Relevance of gestational diabetes mellitus screening and comparison of selective with universal strategies

Diabetes &Metabolism 36 (2010) 575–586

Relevance of gestational diabetes mellitus screening and comparison of selective with universal strategies S. Hiéronimusa, J.-P. Le Meauxb,c,* a

Hôpital l’Archet Centre hospitalier universitaire de Nice, Service d’endocrinologie-maladies métaboliques-médecine de la reproduction, 06202 Nice, France b AP-HP, Groupe hospitalier Cochin-Saint-Vincent de Paul, Service de gynécologie-obstétrique, 75674 Paris cedex 14, France c Faculté de médecine Paris-Descartes, Site Cochin, 75014 Paris, FranceReceived: Accepted November 5th, 20100 Received: 3rd, 2010; Accepted: November 5th, 2010

Abstract

Objectives: To assess the relevance of gestational diabetes mellitus (GDM) screening policies and to compare selective with universal screening. Methods: Systematic review of all French and English language publications in the Medline and Cochrane Databases, published since 1990. Results: Maternal hyperglycaemia is associated with increased maternal and neonatal complications. The 75g OGTT (Oral Glucose Tolerance Test) is a valid and reliable test for GDM diagnosis. Treatment of GDM reduces perinatal complications. Selective screening helps limit false positive rates and concentrate medical resources. Nevertheless, screening could be more difficult and lead to missing up to 45% of GDM cases. Universal screening offers higher sensitivity but leads to more therapeutic interventions whose benefit and cost/effectiveness ratio need to be estimated in low risk women. Conclusion: The benefits of GDM screening and treatment have only been proven for women with GDM risk factors. Their relevance in women without risk factor remains controversial. © 2010 Elsevier Masson SAS. All rights reserved. Keywords: gestational diabetes mellitus, maternal outcome, maternal morbidity, treatment, untreated gestational diabetes, review

Résumé Intérêt du dépistage du diabète gestationnel et comparaison des stratégies ciblée et systématique

Objectifs : Évaluer l’intérêt du dépistage du diabète gestationnel (DG). Comparer les stratégies ciblée et systématique. Méthodes : Recherche systématique des publications en langues française et anglaise dans les bases de données Medline et Cochrane depuis 1990. Résultats : L’hyperglycémie maternelle est associée à une augmentation des complications maternelles et néonatales. L’HGPO (Hyperglycémie provoquée par voie orale) à 75 g est un examen valide et fiable pour le diagnostic de DG. Le traitement du DG permet de diminuer l’incidence de ses complications. Le dépistage ciblé permet de limiter le nombre de faux positifs et de concentrer les moyens. Cependant, il peut s’avérer complexe et conduire à méconnaître jusqu’à 45 % des cas de DG. Le dépistage universel est plus sensible, mais engendre plus d’interventions thérapeutiques dont le bénéfice et le rapport coût-efficacité restent cependant à évaluer dans les populations à faible risque. Conclusions: Les bénéfices d’un dépistage du DG et de la prise en charge qui en résulte n’ont été mis en évidence que chez les patientes présentant des facteurs de risque de DG. Leur intérêt chez les patientes ne présentant aucun facteur de risque reste controversé. © 2010 Elsevier Masson SAS. Tous droits réservés. Mots clés : diabète gestationnel, complications maternelles, pronostic maternel, traitement, diabète gestationnel non traité, revue

* Corresponding author E-mail address: [email protected] (J.-P. Le Meaux) © 2010 Published by Elsevier Masson SAS. All rights reserved.

576

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

1. Introduction

2.1.2. General concept of screening

The evaluation of a screening policy in a population must answer the following questions: should we screen? If yes, who? When? And how? In the case of gestational diabetes mellitus (GDM), the amount of publications, the heterogeneity of definitions used and the diversity of screening methods makes for a complex collection of data, giving rise to diverse and sometimes contradictory answers from the scientific community. Our objectives were to answer two of these questions: are there convincing data that help justify creating a GDM screening policy? And if yes, should it be for the entire population of pregnant woman or just reserved for higher risk patients?

Screening is considered to be pertinent when, along with the resulting intervention, it aims to decrease the morbidity and/or mortality in a population [1]. The criteria that can help increase the validity of a screening policy are disease prevalence in a population, the quality of screening tests and, potentially, diagnostic exams, the efficacy of resulting interventions and, finally, the economic consequences (cost-effectiveness ratio).

2. Materials and methods We carried out systematic searches on Medline and the Cochrane Database for articles in English and French on the epidemiology, complications, screening, treatment and medical-economic data on GDM published between 1990 and 2010. A second search was carried out for selective and universal screenings. Each search was completed with a manual search of referenced articles in the selected articles. For the first part of our research on the relevance of screening, the MeSH expression “gestational diabetes” was repeatedly associated with the terms: “prevalence”, “epidemiology”, “mortality”, “morbidity”, “screening”, “diagnosis”, “reproducibility”, “treatment” and “cost-benefit analysis”. For the second part on selective and universal screening, the expression “gestational diabetes AND screening” was associated with the terms: “selective” and “universal.”

2.1. Relevance of screening gestational diabetes mellitus 2.1.1. Definitions Diagnosis is an individual undertaking for a symptomatic patient in whom you seek to confirm the existence of a curable or incurable disease. Screening is part of public health policy. Its aim is to isolate within a targeted population individuals presenting risk factors or at an asymptomatic stage in a given disease in order to prevent it (primary prevention), treat it, thus lowering its prevalence (secondary prevention) or limit chronic complications or relapses (tertiary prevention). When screening for a particular disease, it can be universal (or “massive”) for all individuals within a targeted population; or targeted (or selective) and only be for individuals with risk factors for the disease.

2.1.3. Epidemiology The prevalence of GDM is estimated to be between 1 and 22% of pregnancies (see question 1 – F. Galtier). This variability in estimates is a reflection of the disparity in populations studied (prevalence of risk factors) and screening methods (tests and thresholds used). In France, the estimation provided by Audipog is around 4% [2]. Estimations of the proportion of diabetes during pregnancy vary between 10% and 40% of cases of diabetes present during pregnancy (see question 5 – A.M. Guedj). In some of these cases, diabetes was unrecognized (type 2 diabetes) and wrongly confused with GDM.

2.1.4. Consequences 2.1.4.1. Perinatal The association of GDM and high perinatal mortality risk remains controversial (see question 3 – D. Mitanchez). The oldest data show a relative significant risk estimated between 2 and 4 [3-5] (EL4). However, the populations studied lacked heterogeneity, mixing preexisting diabetes and true GDM. This association was not confirmed by the HAPO (Hyperglycaemia and Adverse Pregnancy Outcomes) study results on 23,000 pregnant women with no preexisting diabetes and whose results of the 75g oral glucose tolerance test (OGTT) given between the 24th and 32nd weeks of pregnancy remained blind for the entire pregnancy [6]. However, patients with fasting glycaemia over 1.05 g/L, 2 g/L at 2 h or 1.6 g/L with a random dosage were excluded from this study and were given appropriate care. Their results are therefore unavailable and the absence of a link between maternal hyperglycaemia and therefore perinatal mortality cannot be affirmed. The risk could exist in patients with high hyperglycaemia equivalent to cases of preexisting diabetes before pregnancy [7]. However, the association between GDM and neonatal morbidity has been clearly established [6, 8] (EL2). The results of the HAPO study demonstrated a positive correlation between maternal glycaemia during pregnancy and incidence of neonatal morbidity [6]. In this study, maternal glycaemia was positively correlated to a risk of macrosomia (birth weight > 90th percentile for gestational age), neonatal hypoglycaemia, premature birth (< 37 WG), shoulder dystocia, obstetrical trauma, NICU stays and neonatal jaundice. However, macrosomia cannot be

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

considered as intrinsically morbid. It represents an intermediary factor between hyperglycaemia and neonatal morbidities (obstetrical trauma) and maternal morbidities (Caesareans and maternal trauma related to birth) [9]. 2.1.4.2. Maternal The existence of GDM is associated with higher maternal morbidity in the short term, mainly via the intermediary factor of fœtal macrosomia (EL2) (see G. Beucher). An increase in Caesarean risk before and during labour and in severe perineal tearing (grades 3 and 4) [6, 8, 10] has been observed. Furthermore, incidence of preeclampsia is positively correlated with maternal glycaemia [6, 8, 11]. The correlation is rather moderate, particularly after adjusting for body mass index (BMI). As for neonatal morbidity, there is a dose-effect relationship between glycaemia and the incidence of complications, but without being able to specifically define a threshold under which risks are marginal. In the long term, a history of GDM is associated with the onset of type 2 diabetes in the years or decades after the index pregnancy [12-16] (EL1). GDM appears to be the first manifestation of disease, preceded by physiological insulin resistance related to the pregnancy that translates to type 2 diabetes after what can be several years. The diagnosis of GDM could also be an opportunity to begin primary preventative measures that are effective in post partum type 2 diabetes [17]. 2.1.4.3. Socioeconomics A medical economic analysis from the United States in 2007 has reported that the health costs related to GDM and its consequences were estimated to be $635 million (€490 million) with $595 million going toward care for mothers (prenatal and post partum periods) and $40 million for paediatric care during the first year of life [18]. This estimate took into account the costs for hospitalizations, regular and urgent care pregnancy check-ups and prescriptions for mothers and their children. With a prevalence of 180,000 cases (4.5% of pregnancies) in the United States in 2007, the cost was estimated at $3,305 per case of GDM. No similar study in the French population could be found. No data on the social impact and indirect consequences on the global economy (decrease in work time and altered productivity) could be found. It seems probable that the economic consequences of GDM, as with any other disease, go beyond health economics and that its total cost is largely higher than estimates limited to medical care.

577

be considered as a reference test. The performances of the different tests can only be estimated by their ability to predict the onset of perinatal and maternal complications. The only data available are found in the HAPO study that confirms the association between maternal glycaemia and obstetrical and neonatal complications [6]. The secondary analysis of these data helped the IADPSG (International Association of Diabetes and Pregnancy Study Group) to propose the following pathological glycaemia thresholds [19]: fasting glycaemia ≥ 0.92 g/L, post prandial glycaemia at 1 h ≥ 1.80 g/L, post prandial glycaemia at 2 h ≥ 1.53 g/L after a 75g OGTT between 24 and 28 WG (see question 7 – E. Cosson). The thresholds were chosen based on an odds ratio equal to 1.75 for the onset of an unfavourable outcome of the pregnancy in patients with at least abnormal glycaemia levels (the control group was made up of patients with lower than average glycaemia levels). Using these thresholds, 17.8% of patients eligible for the HAPO study had at least one pathological glycaemia level and could be considered to have GDM. Among them, 1.7% of patients were excluded from the study for having over-high glycaemia levels that required informing them and providing appropriate medical care. The combined use of several glycaemia measurements helped increase prediction sensitivity, but at the cost of increasing the rate of false positives. In the HAPO study, the sensitivity of each of the three glycaemia levels used independently did not exceed 20% for predicting macromsomia and 15% for predicting Caesareans before labour and symptomatic neonatal hypoglycaemia. The rate of false positives was estimated to be around 10%. The combination of the results of the three dosages helped increase sensitivities to 27%, 22% and 21%, respectively, for predicting each of the complications (sensitivity calculated by us from the HAPO study results and its secondary analysis) [6, 19]. 2.1.5.2. Reliability The rate of reproducibility (proportion of similarity of glycaemia levels beyond a threshold between two tests carried out at an interval of at least 24 hours) was estimated to be 83% for 50g OGTT and between 76 and 78% for 100g OGTT [20-22] (see question 6 – M. Virally). For 75g OGTT, a quantitative evaluation of 60 patients at 27 WG did not show any significant difference between each glycaemia level (H0, H1 and H2) for two OGTT carried out at three-day intervals [23]. The 75g OGTT therefore seems to have good reproducibility (EL2). 2.1.5.3. Acceptability and side effects

2.1.5. Screening exams 2.1.5.1. Validity In the case of GDM, the evaluation of the validity of screening tests is difficult. The tests used have a dual function of screening and diagnosis; none of the tests available can

The acceptability of a screening policy by a population depends mainly on the simplicity of the screening methods, constraints related to care in case of a positive result and potential adverse events. A positive result for a screening test could be a source of adverse psychological consequences in patients and higher medical costs. The perception of having

578

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

a disease, even if asymptomatic like GDM, can be cause for anxiety and depression. The existence of a pathological state, in cases of a positive screening result, can influence care givers and consequently bias medical care. There is a lot of data concerning GDM screening. For OGTT, the oral dose of the glucose solution provokes vomiting that inhibits patients from finishing the test in fewer than 10% of cases [24] (EL4). Patients with a positive result during screening and considered to have GDM appear to have higher anxiety just after the screening compared to unscreened patients or those with a negative test result. However, this difference seems to be minimal and is not demonstrated in the mid-term (before giving birth) and long-term (post partum) [25, 26] (EL2). In some studies, a positive screening result is experienced by some patients as an alteration in their or their child’s state of health; this perception remains through the mid- and long-term [26-29] (EL2). Medical care (diet or insulin therapy) does not seem to change this perception [30]. A contrario, in the therapeutic randomized trial by Crowther et al. (Australian Carbohydrate Intolerance Study in Pregnant Women trial Group – ACHOIS), patients informed of the GDM diagnosis and given care judged their physical state to be better compared to uninformed or uncared for patients [31] (EL1). The heterogeneity of results could be the result of using different tools to evaluate mental and physical well-being. No study has shown any difference in the incidence of depressive syndromes [25, 26, 28-30, 32]. A positive screening result is associated with an increased risk in Caesarean delivery, independent of fœtal macrosomia (NP2). Naylor et al., in a prospective cohort study of 3,800 screened pregnant patients, showed a significant increase in Caesarean risk estimated to be between 60 and 120% in patients with a positive test result with no fœtal macrosomia, compared to patients with a negative test result [33]. In the randomized ACHOIS study, a positive result was associated with an increase in the number of induced births [31] (EL1).

2.1.6. Intervention efficacy In patients with GDM, medical care (diet, glycaemic monitoring and insulin therapy, if needed) is associated with a decrease in risks of perinatal and maternal complications (EL1) [31, 34-41]. However, the efficacy of medical care on the decrease in Caesarean rates remains controversial. The meta-analysis by Horvath et al. included data from five randomized studies that evaluated the efficacy of specific medical care in patients with GDM compared to unspecific medical care (2,999 patients) [31, 34, 36-39]. In this analysis, specific care was associated, in children, with a decrease in birth weight and a reduction in risks of birth weights higher than the 90th percentile (OR: 0.48 CI 95% [0.38-0.62]), birth weight over 4,000 g (OR: 0.38 CI at 95% [0.30-0.49]) and shoulder dystocia (OR: 0.40 CI 95% [021-0.75]). In women, it was associated with a decrease

in the risk of preeclampsia (OR: 0.46 CI 95% [0.22-0.97]). However, in this meta-analysis, initiating care was not associated with a decrease in the frequency of elective or emergency Caesareans (OR: 0.86 CI 95% [0.72-1.02]). Only the randomized trial by Landon et al., included in this meta-analysis with 958 patients, showed a decrease in Caesarean rates in cases of specific care [39]. Caesarean levels were 33.8% in patients not informed of their GDM and not receiving any specific care and 26.9% in patients aware of their GDM and specifically treated for it (diet and sometimes insulin therapy). Moss et al., in the secondary analysis of data of the randomized ACHOIS trial, calculated that the care for 100 women with GDM could help avoid 2.2 severe neonatal complications and 1 perinatal death [31, 42].

2.1.7. Cost-effectiveness ratio of screening and medical care The medical/economic evaluation of a screening policy must take into account the cost of the screening test itself as well as costs related to potential adverse events. These costs should be compared to those resulting from treating the expected complications in the absence of screening. It should estimate the consequences of screening in terms of improvement or deterioration of the quality of life of screened subjects and, in the case of obstetrical diseases, their children. Studies on the subject are rare, contradictory and not really transposable to the French population. They are on Anglo-Saxon populations with a high prevalence of risk factors. In an economic analysis of a systematic screening policy for GDM in the United States, Nicholson et al. estimated the overall cost of screening GDM to be lower than that resulting from no screening. The cost of screening (OGTT in one or two doses) and its consequences were estimated to be between US$2,836 (€2,100) and US$2,895 (€2,144) per screened subject [43]. The cost of no screening was estimated to be US$2,995 (€2,219) per screenable subject. These estimates included direct health costs related to maternal and paediatric care as well as indirect costs related to loss in productivity or salary for patients. A contrario, the secondary analysis of data from the ACHOIS trial estimated the cost related to medical care of subjects with GDM to be $6, 050 AUS (€3,993) per treated subject (compared to an absence of care for these subjects). These estimates also included costs related to maternal and paediatric care. From a paediatric point of view, the cost of a year of life saved was estimated to be $2,988 AUS (€1,972) [42]. Overall, the increase in risks of neonatal and maternal morbidity in cases of maternal hyperglycaemia (NP2), the existence of valid and reliable screening tests (NP2), tions (NP1) and the acceptable nature of adverse events related to medical care (NP2) favour carrying out GDM screenings.

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

However, given the heterogeneity of the populations and the screening methods between the HAPO study and other clinical studies, the benefits of medical care for patients diagnosed with GDM based on the guidelines proposed by the IADPSG remain unknown. The anticipated prevalence of GDM diagnosed based on these proposed guidelines is considerably higher than what is observed in clinical trial populations as the number of subjects to treat to avoid adverse events could be much higher than what was observed in the trials. Furthermore, most available data on populations different from the French population in terms of prevalence of risk factors (obesity prevalence and ethnic distribution) limits their external validity. A prospective evaluation of the therapeutic efficacy and cost benefit ratio in a screened population following these methods and comparable to ours is necessary.

579

In their meta-analysis, Torloni et al. (NP1) [46] estimated that for each increase of 1 kg/m2 of BMI, the prevalence of GDM increased 0.92% (CI 95% [0.7-1.1]). The relative risk of GDM with obesity (BMI > 25 kg/m2), moderate obesity (BMI > 30 kg/m2) and morbid obesity (BMI > 40 kg/m2) was respectively 1.97 (CI 95% [1.77-2.19]), 3.01 (CI 95% [2.34-387]) and 5.55 (CI 95% [4.27-7.21]). BMI is also associated with a higher risk of GDM in pregnant adolescents. In the study by Khine et al. [47], the threshold of 27 kg/m2 helped diagnose 82% of cases of GDM (sensitivity), with a negative predictive value of 99.4% (EL 3). Aside from BMI, the factors that have the most impact on the risk of GDM are: maternal age (risk multiplied by 3 in the 25-29 age range and by 4 in the 30-34 age range) and a history of gestational diabetes and type 2 diabetes in the immediate family. Ethnic origin only had a moderate impact on the risk of gestational diabetes (see question 1 – F. Galtier).

2.2. Comparison of selective with universal screening methods In selective screening of GDM, patients to be screened are chosen based on their case history. Its goal is to limit screening tests in high risk and high prevalence populations to curb screening costs, decrease the number of false positives and thus increase the positive predictive value of these tests. Universal screening is when all pregnant women are tested. It necessarily implies an increase in the incidence of GDM, at varying rates, depending on the population considered [44].

2.2.1. Risk factors In selective screening, selection of patients is based on searching for risk factors (see F. Galtier). The factors to consider are medical history, the woman’s state of health at the beginning of the pregnancy and complications arising during the pregnancy. They are as follows: family history: type 2 diabetes in one or several first degree relatives; medical history: metabolic syndrome, polycystic ovarian syndrome; obstetrical history: GDM or macrosomia in a previous pregnancy, unexplained in utero fœtal demise; maternal characteristics: age (different thresholds), BMI (different thresholds), ethnic origin; complications in the current pregnancy: hydramnios, suspicion of macrosomia. In the literature, the most often used risk factors in selective screening are a history of GDM and macrosomia, age, BMI, ethnic origin and a first degree history of diabetes. Among these factors, BMI is consistently used in the recommendations for selective screening. There is a strong correlation between being overweight in adulthood and a risk for GDM. In the Nurses’ Health Study (NP2) a BMI > 25 in adults multiplies by 2.36 (CI 95% [2.12-3.77]) the risk of developing GDM [45].

2.2.2. Risk score In order to account for the accumulation of several risk factors in one person, some authors have proposed using the risk scores detailed in Table 1. In the Tri-Hospital Gestational Diabetes Project (EL2) [48] cohort study, a score was established according to the methods detailed in Table 2. The prevalence of GDM increases with the value of the risk score, going from 0.9% for a score ≤ 1 to 18.7% for a score ≥ 6. This score was established based on independent risk factors for the onset of GDM identified in the study, with the number of points attributed to each risk factor being equal to its adjusted odd ratio. A selective screening strategy based on this clinical score was thus developed: • score ≤ 1 – No screening; • score > 1 – Screening carried out with an oral dose of 50 grams of glucose: if the score was between 2 and 3, a threshold of 1.40 g/L(7.8 mmol/L) was used, if the score was ≥ 4, 1.30 g/L was used as the threshold. The authors reported that this selective strategy helped to obtain a sensitivity and specificity identical to those seen in universal screening and avoid screening with 50 g glucose in 34.6% of women [49]. However, it should be noted that this equivalency in performance is the result of an adjustment of diagnostic criteria since the thresholds selected to confirm GDM were lowered for women who had a higher risk level. In this study, the proportion of GDM that would have been missed with specific screening was 7%. In a retrospective study on 1,876 Asian women (NP4), the risk of developing GDM was associated with age, BMI, family history of diabetes and history of macrosomia and abortions (Table 3) [50]. A risk score was established as a regression equation integrating the factors independently associated with GDM:

580

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

Table 1 Risk score in the medical literature Author, year, country

Type of study No. patients Reference Test

Definition criteria of risk score

Naylor [48] 1997, Canada

Retrospective cohort N = 3131 100 g OGTT NDDG criteria

Risk factors Score Age ≤ 30 31-34 ≥ 35 BMI ≤ 22 22.1-25 ≥ 25.1 Ethnic origin Caucasian Afro-American Asian Other

Score performances

Score 0 1 2 0 2 3

GDM Index (%) according to score rating Score ≤ 1 0.9 2 3.7 3 3.9 4-5 7.3 ≥6 18.7

0 0 5 2

Phaloprakarn [50] 2009, Thailand

Retrospective cohort N = 1876 100 g OGTT Carpenter & Coustan criteria

Risk factors Age < 25 25-34 ≥ 35 BMI < 20 20-24.9 25-29.9 ≥ 30 Family history of diabetes yes no* History of macrosomia yes no* History ≥ 2 abortions yes no* *parameter rated 1 if present, 0 if absent Risk score expressed as a regression equation: 6*age + 11*BMI + 109*history diabetes + 42*history macrosomia + 49*history ≥ 2 abortions

Score ≥ 380 Sensitivity Specificity

86.9% 45%

Van Leeuwen [51] 2009, Holland

Prospective cohort N = 995 75 g OGTT WHO criteria

Risk factors Non Caucasian ethnic origin Family history of diabetes Body Mass Index ≥ 22-30 History of GDM Predictive model established using a regression equation integrating risk factors

Sensitivity Specificity

75% 57.8%

Diagnostic criteria of gestational diabetes: NDDG (National Diabetes Data Group): at least 2 threshold values selected: fasting glycaemia ≥ 5.8 mmol/L (1.05 g/L), glycaemia at 1 hour ≥ 10 mmol/L (1.80 g/L), glycaemia at 2 hours ≥ 9.1 mmol/L (1.65 g/L), glycaemia at 3 hours ≥ 8 mmol/L (1.45 g/L) after oral dose of 100 g glucose. Carpenter & Coustan: at least 2 threshold values selected: fasting glycaemia ≥ 5.3 mmol/L (0.95 g/L), glycaemia at 1 hour ≥ 10 mmol/L (1.80 g/L), glycaemia at 2 hours ≥ 8.6 mmol/L (1.55 g/L), glycaemia at 3 hours ≥ 7.8 mmol/L (1.40 g/L) after oral dose of 100 g glucose. WHO (World Health Organization): fasting glycaemia ≥ 7 mmol/L (1.26 g/L) and/or glycaemia -2 h ≥ 7.8 mmol/L (1.40 g/L) after oral dose of 75 g glucose.

Table 2 Risk score created for the Torento Tri-Hospital Gestational Diabetes cohort [48] Risk factors

Score

Odd ratio (CI 95%)

p

0

_

31-34

1

1.0 (0.7-1.5)

0.95

≥ 35

2

1.6 (1.1-2.5)

0.02

0 2

1.8 (1.1-2.7)

0.01

≥ 25.1

3

3.2 (2.1-4.8)

< 0.001

Ethnic origin 0

_

Afro-American

0

0.7 (0.3-1.7)

0.44

Asian

5

4.8 (3.0-7.6)

< 0.001

Other

p

< 25

-

25-34

1.94 (1.51-2.51)

< 0.001

≥ 35

4.27 (2.95-6.20)

< 0.001

< 20

_

22.1-25

Caucasian

OR adjusted (CI 95%)

BMI 1st visit (kg/m2)

BMI (kg/m2) ≤ 22

Risk Factors Age (years)

Age (years) ≤ 30

Table 3 GDM Risk factors in the study by Phaloprakarn et al. [50]

2

1.6 (0.7-3.5)

0.24

-

20.0-24.9

1.68 (1.27-2.22)

< 0.001

25-29.9

3.52 (2.52-4.93)

< 0.001

≥ 30

8.33 (5.14-13.5)

< 0.001

Nulliparous

0.73 (0.51-1.04)

0.08

History diabetes

4.29 (3.07-6.00)

< 0.001

History macrosomia

6.18 (2.40-15.90)

< 0.001

History ≥ 2 abortions

3.50 (1.36-9.02)

0.01

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

Risk score = 6 × age + 11 × BMI + 109 × family history of diabetes + 42 × history macrosomia + 49 × history ≥ 2 abortions. Each of the last three parameters was scored 1 if present and 0 if absent. A ROC curve was created. A score ≥ 380 gave, respectively, sensitivity, specificity, PPV and NPV values of 86.9%, 45%, 41.8% and 88.3%. Using the data from a prospective study of 995 women (NP2), Van Leeuwen et al. [51] developed a predictive model for GDM that integrated the following characteristics: ethnicity, family history, history of GDM and BMI. This model helped identify 75% of women with GDM and avoid screening in 47% of women at low risk.

2.2.3. Performances compared to two screening strategies 2.2.3.1. Performances In order to compare universal and selective screening, it is necessary to be able to estimate the number of cases of GDM that selective screening would miss and the number of unnecessary tests avoided with this strategy. To that end, general population studies are needed to analyze the prevalence of GDM in relation to the presence of risk factors. Table 4 presents a comparison of these studies and their results. The proportion of women with GDM that are not diagnosed by selective screening varies between 3 and 10% in most publications [48, 52-56], and can even reach higher than 30%, particularly in the last three French studies [57-59], or even 50% in some series [58, 60]. This is partially related

581

to the impossibility of using obstetrical history in childless women and unfamiliarity with family history. What is the prognosis of missed cases of GDM in selective screening, i.e. in low risk patients? No study capable of answering this question directly has been found. In the case-control study by Langer et al. [41], unrecognized GDM in women of normal weight (BMI < 25 kg/m2) was not associated with an increase in the incidence of macrosomia or shoulder dystocia (EL4). In the retrospective study by Cosson et al. on 1,515 women (16.6% prevalence of GDM with systematic screening with 75g OGTT), the fœtal/maternal prognosis for women with GDM with no risk factor was similar to that of women without GDM [61]. This is in contrast to the study by Langer et al. [41], in which diabetes in women without risk factors was known and treated. In studies in which selective screening missed few cases of GDM, it is important to consider the proportion of women exempt from screening in the target population. Several studies have reported that screening based on risk factors only averted very few screening tests, which unnecessarily complicated screening [52, 62-64]. On the contrary, other studies report that selective screening has a sensitivity and specificity comparable to those in universal screening but all while avoiding a high number of unnecessary tests [48, 53, 54]. In the retrospective Australian study by Davey et al. [53], comparing selective screening to universal screening, selective screening missed 0.6% of cases of GDM and helped avoid unnecessary screening in 17% of women (EL4). In the Danish prospective multicentre study by Jensen et al. [54], selective screening missed 7% of cases of GDM and helped avoid biological tests in two thirds of women (EL2).

Table 4 Studies comparing selective with universal screening Author year, country

Type of study No. of patients Diagnostic criteria

Prevalence of GDM in the total population

Selection criteria for the population at risk % population exempt if screening selective for risk

Prevalence of GDM in the population at risk

Dietrich [66] 1987, USA

Proportion of GDM missed by selective screening

Authors’ conclusions on selective screening

Prospective cohort N = 2000 100 g OGTT

2.1%

“Standard” risk factors (not specified) 55%

4.2% low risk 0.1 %

4.8%

In a patient pool from a private practice, selective screening helped avoid unnecessary tests in more than half of women

Naylor [48] 1997, Canada

Retrospective cohort N = 3131 100 g OGTT NDDG criteria

3.5%

Age > 30, BMI > 22, ethnicity, at risk 35%

Score ≤ 1 0.9% Score ≥ 6 18.7%

7%

A selective screening strategy selecting for clinical risk score is effective and helps many women avoid unnecessary screening

Moses [56] 1998, Australia

Retrospective cohort N = 2907 75 g OGTT Glycaemia 2 h ≥ 8 mmol/L (1.45 g/L)

6.3%

Selection low-risk population Age < 25, BMI < 25, Caucasian origin 20%

low risk: 2.8%

10%

Selective screening is not “reasonable” since low risk does not mean no risk

582

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

Author year, country

Type of study No. of patients Diagnostic criteria

Prevalence of GDM in the total population

Williams [64] 1999, USA

Retrospective cohort N = 25,118 100 g OGTT NDDG criteria

ND

Selection criteria for the population at risk % population exempt if screening selective for risk

Prevalence of GDM in the population at risk

Proportion of GDM missed by selective screening

Authors’ conclusions on selective screening

Low risk defined as: Age < 25, BMI < 27 Caucasian origin, absence family history of diabetes

ND

4%

No benefit as only 10% of women are exempt

Age ≥ 25 BMI ≥ 27, ethnicity, at risk, family history diabetes 10%

low risk

3%

Few missed GDM but too few tests saved for the extra degree of complexity with selective screening

Age ≥ 25 or ≥ 30, BMI ≥ 27, ethnicity, at risk, 1st degree history diabetes 17% if age ≥ 25 52% if age ≥ 30

low risk 0.2% if age < 25

0.6% if age ≥ 25

1.2% if age < 30

4% if age ≥ 30

Selective screening is effective and helps reduce the number of tests in certain populations

10% Danilenko [52] 1999, USA

Retrospective cohort N = 18,504 100 g OGTT NDDG criteria

3%

Davey [53] 2001, Australia

Retrospective case-control N = 6032 75 g OGTT Fasting glycaemia ≥ 5.5 mmol/L (1 g/L) or glycaemia 2 h ≥ 8 mmol/L (1.45 g/L)

5%

0.9 %

ACOG criteria 11% ADA criteria 3%

The low prevalence of GDM observed in the low-risk group suggests that selective screening is desirable if criteria are selected to define risks

JiménezMoleon [55] 2002, Spain

Retrospective cohort N = 2574 100 g OGTT NDDG criteria

3.31%

Age ≥ 25 (ADA criteria) or ≥ 30 (ACOG criteria) BMI > 27 or ≥ 30 Family history diabetes History macrosomia History GDM History hypertension History foetal death History congenital defects 45% (ACOG criteria) 15.5% (ADA criteria)

With ACOG criteria Low risk (0 factors) 0.6% High risk 4% With ADA criteria Low risk 0.5% High risk 2.9%

Yang [65] 2002, China

Prospective cohort N = 9741 75 g OGTT WHO criteria

1.8%

Age > 25, BMI ≥ 25, family history diabetes 24%

Low risk 0.9%

12%

If selective screening considered in this population, age and BMI limits should be reconsidered

Jensen [54] 2003, Denmark

Prospective cohort N = 5235 75 g OGTT Fasting glycaemia ≥ 6.1 mmol/L (1.11 g/L) or glycaemia 2 h ≥ 9 mmol/L (1.64 g/L)

2.4%

Glycosuria, history GDM, history macrosomia (weight ≥ 4 500 g), BMI ≥ 27, family history diabetes 64%

Population at risk 5.3% Low risk 0. %

7%

Single selective screening helps avoid a test in 2/3 of women with performance comparable to universal screening

Di Cianni [63] 2003, Italy

Retrospective cohort N = 3950 100 g OGTT Carpenter & Coustan criteria

12.3%

Age > 25, BMI > 25, family history diabetes 5.6%

Low risk 2.1%

1.5%

Studies on cost-efficacy ratio required before giving opinion

Corcoy [62] 2004, Spain

Retrospective cohort N = 1635 100 g OGTT NDDG criteria

12.8%

Age > 25, BMI > 25, family history diabetes, personal history of abnormal glucose tolerance, unfavourable obstetrical history 7%

Low risk 2.4%

1.3%

Few missed GDM but too few saved screenings, making selective screening unnecessarily complex

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

Author year, country

Type of study No. of patients Diagnostic criteria

Griffin [60] 2000, Ireland

Prospective randomized trial N = 1853 selective screening N = 1299 universal screening 100 g OGTT NDDG criteria

Cosson [61] 2006, France

Chevalier [57] 2009, France

Prevalence of GDM in the total population

Selection criteria for the population at risk % population exempt if screening selective for risk

Prevalence of GDM in the population at risk

2.7%

Weight > 100 kg 1st degree history diabetes History macrosomia > 4.5 kg History GDM History unexplained fœtal death History major congenital deformities Current pregnancy: glycosuria, macrosomia, hydramnios 86%

Comparative retrospective N = 1909 selective screening N = 2111 universal screening 75 g OGTT Fasting glycaemia > 5.3 mmol/L (0.95 g/L) or glycaemia 2 h > 7.8 mmol/L (1.40 g/L)

12,6 %

Age > 35 BMI > 27 Personal or family history glucose intolerance or diabetes History macrosomia History GDM, preeclampsia History in utero fœtal demise History congenital deformity Current pregnancy: glycosuria, macrosomia, hypertension

Prospective cohort N = 2014 75 g OGTT Fasting glycaemia > 5.3 mmol/L (0.95 g/L) or glycaemia 2 h > 7.8 mmol/L (1.40 g/L)

7.78%

BMI > 27, personal history GDM, macrosomia or IUFD family history diabetes, obesity 20%

The performances of selective screening vary according to the population studied, the frequency of risk factors in the screened population and the thresholds used [65]. In a population of predominantly young, thin, white pregnant women considered to be at low risk for GDM, most women are not candidates for selective screening. Selective screening thus helps many women avoid having to have a test [66]. In a Spanish retrospective study [55], the prevalence of GDM in a low risk obstetrical population that represented 45% of the total population was 6 times lower (0.5%) than in the high risk population (2.9%). Inversely, in an older, overweight obstetrical population, screening based on risk factors would probably lead to subjecting most pregnant women to screening. In North American populations where risk factors are high, selective screening could help detection in 90% of the obstetrical population [64].

583

Proportion of GDM missed by selective screening

Authors’ conclusions on selective screening

1.45%

46%

Universal screening better than selective screening in this Irish population

8.3%

34%

Universal screening could reduce time needed to make diagnosis and begin treatment, it is associated with a better fœtal prognosis

34%

The impact of universal screening associated with intensive therapeutic care in women with no risk factor needs to be specified

on an exhaustive list of risk factors could quickly be discontinued or only save a small percentage of the population from unnecessary tests and be of little interest in terms of a cost-effectiveness ratio. However, a two-part universal screening (O’ Sullivan test followed by a 100 g dose of glucose if results are positive) is constraining and could delay diagnosis and care [66]. Two regional practical enquiries carried out in France five years after the recommendations for universal screening for GDM, show insufficient screening [67-69]. The main limit to a selective approach is its complexity. Nevertheless, maternal age, BMI, a history of GDM or of macrosomia are data that are easily available during the first prenatal consultation [48]. Ethnic origin is slightly more delicate to use in practice (parents of mixed backgrounds, birthplace, etc.). 2.2.3.3. Benefits for complications

2.2.3.2. Feasibility Universal screening is theoretically easier than selective screening. A screening test that is too complex and based

In a Canadian retrospective registry study (NP4) on 1,729,225 women, among whom 38,274 had GDM, Wen et al. [44] concluded that the increase in the prevalence of GDM

584

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

was linked to the establishment of a universal screening policy: 0.3 vs. 2.7% and they suggested that cases diagnosed with universal screening were less severe than those diagnosed with selective screening. The authors compared two regions with different screening practices. In the region with universal screening, the prevalence (2.2%) of GDM was higher than where screening was selective (1%), but the frequency of perinatal complications was the same for macrosomia, hydramnios, preeclampsia and Caesareans, suggesting that universal screening had no obvious benefit in the incidence of complications related to GDM. In a randomized prospective study (EL3) comparing selective and universal screening [60], Griffin et al. reported a significant increase in the prevalence of GDM with universal screening (2.7% vs. 1.45%). They also reported that universal screening helped make an earlier diagnosis: 30 ± 2.6 WG vs. 33 ± 3.7, but the strategy used was the screening defined in their protocol (screening at 32 WG in the risk factor group and 26-28 WG in the universal screening group). Universal screening was associated with a lower prevalence of macrosomia, Caesarean, prematurity, preeclampsia and admission in neonatal wards. The risk factors selected in this study were unusual, such as weighing over 100 kg instead of using the BMI, thus excluding a high proportion of women at risk in cases of selective screening. In a retrospective study, Cosson et al. compared over time a cohort of women who had undergone selective screening and a cohort of women who had undergone universal screening (EL4) [58]. The women did not differ in their demographic, ethnic, BMI or GDM risk factor data. The authors report rates of neonatal complications (macrosomia, prematurity before 37 WG, jaundice and transfer) lower than in the population of diabetics from the second period (universal screening) compared to that of the first period (selective screening). They suggest that universal screening could have a beneficial effect by reducing diagnosis and treatment times. A higher prevalence of GDM should be noted in this study (8.3% with selective screening, 12.6% with universal screening) and a population of diverse ethnic origins, predominately foreign: North African (24%), African (23%), Asian (3.5%), and Indian (3.7%). Data on glycaemic control was unknown. The differences in the results observed could be due to the fact that diabetes detected with universal screening is less severe. The same authors reported that neonatal and maternal prognoses of women with GDM were influenced by the presence of risk factors (family history of diabetes, BMI > 27 kg/m2, age > 35 years, history of GDM, preeclampsia, macrosomia and in utero foetal demise). The presence of risk factors exposed patients with GDM to a higher risk of in utero fœtal demise and a low Apgar score at birth. In the absence of risk factors, the prognosis was similar to that for women without GDM [61]. In the prospective study by Chevalier et al. [68] on 2,014 women in the south west of France who underwent universal screening with the 75 g OGTT (GDM defined as fasting glycaemia > 0.95 g/L and/or glycaemia at 2 h >

1.40), the prevalence of GDM was 7.8% and 34% of the women had no GDM risk factor, defined as: BMI > 27 kg/ m2, personal history of GDM, macrosomia or IUFD, family history of diabetes and obesity. In the group with no risk factors, intensive treatment was associated with a rate of macrosomia (birth weight > 90th percentile) of 3.4% and a rate of hypotrophia (birth weight < 10th percentile) of 20% vs., respectively, 10% and 13.3% in the GDM group with risk factors. It was the only study that reported an association between treatment and neonatal hypotrophia in patients with a low risk for GDM. 2.2.3.4. Cost-effectiveness analysis We could not find any studies with a high level of proof that analyzed cost-effectiveness ratios for universal and selective screening. In an Italian retrospective study by Di Cianni et al. that compared a group of 1,338 women who had undergone universal screening and a group of 4,035 women screened for risk factors, the cost-effectiveness ratio for universal screening was considered to be favourable [70]. There was however a major bias in the universal screening group receiving intensive treatment and the selective screening group receiving a conventional treatment (EL4). In a prospective French study (EL2) carried out in 1999 on 120 women, Poncet et al. compared three screening strategies [71]: - screening for risk factors (50 g glucose then 100 g OGTT for diagnostic purposes). The risk factors selected were as follows: first degree history of diabetes, age > 35 years, BMI > 27 kg/m2, history of GDM, preeclampsia or macrosomia in a previous pregnancy; - 2 step universal screening (50 g then 100 g OGTT for diagnostic purposes); - 1 step universal screening (75 g OGTT). The screening strategy for risk factors had the most favourable cost-effectiveness ratio. Compared to selective screening, a two-step universal screening multiplied costs by 1.1 and the single step universal screening by between 3.27 and 3.75 for an avoided complication (macrosomia, prematurity, perinatal mortality or hypertension). • Overall, ideally, the screening strategy selected should help identify women at high risk for pathological events who would best benefit from intensive medical care and preserve others from the burden of excessive intervention, which could potentially occasion anxiety or Caesareans. In pragmatic terms and in keeping with efficacy, this would help concentrate means and not disperse them. • Selective screening that is too complex and based on an exhaustive list of risk factors could be discontinued by physicians. However, maternal age, 1st degree history of diabetes, BMI, history of GDM or macrosomia is data systematically recorded during the first prenatal consultation.

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

• Among these risk factors, the major impact of BMI should be highlighted in the risk of onset of GDM. It is also an independent risk factor for perinatal complications (macrosomia, preeclampsia). In low-risk populations, selective screening helps avoid testing in a high number of patients. Inversely, in high-risk obstetrical populations, and particularly given the prevalence of overweight or obese women in certain regions in France, it would only save a low proportion of women and have little interest in terms of cost/effectiveness. Selective screening could also lead to missing a high number of GDM for which the prognosis remains to be evaluated. • Universal screening is theoretically easier to organize. There is however an increase in the prevalence of GDM to take into account in estimations of health needs required for the medical care for these patients, a fortiori if the screening thresholds proposed by IADPS are selected (expected prevalence of 16 to 18% of the population). The number of unnecessary tests and the cost-effectiveness ratio of screening low-risk women for which the benefit for perinatal complications remains controversial.

3. Conclusion The benefits of GDM screening and the resulting medical care have only been shown in patients presenting GDM risk factors (older age, high BMI, history of GDM or macrosomia and 1st degree history of diabetes). Its interest for patients with no risk factors remains controversial.

[9]

[10]

[11]

[12]

[13] [14]

[15] [16]

[17]

[18]

[19]

[20]

[21] [22]

4. Conflict of interests No conflict of interests related to the article. Références [1]

[2]

[3] [4]

[5]

[6]

[7] [8]

Haute Autorité de Santé. Guide méthodologique: Comment évaluer a priori un programme de dépistage? 2004;http://www.has-sante.fr/ portail/jcms/c_268095/guide-programmedepistagerappdf. Vendittelli F, Riviere O, Crenn-Hebert C, Claris O, Tessier V, Pinquier D, et al. Réseau sentinelle Audipog 2004–2005. Partie 1 : Résultats des principaux indicateurs périnatals. Gynecol Obstet Fertil 2008;36:1091-100. O’Sullivan JB, Charles D, Mahan CM, Dandrow RV. Gestational diabetes and perinatal mortality rate. Am J Obstet Gynecol 1973;116:901-4. Sepe SJ, Connell FA, Geiss LS, Teutsch SM. Gestational diabetes. Incidence, maternal characteristics, and perinatal outcome. Diabetes 1985;34 Suppl 2:13-6. Johnstone FD, Nasrat AA, Prescott RJ. The effect of established and gestational diabetes on pregnancy outcome. Br J Obstet Gynaecol 1990;97:1009-15. Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, Coustan DR, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 2008;358:1991-2002. Kjos SL, Buchanan TA. Gestational diabetes mellitus. N Engl J Med 1999;341:1749-56. Casey BM, Lucas MJ, McIntire DD, Leveno KJ. Pregnancy outcomes in women with gestational diabetes compared with the general obstetric population. Obstet Gynecol 1997;90:869-73.

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

585

Brody SC, Harris R, Lohr K. Screening for gestational diabetes: a summary of the evidence for the U.S. Preventive Services Task Force. Obstet Gynecol 2003;101:380-92. Stotland NE, Caughey AB, Breed EM, Escobar GJ. Risk factors and obstetric complications associated with macrosomia. Int J Gynaecol Obstet 2004;87:220-6. Yogev, Chen, Hod, Coustan, Oats, McIntyre, et al. Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study: preeclampsia. Am J Obstet Gynecol 2010;202:255 e1-7. Ben-Haroush A, Yogev Y, Hod M. Epidemiology of gestational diabetes mellitus and its association with Type 2 diabetes. Diabet Med 2004;21:103-13. Kim C, Newton KM, Knopp RH. Gestational diabetes and the incidence of type 2 diabetes: a systematic review. Diabetes Care 2002;25:1862-8. Lauenborg J, Hansen T, Jensen DM, Vestergaard H, Molsted-Pedersen L, Hornnes P, et al. Increasing incidence of diabetes after gestational diabetes: a long-term follow-up in a Danish population. Diabetes Care 2004;27:1194-9. O’Sullivan JB. Diabetes mellitus after GDM. Diabetes 1991;40 Suppl 2:131-5. Vambergue A, Dognin C, Boulogne A, Rejou MC, Biausque S, Fontaine P. Increasing incidence of abnormal glucose tolerance in women with prior abnormal glucose tolerance during pregnancy: DIAGEST 2 study. Diabet Med 2008;25:58-64. Ratner RE, Christophi CA, Metzger BE, Dabelea D, Bennett PH, Pi-Sunyer X, et al. Prevention of diabetes in women with a history of gestational diabetes: effects of metformin and lifestyle interventions. J Clin Endocrinol Metab 2008;93:4774-9. Chen Y, Quick WW, Yang W, Zhang Y, Baldwin A, Moran J, et al. Cost of gestational diabetes mellitus in the United States in 2007. Popul Health Manag 2009;12:165-74. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 2010;33:676-82. Catalano PM, Avallone DA, Drago NM, Amini SB. Reproducibility of the oral glucose tolerance test in pregnant women. Am J Obstet Gynecol 1993;169:874-81. Harlass FE, Brady K, Read JA. Reproducibility of the oral glucose tolerance test in pregnancy. Am J Obstet Gynecol 1991;164:564-8. Espinosa de los Monteros A, Parra A, Carino N, Ramirez A. The reproducibility of the 50-g, 1-hour glucose screen for diabetes in pregnancy. Obstet Gynecol 1993;82(4 Pt 1):515-8. Weiss PA, Haeusler M, Kainer F, Purstner P, Haas J. Toward universal criteria for gestational diabetes: relationships between seventy-five and one hundred gram glucose loads and between capillary and venous glucose concentrations. Am J Obstet Gynecol 1998;178:830-5. Agarwal MM, Punnose J, Dhatt GS. Gestational diabetes: problems associated with the oral glucose tolerance test. Diabetes Res Clin Pract 2004;63:73-4. Daniells S, Grenyer BF, Davis WS, Coleman KJ, Burgess JA, Moses RG. Gestational diabetes mellitus: is a diagnosis associated with an increase in maternal anxiety and stress in the short and intermediate term? Diabetes Care 2003;26:385-9. Rumbold AR, Crowther CA. Women’s experiences of being screened for gestational diabetes mellitus. Aust N Z J Obstet Gynaecol 2002;42:131-7. Feig DS, Chen E, Naylor CD. Self-perceived health status of women three to five years after the diagnosis of gestational diabetes: a survey of cases and matched controls. Am J Obstet Gynecol 1998;178:386-93. Kerbel D, Glazier R, Holzapfel S, Yeung M, Lofsky S. Adverse effects of screening for gestational diabetes: a prospective cohort study in Toronto, Canada. J Med Screen 1997;4:128-32. Sjogren B, Robeus N, Hansson U. Gestational diabetes: a case-control study of women’s experience of pregnancy, health and the child. J Psychosom Res 1994;38:815-22. Langer N, Langer O. Emotional adjustment to diagnosis and intensified treatment of gestational diabetes. Obstet Gynecol 1994;84:329-34. Crowther CA, Hiller JE, Moss JR, McPhee AJ, Jeffries WS, Robinson JS. Effect of treatment of gestational diabetes mellitus on pregnancy outcomes. N Engl J Med 2005;352:2477-86.

586

S. Hiéronimus, J.-P.L. Meaux / Diabetes & Metabolism 36 (2010) 575–586

[32] Spirito A, Williams C, Ruggiero L, Bond A, McGarvey ST, Coustan D. Psychological impact of the diagnosis of gestational diabetes. Obstet Gynecol 1989;73:562-6. [33] Naylor CD, Sermer M, Chen E, Sykora K. Cesarean delivery in relation to birth weight and gestational glucose tolerance: pathophysiology or practice style? Toronto Trihospital Gestational Diabetes Investigators. JAMA 1996;275:1165-70. [34] Horvath K, Koch K, Jeitler K, Matyas E, Bender R, Bastian H, et al. Effects of treatment in women with gestational diabetes mellitus: systematic review and meta-analysis. BMJ 2010;340:c1395. [35] Alwan N, Tuffnell DJ, West J. Treatments for gestational diabetes. Cochrane Database Syst Rev 2009:CD003395. [36] O’Sullivan JB, Gellis SS, Dandrow RV, Tenney BO. The potential diabetic and her treatment in pregnancy. Obstet Gynecol 1966;27:683-9. [37] Langer O, Anyaegbunam A, Brustman L, Divon M. Management of women with one abnormal oral glucose tolerance test value reduces adverse outcome in pregnancy. Am J Obstet Gynecol 1989;161:593-9. [38] Bonomo M, Corica D, Mion E, Goncalves D, Motta G, Merati R, et al. Evaluating the therapeutic approach in pregnancies complicated by borderline glucose intolerance: a randomized clinical trial. Diabet Med 2005;22:1536-41. [39] Landon MB, Spong CY, Thom E, Carpenter MW, Ramin SM, Casey B, et al. A multicenter, randomized trial of treatment for mild gestational diabetes. N Engl J Med 2009;361:1339-48. [40] Bancroft K, Tuffnell DJ, Mason GC, Rogerson LJ, Mansfield M. A randomised controlled pilot study of the management of gestational impaired glucose tolerance. BJOG 2000;107:959-63. [41] Langer O, Yogev Y, Most O, Xenakis EM. Gestational diabetes: the consequences of not treating. Am J Obstet Gynecol 2005;192:989-97. [42] Moss JR, Crowther CA, Hiller JE, Willson KJ, Robinson JS. Costs and consequences of treatment for mild gestational diabetes mellitus – evaluation from the ACHOIS randomised trial. BMC Pregnancy Childbirth 2007;7:27. [43] Nicholson WK, Fleisher LA, Fox HE, Powe NR. Screening for gestational diabetes mellitus: a decision and cost-effectiveness analysis of four screening strategies. Diabetes Care 2005;28:1482-4. [44] Wen SW, Liu S, Kramer MS, Joseph KS, Levitt C, Marcoux S, et al. Impact of prenatal glucose screening on the diagnosis of gestational diabetes and on pregnancy outcomes. Am J Epidemiol 2000;152:1009-14; discussion 15-6. [45] Yeung EH, Hu FB, Solomon CG, Chen L, Louis GM, Schisterman E, et al. Life-course weight characteristics and the risk of gestational diabetes. Diabetologia 2010;53:668-78. [46] Torloni MR, Betran AP, Horta BL, Nakamura MU, Atallah AN, Moron AF, et al. Prepregnancy BMI and the risk of gestational diabetes: a systematic review of the literature with meta-analysis. Obes Rev 2009;10:194-203. [47] Khine ML, Winklestein A, Copel JA. Selective screening for gestational diabetes mellitus in adolescent pregnancies. Obstet Gynecol 1999;93(5 Pt 1):738-42. [48] Naylor CD, Sermer M, Chen E, Farine D. Selective screening for gestational diabetes mellitus. Toronto Trihospital Gestational Diabetes Project Investigators. N Engl J Med 1997;337:1591-6. [49] 49. Berger H, Sermer M. Counterpoint: Selective screening for gestational diabetes mellitus. Diabetes Care 2009 Jul;32:1352-4. [50] Phaloprakarn C, Tangjitgamol S, Manusirivithaya S. A risk score for selective screening for gestational diabetes mellitus. Eur J Obstet Gynecol Reprod Biol 2009;145:71-5. [51] van Leeuwen M, Opmeer BC, Zweers EJ, van Ballegooie E, ter Brugge HG, de Valk HW, et al. Estimating the risk of gestational diabetes mellitus: a clinical prediction model based on patient characteristics and medical history. BJOG 2009;117:69-75. [52] Danilenko-Dixon DR, Van Winter JT, Nelson RL, Ogburn PL, Jr. Universal versus selective gestational diabetes screening: application of 1997 American Diabetes Association recommendations. Am J Obstet Gynecol 1999;181:798-802.

[53] Davey RX, Hamblin PS. Selective versus universal screening for gestational diabetes mellitus: an evaluation of predictive risk factors. Med J Aust 2001;174:118-21. [54] Jensen DM, Molsted-Pedersen L, Beck-Nielsen H, Westergaard JG, Ovesen P, Damm P. Screening for gestational diabetes mellitus by a model based on risk indicators: a prospective study. Am J Obstet Gynecol 2003 Nov;189:1383-8. [55] Jimenez-Moleon JJ, Bueno-Cavanillas A, Luna-Del-Castillo JD, Garcia-Martin M, Lardelli-Claret P, Galvez-Vargas R. Prevalence of gestational diabetes mellitus: variations related to screening strategy used. Eur J Endocrinol 2002;146:831-7. [56] Moses RG, Moses J, Davis WS. Gestational diabetes: do lean young caucasian women need to be tested? Diabetes Care 1998;21:1803-6. [57] Chevalier A, Hiéronimus S, Giaume V, Brucker-Davis F, A. B, Fénichel P. Obstetrical outcomes in pregnancies with gestational diabetes: what benefits? Which patients? Diabetologia 2009;52(Suppl 1):S45. [58] Cosson E, Benchimol M, Carbillon L, Pharisien I, Paries J, Valensi P, et al. Universal rather than selective screening for gestational diabetes mellitus may improve fetal outcomes. Diabetes Metab 2006;32:140-6. [59] Vambergue A, Nuttens M, Lemaire C, Goeusse P, Biausque S, Verier-Mine O, et al. Diabète gestationnel : dépistage systématique plutôt qu’orienté par des facteurs de risque. Diabetes Metab 1995;VIII(O24). [60] Griffin ME, Coffey M, Johnson H, Scanlon P, Foley M, Stronge J, et al. Universal vs. risk factor-based screening for gestational diabetes mellitus: detection rates, gestation at diagnosis and outcome. Diabet Med 2000;17:26-32. [61] Cosson E, Benchimol M, Carbillon L, Pariès J, Lormeau B, SandreBacon D, et al. Gestational diabetes mellitus : role of the risk factors on fetal and maternal prognosis. Diabetes Metab 2006;32(suppl. 1):2611-PO. [62] Corcoy R, Garcia-Patterson A, Pau E, Pascual E, Altirriba O, Adelantado JM, et al. Is selective screening for gestational diabetes mellitus worthwhile everywhere? Acta Diabetol 2004;41:154-7. [63] Di Cianni G, Volpe L, Lencioni C, Miccoli R, Cuccuru I, Ghio A, et al. Prevalence and risk factors for gestational diabetes assessed by universal screening. Diabetes Res Clin Pract 2003;62:131-7. [64] Williams CB, Iqbal S, Zawacki CM, Yu D, Brown MB, Herman WH. Effect of selective screening for gestational diabetes. Diabetes Care 1999;22:418-21. [65] Yang X, Hsu-Hage B, Yu L, Simmons D. Selective screening for gestational diabetes in Chinese women. Diabetes Care 2002;25:796. [66] Dietrich ML, Dolnicek TF, Rayburn WF. Gestational diabetes screening in a private, midwestern American population. Am J Obstet Gynecol 1987;156:1403-8. [67] Batallan A, Goffinet F, Paris-Llado J, Fortin A, Breart G, Madelenat P, et al. Macrosomie fœtale : pratiques, conséquences obstétricales et néonatales. Enquête multicentrique cas-témoins menée dans 15 maternités de Paris et d’Île de France. Gynecol Obstet Fertil 2002;30:483-91. [68] Chevalier N, Hiéronimus S, Bongain A, Fénichel P. Évaluation du dépistage systématique du diabète gestationnel au CHU de Nice : à propos d’une cohorte prospective de 1 383 patientes. Diabetes Metab 2008;4(Supplement 3):H65. [69] Touzet S, Rocher L, Dureau-Drevard E, Poncet B, Colin C, Orgiazzi J, et al. Étude d’observation des pratiques de dépistage du diabète gestationnel à partir d’une cohorte de 701 femme. J Gynecol Obstet Biol Reprod (Paris) 2002;31:248-55. [70] Di Cianni G, Volpe L, Casadidio I, Bottone P, Marselli L, Lencioni C, et al. Universal screening and intensive metabolic management of gestational diabetes: cost-effectiveness in Italy. Acta Diabetol 2002;39:69-73. [71] Poncet B, Touzet S, Rocher L, Berland M, Orgiazzi J, Colin C. Costeffectiveness analysis of gestational diabetes mellitus screening in France. Eur J Obstet Gynecol Reprod Biol 2002;103:122-9.