at 28–31 weeks. The risk of neonatal mortality for the earliest PTB is very high (Stoll et al. Pediatrics 2010;126:443–56). Hence, although the VPTB contribution to all births is small, their contributions to both stillbirth and neonatal death are large. In 2011, a PTB classification system recommended that all births at ≥ 16 weeks, both stillborn and liveborn, be counted in order to broaden information available on pregnancy outcomes (Villar et al. Am J Obstet Gynecol 2012;206:119–23). Regardless of the lower GA cut-off chosen, for comparison purposes across sites or over time, a clearly defined and scientifically sound GA cut-off should define PTB. Delnord et al. found variations in the lower GA criteria used to register births as well as reporting of pregnancy terminations. Large differences were found among countries in VPTB as a proportion of all births and in stillbirth as a proportion of VPTB, with even greater reported variations in PTB at 22–23 weeks. To make the data on VPTB more consistent across countries, Delnord et al. conclude that international comparisons of PTB should exclude births before 24 weeks. However, if this were to occur, about 10% of PTB <32 weeks would not be reported, as would 20% of stillbirths. There is no biological reason to use a 24-week cut-off because the aetiologies of PTB and stillbirth do not change dramatically as the GA advances from 23 to 24 weeks. The rationale is that birth data for 22– 23 weeks are not as well collected as data for slightly older pregnancies. The World Health Organization uses a similar justification for recommending a 28-week cut-off for stillbirth for international comparisons (Cartlidge and Stewart. Lancet 1995;346:486–8). Given the resources required and the difficulty in collecting pregnancy data in the poorest countries, this approach may be justifiable. However, in highincome countries, as in this report, with vital registration systems already in place, improved documentation of pregnancy outcomes at 22–23 weeks of gestation, or at even lower GA, is easily feasible. Collecting and reporting appropriate data do not appear to be constrained by inadequate resources, but by lack of will. Rather than recommending not reporting pregnancy outcome data for 22–23 weeks, an alternative recommendation would be to collect it in a standard manner and completely. Hence, if we really want to understand the full extent of adverse pregnancy outcomes in high-income countries, we should strongly advocate for collection of complete and accurate birth outcome data, regardless of the GA or birthweight of the stillborn or liveborn infant.