An annualised linear growth velocity (LGV) reference can identify groups of children at risk of growing poorly. As a single velocity reference for all preschool ages does not exist, we present an interim tool, derived from published, normative growth studies, for detecting growth faltering, illustrating its use in Nepali preschoolers.

The WHO Child Growth Velocity Standard was adapted to derive 12-month increments and conjoined to the Tanner-Whitehouse Height Velocity Reference data yielding contiguous preschool linear growth annualised velocities. Linear restricted cubic spline regressions were fit to generate sex-specific median and standard normal deviate velocities for ages 0 through 59 months. LGV Z-scores (LGVZ) were constructed, and growth faltering was defined as LGVZ < –2.

Use of the reference was illustrated with data from Nepal’s Tarai region.

Children contributing the existing growth references and a cohort of 4276 Nepali children assessed from 2013 to 2016.

Fitted, smoothed LGV reference curves displayed monotonically decreasing 12-month LGV, exemplified by male/female annual medians of 26·4/25·3, 12·1/12·7, 9·1/9·4, 7·7/7·8 and 7/7 cm/years, starting at 0, 12, 24, 36 and 48 months, respectively. Applying the referent, 31·1 %, 28·6 % and 29·3 % of Nepali children <6, 6–11 and 12–23 months of age, and ∼6 % of children 24–59 months, exhibited growth faltering. Under 24 months, faltering velocities were more prevalent in girls (34·4 %) than boys (25·3 %) (P < 0·05) but comparable (∼6 %) in older preschoolers.

A LGV reference, concatenated from extant data, can identify preschool groups at-risk of growth faltering. Application and limitations are discussed.

We examined sex effects on growth faltering in an indigenous population of Bedouin Arab infants born in Israel, an economically developed country.

Retrospective cohort study. Height-for-age Z-score (HAZ) at age 6, 12, 18 and 24 months was calculated for full-term, normal-birth-weight Bedouin infants born during years 2000–2009 and attending maternal and child health (MCH) clinics. Multivariate linear regression analysis (MLRA) was used to calculate the association between sex and HAZ, controlling for year of birth, birth weight (BW) and residence by type of settlement (established settlement (ES); non-established settlement (NES)).

Bedouin are an indigenous poor community of semi-nomadic Arabs, with the highest infant mortality rate in Israel. Fifty per cent of Bedouin infants live in NES with inadequate access to running water, electricity, and rubbish and sewage disposal. All Bedouin receive free well-baby care in community-based MCH clinics.

Full-term, normal Bedouin infants (n 5426) born during 2000–2009 and attending computerized MCH clinics who had growth measurements at age 6, 12, 18 and 24 months.

At all ages, girls had significantly higher mean HAZ than boys (P < 0·05). Increasing birth year, residence in ES and increasing BW were positively associated with HAZ (P < 0·05) at all ages. In MLRA controlling for birth year, BW and type of settlement, sex still had a significant effect, with lower HAZ among boys at 6, 12, 18 and 24 months of age (P < 0·001 at all ages).

Our results indicated that sex effects on growth faltering can occur in an indigenous population with low socio-economic status within an economically developed country.

We characterized post-infancy child growth patterns and determined the incidence of becoming stunted and of recovery from stunting.

Data came from Young Lives, a longitudinal study of childhood poverty in four low- and middle-income countries.

We analysed length/height measurements for children at ages 1, 5 and 8 years.

Children (n 7171) in Ethiopia, India, Peru and Vietnam.

Mean height-for-age Z-score (HAZ) at age 1 year ranged from −1·51 (Ethiopia) to −1·08 (Vietnam). From age 1 to 5 years, mean HAZ increased by 0·27 in Ethiopia (P < 0·001) and decreased among the other cohorts (range: −0·19 (Peru) to −0·32 (India); all P < 0·001). From 5 to 8 years, mean HAZ increased in all cohorts (range: 0·19 (India) to 0·38 (Peru); all P < 0·001). Prevalence of stunting (HAZ<−2·0) at 1 year ranged from 21 % (Vietnam) to 46 % (Ethiopia). From age 1 to 5 years, stunting prevalence decreased by 15·1 percentage points in Ethiopia (P < 0·001) and increased in the other cohorts (range: 3·0 percentage points (Vietnam) to 5·3 percentage points (India); all P ≤ 0·001). From 5 to 8 years, stunting prevalence decreased in all cohorts (range: 5·0 percentage points (Vietnam) to 12·7 percentage points (Peru); all P < 0·001). The incidence of becoming stunted between ages 1 to 5 years ranged from 11 % (Vietnam) to 22 % (India); between ages 5 to 8 years, it ranged from 3 % (Peru) to 6 % (India and Ethiopia). The incidence of recovery from stunting between ages 1 and 5 years ranged from 27 % (Vietnam) to 53 % (Ethiopia); between ages 5 and 8 years, it ranged from 30 % (India) to 47 % (Ethiopia).

We found substantial recovery from early stunting among children in four low- and middle-income countries.