Community Dent Oral Epidemiol 2007; 35: 364–376  2007 The Authors.

Journal compilation
All rights reserved  2007 Blackwell Munksgaard

Maria C.D. Saraiva1, Heliosa Bettiol2,

Are intrauterine growth Marco A. Barbieri2 and Antonio A. Silva3
1
University of São Paulo at Ribeirão Preto,

restriction and preterm birth School of Dentistry, Ribeirão Preto, Brazil,

2
Department of Pediatrics, University of São
Paulo at Ribeirão Preto, School of Medicine,
Ribeirão Preto, Brazil, 3Department of Public

associated with dental caries? Health, Federal University of Maranhão,

Maranhão, Brazil

Saraiva MCD, Bettiol H, Barbieri MA, Silva AA. Are intrauterine growth
restriction and preterm birth associated with dental caries? Community Dent
Oral Epidemiol 2007; 35: 364–376.  2007 The Authors. Journal compilation
 2007 Blackwell Munksgaard

Abstract – Objectives: To assess the association between two intrauterine

growth restriction (IUGR) surrogates – IUGR [small for gestational age birth
(SGA) and fetal growth restriction (FGR)] and preterm birth with dental
caries. Methods: Data from the Third National Health and Nutritional
Examination Survey (1988–1994) were used, including 2- to 5.9-year-old
singletons (n ¼ 3189). Dental caries was defined as presence of any teeth with
dental caries (treated or untreated) and also as presence of at least two teeth
with dental caries. Exposure variables were preterm birth (<37 gestational
weeks), FGR, and SGA. Covariates included were poverty, race/ethnicity, age,
sex, sucrose intake, environmental tobacco smoking, dental visits, education of
head of household, breastfeeding, and use of baby bottle. Separate statistical
analyses were conducted for IUGR and for preterm birth through the estimation
of prevalence ratio (PR), taking complex sampling design into consideration
and adjusting for confounders. Sensitivity analysis was conducted including
and excluding 2-year-old children and also with the two definitions of dental
caries. Results: In general, the inclusion of 2-year-old children and the case
definition of presence of any teeth with dental caries biased the results toward
the null, but with no major changes in the results. In bivariate analysis, SGA and
FGR birth were both negatively but not significantly associated with dental
caries while a significant positive association was found for preterm birth.
Sensitivity analysis showed that the PR for preterm in bivariate analysis varied
from 1.65 (95% CI 1.14–2.40) to 1.84 (95% CI 1.19–2.83). After adjusting for
confounders, the PR for preterm birth varied from 1.38 (95% CI 1.00–1.89) to
1.64 (95% CI 1.22–2.20). After adjustment, the PR for SGA varied from 0.79 (95%
CI 0.56–101) to 0.66 (95% CI 0.33–0.96). For children from 3 to 5.9 years old, the
adjusted PR for FGR using the category ‘none’ as reference were mild (PR 1.10;
95% CI 0.76–1.58), moderate (PR 0.66; 95% CI 0.26–167), and severe (PR 0.59; Key words: dental caries; intrauterine
growth restriction; low birth weight;
95% CI 0.36–0.99). These values for FGR were very similar for the other models preterm
using other classifications of case definition or inclusion of 2-year-old
children. Conclusions: Preterm birth was found to be positively associated with Corresponding author: Dr Maria Saraiva,
Av do Café, s/n Ribeirao Preto, Sao Paulo
dental caries while there is an indication that SGA and FGR are negatively 14040-904, Brazil
associated with dental caries. Although the negative association is Tel.: +55 16 36023995
counterintuitive, it is possible that increased antibiotic use and delayed tooth e-mail: [email protected]
eruption may explain the negative association between IUGR and dental Submitted 8 December 2005;
caries. accepted 18 June 2006

Intrauterine growth restriction (IUGR) is one of the IUGR has been assessed through surrogates such
major perinatal health problems in developing as as low birthweight (LBW), small for gestational age
well as in developed countries (1). The presence of (SGA) birth, fetal growth restriction (FGR), and low

364 doi: 10.1111/j.1600-0528.2006.00345.x

 IUGR, preterm and dental caries

ponderal index (2, 3). In the USA, the rates of 1994), addressed the relationship between LBW
preterm birth and LBW have increased in the past and dental caries among 2- to 6-year-old children
decades (1). (19). The authors evaluated data on gestational age
In addition to perinatal mortality, SGA birth and and concluded that neither LBW nor preterm birth
LBW have been associated with chronic disorders was associated with dental caries.
such as respiratory diseases (4, 5), cognitive Although gestational ages and birth weights are
impairments (6), and risk factors for cardiovascular available from NHANES III data, no evaluation of
disease (7). However, the oral health of preterm the relationship between IUGR and dental caries
and intrauterine growth-restricted children has has been performed to date. Therefore, given the
received little attention, and the risk of dental significant impact of IUGR and dental caries on
caries among such children is still a matter of public health, together with the lack of information
debate (8). regarding such a relationship in the American
Dental caries is a disease with chronic character- population, the objective of this study was to assess
istics modulated by behavior and involving colon- the association between IUGR surrogates (birth-
ization by Streptococcus mutans (9). Although its weight adjusted for preterm birth, SGA birth, and
prevalence has decreased, dental caries is still the FGR) and preterm birth with dental caries in
most prevalent disease among American children primary teeth, limiting its scope to the population
(10). The biological explanation for the association of the USA.
of IUGR and dental caries is based on the increased
incidence of enamel defects (hypoplasia and/or
hypomineralization) (11–14). These enamel defects
have been attributed to hypocalcemia and to the
Materials and methods
physiological stress associated with perinatal out- This study used as a data source the NHANES III
comes, which would interfere with the secretion (1988–1994) – a cross-sectional survey of a complex,
and maturation of enamel tissue (11, 15, 16). multistage, stratified, clustered sample intended to
Enamel defects would hypothetically predispose be representative of the civilian non-institutional-
to early colonization by S. mutans and to increased ized American population more than 6 months of
formation of dental biofilm (9, 17). age (24). This analysis included singleton White,
Despite the biological plausibility, there is a lack African–American, and Mexican–American chil-
of empirical evidence that IUGR or preterm birth dren 2–5.9 years of age having undergone com-
are associated with dental caries (8). This lack of plete examinations for dental caries and for whom
evidence has primarily been attributed to the fact birth certificates were available. Information on
that most studies have been poorly designed and gestational age and birthweight was taken from
did not adjust for confounders (8). In fact, two birth certificates available in the ‘Natality Data’
recent studies involved adequate adjustment for section of the NHANES III database (25), which
confounders, but neither demonstrated a correla- contains information from the birth certificates of
tion between LBW and dental caries in primary 94% of all children under the age of 6 years.
dentition (18, 19). A significant issue that has not For the study population defined, birth certifi-
been addressed is the imprecision of using LBW as cates were available for 92.7% (n ¼ 3888). Only
a surrogate for IUGR. Using LBW is imprecise singletons were included in our study (n ¼ 3788),
because it can represent both preterm birth and and information on both gestational age and
rate of fetal growth (3). Although these two birthweight was available for 97.8% of those
outcomes may occur concomitantly, they are (n ¼ 3621). To avoid misclassification of IUGR,
known to have different risk factors (2, 20, 21) as we excluded children presenting implausible val-
well as different effects on the health of the child (4, ues for gestational age (>44 weeks; n ¼ 84) (26)
22). In addition, the hypothesis that enamel defects and birth weights that were improbable for the
are increased is based on studies of preterm LBW gestational ages (n ¼ 61) (26, 27). We also exclu-
and very LBW infants rather than on those invol- ded children whose gestational age and birth-
ving LBW infants exclusively (12, 13, 23). There- weight were recorded as 26 weeks (n ¼ 17) and
fore, these studies did not differentiate between 1200 g (n ¼ 17), respectively. These exclusions
IUGR and preterm birth. A recent study, conducted were performed because these values grouped
as part of the Third National Health and Nutri- children with less than 26 weeks and less than
tional Examination Survey (NHANES III, 1988– 1200 g which would result in misclassification of

365
Saraiva et al.

IUGR. Complete dental examination was available intake, and environmental exposure to tobacco
for 3189 children. smoke, as well as breastfeeding. The possible effect
Dental examinations were performed by six modifiers evaluated were fluoride supplement
trained and calibrated examiners at the Medical intake, frequency of dental visits, and urbanization
Examination Center of NHANES III. Dental caries classification (rural or urban residence). Residing in
was defined according to decayed and filled an urban area can be considered a weak but
(primary) teeth (dft) index. Detailed information acceptable surrogate for water fluoridation.
on the examination procedures used in the NHA- Race/ethnicity was categorized as Whites, Afri-
NES III can be found elsewhere (28). We classified can–Americans, and Mexican–Americans. Head of
children with dental caries using two case defini- household education level at the time of interview
tions: (i) children with more than one tooth with was categorized as <12, 12 or >12 years of school-
dental caries and (ii) children with any teeth with ing. Poverty level was defined as the ratio of
dental caries. The reasons for using two case observed annual family income by the poverty
definitions are the following. The first case defini- threshold (>3.5, 1.301–3.5 or <1.301) (31).
tion was used because having a single tooth with Daily intakes of sucrose and carbohydrates were
dental caries does not necessarily indicate that a obtained from a 24-h food-frequency questionnaire.
child is at increased risk of dental caries. A single Carbohydrate intake (g/day) and sucrose intake
tooth with dental caries may represent an incorrect (g/day) were categorized by their median values.
placement of a restoration or a unique lesion Categorization into quartiles did not change our
caused by a specific enamel defect or dental fissure. results. Environmental exposure to tobacco smoke,
Therefore, using a case definition of at least two which has been associated with dental caries (32),
teeth with dental caries will avoid misclassification. was measured by means of a proxy by determining
The second definition was used for comparison the number of smokers who smoke inside the home
with other studies. where the child lives (0, 1–2 or >2). A proxy was
We identified IUGR by assessing LBW adjusted used because data on serum cotinine levels, a more
for preterm birth, SGA birth, and FGR. We defined objective measure of environmental tobacco smoke
LBW as <2500 g, and preterm birth was defined as exposure (32), were not available for the age range
being born at <37 weeks of gestational age. SGA we evaluated in this study. Breastfeeding was also
birth was calculated by using the reference values considered as a possible confounder because it has
for the 10th percentile of birthweight reference data been associated with dental caries (19) and because
described by Alexander et al. (29). As those authors preterm infants are less likely to be breastfed
did not provide information on the 50th percentile because of difficulty in suction (35). The breast-
of growth distribution, the FGR – measured feeding variable was categorized according to
through the fetal growth ratio, which describes Schulman (19) as bottle-feeding until 19 months
the relationship between birthweight and the 50th and bottle-feeding after 19 months. Fluoride sup-
percentile of birthweight for each gestational age plementation was defined as the intake of any
from reference data (32) – was calculated using the vitamin supplement containing fluoride or pre-
reference data presented by Zhang and Bowes (27). scription fluoride within the preceding month,
Based on the calculation of the fetal growth ratio, either or both having been ingested regularly for
FGR is classified as none (0.85+), mild (0.85–0.80), more than 3 months. The frequency of dental visits
moderate (0.79–0.75), or severe (<0.75) (30). was categorized as never having been to a dentist,
going to a dentist only when necessary, or visiting
Covariates a dentist on a regular basis (every year or every
Covariate information was taken from question- 2 years). We also included in the analysis the
naires completed during interviews of parents or census region of the country (northeast, central-
guardians and from dental examination records. west, south or west) and urbanization classification
Covariates included known determinants of dental (urban or rural).
caries and possible confounders, as well as possible
effect modifiers. Known determinants of dental Statistical analysis
caries were race/ethnicity and age, as well as One of our concerns about Schulman’s study (19)
socioeconomic variables such as poverty level and was that inclusion of 2-year-old children would
education of head of household. Possible con- have biased his results toward the null. The bias
founders included sucrose intake, carbohydrate would happen because recently erupted teeth and

366
 IUGR, preterm and dental caries

teeth still erupting would have a very short period caries. For this age group, children had on average
of time at risk to develop the disease. Therefore, we 1.0 (SE 0.1) tooth and 1.7 surfaces (SE 0.2) with
performed a sensitivity analysis including and dental caries. There were 2341 3- to 6-year-old
excluding 2-year-old children, and with the two children with a prevalence of dental caries of 25.9%
dental caries definitions. (SE 1.9). The average number of teeth affected by
All analysis was performed for each IUGR dental caries was 1.5 with surface average of 2.2 (SE
surrogate separately. Bivariate and stratified ana- 0.3).
lyses were followed by Poisson regression mode- Background characteristics of the study popula-
ling of PRs in order to assess the association tion are described in Table 1. Bivariate analysis for
between each IUGR surrogate and dental caries. dental caries, including both classifications of
The PR is the correct measurement of association dental caries, is shown in Tables 2 and 3 for 2–5-
for a cross-sectional survey. In particular, for and 3–5-year-old children, respectively. Independ-
common events, such as the prevalence of dental ent of the age group or dental caries definition,
caries in this study (from 15.4% to 21.17%, depend- these tables show that preterm birth was positively
ing on the case definition and the study population and statistically associated with dental caries.
included), logistic regression leads to an overesti- Moreover, low birthweight (LBW) although posi-
mation of the measure of association provided tively associated with dental caries did not reach
(odds ratio) compared with the desired PR. Such statistical significance. SGA and FGR, although not
overestimation can also interfere with the correct statistically significant, tended to be negatively
modeling process because the selection of con- associated with dental caries. Moreover, including
founders is based on the effect size (34). We also 2-year-old children and use of a more loose
conducted an analysis of the dft index as a definition of dental caries can lead the point
continuous variable using negative binomial estimators toward the null.
regression. However, as the findings were very Table 4 shows the final model for the analysis of
similar, we decided to present only the estimation preterm birth and LBW. After adjusting for major
of PR, which is more easily understood. Poisson confounders, independent of inclusion of 2-year-
regression models were built using a stepwise old children, and independent of the dental caries
procedure with backward elimination of variables. definition, preterm birth was statistically signifi-
All analyses were performed using the ‘survey cantly associated with dental caries. An important
procedure’ command in the stata 8.0 program observation is that LBW reversed its association
(35), thereby taking the sampling design into when adjusted by preterm, becoming negatively
account. associated with dental caries. However, this associ-
All covariates associated with dental caries at ation was not statistically significant.
values of P £ 0.2 were included in the models. We To ensure that multicollinearity between preterm
also included known confounders and possible birth and IUGR surrogates did not interfere with
effect modifiers, even those not reaching the 0.2 our results, we conducted several analyses. Adding
cut-off point. Known risk factors for dental caries preterm birth to the models did not lead to inflation
and confounders for IUGR surrogates or preterm of P-values or to a considerable change in the
that changed coefficients more than 10% were kept confidence intervals, which would provide evi-
in the models. Plausible interactions were also dence of multicollinearity.
tested: race/ethnicity, education of head of house- Table 5 shows the models for SGA. In these
hold, dental visits, use of bottle-feeding after models preterm birth was only significant when
19 months, poverty, and fluoride supplement excluding 2-year-old children. However, we can
intake. Additional models including FGR as a observe that the point estimators for preterm birth
continuous explanatory variable were also carried increase as the case definition gets more stringent
out in order to test for linear trend. and with the exclusion of 2-year-old children. SGA
maintains its trend of negative association with
dental caries. A similar analysis for FGR is shown
in Table 6 with similar pattern of results. There is a
Results tendency that moderate and severe FGR are neg-
The final analytical sample consisted of 3189 atively associated with dental caries compared
children 2–5.9 years old (which we will refer to with no and mild FGR. As mild FGR is considered
5 years old), 21.2% (SE 1.5) of them with dental a borderline category near normality, and because

367
Saraiva et al.

Table 1. Background characteristics of children aged 3–5 years (n ¼ 3189), NHANES III (1988–1994)
2- to 5-year olds (n ¼ 3189) 3- to 5-year olds (n ¼ 2341)
a b c
Characteristics n % SE na %b SEc
Race/ethnicity
White 1031 72.6 1.4 720 73.1 1.5
African–American 1083 17.7 1.3 815 17.6 1.4
Mexican–American 1075 9.6 0.9 806 9.3 0.9
Poverty ratio
>3.500 326 19.3 1.6 224 18.3 1.7
1.301–3.500 764 34.6 1.6 896 36.1 1.8
<1.301 1867 46.2 1.6 1058 45.6 1.8
Head of household education (years)
>12 945 43.4 1.8 670 42.0 2.0
¼ 12 1053 35.0 1.3 783 36.0 1.4
<12 1086 21.7 1.1 812 22.0 1.3
Number of smokers in the home
None 1976 61.2 2.1 1438 61.8 2.1
1–2 790 22.2 1.6 600 22.8 1.6
>2 421 16.6 1.5 305 15.4 1.5
Dental visits
Never 1867 52.0 1.5 1142 41.4 1.9
As needed 349 8.4 0.8 309 9.7 0.9
Regularly 953 39.6 1.5 881 48.9 1.9
Fluoride supplementation
No 3123 96.4 1.3 2289 96.9 1.5
Yes 66 3.6 1.3 52 3.9 1.5
Sucrose intake (g/day)
<35.0 1451 47.7 1.6 963 44.4 1.8
‡35.0 1416 52.3 1.6 1119 55.6 1.8
Breast-feeding
No 1777 46.1 1.7 1325 46.0 2.0
Yes 1411 53.9 1.7 1018 54.0 2.0
Bottle-feeding after 19 months
No 2157 72.2 1.5 1614 73.3 1.5
Yes 1032 27.8 1.5 727 26.7 1.5
Carbohydrate intake (g/day)
<161.0 838 25.1 1.6 535 22.5 1.8
161.0–202.2 672 25.1 1.2 458 22.9 1.4
202.3–249.1 640 24.7 1.2 493 26.6 1.4
‡249.2 717 25.1 1.4 596 27.9 1.5
Dental caries (dft > 0)
No 2284 78.8 1.5 1529 74.1 1.9
Yes 907 21.2 1.5 815 25.9 1.9
Dental caries (dft > 1)
No 2502 84.5 1.4 1725 81.2 1.8
Yes 689 15.5 1.4 619 18.9 1.8
Preterm birth (weeks)
‡37 2914 93.4 0.7 2137 93.5 0.8
<37 275 6.6 0.7 204 6.5 0.8
Low birthweight (g)
‡2500 2997 94.8 0.6 2198 94.5 0.7
<2500 192 5.3 0.6 143 5.5 0.7
Small for gestational age birth
No 2747 88.2 1.0 2010 87.7 1.2
Yes 442 11.8 1.0 331 12.3 1.2
Fetal growth restriction
None 2751 87.7 0.9 2003 86.7 1.1
Mild 179 5.0 0.6 136 5.3 0.7
Moderate 122 3.6 0.5 97 4.1 0.7
Severe 137 3.7 0.5 105 3.9 0.6
a
Total number might vary due to missing values for independent variables.
b
Weighted percentage.
c
SE, standard error.

368
 IUGR, preterm and dental caries

Table 2. Bivariate analysis of background variables, oral health-related characteristics, IUGR surrogates and preterm
birth in relation to dental caries among children aged 2–5 years NHANES III (1988–1994)
dft > 0a dft > 1
b c d e
% SE PR 95% CI P-value %b SEc PRd 95% CIe P-value
Age (years)
2 7.2 0.8 1.00 <0.01 4.8 0.7 1.00 <0.01
3 12.6 1.8 1.99 1.32–3.00 9.4 1.5 1.93 1.21–3.09
4 27.9 2.6 4.42 3.19–6.13 18.4 1.9 3.80 2.59–5.57
5 37.2 3.2 5.90 4.25–8.18 28.6 3.7 5.91 3.90–8.96
Race/ethnicity
White 17.2 1.8 1.00 <0.01 12.2 1.7 1.00 <0.01
African–American 28.1 2.1 1.63 1.27–2.09 21.5 1.9 1.76 1.30–2.37
Mexican–American 38.3 2.4 2.22 1.74–2.84 29.2 1.8 2.39 1.78–3.20
Poverty ratio
>3.500 8.1 2.0 1.00 <0.01 5.7 1.8 1.00 <0.01
1.301–3.500 17.9 1.6 2.22 1.42–3.47 11.0 1.5 1.92 1.12–3.30
<1.301 28.9 25 3.59 2.14–6.01 22.7 2.4 3.96 2.11–7.45
Maternal age at birth (years)
<20 29.9 3.5 1.00 <0.01 21.4 2.9 1.00 <0.01
20–29 22.2 1.7 0.74 0.57–0.97 16.6 1.6 0.78 0.57–1.05
>29 15.5 2.0 0.52 0.40–0.67 10.7 1.7 0.50 0.36–0.69
Head of household education (years)
>12 15.2 1.7 1.00 <0.01 9.5 1.4 1.00 <0.01
¼ 12 21.7 2.1 1.43 1.11–1.85 16.3 1.9 1.72 1.25–2.37
<12 31.7 2.7 2.09 1.68–2.59 25.5 2.9 2.68 1.97–3.65
Number of smokers in the home
None 17.6 1.3 1.00 <0.01 12.2 1.0 1.00 <0.01
1–2 28.6 3.2 1.62 1.32–2.00 22.7 3.3 1.86 1.43–2.40
>2 24.6 4.1 1.40 1.10–1.78 18.0 2.6 1.47 1.06–2.04
Dental visits
Never 14.2 1.1 1.00 <0.01 9.7 0.9 1.00 <0.01
As needed 36.0 4.3 2.53 1.93–3.31 27.7 3.6 2.87 2.12–3.88
Regularly 26.8 2.8 1.88 1.51–2.34 20.4 2.6 2.11 1.67–2.66
Fluoride supplementation
No 21.0 1.3 1.00 0.52 15.4 1.3 1.00 <0.57
Yes 26.9 10.6 1.28 0.59–2.77 19.1 7.5 1.24 0.58–2.67
Carbohydrate intake (g/day)
<161.0 17.7 2.2 1.00 0.03 14.0 2.0 1.00 0.2319
161.0–202.2 15.5 1.8 0.88 0.66–1.18 11.7 2.1 0.79 0.55–1.14
202.3–249.1 23.5 2.8 1.33 1.01–1.75 17.5 2.5 1.08 0.78–1.51
‡249.2 25.3 2.4 1.44 1.05–1.96 17.0 2.0 1.02 0.69–1.51
Sucrose intake
<35 g/day 15.6 1.3 1.00 <0.01 12.7 1.3 1.00 0.06
‡35 g/day 24.9 2.7 1.60 1.23–2.08 17.2 2.3 1.35 0.98–1.87
Bottle-feeding
£19 months 19.5 1.2 1.00 0.02 14.0 1.0 1.00 0.02
>19 months 25.4 2.9 1.30 1.05–1.61 19.4 2.9 1.39 1.06–1.82
Breast-feeding
No 25.6 1.8 1.00 <0.01 12.3 1.6 1.00 <0.01
Yes 17.4 1.9 0.68 0.54–0.86 19.3 1.5 0.64 0.51–0.80
Preterm birth
‡37 weeks 20.3 1.3 1.00 <0.01 14.7 1.3 1.00 0.01
<37 weeks 33.5 6.8 1.65 1.14–2.40 26.3 5.9 1.78 1.14–2.78
Low birthweight
‡2500 g 20.8 1.5 1.00 0.09 15.3 1.4 1.00 0.12
<2500 g 28.4 5.2 1.37 0.95–1.97 22.0 4.8 1.46 0.90–2.35
Small for gestational age birth
No 21.6 1.5 1.00 0.96 15.6 1.4 1.00 0.82
Yes 21.0 2.7 0.99 0.76–1.29 15.1 2.3 0.97 0.74–1.28

369
Saraiva et al.

Table 2. Continued
dft > 0a dft > 1
b c d e
% SE PR 95% CI P-value %b SEc PRd 95% CIe P-value
Fetal growth restriction
None 20.8 1.5 1.00 0.66 15.4 1.3 1.00 0.66
Mild 27.5 5.9 1.32 0.84–2.07 18.4 5.2 1.19 0.70–2.03
Moderate 22.4 6.5 1.06 0.58–1.94 17.3 3.3 1.12 0.76–1.65
Severe 19.7 3.5 0.95 0.67–1.34 11.8 4.3 0.77 0.37–1.59
a
Dft, decayed and filled (primary) teeth index.
b
Weighted percentage.
c
SE, standard error.
d
PR, prevalence ratio.
e
95% CI, 95% confidence interval.

Table 3. Bivariate analysis of background variables, oral health-related characteristics, IUGR surrogates and preterm
birth in relation to dental caries among children aged 3–5 years NHANES III (1988–1994)
dft > 0a dft > 1
b c d e
% SE PR 95% CI P-value %b SEc PRd 95% CIe P-value
Age (years)
3 12.6 1.8 1.00 <0.01 9.4 1.5 1.00 <0.01
4 27.9 2.6 2.22 1.58–3.12 18.4 1.9 1.97 1.37–2.83
5 37.2 3.2 2.96 2.22–3.95 28.6 3.7 3.06 2.18–4.30
Race/ethnicity
White 21.4 2.4 1.00 <0.01 15.1 2.3 1.00 <0.01
African–American 34.1 2.4 1.60 1.23–2.08 25.6 2.3 1.71 1.24–2.36
Mexican–American 45.8 2.9 2.14 1.64–2.80 35.6 2.2 2.36 1.72–3.25
Poverty ratio
>3.500 10.3 2.7 1.00 <0.01 7.7 2.5 1.00 0.01
1.301–3.500 22.2 2.0 2.03 1.29–3.19 16.0 1.6 2.07 1.16–3.70
<1.301 34.6 3.2 3.17 1.87–5.36 29.2 3.9 3.79 1.94–7.40
Maternal age at birth (years)
<20 35.1 4.4 1.00 <0.01 25.1 3.5 1.00 0.01
20–29 27.3 2.3 0.78 0.59–1.03 20.3 2.2 0.81 0.59–1.10
>29 19.2 2.5 0.55 0.41–0.74 13.5 2.2 0.54 0.39–0.75
Head of household education (years)
>12 19.9 2.3 1.00 0.01 12.4 2.0 1.00 0.01
¼ 12 25.9 2.6 1.30 1.01–1.68 19.2 2.4 1.54 1.12–2.13
<12 36.5 3.6 1.83 1.46–2.31 29.7 3.8 2.38 1.70–3.35
Number of smokers in the home
None 21.7 1.7 1.00 <0.01 15.0 1.4 1.00 <0.01
1–2 34.9 4.0 1.61 1.31–1.98 27.6 4.3 1.84 1.41–2.41
>2 29.3 3.8 1.35 1.02–1.80 21.3 3.6 1.42 0.99–2.04
Dental visits
Never 19.7 1.6 1.00 <0.01 13.2 1.3 1.00 <0.01
As needed 38.8 4.7 1.96 1.51–2.55 29.6 3.9 2.25 1.66–3.06
Regularly 28.2 3.1 1.43 1.13–1.81 21.5 2.8 1.64 1.28–2.09
Fluoride supplementation
No 25.6 1.7 1.00 0.60 18.7 1.7 1.00 0.71
Yes 31.9 12.8 1.25 0.57–2.70 22.9 8.9 0.82 0.28–2.37
Carbohydrate intake (g/day)
<161.0 23.4 3.3 1.00 0.19 18.8 3.1 1.00 0.36
161.0–202.2 19.7 2.5 0.84 0.64–1.12 14.8 2.6 0.79 0.55–1.14
202.3–249.1 27.3 3.3 1.17 0.87–1.58 20.3 3.0 1.08 0.78–1.51
‡249.2 28.5 2.5 1.22 0.89–1.67 19.6 2.3 1.02 0.69–1.51
Sucrose intake
<35 g/day 19.6 1.9 1.00 <0.01 11.0 1.6 1.00 0.19
‡35 g/day 29.4 3.2 1.50 1.14–1.98 13.1 1.7 1.27 0.89–1.80

370
 IUGR, preterm and dental caries

Table 3. Continued
dft > 0a dft > 1
b c d e
% SE PR 95% CI P-value %b SEc PRd 95% CIe P-value
Bottle-feeding
£19 months 24.0 1.6 1.00 0.02 17.0 1.5 1.00 0.02
>19 months 31.1 3.8 1.29 1.04–1.62 24.0 3.9 1.42 1.07–1.88
Breast-feeding
No 30.6 2.3 1.00 <0.01 23.0 2.0 1.00 <0.01
Yes 21.9 2.5 0.71 0.57–0.90 15.3 2.1 0.67 0.54–0.83
Preterm birth
‡37 weeks 24.7 1.7 1.00 <0.01 17.9 1.7 1.00 0.06
<37 weeks 42.4 8.3 1.72 1.22–2.42 32.9 7.4 1.84 1.19–2.83
Low birthweight
‡2500 g 25.5 1.9 1.00 0.23 18.6 1.9 1.00 0.26
<2500 g 32.1 6.5 1.26 0.86–1.84 24.2 5.4 1.30 0.82–2.07
Small for gestational age birth
No 26.2 2.0 1.00 0.55 19.2 1.8 1.00 0.35
Yes 23.9 3.6 0.92 0.68–1.23 16.5 3.1 0.86 0.63–1.18
Fetal growth restriction
None 25.7 2.0 1.00 0.43 19.2 1.8 1.00 0.55
Mild 33.7 6.5 1.31 0.88–1.95 22.4 6.3 1.18 0.70–1.98
Moderate 24.3 7.5 0.95 0.50–1.80 16.9 4.2 0.89 0.58–1.37
Severe 19.9 4.7 0.77 0.51–1.18 12.4 4.9 0.65 0.29–1.44
a
dft, decayed and filled (primary) teeth index.
b
Weighted percentage.
c
SE, standard error.
d
PR, prevalence ratio.
e
95% CI, 95% confidence interval.

the sample size for moderate and severe FGR are old children and a more loose definition of dental
small we also performed the analysis collapsing no caries seem to bias the results toward the null. An
and mild category and the moderate and severe explanation for these biases would be a systematic
categories that we call low FGR and high FGR. For selection bias when excluding 2-year-old children
models including 2- to 5-year-old children, the PRs or a directional misclassification bias by using a
comparing low FGR with high FGR were 0.65 (95% more stringent case definition. However, we have
CI 0.34–1.27) for the presence of any tooth with no reason to believe that excluding 2-year-old
dental caries and 0.78 (95% CI 0.57–1.09) for case children would result in selection bias, nor that
definition of at least two teeth with dental caries; using a more stringent case definition would create
the PR values were 0.59 (95% CI 0.60–0.28) and 0.40 a directional misclassification. The inclusion of 2-
(95% CI 0.23–0.81; P ¼ 0.0102), for 3- to 5-year-old year-old children represents an entrance 1/4 of the
children, respectively. study sample with a very short period of time at
risk because of their recent erupted teeth and very
low prevalence of dental caries. A more loose
definition for dental caries and inclusion of very
Discussion young children may be the reason why no statis-
In the present study, dental caries was associated tical association was found in other studies.
positively with preterm birth but tends to be As the negative association between dental caries
negatively associated with IUGR. The positive and IUGR suggested in this study is counterintu-
association between dental caries and preterm itive, we should interpret this finding with caution.
birth was significant independent of the inclusion It is possible that this negative association is the
of 2-year-old children or the case definition. A result of chance, being a particular characteristic of
similar independence of results could be observed this data set. This negative association does not
for the negative association, although not always seem to be a result of any interaction, at least with
statistically significant, between dental caries and the information available from NHANES III. In
SGA or FGR. As we predicted, inclusion of 2-year- addition, the imprecision of the specific reference

371
Saraiva et al.

Table 4. Models for dental caries according to low birthweight and preterm birth among children aged 2–5 years,
NHANES III (1988–1994)
2- to 5-year olds 3- to 5-year olds
a
dft > 0 dft > 1 dft > 0 dft > 1
PRb 95% CIc PRb 95% CIc PRb 95% CIc PRb 95% CIc
Race/ethnicity
White 1.00 1.00 1.00 1.00
African–American 1.58 1.23–2.03 1.52 1.14–2.04 1.60 1.21–2.13 1.54 1.11–2.13
Mexican–American 1.86 1.29–2.67 1.75 1.18–2.60 1.86 1.25–2.77 1.75 1.15–2.65
Poverty ratio
>3.500 1.00 1.00 1.00 1.00
1.301–3.500 1.83 1.10–3.03 1.58 0.88–2.84 1.82 1.09–3.03 1.62 0.89–2.96
<1.301 2.61 1.38–4.94 2.81 1.35–5.83 2.44 1.28–4.64 2.71 1.28–5.71
Number of smokers in the home
None 1.00 1.00 1.00 1.00
1–2 1.42 1.13–1.78 1.69 1.29–2.20 1.42 1.12–1.81 1.68 1.25–2.26
>2 1.39 1.02–1.89 1.65 1.13–2.40 1.34 0.96–1.89 1.60 1.08–2.39
Dental visits
Never 1.00 1.00 1.00 1.00
As needed 1.50 1.15–1.94 1.79 1.32–2.43 1.44 1.10–1.87 1.70 1.24–2.33
Regularly 1.63 1.31–2.03 2.16 1.74–2.69 1.60 1.26–2.03 2.14 1.71–2.68
Bottle-feeding
£19 months 1.00 1.00 1.00 1.00
>19 months 1.44 1.21–1.71 1.52 1.22–1.90 1.41 1.19–1.68 1.47 1.19–1.82
Low birthweight
No 1.00 1.00 1.00 1.00
Yes 0.79 0.53–1.19 0.78 0.45–1.04 0.71 0.51–0.98 0.67 0.44–1.02
Preterm birth
No 1.00 1.00 1.00 1.00
Yes 1.38 1.00–1.89 1.49 1.01–2.19 1.52 1.18–1.94 1.64 1.22–2.20
Note: Models were also adjusted for child’s sex and age, maternal age at birth, fluoride supplementation, carbohydrate
intake, and head of household education.
a
dft, decayed and filled (primary) teeth index.
b
PR, prevalence ratio.
c
95% CI, 95 percent confidence interval.

data for estimating SGA birth does not seem to be with the occurrence of IUGR, indicating that it
responsible for our results because independent of could not be a confounder. As IUGR is primarily
the three reference data used (27, 29, 36), the results associated with low socioeconomic level, it seems
were very similar (data not shown). The results unreasonable that the poorest communities in the
were also similar when low FGR, which has been US would be more likely to have water fluoridation
proposed to be a better proxy for IUGR than is SGA than the others. If so, underprivileged groups also
birth (30), was used. Furthermore, when we tested would tend to have less dental caries than the
the interaction between LBW and preterm birth, we richest ones, which is contrary to what has been
observed that children who were not born preterm observed (37). However, if there is any socioeco-
but presented LBW (possibly because of IUGR) nomic pattern associated with water fluoridation in
were less likely to have dental caries than the the US, we suppose this could at least be in part
others. Additional support for our conclusion is controlled by the fact that we took ethnicity and
found in the trend analysis performed for FGR. The certain socioeconomic variables into consideration.
higher the severity of the FGR, the lower was the Another potential limitation of our study is the
risk of dental caries, suggesting a dose–response lack of information on serum levels of cotinine (32)
gradient. and blood lead levels (38), both of which have been
One limitation of our study is the lack of associated with dental caries. Blood lead levels
information on oral hygiene and water fluorid- were only available for a small part of our sample.
ation. However, we could not find evidence in the The biological explanation for these associations
literature that water fluoridation was associated relies on impairment of enamel deposition and

372
 IUGR, preterm and dental caries

Table 5. Models for dental caries according small for gestational age (SGA) birth among children aged 2–5 years,
NHANES III (1988–1994)
2- to 5-year olds 3- to 5-year olds
a
dft > 0 dft > 1 dft > 0 dft > 1
PRb 95% CIc PR 95% CI PR 95% CI PR 95% CI
Race/ethnicity
White 1.00 1.00 1.00 1.00
African–American 1.57 1.31–2.02 1.51 1.13–2.01 1.59 1.20–2.10 1.55 1.12–2.15
Mexican-American 1.88 1.31–2.68 1.77 1.20–2.62 1.88 1.28–2.78 1.77 1.17–2.66
Poverty Ratio
>3.500 1.00 1.00 1.00 1.00
1.301–3.500 1.83 1.10–3.05 1.58 0.88–2.85 1.82 1.08–3.06 1.62 0.88–2.95
<1.301 2.63 1.38–5.00 2.83 1.35–5.92 2.47 1.29–4.74 2.75 1.29–5.87
Number of smokers in the home
None 1.00 1.00 1.00 1.00
1–2 1.43 1.14–1.79 1.70 1.31–2.22 1.44 1.13–1.83 1.69 1.26–2.28
>2 1.42 1.05–1.94 1.69 1.16–2.46 1.38 0.98–1.95 1.67 1.12–2.49
Dental visits
Never 1.00 1.00 1.00 1.00
As needed 1.48 1.14–1.91 1.76 1.30–2.38 1.41 1.08–1.84 1.68 1.24–2.28
Regularly 1.61 1.30–2.00 2.13 1.72–2.64 1.57 1.24–1.98 2.10 1.70–2.59
Bottle-feeding
£19 months 1.00 1.00 1.00
>19 months 1.42 1.20–1.69 1.50 1.22–1.86 1.39 1.17–1.65 1.47 1.19–1.81
SGA birth
No 1.00 1.00 1.00 1.00
Yes 0.79 0.56–1.01 0.72 0.52–0.99 0.73 0.50–1.01 0.66 0.33–0.96
Preterm birth
No 1.00 1.00 1.00 1.00
Yes 1.25 0.97–1.62 1.34 0.98–1.84 1.32 1.03–1.69 1.38 1.03–1.85
Note: models were also adjusted for child’s sex and age, maternal age at birth, fluoride supplementation, carbohydrate
intake, and head of household education.
a
dft, decayed and filled (primary) teeth index.
b
PR, prevalence ratio.
c
95% CI, 95 percent confidence interval.

maturation (32, 38). However, only smoking has not find any major differences between the original
been associated with perinatal outcomes, and there models and the reconstructed models. Therefore,
is little evidence of a causal association between we decided to present the original models that
lead exposure and perinatal outcomes (39–41) to be included the variables with missing values.
considered a confounder for dental caries. In order An explanation for the negative association
to control for environmental exposure to tobacco between IUGR and dental caries is the possible
smoke, we used the number of smokers in the delay in tooth eruption, which has been reported
home. among preterm LBW children (15, 42). Unfortu-
An important concern in any NHANES III data nately, we were not able to find studies dissociating
analysis is the possibility of missing values for the effect of IUGR from preterm birth on tooth
covariates, which results in exclusion of the indi- eruption. Delay in tooth eruption of primary teeth
vidual from the multivariate analysis. In an attempt would delay colonization with S. mutans, resulting
to understand the possible impact of the missing in lower prevalence of dental caries. It can also be
values in our final models, the covariates with hypothesized that increased utilization of antibiot-
missing information were recoded to include a ics contributes to the lower prevalence of dental
category discriminating the missing values. We then caries among growth-retarded children as pulmon-
reconstructed the models, each time replacing one ary infections in the first years of life are more
variable with missing values with the correspond- severe in such children (22, 43). Antibiotic use is
ing recoded variable as well as using simultaneous thought to reduce dental caries by inhibiting
substitution. After exhaustive analysis, we could colonization by cariogenic bacteria (44).

373
Saraiva et al.

Table 6. Models for dental caries according the fetal growth restriction (FGR) among children aged 2–5 years, NHANES
III (1988–1994)
2- to 5-year olds 3- to 5-year olds
a
dft > 0 dft > 1 dft > 0 dft > 1
PRb 95% CIc PR 95% CI PR 95% CI PR 95% CI
Race/ethnicity
White 1.00 1.00 1.00 1.00
African-American 1.59 1.24–2.04 1.53 1.15–2.04 1.61 1.22–2.14 1.57 1.13–2.19
Mexican-American 1.86 1.30–2.68 1.77 1.20–2.62 1.87 1.26–2.78 1.80 1.18–2.75
Poverty Ratio
>3.500 1.00 1.00 1.00 1.00
1.301–3.500 1.83 1.10–3.05 1.56 0.86–2.81 1.81 1.07–3.06 1.59 0.86–2.92
<1.301 2.63 1.38–5.00 2.79 1.34–5.80 2.45 1.27–4.73 2.69 1.27–5.67
Number of smokers in the home
None 1.00 1.00 1.00 1.00
1–2 1.43 1.14–1.79 1.69 1.30–2.20 1.44 1.13–1.82 1.69 1.27–2.25
>2 1.42 1.04–1.93 1.68 1.15–2.44 1.37 0.97–1.94 1.66 1.11–2.49
Dental visits
Never 1.00 1.00 1.00 1.00
As needed 1.49 1.15–1.92 1.77 1.31–2.39 1.43 1.10–1.85 1.68 1.24–2.29
Regularly 1.62 1.31–2.01 2.14 1.73–2.66 1.59 1.26–2.00 2.13 1.68–2.69
Bottle-feeding
£19 months 1.00 1.00 1.00 1.00
>19 months 1.43 1.20–1.69 1.51 1.22–1.86 1.39 1.17–1.65 1.51 1.24–1.84
FGR
None 1.00 1.00 1.00 1.00
Mild 0.98 0.71–1.36 1.06 0.74–1.52 1.04 0.75–1.44 1.10 0.76–1.58
Moderate 0.94 0.50–1.75 0.71 0.31–1.62 0.92 0.47–1.78 0.66 0.26–1.67
Severe 0.70 0.48–1.01 0.74 0.49–1.12 0.58 0.38–0.90 0.59 0.36–0.99
Preterm birth
No 1.00 1.00 1.00 1.00
Yes 1.27 0.97–1.66 1.37 1.00–1.89 1.34 1.03–1.74 1.43 1.05–1.94
Note: models were also adjusted for child’s sex and age, maternal age at birth, fluoride supplementation, carbohydrate
intake, and head of household education.
a
dft, decayed and filled (primary) teeth index.
b
PR, prevalence ratio.
c
95% CI, 95 percent confidence interval.

We can further hypothesize that the well-known find IUGR to correlate with low sucrose, carbohy-
increased lung maturation among IUGR fetuses drate intake or frequency of dental visits. Having
compared with preterm children is analogous to been breastfed is in fact associated with neonatal
dental development (45, 46). If a similar increased conditions as most preterm children have difficul-
maturation also affects tooth germs, IUGR children ties in suction and are thus introduced to exclusive
might present enamel structure that is more nearly bottle-feeding very early in life (47). Nevertheless,
normal in comparison with that seen in preterm the duration of bottle-feeding was no greater
children. among such children.
Unfortunately, we could not find any human or The study of the association of dental caries and
animal studies of tooth development in which the perinatal outcomes is challenging because of the
effect of IUGR was dissociated from preterm birth. presence of several confounders, especially those
Therefore, it is possible that our results are attrib- associated with socioeconomic status and health
utable to a sum of events, such as delay in tooth behaviors. These confounders are very difficult to
eruption together with increased antibiotic intake measure, in particular because they rely on
and accelerated enamel maturation among IUGR personal information and change over time. In
children. addition, there are common risk factors for both
Another possible explanation for our results is conditions, related to maternal behaviors (e.g.
the improvement in positive health behaviors smoking), prenatal care and socioeconomic status,
among IUGR children. However, we could not which may reveal, in a broad sense, negative

374
 IUGR, preterm and dental caries

maternal health behaviors (18, 48–50). Therefore, it 4. Steffensen F, Sorensen H, Gilman M, Rothman RJ,
is possible that IUGR and preterm birth are only Sabroe S, Fisher D, et al. Low birthweight and
preterm delivery as risk factors for asthma and atopic
markers for negative health behaviors commonly dermatitis in young adult males. Epidemiology
associated with dental caries. 2000;11:185–8.
We found no data in the literature regarding the 5. Brooks A, Byrd R, Weitzman M, Auinger P, McBride
independent effect of IUGR on dental caries with J. Impact of low birthweight on early childhood
asthma in United States. Arch Pediatr Adolesc Med
which we could compare our results. Schulman
2001;155:401–6.
(19) recently conducted an analysis of preterm birth 6. Thompson J, Carter R, Edwards A, et al. A popula-
using NHANES III data but found no association tion-based study of the effects of birthweight on early
between preterm birth and dental caries. Despite developmental delay or disability children. Am J
Perinatol 2003;20:321–32.
the fact that the same data set was used, there are
7. Mzayek F, Sherwin R, Fonseca V, et al. Differential
some differences in our analysis that might explain association of birthweight with cardiovascular risk
the contradictory findings. For example, the classi- variables in African-Americans and Whites: the
fication of gestational age in four categories (>44, Bogalusa heart study. Ann Epidemiol 2004;14:258–
44–37, 33–36, and <33 weeks), with those with 64.
8. Burt BA, Pai S. Does low birthweight increase the
<33 weeks with a very small sample size. In spite risk of caries? A systematic review. J Dental Educ
of the small sample size, the bivariate analysis 2001;65:1024–7.
clearly showed an increased prevalence of dental 9. Seow WK. Biological mechanisms of early childhood
caries among preterm children. Moreover, whereas caries. Community Dent Oral Epidemiol 1998;26:8–
27.
Schulman (19) modeled decayed and filled surfaces 10. Brown L, Wall T, Lazar V. Trends in total caries
as a continuous variable using Poisson regression, experience: permanent and primary teeth. J Am
we estimated PR. Dental Assoc 2000;131:223–31.
In spite of some limitations such as the cross- 11. Seow WK, Humphrys C, Tudehope DI. Increased
prevalence of developmental dental defects in low
sectional design, this study has several strengths. birth-weight, prematurely born children: a controlled
One advantage was the large sample size and the study. Pediatr Dent 1987;9:221–5.
availability of data on the most significant risk 12. Aine L, Backstrom MC, Maki R, et al. Enamel
factors for dental caries, on gestational age and on defects in primary and permanent teeth of chil-
dren born prematurely. J Oral Pathol Med
the most significant confounders. In addition, data
2000;29:403–9.
on LBW taken from birth certificates avoid infor- 13. Fearne JM, Bryan EM, Elliman AM, Brook AH,
mation bias related to the faulty recollections of Williams DM. Enamel defects in the primary denti-
parents or guardians. Furthermore, this is the first tion of children born weighing less than 2000 g. Br
Dental J 1990;168:433–7.
study to address the various effects of IUGR and
14. Grahnén H, Sjolin S, Stentronm A. Mineralization
preterm birth on dental caries. defects of primary teeth in children born preterm.
In conclusion, we found a positive association Scand J Dent Res 1974;82:396–400.
between preterm birth and dental caries, and a 15. Seow WK. Effects of preterm birth on oral growth
negative association between IUGR and dental and development. Austr Dental J 1997;42:85–91.
16. Noren JG. Enamel structure in deciduous teeth from
caries. As preterm delivery has been increasing in low-birth-weight infants. Acta Odontol Scand
the US, further studies are needed in order to 1983;41:355–62.
confirm and better understand the associations 17. Li Y, Navia JM, Caufield PW. Colonization by
identified in this study. mutans streptococci in the mouths of 3- and 4-year-
old Chinese children with or without enamel
hypoplasia. Arch Oral Biol 1994;39:1057–62.
18. Peres M, Latorre M, Sheiham A, et al. Social and
biological early life influences on severity of dental
References caries in children aged 6 years. Community Dent
1. Demissie K, Rhoads GG, Ananth CV, Alexander GR, Oral Epidemiol 2005;33:53–63.
Kramer MS, Kogan MD, et al. Trends in preterm 19. Shulman JD. Is there association between low birth-
birth and neonatal mortality among blacks and weight and dental caries in the primary dentition?
whites in the United States from 1989 to 1997. Am J Caries Res 2005;39:161–7.
Epidemiol 2001;154:307–15. 20. Kramer MS. Socioeconomic determinants of
2. Kramer MS. Intrauterine growth and gestational intrauterine growth retardation. Eur J Clin Nutr
duration determinants. Pediatrics 1987;80:502–11. 1998;52:S29–32.
3. Kramer MS. Determinants of low birthweight: meth- 21. Kramer MS, Platt R, Yang H, McNamara H, Usher
odological assessment and meta-analysis. Bull WHO RH. Are all growth-restricted newborns created
1987;65:663–737. equal(ly)? Pediatrics 1999;103:599–602.

375
Saraiva et al.

22. Vik T, Markestad T, Ahlsten G, Jacobsen G, Bak- NES III, 1988–1994. J Am Dental Assoc
keteig L. Morbidity during the first year of life in 1988;129:1229–38.
small for gestational age infants. Arch Dis Child Fetal 38. Moss M, Lanphear B, Auinger P. Association of
Neonatal 1996;75:F33–7. dental caries and blood lead levels. J Am Med Assoc
23. Seow WK. A study of the development of the 1999;281.
permanent dentition in very low birthweight chil- 39. Torres-Sanchez L, Berkowitz G, Lopez-Carrilho L,
dren. Pediatr Dent 1996;18:379–84. Torres-Arreola L, Rios C, Lopez-Cervantes M. Intra-
24. U.S. Department of Health & Human Services uterine lead exposure and preterm birth. Environ Res
(DHHS) NCfHS. NHANES III Reference Manual 1999;81:297–2001.
and Reports (CD-ROM). Hyattsville, MD: Centers for 40. Lin S, Hwang S, Marshall E, Marion D. Does paternal
Disease Control and Prevention; 1996. occupational lead exposure increase the risk of low
25. NNCHS. NHANES III Series 11, N0. 8a. Natality birthweight or prematurity? Am J Epidemiol
Data 2003: National Health Statistics. Hyattsvile, MD: 1998;148:173–81.
NNCHS; 2003. 41. Andrews K, Savitz DA, Hertz-Picciotto I. Prenatal
26. Hediger ML, Overpeck MD, Maurer K, et al. Growth lead exposure in relation to gestational age and
of infants and young children born small or large for birthweight: a review of epidemiological studies. Am
gestational age: findings from the Third National J Ind Med 1994;26:13–32.
Health and Nutritional Examination Survey. Arch 42. Viscardi R, Romberg E, Abrams R. Delayed primary
Pediatr Adolesc Med 1998;152:1225–31. tooth eruption in premature infants: relationship to
27. Zhang J, Bowes WA Jr. Birth-weight-for-gestational- neonatal factors. Pediatr Dent 1994;16:23–8.
age patterns by race, sex, and parity in the United 43. Barros AJ, Huttly S, Victora CG, Kirkwood BR,
States population. Obstet Gynecol 1995;86:200–8. Vaughan JP. Comparison of the causes and conse-
28. National Health for Statistics. National Health and quences of prematurity and intrauterine growth
Nutrition Examination Survey III Oral Examination retardation: a longitudinal study in southern Brazil.
Component. Hyattsville, MD: Westat Inc.; 1992. Pediatrics 1992;90:238–44.
29. Alexander GR, J.H. H, Kaufman RB, Kogan M. A 44. Fukuda JT, Sonis AL, Kurth S. Acquisition of mutans
United States national reference for fetal growth. streptococci and caries prevalence in pediatric sickle
Obstet Gynecol 1996;87:163–8. cell anemia patients receiving long term antibiotic
30. Kramer MS, McLean FH, Olivier M, Willis DM, therapy. Pediatr Dent 2005;27:186–90.
Usher R. Body proportionality and head and lenght 45. Cunningham F, MacDonald P, Gant N, Levenco K,
sparing in growth-retarded neonates: a critical Gilstrap L. Preterm and post-term pregnancy and
appraisal. Pediatrics 1989;84:717–23. fetal growth retardation. In: Cunningham F, Wil-
31. National Center for Health Statistics CfDCaP.Ana- liams J, editors. Williams Obstetrics. 19th edition.
lytic and Reporting Guidelines. The Third National Norwalk, CT: Appleton & Lange, 1993.
Health and Nutrition Examination Survey NHANES 46. Procianoy R, Garcia-Prats J, Adams JM, Silvers A,
III (1988–94). Hyattsville, MD: CfDCaP; 1996. Rudolph A. Hyaline membrane disease and intra-
32. Aligne C, Moss M, Auinger P, Weitzman M. Associ- ventricular hemorrage in small-for-gestational-age
ation of pediatric dental caries with passive smoking. infants. Arch Dis Child 1980;55:502–5.
J Am Med Assoc 2003;289:1258–64. 47. BuLock F, Woolridge M, Baum J. Development of co-
33. Callen J, Pinelli J. A review of literature examining ordination of sucking, swallowing and breathing:
the benefits and challenges, incidence and duration, ultrasound study of term and preterm infants. Dev
and barriers to breastfeeding in preterm infants. Adv Med Child Neurol 1990;32:669–78.
Neonatal Care 2005;5:72–88. 48. Kramer MS, Olivier M, McLean FH, Dougherty GE,
34. Barros AJ, Hirakata VN. Alternatives for logistic Willis DM, Usher RH. Determinants of fetal growth
regression in cross-sectional studies: empirical com- and body proportionality. Pediatrics 1990;86:18–26.
parison of models that directly estimate the preval- 49. Kramer MS, Seguin L, Lydon J, Goulet L. Socio-
ence ratio. BMC Med Res Methodol 2003;20:21. economic disparities in pregnancy outcome: why do
35. Stata Corporation. Intercooled Stata 8. College Sta- the poor fare so poorly? Paediatr Perinatal Epidemiol
tion, TX: Stata Corporation; 2004. 2000;14:194–210.
36. Williams RL, Creazy R, Cunningham G, Hawes W, 50. Savitz D, Kaufman J, Dole N, Siega-Riz A, Thorp JJ,
Norris F, Tashiro M. Fetal growth and perinatal Kaczor D. Poverty, education, race, and pregnancy
viability in California. Obstet Gynecol 1982;59:624–32. outcome. Ethn Dis 2004;14:322–9.
37. Vargas CM, Crall JJ, Scheneider DA. Sociodemo-
graphic distribution of pediatric dental caries: NHA-

376