Key words epidemiology - pregnancy - growth retardation
Schlüsselwörter Epidemiologie - Schwangerschaft - Wachstumsretardierung
Introduction
The weight gain of women during pregnancy is a key perinatal parameter; it is easily
measured and there is now a wealth of evidence showing that weight gain has an
impact on a number of important maternal and neonatal outcomes, including the
preterm birth rate and the duration of pregnancy [1 ], [2 ], [3 ], [4 ], [5 ], [6 ], [7 ], [8 ], [9 ], [10 ], [11 ], [12 ], [13 ], [14 ], [15 ],
the birth weight [7 ], [8 ], [9 ], [10 ], [11 ], [12 ], [13 ], [15 ], [16 ], [17 ], [18 ], [19 ], [20 ],
and the somatic classification of neonates [7 ], [8 ], [9 ], [15 ], [21 ], [22 ], [23 ], [24 ].
For this somatic classification neonates are typically classified using the 10th and
90th birth weight percentiles, calculated according to gestational age; neonates
below the 10th birth weight percentile are “small for gestational age” (SGA), those
above the 90th birth weight percentile are “large for gestational age” (LGA), and
all those in-between are “appropriate for gestational age” (AGA).
We previously investigated weight gain in pregnancy in relation to maternal
anthropometric measurements, finding an increase in the weight gained during
pregnancy with increasing maternal height and, for women weighing more than about
63 kg at the first obstetric consultation (for the pregnancy in question), an
inverse relationship between maternal weight at the first consultation and the
weight gained during pregnancy [25 ]. Maternal
anthropometric measurements of course also influence neonatal anthropometric
measurements and therefore the somatic classification of neonates, when this is
performed using percentiles of the total study population [26 ], [27 ].
The classification of weight gain during pregnancy and the somatic classification of
neonates can therefore be expected to be more accurate when maternal body height and
weight are taken into account, rather than when percentiles calculated from the
total study population are used, irrespective of maternal anthropometric
measurements. As body height and weight can be combined into a single measure such
as the body mass index (BMI), the question arises whether the classification of
maternal weight gain and the neonatal somatic classification should be undertaken
according to maternal BMI, e.g. whether the ranges for weight gain should be
specified for different BMI ranges. Indeed, the Institute of Medicine recommends
assessing weight gain during pregnancy by maternal BMI [28 ]. However, we have shown that both maternal weight gain [25 ] and the somatic classification of neonates [29 ] can differ substantially between women who have the
same BMI but different body heights and weights.
An alternative approach, i.e. taking account of maternal anthropometric measurements
by grouping women according to their height and weight, therefore deserves
investigation. We previously presented norm values for weight gain during pregnancy
for different maternal height and weight groups [30 ]. In
the present analysis, we investigate the classification of weight gain during
pregnancy and the somatic classification of neonates: we compare classifications
based on weight gain and birth weight percentiles that were calculated from the
total study population with classifications based on percentiles that were
calculated separately for the different maternal height and weight groups.
The question we wanted to answer was whether these classifications differ
substantially, because if they do not, it would be hard to see any advantage of
using classification systems that take maternal height and weight into account.
Material and Methods
The data on which this analysis is based were taken from the routine data collection
of the German Perinatal Survey that is done throughout Germany. Data were kindly
made available to Dr. Voigt by the Chambers of Physicians of the States of Germany.
The data were collected in the years 1995 to 2000. For the years 1995–1997 all
States of Germany except Baden-Württemberg provided data; thereafter only Bavaria,
Brandenburg, Hamburg, Mecklenburg-Western Pomerania, Lower Saxony, Saxony,
Saxony-Anhalt, and Thuringia provided data, not necessarily for all years. Overall,
our database contains datasets from more than 2.2 million singleton pregnancies;
this is our total study population. Because not all datasets were complete with
regard to all maternal and neonatal parameters collected, the sample sizes vary
between analyses. The figures presented in this paper and the supplemental online
figures contain information on the case numbers included in individual analyses.
We calculated weight gain during pregnancy from the weights recorded at the first
obstetric consultation and at the end of pregnancy; weight gain was classified using
the 10th and 90th weight gain percentiles; “low weight gain” was defined as a weight
gain below the 10th percentile, “high weight gain” as above the 90th percentile, and
“medium weight gain” was between the 10th and 90th percentiles.
As described in the introduction, the neonatal somatic classification was based on
birth weight percentiles specified according to gestational age, using the 10th and
90th percentiles; neonates with a birth weight below the 10th percentile were SGA,
those with a birth weight greater than the 90th percentile were LGA, and those
in-between were AGA.
The classification of weight gain during pregnancy and the somatic classification of
neonates were done using either the 10th and 90th percentiles calculated from the
total study population or the 10th and 90th percentiles calculated specifically for
the groups of mothers and neonates compiled based on maternal height and weight. We
compiled 12 such groups based on a division of maternal height into three groups
(≤ 161 cm, 162–171 cm, ≥ 172 cm) and a division of maternal weight into four groups
within the height groups, as described previously [30 ].
In addition to presenting the classification of maternal weight gain during
pregnancy and the somatic classification of neonates separately, we also show
combined classifications: neonatal somatic classifications are presented separately
for low, medium, and high weight gain.
The chi-squared test was used to establish the statistical significance of
differences between neonatal somatic classifications of different maternal weight
gain groups. Data analysis was performed using the computer programme SPSS (version
20) in the computing centre of the University of Rostock, Germany.
Results
Using percentile values for maternal weight gain and birth weight calculated
from the total study population
In the printed version of this article we show data for women with a height
≤ 161 cm and their neonates. Similar findings were obtained for the other
maternal height groups: 162–171 cm and ≥ 172 cm (see supplemental online
figures). [Fig. 1 ] and supplemental online
Figs. 1S and 2S illustrate the classification of maternal
weight gain during pregnancy and the neonatal somatic classification as well as
the combination of these classification systems (neonatal somatic
classifications presented separately for low, medium, and high weight gain, as
described above) when the percentiles used for the classification of weight gain
and the somatic classification of neonates (i.e. the 10th and 90th percentiles)
were calculated from the total study population of women or neonates,
respectively.
Fig. 1 Classifications based on weight gain and birth weight
percentiles calculated from the total study population for women with a
height ≤ 161 cm and their neonates. The classification of maternal weight
gain during pregnancy for women with a height ≤ 161 cm and a weight of
≤ 57 kg, 58–73 kg, 74–89 kg, or ≥ 90 kg (Figs. 1 a–d , respectively)
and the somatic classification of infants born to them are shown in the two
columns on the left of each figure. The right side of each figure shows a
combination of the two: the somatic classification of neonates within the
three weight gain groups (< 10th, 10–90th and > 90th weight gain
percentile). Neonates were classified as small for gestational age (SGA,
< 10th birth weight percentile), appropriate for gestational age (AGA,
10–90th birth weight percentile), or large for gestational age (LGA,
> 90th weight gain percentile). Contributing case numbers are shown on
top of the columns. The p-values were calculated using the chi-squared
test.
Using percentile values for maternal weight gain and birth weight calculated
separately for the 12 groups of women compiled according to height and
weight
[Fig. 2 ] and supplemental online Figs. 3S and
4S show the classification of maternal weight gain during pregnancy
and the neonatal somatic classification as well as the combination of these
classification systems (as above) but this time using group-specific percentiles
for weight gain during pregnancy and also group-specific percentiles of birth
weight for gestational age calculated separately for the 12 groups of women and
their neonates formed according to maternal height and weight.
Fig. 2 Classifications based on weight gain and birth weight
percentiles calculated specifically for the maternal height and weight
groups in question for women with a height ≤ 161 cm and their neonates.
Details are analogous to the description in the legend to [Fig. 1 ].
When percentiles based on the total study population were used, there was a large
variability in the percentages of women with low and high weight gain and in the
percentages of SGA and LGA neonates. When group-specific percentiles were used,
this variability was much lower. For example, for percentiles calculated from
the total study population the percentage of women with a low weight gain (below
the 10th weight gain percentile) ranged between 5.7 and 39.4 % in women with a
height ≤ 161 cm, depending on their weight group ([Fig. 1 ]). Using group-specific percentiles the variation was
considerably lower and ranged between 7.7 and 8.5 % ([Fig. 2 ]).
The proportion of SGA neonates born to women with a height ≤ 161 cm varied
between 8.0 and 17.6 %, depending on maternal weight, when percentiles
calculated from the total study population were used for the somatic
classification of neonates; when group-specific percentiles were used, the
variation was between 9.5 and 9.8 %, again substantially lower. The same pattern
was observed for the other maternal height and weight groups.
When the combined classifications of weight gain during pregnancy and birth
weight for gestational age were examined, considerable differences were also
found between classifications using percentiles calculated from the total study
population and those using group-specific percentiles. When percentiles
calculated from the total study population were used, there was considerable
variability between maternal weight groups with regard to the influence of
maternal weight gain on the somatic classification of neonates. Using
group-specific percentiles there was less variability between the maternal
weight groups in this regard.
Discussion
We were fortunate to have a large amount of data available which permitted reliable
calculations of percentiles for weight gain during pregnancy and birth weight for
gestational age even in sub-populations of our total study population, i.e. in the
groups of women and neonates compiled according to maternal height and weight. Our
key finding is that the classifications of maternal weight gain during pregnancy by
weight gain percentiles and the somatic classifications of neonates as SGA, AGA, or
LGA differ substantially depending on whether these classifications were done using
percentiles calculated from the total study population or using group-specific
percentiles calculated specifically for the groups of women and neonates compiled
based on maternal height and weight. The impact of using classification systems that
take account of maternal height and weight on the medical care and health of
neonates still needs to be investigated and this remains a task for future
research.
Limitations of our study include a degree of uncertainty with regard to the
calculation of weight gain during pregnancy: weight at the first obstetric
consultation depends on when the first consultation occurs and weight at the end of
pregnancy varies depending on the length of gestation. Moreover, the grouping of
women by maternal height and weight was arbitrary and different cut-off points and
different numbers of groups would also have been possible.
Because the classification of neonates as small, appropriate, or large for
gestational age may well have consequences for the care of these neonates and with
regard to the use of resources, it is important to use as accurate a classification
system as possible. A classification system that takes account of maternal
anthropometric measurements should more accurately identify small or large neonates
due to abnormal intrauterine growth as compared to constitutionally small or large
neonates.
Acknowledgements
We would like to thank the Chambers of Physicians of the States of Germany for
contributing data to our database. We also thank Christel Fernow for her help with
preparing the figures.
