Dear Editor,
In our country (1),comparing local neonatal standards with a
fractional growth curve using Mongelli’s formula, at the 10th percentile
level (used to define small for gestational age –SGA-), neonatal-derived
fetal growth standards of the preterm period (27th - 34th weeks) were
significantly lower than the estimated fractional curve and with the
Hadlock’s fetal ultrasonographic standards, confirming the so called
“hidden” fetal growth restriction.
In recent Latin-American studies (2,3) , SGA (mostly symmetric) is a
frequent condition in infants with VLBW, and it is associated with higher
mortality rates, NEC, severe IVH, BPD, O2 requirement and malnutrition at
36 post-conceptional weeks.
As time may be a confounder because the long period included (25
years), other available analytical techniques can address several
questions that arise from the study:
1) Splitting up the study period into five year categories may show
different results in the relationship of FGR with mortality and morbidity
in preterm infants, because several clinical practices have came up (e.g.
surfactant).
2) Which is the cover rate of prenatal steroids across the period?
This may influence neonatal survival (“mortality effect”) and, on the
other hand, incidence and severity of several morbidities (e.g. PVH).
3) Changes in diagnostic criterions across time period.
4) Because both FGR and outcomes are time dependent, in addition to
unconditional logistic regression (adjusted for year of birth or year
categories), it could be useful to employ stratified analysis by time
period and/or survival analysis methods (specifically Cox regression) that
addresses time-dependent exposures and differences in observation period.
Sincerely yours
Carlos Grandi, MD, MS, PhD
Perinatal Epidemiology Unit and Biostatistics
Sarda Maternity Hospital, Buenos Aires, Argentina
cgrandi@intramed.net
Competing interests: none
References:
1. Grandi C, Luchtenberg G, Rojas E. ¿Are neonatal growth standards
adequate tools for the diagnosis of fetal growth restriction in preterm
babies? (see also Editorial). Arch argent pediatr 2003; 101: 357-364
(available at www.sap.org.ar/publicaciones).
2. Grandi C, Balanian N, Benítez A, Brundi M, Larguía M, Solana C and
Grupo Colaborativo NEOCOSUR. Intrauterine growth retardation (IUGR):
Effects on Morbimortality of very low birth weight infants. A multicenter
South American analysis. Pediatric Research 2004; 55: 530 (Abstract).
3. Grandi C, Tapia J, Marsall G , Grupo Colaborativo N EOCOSUR. Severity,
proportionality and risk of neonatal mortality of very low birth weight
infants with fetal growth restriction. A multicentric South American
analysys ( see also Editorial). J Pediatr (Rio J) 2005; 81(3): 198 – 204
(Medline).
Dear Editor,
In our country (1),comparing local neonatal standards with a fractional growth curve using Mongelli’s formula, at the 10th percentile level (used to define small for gestational age –SGA-), neonatal-derived fetal growth standards of the preterm period (27th - 34th weeks) were significantly lower than the estimated fractional curve and with the Hadlock’s fetal ultrasonographic standards, confirming the so called “hidden” fetal growth restriction.
In recent Latin-American studies (2,3) , SGA (mostly symmetric) is a frequent condition in infants with VLBW, and it is associated with higher mortality rates, NEC, severe IVH, BPD, O2 requirement and malnutrition at 36 post-conceptional weeks.
As time may be a confounder because the long period included (25 years), other available analytical techniques can address several questions that arise from the study: 1) Splitting up the study period into five year categories may show different results in the relationship of FGR with mortality and morbidity in preterm infants, because several clinical practices have came up (e.g. surfactant).
2) Which is the cover rate of prenatal steroids across the period? This may influence neonatal survival (“mortality effect”) and, on the other hand, incidence and severity of several morbidities (e.g. PVH).
3) Changes in diagnostic criterions across time period.
4) Because both FGR and outcomes are time dependent, in addition to unconditional logistic regression (adjusted for year of birth or year categories), it could be useful to employ stratified analysis by time period and/or survival analysis methods (specifically Cox regression) that addresses time-dependent exposures and differences in observation period.
Sincerely yours
Carlos Grandi, MD, MS, PhD
Perinatal Epidemiology Unit and Biostatistics Sarda Maternity Hospital, Buenos Aires, Argentina
cgrandi@intramed.net
Competing interests: none
References:
1. Grandi C, Luchtenberg G, Rojas E. ¿Are neonatal growth standards adequate tools for the diagnosis of fetal growth restriction in preterm babies? (see also Editorial). Arch argent pediatr 2003; 101: 357-364 (available at www.sap.org.ar/publicaciones).
2. Grandi C, Balanian N, Benítez A, Brundi M, Larguía M, Solana C and Grupo Colaborativo NEOCOSUR. Intrauterine growth retardation (IUGR): Effects on Morbimortality of very low birth weight infants. A multicenter South American analysis. Pediatric Research 2004; 55: 530 (Abstract).
3. Grandi C, Tapia J, Marsall G , Grupo Colaborativo N EOCOSUR. Severity, proportionality and risk of neonatal mortality of very low birth weight infants with fetal growth restriction. A multicentric South American analysys ( see also Editorial). J Pediatr (Rio J) 2005; 81(3): 198 – 204 (Medline).