European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491

Contents lists available at ScienceDirect

European Journal of Obstetrics & Gynecology and

Reproductive Biology
journal homepage: www.elsevier.com/locate/ejogrb

Glycemic control and maternal and fetal outcomes in pregnant women

with type 1 diabetes according to the type of basal insulin
A. Chicoa,b,c,* , L. Herranzd,1, R. Corcoya,b,c , O. Ramreze,2 , M.M. Goyaf,3 , J. Bellartg,4 ,
S. Gonzlez-Romeroh,5 , M. Codinai,6 , P. Snchezj,7, A. Cortzark,l,8 , D. Acostam,9 ,
M.J. Picnn,10 , J.A. Rubioo,11, A. Megal,p,q,r,12 , M.A. Sanchos,13 , M. Balsellst,14 , E. Solu,15 ,
N.L. Gonzlezv,16 , J. Lpez-Lpezw,17, on behalf of the GEDE (Group of Diabetes and
Pregnancy of the Spanish Diabetes Association)
a
Department of Endocrinology, Hospital Santa Creu i Sant Pau, Av. Sant Antoni Maria Claret 167, 08025 Barcelona, Spain
b
CIBER-BBN, Spain
c
Departament de Medicina, Universitat Autnoma de Barcelona, Av. Sant Antoni Maria Claret 167, 08025 Barcelona, Spain
d
Department of Endocrinology, Hospital La Paz, Paseo de la Castellana 261, 28046 Madrid, Spain
e
Department of Obstetrics and Gynecology, Hospital Materno-Infantil de Canarias, Av. Martima Sur, s/n, 35001 Las Palmas de Gran Canaria, Las Palmas,
Spain
f
Department of Obstetrics, Hospital Vall dHebron, Passeig Vall d'Hebron, 119-129, 08035 Barcelona, Spain
g
Department of Obstetrics and Gynecology, Hospital Clinic, Sabino Arana, 1,08028 Barcelona, Spain
h
Department of Endocrinology, Hospital Regional de Mlaga, Av. Carlos Haya s/n, 29010 Malaga, Spain
i
Department of Endocrinology, Hospital Son Espases, Carretera de Valldemossa, 79, 07120 Palma, Illes Balears, Spain
j
Department of Endocrinology, Hospital Gregorio Maran, Calle del Dr. Esquerdo, 46, 28007 Madrid, Spain
k
Department of Endocrinology, Hospital Cruces, Bilbao, Plaza de Cruces, 12, 48903 Barakaldo, Vizcaya, Spain
l
CIBERDEM, Spain
m
Department of Endocrinology, Hospital Virgen del Roco, Av. Manuel Siurot, s/n, 41013 Sevilla, Spain
n
Department of Endocrinology, Hospital Virgen de la Victoria, Campus de Teatinos, S/N, 29010 Mlaga, Spain
o
Department of Endocrinology, Hospital Prncipe de Asturias, Carretera Alcal-Meco, s/n, 28805 Alcal de Henares, Madrid, Spain
p
Department of Endocrinology, Hospital Universitari de Tarragona Joan XXIII, C/Dr. Mallafr Guasch, 4, 43005 Tarragona, Spain
q
IISPV, Spain
r
Universitat Rovira I Virgili, C/Dr. Mallafr Guasch, 4, 43005 Tarragona, Spain
s
Department of Endocrinology, Hospital Clinico Lozano Blesa, Avenida San Juan Bosco, 15, 50009 Zaragoza, Spain
t
Department of Endocrinology, Hospital Mutua de Terrassa, Plaa del Doctor Robert, 5, 08221 Terrassa, Barcelona, Spain

Abbreviations: T1DM, type 1 diabetes; CSII, continuous subcutaneous insulin infusion; HbA1c, glycated haemoglobin A1c; SGA, small-for-gestational age; MDI, multiple
daily injections of insulin; RCT, randomized control trials; PIH, pregnancy-induced hypertension; SD, standard deviations; DCCT, Diabetes Complications Control Trial; LGA,
large-for-gestational age; OR, odds ratio.
* Corresponding author at: Department of Endocrinology, Hospital Santa Creu i Sant Pau, Av. Sant Antoni Maria Claret 167, 08025 Barcelona, Spain. Fax: +34 93 556 56 02.
E-mail addresses: achicob@santpau.cat (A. Chico), lucherranz.hulp@salud.madrid.org (L. Herranz), rcorcoy@santpau.cat (R. Corcoy), oramirezg@sego.es (O. Ramrez),
mariagoya@mac.com (M.M. Goya), jbellart@clinic.ub.es (J. Bellart), stellagoro@movistar.es (S. Gonzlez-Romero), mcodinam@binimelis.jazztel.es (M. Codina),
psanchezga.hgugm@salud.madrid.org (P. Snchez), alicia.cortazargalarza@osakidetza.net (A. Cortzar), dacostadelgado@gmail.com (D. Acosta), mjpiconcesar@gmail.com
(M.J. Picn), jarubiogarcia@gmail.com (J.A. Rubio), ana.megia@gmail.com (A. Mega), ma.sancho@comz.org (M.A. Sancho), 23591mbc@comb.cat (M. Balsells),
eva.sola@uv.es (E. Sol), ngonzalezg@sego.es (N.L. Gonzlez), jololo@sescam.jccm.es (J. Lpez-Lpez).
1
Fax: +34 91 727 70 50.
2
Fax: +34 928 444 732.
3
Fax: +34 93 489 31 29.
4
Fax: +34 93 227 56 05.
5
Fax: +34 951 29 11 21.
6
Fax: +34 871 205 500.
7
Fax: +34 91 586 66 69.
8
Fax: +34 944 99 29 45.
9
Fax: +34 955 01 34 73.
10
Fax: +34 951 03 21 88.
11
Fax: +34 918 80 18 25.
12
Fax: +34 977 22 40 11.
13
Fax: +34 976 56 59 95.
14
Fax: +34 93 736 50 59.
15
Fax: +34 96 386 25 01.
16
Fax: +34 922 65 38 08.
17
Fax: +34 925 21 48 36.

http://dx.doi.org/10.1016/j.ejogrb.2016.07.490
0301-2115/ 2016 Elsevier Ireland Ltd. All rights reserved.
 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491 85

u
Department of Endocrinology, Hospital Dr Peset, Gaspar Aguilar 90, 46017 Valencia, Spain
v
Department of Obstetrics and Gynecology, Hospital Clinico, Ctra. Ofra S/N La Cuesta 38320 La Laguna, Tenerife, Spain
w
Department of Endocrinology, Hospital Virgen de la Salud, Avenida de Barber, 30, 45071 Toledo, Spain

A R T I C L E I N F O A B S T R A C T

Article history: Objective: To examine the potential role of the type of basal insulin on glycemic control and maternal and
Received 29 October 2015 foetal outcomes in pregnant women with type 1 diabetes (T1DM).
Received in revised form 8 July 2016 Study design: Retrospective cohort study of pregnancies attended at 18 Spanish tertiary hospitals.
Accepted 26 July 2016
Inclusion criteria: T1DM, singleton pregnancies, delivery between 20022010, and use of the same basal
Available online xxx
and prandial insulin from before pregnancy until delivery.
Results: A total of 1534 pregnancies were included. The basal insulin most commonly used was Neutral
Keywords:
Protamine Hagedorn (NPH) (51.7%), followed by glargine (23.2%) and continuous subcutaneous insulin
Insulin analogue
Continuous subcutaneous insulin infusion
infusion (CSII) (21.1%). CSII users had longer diabetes duration. Multiple logistic regression analysis
Glycated haemoglobin showed that CSII was independently associated with lower doses of insulin, higher glycated haemoglobin
Perinatal outcomes (HbA1c) in all trimesters, and higher rates of miscarriage, preterm birth and neonatal hypoglycemia.
Type 1 diabetes Glargine was related to a higher risk of preterm birth and a small-for-gestational age infant (SGA). The
Pregnancy odds ratios (OR) of the associations between insulin type and clinical outcomes (from 0.642 to 4.894)
have a relevant magnitude.
Conclusions: In this observational study of pregnant women with T1DM, the type of basal insulin was
independently associated with metabolic variables and foetal outcomes.
2016 Elsevier Ireland Ltd. All rights reserved.

Introduction
A strict glycaemic control from before pregnancy until delivery
is essential to reduce the risk of maternal and foetal complications
in women with type 1 diabetes (T1DM) [1]. The type of basal
insulin could play a role in both the glycaemic control and in
pregnancy outcomes.
Outside pregnancy, continuous subcutaneous insulin infusion
therapy (CSII) has proven to be superior to multiple daily injections
(MDI) in terms of glycaemic control and hypoglycaemia [2] but
randomized control trials (RCT) have shown no clear benet of CSII
in pregnancy. However, these trials included a limited number of
patients and compared old CSII systems with human insulins [3].
Despite lack of evidence, this treatment is still used as a last option
when glycaemic goals are not achieved with MDI in women who
are pregnant or in pre-pregnancy care.
MDI is the standard treatment for T1DM subjects and
consequently for women in childbearing age. The introduction
of new long-acting insulin analogues such as insulin detemir and
glargine with its characteristic non-peaking action prole could
reduce the CSII indications. Nevertheless, little information is
available on safety and efcacy of these analogues compared to
other types of basal insulin in pregnancy. Glargine was the rst
marketed long-acting analogue, and no increase in adverse events
in pregnancy has been reported in case reports, case-series and
meta-analyses published to date [4,5]. The safety of insulin detemir
in pregnancy has been demonstrated in a RCT [6], but no RCT
addressing the use of glargine has been performed. Further
information is therefore needed on glycaemic control and
pregnancy outcomes achieved with different basal insulins in
pregnancy.
The objective of the present study was to compare three basal
insulin regimens in pregnant women with T1DM in terms of
glycaemic control and maternal and foetal outcomes.
Patients and methods
We performed an observational, retrospective, multicentre
cohort study in women with T1DM attended at 18 tertiary
university hospitals in Spain between 2002 and 2010. Inclusion
criteria were: (1) T1DM; (2) singleton pregnancy; (3) MDI or CSII;
(4) entire pregnancy followed at the same hospital; and (5) the
same basal and prandial insulin from before pregnancy until
delivery. Use of detemir and glulisine was not considered due to
infrequent use and lack of authorization for pregnancy respectively
in the study period. No additional exclusion criteria were used. The
study was approved by the ethics committee at each participating
centre. Information was obtained from medical records or data-
bases in place. Most of these databases had a prospective design.
Variable denitions
We assessed baseline demographic (age at time of delivery, self-
reported prepregnancy weight and body mass index (BMI), height
at booking), and diabetes characteristics (diabetes duration at
booking), smoking habit (none, 1 cigarette/day at the beginning
of pregnancy and discontinued, 1 cigarette/day at the beginning
of pregnancy and continued) and additional treatment (folic acid
supplementation, metformin, statins, angiotensin converter en-
zyme inhibitors/angiotensin receptor antagonists).
Maternal outcomes were: glycaemic control (glycated haemo-
globin (HbA1c), insulin dose and severe hypoglycaemia (events
requiring third party assistance)), weight gain (difference between
prepregnancy weight and weight at the end of pregnancy),
pregnancy-induced hypertension (PIH, blood pressure  140/
90 mmHg [7]), preeclampsia (blood pressure  140/90 mmHg plus
proteinuria  300 mg/day [7]) and caesarean section.
Gestational age at delivery was dened as the number of
completed weeks based on the last menstrual period or on the
earliest ultrasound assessment if discordant. Foetal outcomes
were: miscarriage (foetal death < 22 weeks [8]), preterm birth
(delivery < 37 weeks [8]), small-for-gestational age infant (SGA,
birth weight <10th centile according to Spanish foetal growth
charts that take into account sex and gestational age [9]), large-for-
gestational age infant (LGA, birth weight >90th centile), macro-
somia (birth weight  4000 g [8]), stillbirth (foetal death  22
weeks [8]), perinatal mortality (foetal and infant death from
22 weeks of gestation to 4 weeks after birth [8]), neonatal
hypoglycaemia (glycemia < 40 mg/dl in the rst 24 h after delivery
requiring treatment [10]), respiratory distress (any distress
requiring treatment [8]), neonatal sepsis (conrmed or suspected
systemic infection treated with antibiotics), umbilical cord blood
86 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491

Table 1
Baseline characteristics of the whole group of pregnant women with type 1 diabetes, and according to the type of basal insulin.

Total (n = 1534) NPH (n = 854) Glargine (n = 356) CSII (n = 324) Overall p

Year of delivery 2007 (2004,2008) 2006 (2003,2007) 2008a (2007,2009) 2007b,c (2005,2008) 0.001
Age (years) 31 (29,34) 31 (28,34) 32 (29,35) 32 (30,35)b 0.0001
Diabetes duration (years) 14 (7,20) 13 (7,19) 14 (7,20) 16 (10,23)b,c 0.0001
Non-Caucasian (%) 5.4 7 5.6 0.8b 0.001
BMI (kg/m2) 23.91 (22,26.40) 23.95 (22,26.47) 23.53 (21.9,26.22) 24.21 (22.23,26,7) 0.180
Chronic hypertension (%) 6.4 6.8 4.2 7.7 0.133
Smoking habit (%)
Stop in pregnancy 14.9 17 14.6 10.3 0.069
Active in pregnancy 3.6 3.7 2.6 4.5
Retinopathy (%) 18.9 18 17.7 22.6 0.156
Nephropathy (%) 6.8 5.9 9.3 6.8 0.100
Preconception care (%) 55.6 49.2 46.9a 82.3b,c 0.0001
Folic acid supplementation (%) 57.1 50.7 52 79.6b 0.0001
Metformin (%) 1.4 1.4 1.7 1.2 0.880
Statins (%) 0.8 0.2 2.4a 0.7 0.001
ACEI/ARA2 (%) 1.8 1.6 3.4 0.4c 0.02
Prandial insulin (%)
Regular 32.9 49.3 2.4a 7.6b,c 0.001
Lispro 45 35.9 37a 83.6b,c 0.001
Aspart 22.1 14.8 60.6a 3.1b,c 0.001

Qualitative variables are expressed as % and quantitative variables as P50 (P25,P75). CSII: continuous subcutaneous insulin infusion; ACEI: angiotensin converter enzyme
inhibitors; ARA2: angiotensin receptor antagonists. a,bStatistically signicant vs. NPH. cStatistically signicant vs. glargine.

pH, Apgar score at 1 and 5 min, and major and minor
malformations classied according the Spanish Collaborative
Study of Congenital Malformations (ECEMC) [11], based on the
International Working Group [12]. The primary outcome was
dened as the combination of perinatal mortality and/or major
malformation. This outcome was dened to capture two severe
foetal outcomes with a single variable.
HbA1c measured during the study period at all centres was
expressed as SDs around the mean for the method used. SDs were
then transformed to % values using DCCT (Diabetes Control and
Complications Trial) mean and standard deviation.
Statistical analysis was performed with SPSS v. 19. Qualitative
variables are described as % and quantitative variables (not
normally distributed) as median (P25, P75). Bivariate analysis
compared overall differences in baseline characteristics and
outcome variables across the groups of basal insulin using chi-
square for qualitative variables and KruskallWallis/MannWhit-
ney U test for quantitative variables.
To address the predictive ability of basal insulin on outcomes,
multiple regression analyses or logistic regression analyses
(backward method) were performed according to the outcome
and including basal insulin as a predictor. Qualitative variables
with more than two categories (i.e. basal insulin) were trans-
formed to dummy variables for multiple regression analysis
(glargine no/yes, CSII no/yes). A rst set of models was adjusted for
maternal baseline characteristics (ethnicity, age, BMI, smoking,
chronic hypertension) and T1DM features (diabetes duration,
prandial insulin, pregnancy planning, gestational age at 1st visit,
delivery year, centre, number of patients attended at the centre,
retinopathy, nephropathy). Folic acid was included as a predictor of
congenital malformations and foetal sex as a predictor of perinatal
outcomes. A second set of models was fully adjusted adding
intermediate variables to potential predictors (HbA1c, insulin doses
and severe hypoglycaemia in individual trimesters, and maternal
weight gain) and substituting chronic hypertension for hyperten-
sion during pregnancy (chronic or pregnancy-induced). Results are
expressed as coefcients or OR obtained for glargine or CSII using
Neutral Protamine Hagedorn (NPH) as the reference category.
Information regarding predictors other than basal insulin is not
given.
The association between type of basal insulin and time to
delivery was also analysed with KaplanMeier and Cox regression
analysis using the same predictors used for logistic regression
analysis.
A bilateral p < 0.05 was considered as the level of signicance.
Results
A total of 1534 pregnancies were included. NPH was used as
basal insulin by 55.6% of women (n = 854), glargine by 23.2%
(n = 356), and CSII by 21.1% (n = 324).
Table 1 summarises baseline characteristics of women includ-
ed. Depending on basal insulin, women showed differences in year
of delivery, age, diabetes duration, ethnicity, preconception care,
folic acid supplementation, use of potentially teratogenic drugs in
early pregnancy and type of prandial insulin. The most frequent
prandial insulin was lispro in the whole group (45%) and in CSII
(83.6%), regular insulin in women using NPH (49.3%) and as part in
those treated with glargine (60.6%). Differences were essentially
due to the CSII group.
Table 2 shows glycemic control before and during pregnancy.
HbA1c differed according to basal insulin in three of the four
periods analysed (before pregnancy, rst and third trimester).
Figures of HbA1c were lower for CSII users before pregnancy and in
the rst trimester and higher in the last trimester. Insulin doses
differed signicantly according to basal insulin in the four periods.
Compared to NPH group, insulin doses were lower in CSII in the
four periods and in glargine users in the third trimester. Severe
maternal hypoglycaemia was similar in the three groups, with the
exception of the second trimester at the expense of lower rates in
women with CSII compared to women with glargine. Weight gain
was similar in all groups.
Table 3 shows pregnancy outcomes according to insulin type.
According to overall p, groups were comparable in 17 of the
23 analysed variables including the combined outcome. However,
miscarriages, gestational age at delivery, caesarean section, minor
malformations and neonatal hypoglycaemia differed across
groups. Miscarriages and gestational age at delivery were less
favourable in CSII vs. NPH users. Minor malformations were higher
in glargine vs. NPH group. For caesarean section and neonatal
hypoglycaemia, despite overall signicance across the groups, no
differences were found in head-to-head comparisons. Kaplan
Meier analysis conrmed the association between type of basal
insulin and time to delivery (median week of delivery was
 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491 87

Table 2
Metabolic parameters before and during pregnancy of the whole group of pregnant women with type 1 diabetes mellitus, and according to the type of basal insulin.

Total (n = 1534) NPH (n = 854) Glargine (n = 356) CSII (n = 324) Overall p

HbA1c (%)
Before pregnancy 6.36 (5.79,7.25) 6.44 (5.79,7.63) 6.38 (5.79,7.15) 6.24b (5.83,6.92) 0.036
First trimester 6.24 (5.7,7.03) 6.33 (5.71,7.28) 6.16 (5.63,6.97) 6.10b (5.69,6.83) 0.020
Second trimester 5.75 (5.29,6.29) 5.79 (5.25,6.32) 5.70 (5.29,6.23) 5.76 (5.36,6.31) 0.508
Third trimester 5.83 (5.34,6.38) 5.83 (5.3,6.38) 5.70 (5.32,6.33) 5.84c (5.49,6.48) 0.043

Insulin dose (IU/kg/day)

Before pregnancy 0.66 (0.54,0.80) 0.70 (0.57,0.86) 0,69 (0.55,0.79) 0.56b,c (0.49,0.66) 0.001
First trimester 0,64 (0.51,0.78) 0.68 (0.53,0.81) 0.64 (0.50,0,78) 0.56b,c (0.47,0,67) 0.001
Second trimester 0.74 (0.59,0.92) 0.77 (0.61,0.95) 0,750 (0.61,0.93) 0.63b,c (0.53,0.79) 0.001
Third trimester 0.86 (0.69,1.07) 0.90 (0.72,1.11) 0.84a (0.66,1.05) 0.76b,c (0.63,0.95) 0.001

Severe hypoglycemia (%)

Total 12.0 12.9 11.6 9.2 0.327
First trimester 7.4 8.1 7.0 5.4 0.347
Second trimester 3.3 3.1 5.5 1.4c 0.039
Third trimester 1.9 1.8 2.4 1.8 0.810

Weight gain (kg) 12.5 (9.7,15.7) 12.5 (9.8,15.7) 12.0 (9.0,15.5) 13.0 (10.0,16.0) 0.163

Qualitative variables are expressed as % and quantitative variables as P50 (P25,P75).

a,b
HbA1c: glycated haemoglobin A1c, CSII: continuous subcutaneous insulin infusion; BMI: body mass index. Statistically signicant vs. NPH. cStatistically signicant vs.
glargine.

38 weeks, 38 weeks and 37 weeks for NPH, glargine and CSII users,
respectively; log rank test 0.006).
Table 4 shows the rst set of multivariate analyses to predict
glycaemic control and pregnancy outcomes after basal insulin,
adjusting for the variables indicated in the methods section. Basal
insulin was an independent predictor in 6 out of 12 metabolic
characteristics: CSII was associated with higher levels of HbA1c and
lower insulin doses in the three trimesters. Basal insulin was also
an independent predictor in 4 out of 20 pregnancy outcomes
(miscarriage, preterm birth, SGA infants and neonatal hypoglyce-
mia). The OR for miscarriage was higher in CSII users, the OR for
preterm birth was lower in glargine and higher in CSII, the OR for
SGA infant was higher in glargine and the OR for neonatal
hypoglycaemia was higher in CSII. All the associations in the
logistic regression analysis in the fully adjusted model (adjusting
both for basal characteristics and intermediate variables) were
consistent with those described in the former paragraph.
Cox regression analysis conrmed the relationship between the
type of basal insulin and time to delivery. Hazard ratio vs. NPH
insulin was for glargine 0.899 (CI95 0.7591.065) and 1.237 for CSII
(CI95 1.0031.525); overall p = 0.026.
In summary, in terms of foetal outcomes, glargine was
unfavourably associated with SGA and favourably with preterm
birth while CSII therapy is unfavourably associated with miscar-
riages, preterm birth, and neonatal hypoglycaemia.
Discussion
The main observation in this retrospective cohort study of
1534 T1DM women was that the type of basal insulin was
associated with parameters of glycaemic control and maternal and
foetal outcomes. The main strengths of the study are its large size,
the use of the same type of basal and prandial insulin since before
and throughout pregnancy and the comprehensive statistical
analysis. The limitations are its retrospective design and the lack of
information about previous obstetric history of the women.
As basal characteristics differed according to the type of insulin,
differences in metabolic parameters could be due to the type of
basal insulin or to basal characteristics. Similarly, differences in
pregnancy outcomes across groups of basal insulin could be due to
baseline characteristics or to the type of insulin acting either per se
or through intermediate metabolic parameters. Thus, discussion
and conclusions are based on the results of the multivariate
regression analysis, applying the best possible adjustment.
In terms of metabolic outcomes, we evaluated HbA1c, insulin
doses, severe hypoglycaemia and weight gain. We observed that
CSII was related to higher HbA1c in the three trimesters. These
results of the multiple regression analysis are apparently at odds
with the bivariate analysis (i.e. rst trimester HbA1c is lower in
women using CSII in the bivariate analysis whereas the coefcient
of the multivariate is positive). This emerged due to the negative
association of prepregnancy care (high in women with CSII) with
rst trimester HbA1c. One possible explanation for higher HbA1c in
CSII users is that CSII is often initiated in patients whose diabetes is
difcult to manage. However, as we did not measure management
difculty we were unable to adjust for this characteristic. Similar
results (lower HbA1c at the beginning of pregnancy and higher at
the end in patients using CSII and attending prepregancy care) have
been previously reported [13]. We also observed that CSII was
associated with lower doses of insulin throughout pregnancy. This
has been reported previously in pregnancy [1416] and in other
scenarios [8,17,18]. In the present study there was no association
between basal insulin and maternal severe hypoglycaemia. This was
unexpected since in T1DM patients, CSII has been related with
lower risk of hypoglycaemia both outside [2] and in a few studies
during pregnancy [19].
In terms of pregnancy outcomes, we observed that spontaneous
abortions were independently associated with basal insulin at the
expense of a higher rate in CSII users (OR 2.162). This cannot be
attributed to other measured baseline or intermediate character-
istics such as older age, longer diabetes duration, higher HbA1c or
earlier booking since we have adjusted for them. A potential
explanation could be the indication for CSII (either difculty in
management and/or poor obstetric history). As in the case of
HbA1c, the retrospective nature of the study does not allow us to
adjust for the difculty in management and we have acknowl-
edged that we do not have information on obstetric antecedents.
Preterm birth was associated with basal insulin, the rates being
higher for CSII (OR 2.158) and lower for glargine (OR 0.642),
association that was conrmed with KaplanMeier and Cox
regression analysis. For CSII the explanation could be again
difculty in management and prior unfavourable obstetric history.
For long-acting insulin, the study of Hod et al., supports an effect of
basal insulin on gestational age at birth [6]. They reported that
patients treated with insulin detemir delivered at a higher
 88
Table 3
Maternal and fetal outcomes of pregnancies of the whole group of women with type 1 diabetes, and according to the type of basal insulin.

Total NPH n = 854 Glargine Glargine Glargine 95% CSII n = 324 CSII CSII 95% p glargine vs. p CSII vs. p CSII vs. Overall P
n = 1534 n = 356 RR CI RR CI NPH NPH glargine
Abortions (%)
Spontaneous 10.2 7.6 9.6 1.25 0.841.86 17.6 2.31 1.66 0.300 <0.001 0.002 <0.001
3.22
Non spontaneous 1.8 2.2 1.4 0.63 0.241.68 1.2 0.55 0.191.62 0.498 0.350 1.00 0.418

A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491
PIH (%) 12.5 13.6 12.6 0.93 0.661.30 9.2 0.68 0.44 0.695 0.066 0.231 0.179
1.03
Preeclampsia (%) 9.9 11.2 7.3 0.65 0.421.02 9.2 0.82 0.50 0.057 0.516 0.504 0.143
1.33
Preterm birth (%) 22.7 21.6 20.8 0.97 0.751.24 28.1 1.31 1.03 0.808 0.034 0.041 0.059
1.65
Gestational age at delivery (weeks) 38 (37,38) 38 (37,38) 38 (37,38) 37 (36,38) 0.870 0.007 0.028 0.022
Shoulder dystocia (%) 2.6 2.5 2.0 0.78 0.311.94 3.8 0.52 0.181.53 0.660 0.328 0.255 0.473
Cesarean section (%) 49.6 45.7 50.5 1.10 0.971.26 59.7 1.31 1.151.48 0.161 <0.001 0.029 <0.001
Birth weight (g) 3500 3495 3495 3520 0.761 0.479 0.655 0.763
(3120,3910) (3097,3920) (3120,3905) (3150,3885)
Macrosomia (%) 20.5 21.3 19.2 0.90 0.691.18 19.8 0.93 0.71 0.460 0.661 0.916 0.713
1.23
LGA (%) 46.6 46.1 45.4 0.99 0.851.14 49.6 1.08 0.93 0.841 0.352 0.317 0.546
1.24
SGA (%) 1.9 2.1 2.5 1.21 0.532.81 0.4 0.18 0.02 0.653 0.088 0.045 0.128
1.38
Major malformations (%) 4.9 4.9 4.0 0.83 0.451.55 6.0 1.23 0.70 0.638 0.521 0.340 0.552
2.18
Minor malformations (%) 1.9 1.1 3.7 3,25 1.387.64 1.9 1.63 0.55 0.007 0.364 0.221 0.017
4.82
Stillbirth (%) 1.3 1.7 0.6 0.37 0.081.65 0.8 0.45 0.101,98 0.354 0.379 1.00 0.263
Perinatal mortality (%) 1.8 2.1 1.3 0.61 0.201.80 1.5 0.73 0.25 0.462 0.796 1.00 0.613
2.17
Perinatal mortality and/or major malformations 6.5 6.9 5.0 0.72 0.421.24 7.1 1.03 0.62 0.277 0.890 0.297 0.453
(%) 1.71
Neonatal hypoglycemia (%) 23.2 20.7 23.6 1.14 0.891.46 30.1 1.45 1.151.84 0.318 0.003 0.098 0.010
Respiratory distress (%) 10.4 11.5 9.0 0.78 0.521.17 8.8 0.77 0.501.19 0.234 0.249 1.00 0.305
Neonatal sepsis (%) 1.8 1.6 2.3 1.40 0.631.53 1.9 1.19 0.42 0.455 0.781 1.00 0.766
3.35
Apgar score
Minute 1 9 (8,9) 9 (8,9) 9 (8,9) 9 (8,9) 0.269 0.137 0.617 0.258
Minute 5 9 (9,9) 9 (9,9) 9 (9,9) 9 (9,10) 0.859 <0.001 <0.001 <0.001
<7 at minute 1 (%) 10.2 10 9.5 0.89 0.591.34 11.6 1.08 0.71 0.906 0.521 0.473 0.721
1.65
<7 at minute 5 (%) 2.4 2.2 2.3 0.79 0.351.76 2.7 0.91 0.39 1.00 0.796 1.00 0.934
2.13
Cord arterial pH 7.23 7.24 7.23 (7.18,7.29) 7.23 (7.18,7.26) 0.211 0.021 0.357 0.064
(7.18,7.29) (7.17,7.30)

Qualitative variables are expressed as % and quantitative variables as P50 (P25,P75); RR: relative risk; 95%CI: condence interval; CSII: continuous subcutaneous insulin infusion; PIH: pregnancy-induced hypertension; LGA: large-
for-gestational age infant; SGA: small-for-gestational age infant.
 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491 89

Table 4
Multivariate prediction of glycemic control and pregnancy outcomes in women with type 1 diabetes mellitus.a

Outcome variables Basal insulin

Type Coefcient (CI, multiple regression analysis) p

HbA1c
Before pregnancy Glargineb Not in the last step
CSIIc Not in the last step
1st trimester Glargineb Not in the last step
CSIIc 0.286 (0.0790.493) 0.007
2nd trimester Glargineb Not in the last step
CSIIc 0.244 (0.0590.429) 0.010
3rd trimester Glargineb Not in the last step
CSIIc 0.350 (0.1610.539) 0.001

Insulin dose
1st trimester Glargineb Not in the last step
CSIIc 0.088 ( 0.125 to 0.05) 0.005
2nd trimester Glargineb Not in the last step
CSIIc 0.105 ( 0.147 to 0.0629) 0.005
3rd trimester Glargineb Not in the last step
CSIIc 0.067 ( 0.122 to 0.012) 0.017

Weight gain Glargineb Not in the last step

CSIIc Not in the last step

Type Odds ratio (CI, logistic regression) p

Severe hypoglycemia (%)
Total NPH 1.00
Glargine Not in the last step
CSII Not in the last step
First trimester NPH 1.00
Glargine Not in the last step
CSII Not in the last step
Second trimester NPH 1.00
Glargine Not in the last step
CSII Not in the last step
Third trimester NPH 1.00
Glargine Not in the last step
CSII Not in the last step

Spontaneous abortion 0.019

NPH 1.00
Glargine 1.673 (0.9762.867) 0.061
CSII 2.162 (1.1193.927) 0.011

Non spontaneous abortion 1.000

NPH 1.00
Glargine Odd coefcient 0.999
CSII Odd coefcient 0.999

PIH NPH 1.00

Glargine Not in the last step
CSII Not in the last step

Preeclampsia 1.000
NPH 1.00
Glargine Odd coefcient 0.999
CSII Odd coefcient 0.999

Preterm birth 0.001

NPH 1.00
Glargine 0.642 (0.4130.998) 0.049
CSII 2.158 (1.2953.597) 0.003

Shoulder dystocia NPH 1.00

Glargine Not in the last step
CSII Not in the last step

Cesarean section NPH 1.00

Glargine Not in the last step
CSII Not in the last step

Macrosomia NPH 1.00

Glargine Not in the last step
CSII Not in the last step

LGA NPH 1.00

Glargine Not in the last step
90 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491

Table 4 (Continued)
Outcome variables Basal insulin

Type Coefcient (CI, multiple regression analysis) p

CSII Not in the last step

SGA 0.032
NPH 1.00
Glargine 4.894 (1.1361.08) 0.033
CSII 0.306 (0.0293.209) 0.324

Major malformations 1.000

NPH 1.00
Glargine Odd coefcient 0.992
CSII Odd coefcient 0.991

Minor malformations
NPH 1.00
Glargine Not in the last step
CSII Not in the last step

Stillbirth 0.053
NPH 1.00
Glargine 0.091 (0.110.765) 0.027
CSII 0.294 (0.0471.821) 0.188

Perinatal mortality 0.084

NPH 1.00
Glargine 0.184 (0.0380.885) 0.035
CSII 0.448 (0.1021.962) 0.286

Perinatal mortality and/or major malformations NPH 1.00

Glargine Not in the last step
CSII Not in the last step

Neonatal hypoglycemia 0.029

NPH 1.00
Glargine 1.089 (0.7091.672) 0.697
CSII 1.928 (1.1813.150) 0.009

Respiratory distress NPH 1.00

Glargine Not in the last step
CSII Not in the last step

Neonatal sepsis 0.058

NPH 1.00
Glargine 17.346 (1.119268.8) 0.041
CSII 12.326 (1.225124.003) 0.033

Apgar score
<7 at minute 1 NPH 1.00
Glargine Not in the last step
CSII Not in the last step
<7 at minute 5 NPH 1.00
Glargine Not in the last step
CSII Not in the last step

CI: condence interval; HbA1c: glycated haemoglobin A1c; CSII: continuous subcutaneous insulin infusion therapy; SGA: small-for-gestational age infant.
a
Backward method was used; the coefcients for basal insulin in the last step are displayed.
b
The reference category is no glargine.
c
The reference category is no CSII.

gestational age than those treated with NPH and this difference
was not attributable to glycaemic control since even when fasting
glucose was lower in the detemir arm, HbA1c did not differ.
We also observed an association between basal insulin and the
birth of an SGA infant (increased risk with glargine, OR 4.894).
Information on the association between basal insulin and SGA is
limited in the literature since most reports do not address this
outcome. An observational study described fewer SGA infants in
women using CSII [20] whereas no difference was reported in a
meta-analysis of RCTs comparing CSII vs. NPH [3]. The information
with glargine is also limited: this outcome has not been addressed
[21] or reported as non different [4]. In the current study, these
results are not explained by glycaemic control and since there are
data indicating that in animal models, insulin reaches the foetus in
early pregnancy [22] and insulin has growth-promoting effects at
that time [23], glargine could potentially have a direct effect on the
human foetus.
We also found that basal insulin was associated with neonatal
hypoglycaemia with higher rate in CSII users (OR 1.928). Although
the mechanism is not clear, this observation has been previously
described [24].
The combined outcome perinatal mortality and/or severe
congenital malformation was not associated with basal insulin.
We cannot compare this outcome with other reports because it has
not been previously addressed.
In this large retrospective cohort of T1DM women, basal insulin
was associated with more than 50% of the variables addressing
glycaemic control and about 20% of pregnancy outcomes. Overall,
 A. Chico et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 206 (2016) 8491
and compared with NPH insulin, glargine and CSII were not
associated with improved pregnancy outcomes; in fact, CSII was
related with a few worse outcomes. The number and magnitude of
some of the associations (OR around 2.00 for CSII and near 5.0 for
glargine) suggest a real effect of the type of insulin on glycaemic
and pregnancy outcomes. In statistical terms, relative risk >2530%
is considered relevant [25] and similarly important magnitudes
have been previously reported in pregnancy [6,24]. The potential
causality could be due to the different structures of insulins. As CSII
therapy uses short acting insulin and since we adjusted for this
variable, if looking for potential causes we should turn to other
components of the therapy.
In summary, we observed that the type of basal insulin in
pregnant women with T1DM was independently associated with
metabolic variables and some fetal outcomes. However, because of
the observational nature of this study, we cannot be certain that
the observations are due to incomplete adjustment for additional
characteristics (difculty in diabetes management and previous
obstetric history). Trials addressing insulin in pregnant women
with T1DM and sufciently powered for pregnancy outcomes are
needed, specically for glargine and CSII.
According to the present study on three cohorts of women with
T1DM treated with the same basal insulin throughout pregnancy,
we conclude that glargine is preferable to NPH taking into account
both patient convenience and associated pregnancy outcomes.
Although CSII therapy is independently associated to some adverse
outcomes, its discontinuation in women with clear CSII indication
before pregnancy (such as severe hypoglycaemia or dawn
phenomenon), does not seem advisable.
Conicts of interest
A. Chico declares that she has received honoraria and/or
nancial support to attend scientic meetings from Sano,Novo-
Nordisk, Lilly, Medtronic, Roche Diagnostics and Novalab and is
member of an Advisory Board of Abbott.
R. Corcoy declares that she has received honoraria and/or
nancial support to attend scientic meetings from Sano, Novo-
Nordisk, Roche and Medtronic.
J. Lpez-Lpez have received support for attending scientic
meetings and phase III studies from Sano, Lilly, Novo-Nordisk,
Abbott, Ferrer, MSD Almirall.
L. Herranz, O. Ramrez, M.M. Goya, J. Bellart, S. Gonzlez-
Romero, M. Codina, P. Snchez, A. Cortzar, D. Acosta, M.J. Picn, J.A.
Rubio, A. Mega, M.A. Sancho, M. Balsells, E. Sol, N.L. Gonzlez
declare not having conicts of interest.
Acknowledgement
Sano is currently supporting the GEDE group of the Spanish
Diabetes Association.
References
[1] Kitzmiller JL, Block JM, Brown FM, et al. Managing preexisting diabetes for
pregnancy: summary of evidence and consensus recommendations for care.
Diabetes Care 2008;31(5):106079.
[2] Misso ML, Egberts KJ, Page M, et al. Continuous subcutaneous insulin infusion
(CSII) versus multiple insulin injections for type 1 diabetes mellitus. Cochrane
91
Database Syst Rev 2010(1), doi:http://dx.doi.org/10.1002/14651858.
CD005103.pub2 Art. No.: CD005103.
[3] Mukhopadhyay A, Farrell T, Fraser RB, et al. Continuous subcutaneous insulin
infusion vs. intensive conventional insulin therapy in pregnant diabetic
women: a systematic review and metaanalysis of randomized, controlled
trials. Am J Obstet Gynecol 2007;197(5):44756.
[4] Imbergamo MP, Amato MC, Sciortino G, et al. Use of glargine in pregnant
women with type 1 diabetes mellitus: a case-control study. Clin Ther 2008;30
(8):147684.
[5] Lepercq J, Lin J, Hall GC, et al. Meta-analysis of maternal and neonatal
outcomes associated with the use of insulin glargine versus NPH insulin
during pregnancy. Obstet Gynecol Int 2012;649070, doi:http://dx.doi.org/
10.1155/2012/649070 Epub ahead of print 2012 May 16.
[6] Hod M, Mathiesen ER, Jovanovi9 c L, et al. A randomized trial comparing
perinatal outcomes using insulin detemir or neutral protamine Hagedorn in
type 1 diabetes. J Matern Fetal Neonatal Med 2014;27(1):713.
[7] Tranquili AL, Dekker G, Magee L, et al. The classication, diagnosis and
management of the hypertensive disorders of pregnancy: a revised statement
from the ISSHP. Pregnancy Hypertens 2014;4:97104.
[8] Chico A, Saig I, Garca-Patterson A, et al. Glycemic control and perinatal
outcomes of pregnancies complicated by type 1 diabetes: inuence of
subcutaneous insulin infusion and lispro insulin. Diabetes Technol Ther
2010;12(12):93745.
[9] Carrascosa A, Fernndez A, Yeste D, et al. Estudios espaoles de crecimiento
2008. Nuevos patrones antropomtricos. Endocrinol Nutr 2008;55(10):484
506.
[10] Metzger BE, Persson B, Lowe LP, et al. Hyperglycemia and adverse pregnancy
outcome study: neonatal glycemia. Pediatrics 2010;126:154552.
[11] Martnez-Fras ML. In: Martnez-Fras, Bermejo, editors. Manual operacional
del estudio colaborativo espaol de malformaciones congnitas. Madrid:
ECEMC, CIAC, Instituto de Salud Carlos III; 2003.
[12] Spranger J, Benirschke K, Hall JG, et al. Errors of morphogenesis: concepts and
terms. Recommendations of an international working group. J Pediatr
1982;100(1):1605.
[13] Gonzlez-Romero S, Gonzlez-Molero I, Fernndez-Abelln M, et al. Continu-
ous subcutaneous insulin infusion versus multiple daily injections in pregnant
women with type 1 diabetes. Diabetes Technol Ther 2010;12(4):2639.
[14] Wender-Ozegowska E, Zawiejska A, Ozegowska K, et al. Multiple daily
injections of insulin versus continuous subcutaneous insulin infusion for
pregnant women with type 1 diabetes. Aust N Z J Obstet Gynaecol 2013;53
(2):1305.
[15] Garca-Domnguez M, Herranz L, Hillman N, et al. Use of insulin lispro during
pregnancy in women with pregestational diabetes mellitus. Med Clin 2011;137
(13):5816.
[16] Mathiesen ER, Kinsley B, Amiel SA, et al. Maternal glycemic control and
hypoglycemia in type 1 diabetic pregnancy: a randomized trial of insulin
aspart versus human insulin in 322 pregnant women. Diabetes Care 2007;30
(4):7716.
[17] Torres I, Ortego J, Valencia I, et al. Benets of continuous subcutaneous insulin
infusion in type 1 diabetes previously treated with multiple daily injections
with once-daily glargine and pre-meal. Exp Clin Endocrinol Diabetes 2009;117
(8):37885.
[18] Chico A, Tundidor D, Jordana L, et al. Changes in insulin requirements from the
onset of continuous subcutaneous insulin infusion (CSII) until optimization of
glycemic control. J Diabetes Sci Technol 2014;8(2):3717.
[19] Garg S, Ampudia FJ, Pfohl M. Rapid-acting insulin in basal-bolus regimens in
type 1 diabetes mellitus. Endocr Pract 2010;16(3):486505.
[20] Bruttomesso D, Bonomo M, Costa S, et al. Type 1 diabetes control and
pregnancy outcomes in women treated with continuous subcutaneous insulin
infusion (CSII) or with insulin glargine and multiple daily injections of rapid-
acting insulin (glargine-MDI). Diabetes Metab 2011;37(5):42631.
[21] Egerman RS, Ramsey RD, Kao LW, et al. Perinatal outcomes in pregnancies
managed with antenatal insulin glargine. Am J Perinatol 2009;26(8):5915.
[22] Heyner S, Rao LV, Jarett L, et al. Preimplantation mouse embryos internalize
maternal insulin via receptor-mediated endocytosis: pattern of uptake and
functional correlations. Dev Biol 1989;134(1):4858.
[23] Herrler A, Krusche CA, Beier HM. Insulin and insulin-like growth factor-I
promote rabbit blastocyst development and prevent apoptosis. Biol Reprod
1998;59(6):130210.
[24] Chen R, Ben-Haroush A, Weismann-Brenner A, et al. Level of glycemic control
and pregnancy outcome in type 1 diabetes: a comparison between multiple
daily insulin injections and continuous subcutaneous insulin infusions. Am J
Obstet Gynecol 2002;4(404):e15.
[25] Argimon JM. Condence intervals: something more than a statistical
signicance test. Med Clin 2002;118:3824.