Diabetes in Pregnancy and Childhood Cognitive Development

17 Pages·2016·0.87 MB·English

Diabetes in Pregnancy and
Childhood Cognitive Development:
A Systematic Review
Akilew Awoke Adane, MPH, Gita D. Mishra, PhD, Leigh R. Tooth, PhD
abstract
CONTEXT: The effect of diabetes during pregnancy on the cognitive development of offspring is
unclear because of inconsistent findings from limited studies.
OBJECTIVE: This review was aimed to provide the best available scientific evidence on the
associations between maternal pregnancy diabetes and the cognitive development of
offspring.
DATA SOURCES: A search was conducted in the Embase, CINAHL, PubMed, PsycINFO, and Scopus
databases.
STUDY SELECTION: Studies addressing the cognitive development of offspring (aged ≤12 years) as
outcome and any diabetes in pregnancy as an exposure were included.
DATA EXTRACTION: Data were extracted and evaluated for quality by 2 independent reviewers.
RESULTS: Fourteen articles were eligible for the review. Ten studies investigated the
associations between maternal pregestational diabetes or both pregestational and
gestational diabetes and offspring’s cognitive development; 6 studies found at least
1 negative association. Four studies exclusively examined the relationships between
gestational diabetes and offspring’s cognitive development; 2 studies found a negative
association, 1 a positive association, and 1 a null association. The use of diverse cognitive
and diabetes assessment tools/criteria, as well as statistical power, contributed to the
inconsistent findings.
LIMITATIONS: The English-language restriction and publication bias in the included studies are
potential limitations.
CONCLUSIONS: Although there are few data available regarding the associations between
maternal pregnancy diabetes and offspring’s cognitive development, this review found that
maternal diabetes during pregnancy seems to be negatively associated with offspring’s
cognitive development. Large prospective studies that address potential confounders are
needed to confirm the independent effect of maternal diabetes during pregnancy.
Centre for Longitudinal and Life Course Research, School of Public Health, the University of Queensland, Australia
Mr Adane designed the study, performed the systematic review, and drafted the manuscript; Prof Mishra and A/Prof Tooth contributed to the design of the study,
interpretation of the results, and critical revision of the manuscript for important intellectual content; and all authors approved the fi nal manuscript as submitted
and agree to be accountable for all aspects of the work.
DOI: 10.1542/peds.2015-4234
Accepted for publication Feb 19, 2016
To cite: Adane AA, Mishra GD, Tooth LR. Diabetes in Pregnancy and Childhood Cognitive Development: A Systematic Review. Pediatrics. 2016;137(5):e20154234
Downloaded from www.aappublications.org/news by guest on November 19, 2018
PEDIATRICS Volume 137 , number 5 , May 2016 :e 20154234 REVIEW ARTICLE
Globally, the prevalence of both models have shown that maternal child*)) AND (((((((((((((“cognitive
pregestational diabetes mellitus (PDM) diabetes during pregnancy is usually development”) OR “cognitive
and gestational diabetes mellitus associated with hyperglycemia and functions”) OR “school performance”)
(GDM) are increasing significantly due that more glucose will pass to the OR “educational outcomes”) OR
to the rising rates of obesity, aging, fetus, and this action may hinder “language development”) OR
and population growth.1–4 In women, the cognitive development of the “neuropsychological tests”)
the global prevalence of age-adjusted offspring.25 Although not always OR “mental development”) OR
diabetes increased from 7.5% in 1980 found, 26 metabolic complications neurodevelopment) OR “child
to 9.2% in 2008.4 An increasing future related to diabetes during pregnancy development”) OR cognition)
trend in the prevalence of diabetes such as acetonuria,2 7,2 8 ketoacidosis,2 0 OR intelligence) OR ” IQ”) OR
is also projected across all countries and glycosuria16 are also associated “intelligence tests”)) AND (((mother*)
worldwide. For instance, in 2011, there with offspring’s cognitive OR pregnan*) OR women)) AND
were 366 million people with diabetes, impairment. It has been suggested ((((((“type I diabetes”) OR “type Il
and over the next 19 years, this figure by some investigators that offspring diabetes”) OR “gestational diabetes”)
is expected to rise to 552 million. born to either mothers with PDM or OR “diabetes insipidus”) OR “diabetes
Substantial increases are expected in GDM are more likely to have poorer mellitus”) OR diabetes). Similarly,
developing countries related to lifestyle cognitive and language development systematic searches were conducted
changes after rapid urbanization, than those born to nondiabetic in the Embase, CINHAL, PsycINFO,
globalization, and aging.5 mothers.28–33However, other and Scopus databases using the same
investigators have either found no key word combinations tailored to
Any type of diabetes during pregnancy
such associations17, 18, 20, 22 or report each database until June 2015. These
is associated with poor maternal and
contradictory findings.23 searches were limited to studies in the
perinatal outcomes. For instance, con-
English language with human domain
genital malformation, 6 preterm birth,7 ,8 To the best of our knowledge,
restrictions. Moreover, the reference
large for gestational age at birth/ systematically synthesized
list of all identified relevant records
macrosomia,7 ,9 –11 and stillbirth6,1 0,1 1 information on the associations
were searched for additional studies.
are significantly more common in between maternal diabetes
The authors were not contacted for
offspring of diabetic mothers than during pregnancy and offspring’s
additional studies or data.
nondiabetic mothers. Even offspring cognitive development, particularly
of mothers with borderline GDM have in childhood, is lacking. The
Study Selection
been found to be born too early and to implications of such information
be macrosomic.12 are twofold: (1) it would provide Studies addressing the cognitive
decision-makers/clinicians with and/or language development
Due to advances in medical care,
comprehensive information to of offspring (aged ≤12 years)
offspring of mothers with diabetes
deliver preconception counseling; as outcome and either PDM or
are increasingly surviving the
and (2) it would enable the early GDM or both as a main exposure
perinatal period. Consequently, the
identification of infants at future risk or confounder were included in
long-term effects of diabetes during
of cognitive development. It would this systematic review. Reviews,
pregnancy on later childhood health
also avail information for researchers commentaries, and case or
such as obesity/adiposity, mainly via
to identify research gaps and guide descriptive studies lacking relevant
large for gestational age at birth, has
future research development. control groups were excluded.
been clearly established.13–15
Thus, the goal of the present
In contrast, data regarding the long-
Records deemed relevant from the
systematic review was to provide
term effects of maternal diabetes
title screening were further evaluated
the best available scientific evidence
during pregnancy on the offspring’s
by using the information available
regarding the possible associations
cognitive development are relatively
in the abstract by 2 independent
between maternal diabetes in
limited and, when data are available,
reviewers (A.A.A. and L.R.T.) using
pregnancy and the cognitive
the findings are inconclusive.16–23
the eligibility criterion. The same
development of offspring.
The mechanisms by which maternal authors further evaluated the full
diabetes during pregnancy is records of those eligible articles from
associated with the offspring’s METHODS the abstract review. Disagreements
cognitive development are unclear. were resolved by face-to-face
Data Sources and Search Strategies
Excess glucose levels in patients discussions and through joint review
with diabetes, a condition known as The PubMed database was of the records. The third author
hyperglycemia, is hypothesized to searched by using combinations helped resolve any disagreement
cause cognitive impairment.24 Animal of key words: (((((offspring) OR between the first 2 reviewers (Fig 1).
Downloaded from www.aappublications.org/news by guest on November 19, 2018
2 ADANE et al
Data Extraction
Data on the first author, publication
year, country, study design, sample
size, offspring’s age at follow-up,
diabetes type and assessment,
cognitive type and measurement,
confounders accounted/adjusted for,
and main findings between maternal
pregnancy diabetes and offspring’s
cognitive development were
extracted by the reviewers by using a
standardized data extraction format
(Table 1).
Quality Assessment
The Newcastle-Ottawa quality
assessment scale for case-control
and cohort studies34 was used to
assess the quality of included studies.
Two modifications were made: a
“nonblind standard assessment”
option was included in the outcome FIGURE 1
Flow diagram of studies included in the systematic review of maternal diabetes during pregnancy
subsection of the cohort studies
and childhood cognitive development.
assessment scale and “the length of
follow-up for the outcome to occur”
847 articles were screened; 679 for Children) in the review. Overall,
criteria was omitted because the were identified as not relevant or 14 studies16–23, 28–33 were eligible for
goal of this systematic review was
as further duplicates. The abstracts the present review.
to examine the effect of any diabetes
of 168 articles were then evaluated
during pregnancy on cognitive independently, and 111 records Quality Assessment
development of offspring to the age
were excluded. Subsequent full-
of 12 years. Consequently, a slightly record evaluations of 68 articles Using the adapted Newcastle-
modified score ranging from 0 (most resulted in a total of 19 eligible Ottawa quality assessment scale, the
likely biased) to 8 (highly unlikely to studies. However, 2 studies35, 36 were scores for each study ranged from 2
be biased) was calculated for each further excluded because they were to 6 (from a total of 8). In addition
study. published with the same populations to low overall scores, the following
used in 2 included studies.18, 20 quality concerns were observed:
Data Synthesis and Analysis In addition, 5 relatively different one-half of the studies reported
studies22, 32, 37–39 were conducted by unadjusted results; in all studies
Meta-analysis was not performed
using the same cohort study with (except 1 report30) the population
because of heterogeneous
varied sample sizes. All reported was obtained from unrepresentative
cognitive outcomes and having
global cognitive development as a populations, mainly from
too few studies that adjusted for
secondary outcome or confounder at hospital facilities; and 7 of 14
potential confounders. Instead, a
the age of 1 year32, 37–39 and 4 years.22 studies16, 17, 22, 23,2 9, 31, 33 reported
narrative review and qualitative
None of these studies primarily the use of standard cognitive
summarizations were undertaken
aimed to evaluate the global cognitive development assessment tools but
(Table 2).
development of offspring of mothers failed to provide or report blind
with diabetes mellitus (ODM). assessment.
RESULTS Therefore, we included Nelson et al32
Study Characteristics
(ie, better sample size and relatively
A total of 1295 records were recent) and Townsend et al22 The 14 studies were conducted
identified with English-language and (ie, cognitive development was between 1969 and 2015. The
human domain restrictions. After measured at the age of 4 years by offspring’s age at cognitive
removal of duplicates, the titles of using the Wechsler Intelligence Scale assessment ranged from 6 months
Downloaded from www.aappublications.org/news by guest on November 19, 2018
PEDIATRICS Volume 137 , number 5 , May 2016 3
7
2
d 0.
Effect Size; Cohen’s (95% CI) No suffi cient data No suffi cient data Expressive vocabulary Effect size varies from to 0.41 Expressive grammar Effect size varies from –0.37 to –0.30 Expressive vocabularyEffect size varies from –0.34 to –0.17 Receptive vocabulary
nt of Offspring Main Findings on Maternal Pregnancy Diabetes and Offspring’s Cognitive Development (95% CI) DM associated with lower IQ (MD), –3.5 −P(5.6, –1.5); = .001)PDM associated with lower IQ (96 vs 103; = P.001) and MDI (78 vs 81; = .001)DM without acetone has no association with MDI (81.0 vs 81.0), PDI (33.4 vs 34.2), Pand IQ (101.3 vs 100.8); all, > .05QNTS GDM associated with lower (mean); expressive vocabulary at 18 mo (–0.25 ± 0.96 vs 0.02 ± 0.99) and at 30 mo (–0.33 ± 1.22 vs 0.04 ± 0.96), expressive grammar at 30 mo (–0.28 ± 1.18 vs 0.04 ± 0.96) and oral communication at 72/84 Pmo (–0.30 ± 1.21 vs 0.05 ± 0.96); all, < .05No association with (mean) receptive vocabulary (–0.06 ± 1.00 vs 0.00 ± 0.99) and receptive grammar (–0.10 ± 1.16 vs 0.02 ± 0.98) at 18 mo, receptive grammar at 30 mo (0.04 ± 0.97 vs 0.01 ± 1.01), expressive vocabulary (–0.12 ± 0.90 vs 0.04 ± 0.99) and receptive vocabulary at 60 mo (–0.09 ± 0.86 vs 0.01 ± 1.01), math (–0.10 ± 1.02 vs 0.05 ± 0.93) and reading (–0.04 ± 0.96 vs 0.03 ± P0.99) scores at 72/84 mo; all, > .05QLSCDGDM associated with lower (mean); expressive vocabulary (–0.28 ± 1.09 vs 0.02 ± 0.99) and grammar at 30 mo (–0.38 ± 1.21 vs 0.03 ± 0.97) but not with receptive vocabulary at 30 mo (–0.18 ± 1.10 vs 0.01 ± 0.98) and at 42 mo (–0.16 ± 0.81 vs 0.02 ± 1.00)
e
m
nitive Develop Cognitive Outcome IQ nIQ ( = 121) nand MDI ( = 231) Language, NVIQ, and RM
g
Pregnancy and Co Cognitive Assessment WISC-III BSID at 8 mo and SBIS at 4 y MCDI at 18 and 30 mo, PPVT at 48 mo, expressive and receptive vocabulary at 60 mo, and EDI teacher-assessed communication at 72 and 84 mo
Diabetes During Diabetes Diagnosis Record- basedSomogyi- Nelson glucose tolerance curves OGTT
al s
Matern DiabeteType PDM + GDMPDM + GDM GDM
n
e
ation Betwe Age at Follow-up 8 y 8 mo and 4 y 1.5–7 y
rting the Associ Sample Size 6272 mother–child pairs237 ODM and control subjects 221 OGDM and 2612 control subjects nQNTS ( ranges from 721 to 861) and QLSCD n( varies from 955 to 1728)
o
of Studies Rep Study Design Cohort study Retrospective cohort study Retrospective cohort study
s
cteristic Country UK US Canada
a
r
TABLE 1 Cha First Author/Year Bonilla et al, 292012Churchill et 28al, 1969 Dionne et al, 302008
Downloaded from www.aappublications.org/news by guest on November 19, 2018
4 ADANE et al
dEffect Size; Cohen’s (95% CI) Effect size varies from –0.18 to –0.06Receptive grammarEffect size varies from –0.11 to 0.03Oral communicationdEffect size () = –0.32MathdEffect size () = –0.15ReadingdEffect size () = –0.07For the others, no suffi cient dataPDM SEA, –0.13 (–0.51, 0.26) FIQ, –0.13 (–0.51, 0.26)VIQ, –0.01 (–0.40, 0.37)PIQ, –0.27 (–0.66, 0.11)GDMSEA, –0.03 (–0.37, 0.31)FIQ, –0.40 (–0.75, –0.06)VIQ, –0.43 (–0.78, –0.09)PIQ, –0.19 (–0.53, 0.15)MDI, –0.64 (–1.12, –0.17) MDI, –0.49 (–0.94, –0.03) FIQ, –1.30 (–2.01, –0.60)VIQ, –0.76 (–1.46, –0.07)Language, –1.09 (–1.79, –0.39)Memory, –1.23 (–1.93, –0.52)PIQ, –0.57 (–1.26, 0.12)
Main Findings on Maternal Pregnancy Diabetes and Offspring’s Cognitive Development (95% CI)DM associated with language impairment P(OR), 2.2 (1.4, 3.5), RM (F = 5.31; = .02), Pand NVIQ (F = 3.73; = .05) but not with Pshort-term memory (F = 2.86; = .09) DM has no association with (MD) SEA, −−−0.04 (1.68, 1.75); FIQ, 0.54 (9.61, 8.52); −−VIQ, 1.5 (7.63 , 10.74); and PIQ, 4.01 −(13.80, 5.78)−DM associated with (MD) lower VIQ, 9.92 −−(18.34, 1.50); not associated with FIQ, −−−−5.93 (14.24, 2.38); PIQ, 0.19 (9.17, −8.78); or SEA, 0.30 (1.12, 1.72) DM associated with lower MDI (91 ± 9.0 vs P98 ± 12.1; < .05) M associated with lower MDI (100 ± 9 vs P104 ± 8; < .03) DM associated with lower FIQ (109.2 ± 1.4 vs 113.6 ± 3.5), VIQ (110.5 ± 1.5 vs 113.6 ± 4.2), language (108.8 ± 1.4 vs 112.9 ± 3.9), and memory (96.6 ± 1.4 vs 101.1 ± P3.8); all, < .05; No association with PIQ P(109.8 ± 1.5 vs 111.9 ± 3.8); = .14
G P G P D G
Cognitive Outcome Educational attainment and IQ MDI MDI IQ, language
c
Cognitive Assessment SEA at 4 y and WISC-III at 8 y BSID-II BSID-II WPPSI-III, developmental; neuropsychologiassessment
Diabetes Diagnosis ecord based GTT GTT elf-reported
R O O S
s
DiabeteType PDM + GDM PDM PDM + GDM GDM
p
Age at Follow-u 4 and 8 y 1 y 1 y 3–4 y
Sample Size SEA 24 OGDM, 21 OPDM, and 5804 control subjectsIQ23 OGDM, 20 OPDM, and 5079 control subjects 31 OPDM and 41 control subjects52 ODM and 75 control subjects21 OGDM and 191 control subjects
n
Desig study study study study
Study Cohort Cohort Cohort Cohort
y
r
nued Count UK Israel US US
TABLE 1 Conti First Author/Year Fraser et al, 162012 Hod et al, 311999 Nelson et al, 322003 Nomura et 17al, 2012
Downloaded from www.aappublications.org/news by guest on November 19, 2018
PEDIATRICS Volume 137 , number 5 , May 2016 5
dEffect Size; Cohen’s (95% CI) Young groupIQ, –0.95 (–1.40, –0.49) VIQ, –0.73 (–1.17, –0.28)PIQ, –0.67 (–1.11, –0.23)Older groupFIQ, –0.08 (–0.51, 0.35)VIQ, –0.32 (–0.75, 0.11)PIQ, 0.15 (–0.28, 0.58)PDM IQ, –0.08 (–0.44, 0.30)VIQ, –0.17 (–0.53, 0.20)PIQ, 0.05 (–0.32, 0.41)GDMIQ, –0.41 (–0.84, 0.03)VIQ, –0.54 (–0.98, –0.10)PIQ, –0.28 (–0.71, 0.16)PDM MDI, 0 (–0.43, 0.43)IQ, –0.23 (–0.67, 0.20)GDMMDI, 0.07 (–0.35, 0.50)IQ, 0.08 (–0.36, 0.52)
Main Findings on Maternal Pregnancy Diabetes and Offspring’s Cognitive Development (95% CI)Young age (5–8 y)GDM associated with lower FIQ (111 ± 14 vs 121 ± 8) and VIQ (107 ± 11 vs 115 ± 11); Pboth, < .05; Not associated with PIQ P(114 ± 17 vs 123 ± 11); > .05Older age (9–12 y)No association with FIQ (115 ± 13 vs 116 ± 12), VIQ (109 ±12 vs 113 ± 13), or PIQ P(119 ± 15 vs 117 ± 12); all, > .05 PDM has no association with FIQ (117.7 ± 12 vs 118.5 ± 11), VIQ (112.4 ± 12 vs 114.4 ± 12), and PIQ (120.4 ± 19 vs 119.7 ± 11.5)GDM has no association with FIQ (113.5 ± 14.3 vs 118.5 ± 11), VIQ (108.0 ± 11.5 vs 114.4 ± 12), and PIQ (116.0 ± 16.0 vs P119.7 ± 11.5); all, > .05 PDM has no association with MDI (89 ± 18 vs 89 ± 13) and IQ (89 ± 14 vs 92 ± 10)GDM has no association with MDI (90 ± 14 vs 89 ± 13) and IQ (93 ± 13 vs 92 ± 10)
e e
CognitivOutcom DI, IQ
Q Q M
I I
Cognitive Assessment WISC-Revised WISC-Revised BSID at 2 y SBIS at 3–5 y
Diabetes Diagnosis Blood glucose concentration test Blood glucose concentration test OGTT-O’Sullivan and Mahan criteria
s
DiabeteType GDM PDM + GDM PDM + GDM
p
Age at Follow-u 5–12 y 5–12 y 2–5 y
Sample Size 2 OGDM and 57 control subjects 7 OPDM, 32 OGDM, and 57 control subjects 0 OPDM, 82 OGDM, and 29 control subjects
3 5 8
Study Design Retrospective cohort study Retrospective cohort study Cohort study
y
r
nued Count Israel Israel USA
TABLE 1 Conti First Author/Year Ornoy et al, 191999 Ornoy et al, 182001 Rizzo et al, 201991
Downloaded from www.aappublications.org/news by guest on November 19, 2018
6 ADANE et al
dEffect Size; Cohen’s (95% CI) PPVT Early entry, –0.28 (–0.65, 0.09) Later entry, –0.81 (–1.30, –0.32) CommunicationEarly entry, –0.10 (–0.48, 0.27)Later entry, –0.76 (–1.25, –0.27)Length of UtteranceEarly entry, –0.11 (–0.48, 0.27)Later entry, –0.41 (–0.89, 0.07)MDIEarly entry, ranges –0.29 to 0Later entry, ranges –0.33 to –0.08SBISEarly entry, ranges –0.35 to 0Later entry, ranges –0.70 to –0.09FIQ, –0.16 (–0.88, 0.56)VIQ, 0.06 (–0.66, 0.77)PIQ, –0.31 (–1.03, 0.41)
Main Findings on Maternal Pregnancy Diabetes and Offspring’s Cognitive Development (95% CI)PDM associated with lower PPVT scores (114 ± 11.0 vs 111 ± 10.1 vs 105 ± 11.4) and Vineland Communication subscale (111 ± 10.2 vs 110 ± 9.2 vs 103 ± 11.5) PDM has no association with mean length of utterance (4.1 ± 1.0 vs 4.0 ± 0.86 vs 3.7 ± 0.91)MDI at 6 mo (107 ± 12.1 vs 104 ± 14.4 vs 106 ± 13.7), at 12 mo (117 ± 12.5 vs 113 ± 15.3 vs 112 ± 3.5) and at 24 mo (118 ± 18.4 vs 118 ± 19.4 vs 112 ± 16.8)SBISVerbal reasoning (114 ± 12.4 vs 110 ± 10.2 vs 105 ± 14.2), abstract reasoning (104 ± 11.8 vs 104 ± 8.7 vs 103 ± 9.7), quantitative reasoning (105 ± 9.5 vs 104 ± 9.1 vs 96.5 ± 8.7), short-term memory (110 ± 12.1 vs 110 ± 10.5 vs 108 ± 9.6) and composite score (110 ± 9.6 vs 109 ± 7.9 vs 103 ± 11.0). All comparisons are control versus early entry versus late entry DM has no association with FIQ (118 ± 16 vs 121 ± 21), VIQ (116 ± 16 vs 115 ± 19), or PIQ (116 ± 14 vs 121 ± 18)
Cognitive Outcome DI, IQ, and language
M Q
I
Cognitive Assessment BSID at 6, 12, and 24 mo; SBIS; spontaneous language sample; and PPVT at 36 mo revisedVineland Adaptive Behavior scales at 12, 24, and 36 mo WPPSI-Revised at 4 y
es sis d d
DiabetDiagno ot state ot state
N N
s
DiabeteType PDM PDM + GDM
Age at Follow-up 6, 12, 24, and 36 mo 4 y
Sample Size 09 OPDM (70 early entry and 39 late entry), and 90 control subjects 5 ODM and 15 control subjects
1 1
n
Desig study study
Study Cohort Cohort
y
r
nt
d u
nue Co US US
TABLE 1 Conti First Author/Year Sells et al, 211994 Townsend et 22al, 2005
Downloaded from www.aappublications.org/news by guest on November 19, 2018
PEDIATRICS Volume 137 , number 5 , May 2016 7
TABLE 1 Continued dFirst Author/CountryStudy DesignSample SizeAge at Diabetes Diabetes Cognitive Cognitive Main Findings on Maternal Pregnancy Effect Size; Cohen’s YearFollow-upTypeDiagnosisAssessmentOutcomeDiabetes and Offspring’s Cognitive (95% CI)Development (95% CI)Veena et al, IndiaCohort study32 OGDM and 9–10 yGDMOGTTKABC-II (learning, Learning GDM associated with higher mean Long-term retrieval/232010483 control long- term long-term cognitive scores of long-term retrieval/storage, 0.49 (0.13, 0.85)βsubjectsretrieval/retrieval/storage (), 0.38 (0.01, 0.75), verbal storage)storageability–names, 0.46 (0.09, 0.83)Word orderShort-term GDM has no association with: attention Verbal ability– names, 0.45 −memoryand concentration, 0.32 (0.04, 0.67); (0.09, 0.81) −Pattern reasoningReasoning Attention and visuospatial ability, 0.01 (0.36, 0.39); −abilityconcentration, 0.54 verbal ability–animals, 0.22 (0.16, 0.61); −(0.19, 0.90)reasoning ability, 0.14 (0.23, 0.51); and −Verbal abilityReasoning Verbal ability—animals, short-term memory, 0.03 (0.35, 0.40)ability0.34 (–0.02, 0.69)Kohs block-designVerbal fl uencyShort-term memory, 0.23 (–0.13, 0.59)Coding WISC-IIIVisuospatial All effect sizes are abilityunadjustedAttention and concentrationYamashita et JapanCohort study33 OPDM, 3 3–4 yPDM + OGTTTanaka-Binet IQDM associated with lower IQ (98 ± 17 vs IQ, –0.93 (–1.43, –0.44)33al, 1996OGDM, and GDMintelligence 113 ± 15)34 control scalesubjects BSID, Bayley Scales of Infant Development; CI, confi dence interval; DM, diabetes mellitus; EDI, Early Development Instrument; FIQ, full IQ; KABC, Kaufman Assessment Battery for Children; MCDI, MacArthur Communicative Development Inventory; MD, mean difference; NVIQ, nonverbal IQ; OGTT, oral glucose tolerance test; OR, odds ratio; PIQ, performance IQ; PPVT, Peabody Picture Vocabulary Test; QLSCD, Quebec Longitudinal Study of Child Development; QNTS, Quebec Newborn Twin Study; RM, recognition memory; SBIS, Stanford-Binet Intelligence Scale; SEA, school entry assessment; VIQ, verbal IQ; WISC, Wechsler Intelligence Scale for Children; WPPSI-III, Wechsler Preschool and Primary Scale of Intelligence, Third Edition.
Downloaded from www.aappublications.org/news by guest on November 19, 2018
8 ADANE et al
to 12 years. Most of the studies (OGDM), and 35 control subjects investigators administered various
were conducted in the United States, aged 2 to 5 years. The study cognitive and language development
Israel, and the United Kingdom; 3 examined the correlation between measures at 6 to 36 months of age.
were conducted in India, Japan, and maternal metabolism during In this study, PDM was significantly
Canada. The majority of the studies pregnancy and offspring’s cognitive associated with lower scores of
(n = 10) were prospective cohorts. and behavioral functioning. language development in late-entry
Although the measurement tools Accordingly, the MDI scores of groups, although this association was
were diverse, offspring’s IQ according the offspring correlated inversely not observed between maternal PDM
to several different IQ scales was the with the mother’s third-trimester and offspring’s MDI and the Binet IQ
most commonly assessed cognitive plasma β-hydroxybutyrate levels, scores.
outcome, followed by the mental and the Binet’s IQ scores correlated
Hod et al31 confirmed the negative
development index (MDI) of the inversely with the mother’s third-
association in 31 OPDM and 41
Bayley Scales of Infant trimester plasma β-hydroxybutyrate
control subjects at 1 year of age
Development. Overall, 10 of and free fatty acid levels. However,
between maternal PDM and offspring
14 studies16,1 8,2 0–22,2 8,2 9,3 1–33 the investigators found similar MDI
MDI; OPDM scored 7 points below
investigated the associations and Binet’s IQ scores among groups.
control subjects (91 ± 9 vs 98 ± 12;
between maternal PDM or both In support of this outcome, Ornoy
P < .05). Similarly, Yamashita et al33
PDM and GDM with cognitive et al, 18 using 32 OGDM, 57 OPDM,
found significantly lower Tanaka-
development of offspring. The and 57 control subjects matched
Binet IQ scores in 15 OPDM (98 ± 17)
remaining 4 studies17,1 9,2 3,3 0 on age, socioeconomic status
compared with 15 control subjects
exclusively examined the (SES), gestational age, birth order,
(113 ± 15; P < .0001). However, both
associations between maternal and family size, reported similar
studies failed to account for any
GDM and offspring’s cognitive Wechsler’s IQ scores across these
potential confounder.
development (Table 1). groups. Despite lack of statistical
significance, OGDM scored 5 points Bonilla et al29 (n = 6272) and Fraser
below control subjects. et al16 examined the cognitive
Association Between PDM and
development of offspring born to
Offspring’s Cognitive Development Nelson et al32 aimed to evaluate
women with PDM and GDM at the
In 1969, Churchill et al28 reported cross-modal recognition memory of
age of 4 and 8 years in 2 different
the negative associations between infants, and they provided a simple
large cohort studies. Bonilla et al
maternal diabetes in pregnancy MDI score comparison (unadjusted)
reported that maternal diabetes
(ie, both PDM and GDM) and between 52 ODM and 75 control
during pregnancy was negatively
offspring’s cognitive development. subjects at the age of 1 year. They
associated with Wechsler’s IQ
They administered the MDI at found a significant difference
scores; ODM scored 3.5 (95%
age 8 months (n = 121) and the between these groups (mean ± SD,
confidence interval, –5.6 to –1.5;
Stanford-Binet Intelligence Scale 100 ± 9 vs 104 ± 8; P < .03). However, P = .001) points lower compared
at age 4 years (n = 231). In this using a small subset (15 ODM and
with offspring of women without
study, maternal diabetes during 15 control subjects) of the same
diabetes. The investigators did not
pregnancy with acetonuria was population but at the age of 4 years,
report the outcomes according to
associated with lower MDI and IQ Townsend et al22 supported the
diabetes type, failed to account
scores. ODM, complicated null association between maternal
for potential confounders such as
by acetonuria, scored ∼2.5 diabetes during pregnancy and
SES and prepregnancy BMI, and
and 7 points lower in MDI offspring’s cognitive development
did not provide information about
and IQ scores, respectively, as measured by using Wechsler’s
the nature and size of the control
compared with offspring of mothers IQ scale. The investigators failed
group. Fraser et al also reported
without diabetes. However, in the to provide explanations for this
that both PDM and GDM were
absence of acetonuria, ODM had discrepancy except for the cognitive
associated with lower offspring
MDI and IQ scores similar to control measurement variations.
school entry assessment (age
subjects. In another study, Sells et al21 4 years, ∼5849 children) and
In 1991, Rizzo et al20 conducted examined the neurodevelopmental Wechsler’s IQ (age 8 years, ∼5124
a cohort study of 89 offspring outcomes of offspring born to children) scores. However, after
of mothers with pregestational mothers with insulin-dependent full adjustments were made for
diabetes mellitus (OPDM), PDM. The study included 109 OPDM various confounders, the negative
99 offspring of mothers with (70 early entry and 39 late entry) associations persisted only between
gestational diabetes mellitus and 90 control subjects, and the maternal GDM and verbal IQ (mean
Downloaded from www.aappublications.org/news by guest on November 19, 2018
PEDIATRICS Volume 137 , number 5 , May 2016 9
r difference, 9.92) scores. Despite
e
d √√ √√ √ √
en full model adjustment, Fraser et
G
al acknowledged the presence of
ng Age √ √√ √ tdhiaeb semteasl l( n∼u4m4)b earn do fa m siogtnhiefircsa wnti th
ri
p
s loss to follow-up (ie, IQ data were
Off GA √ √√ √√ √ available for ∼49% of the cohort).
h ht
BirtWeig √ √√ √ √ Association Between GDM and
Offspring’s Cognitive Development
Duration of reastfeeding √ √ Abofelfttshwporeuiengnhg ’m sth aceot egarnsnsiatoilcv GieaD tdiMeovn aesnl odp ment
B was separately reported in
7 studies, 16–20, 23, 30 only
s
ariate s Parity √ 4ad sdturedsieses1d7 ,G 19D, 2M3, 3. 0H eexncclue,s tivoe alvyo id
Development, and Confounders and Cov Potential Confounders/Covariate Maternal DepressionLifestyle Behaviors √√ √ e; HDP, hypertensive disorder of pregnancy. n3maarpIOnenc2rero pceudna stoes r ohcr5ertoyheeid7npt troiei trsncsoeotoy og ndwlasc-n , l tp hhiat1ootero9eghn o calrec lotlwegydo igs.ev m eticu,ehc lhbbpl aec-iijalcl ore ldroftahcuerhnstoends ttorn cr 4mttrot hb disslaoeltoetetun urrcdd n,dho ay Gieften, odD sd M ar e
e ag parental SES. Although the study
ng’s Cognitiv HDP √ GA, gestational waydoaujsun sugtnmedre eanrgpte ofdow r( eicero,e n5df– oa8un ynded laearrcssk),e tOdhG eaD nMy
TABLE 2 Summary of the Associations Between Diabetes in Pregnancy and Offspri Study/YearAssociation AgeBMISESEducation ↓√29Bonilla et al, 2012↓√√Churchill et al, 281969↓30Dionne et al, 2008↓√√√√16Fraser et al, 2012↓√31Hod et al, 1999↓32Nelson et al, 2003↔√17Nomura et al, 2012↔√19aOrnoy et al, 1999↔√18Ornoy et al, 2001↔√20Rizzo et al, 1991↓√21Sells et al, 1994↔Townsend et al, 222005↑√√√√23Veena et al, 2010↓Yamashita et al, 331996 ↓↑↔, signifi cant negative association; , signifi cant positive association; , no signifi cant association; a Found signifi cant association between GDM and IQ scores in young age (5–8 years) groups only. s18fcw(pDcimsSea8asbfwnuoacoccoxtlD1)eo4ieesel gor oo nopnr1)vtl l ori aem-robfrrh tngrtgae8aso edrseerweea enhrc no ocrduonds alssnia9temeadnol orlo ssn,8 lid– br s s ewetniueostA1ad 1eevhue s piut csnoI0n p osr2e.bts cQ3o ib abc w,, lgI,a r jop Wi0 vaye ljnestgamnaste3noehno ee cee clI0rtmectiet artcQeagsdo s nchtcrscsvnra tutorso h(ot;miocsseewab e ss1anms0sl)o(cmdrt s u og0u liat(aiss.nro ,et2mho1le idp7cgtluehraarnreu7n sh1ee a,eb ’dulr ± . os.dl a1odglsr yjaaGoen Anepleyvl1e w snv±w di1Da ctdea,es1sct ed ean.thtOrcgM1oe 7a5 lsOrasv bdteeocr4t02G– i ts iacoG i a,vwp .i7 ogm aDv4hla1obndeDm rt nsaIi1nMy1nl ieijalQ Myr.osud1 ede a5am t I, nsn r2a n al hnta± tesmrh1tn en nsa dd±a)d n r
Downloaded from www.aappublications.org/news by guest on November 19, 2018
10 ADANE et al