The presence of diabetes before pregnancy is well known to be a risk factor for adverse neonatal outcomes, including increased rates of perinatal mortality, congenital anomaly, and macrosomia.1 In 1989, the St. Vincent Declaration in Europe made it a healthcare goal to improve outcomes of diabetic pregnancies such that the incidence of adverse outcomes approached those of the general population.2 Since 1989, care of diabetes in general and during pregnancy has changed; however, population-based studies show that the goals of the St. Vincent Declaration have not been reached.3–14 However, in selected populations with intensive diabetes management before conception and during pregnancy, such as participants in the Diabetes Control and Complications Trial (1996), rates of spontaneous abortion and congenital malformation have approached rates in the nondiabetic population.15 Because good glycemic control is necessary in the very early stages of pregnancy to reduce the incidence of congenital anomalies, prepregnancy counseling has been found to improve outcomes significantly.16 However, with the exception of some reports from Scandinavia that did not find a significant increase in death rates of infants born to diabetic women,17,18 most population-based studies of nonselected diabetic mothers have shown little improvement in pregnancy outcomes.3–14 The published population-based data regarding outcome of diabetic pregnancy varies among reporting centers in Europe, and no population-based North American data are available.
We report here the results of a large population-based study of diabetic pregnancy outcomes based on North American data. The objectives of this cohort study were to 1) compare rates of perinatal mortality, congenital malformations, and growth disturbances between women with pregestational diabetes and nondiabetic women and 2) determine whether there was a change in incidence of these adverse outcomes between 1988 and 2002 in infants of diabetic mothers.
MATERIALS AND METHODS
A population-based cohort study design was used to compare pregnancy outcomes in all nondiabetic mothers and mothers with pregestational diabetes in Nova Scotia, Canada delivering between January 1, 1988 and December 31, 2002. Outcomes in infants of diabetic mothers were examined in two epochs, 1988 to 1995, and 1996 to 2002, to determine whether changes in diabetic pregnancy management in Nova Scotia instituted in the mid 1990s translated to change in pregnancy outcomes.
The population of Nova Scotia is approximately 900,000 and is predominantly white, with only 4% of the population being visible minorities.19 Births in Nova Scotia take place in 11 maternity units throughout the province, with approximately one half occurring at the Izaak Walton Killam Health Centre in Halifax, the only tertiary care center in the province. The prevalence of diabetes in the Nova Scotia adult population was estimated to be 4.06% in 1995, indicating there were approximately 36,540 diabetic patients in the province in that year (LeBlanc J. The prevalence of diabetes mellitus in Nova Scotia: what can we learn from routinely collected health data? Diabetes Care in Nova Scotia: a newsletter of the Diabetes Care Program of Nova Scotia. 1998;8:1–3). Changes in the organization of care of diabetes in pregnancy occurred in the mid 1990s including the establishment of a multidisciplinary pregnancy and diabetes clinic and the establishment of local guidelines. In addition, general diabetes care was changing after publication of the results of the Diabetes Control and Complications Trial in 1993.20 The Pregnancy and Diabetes Education Centre at the Izaak Walton Killam Health Centre opened in 1995 and currently follows approximately 40 pregnancies per year in women with preexisting diabetes from throughout the province. Most women are not referred to the clinic until several weeks into the pregnancy and some are fully managed in their local centers. In 1993, the Diabetes Care Program of Nova Scotia, a provincial organization that sets standards for diabetes care in the province, published guidelines for management of diabetic pregnancy. These were disseminated throughout the province and an update was produced in 2000.21
Data were obtained from the Nova Scotia Atlee Perinatal Database, which is located at the Izaak Walton Killam Health Centre and maintained and administered by the Reproductive Care Program of Nova Scotia. Since 1988, the Nova Scotia Atlee Perinatal Database has collected data on standardized clinical forms used uniformly across the province. Information collected includes demographic variables, detailed clinical information on maternal medical conditions, labor and delivery events, and neonatal outcomes for all pregnancies reaching 20 weeks of gestation and 500 g at birth in Nova Scotia, Canada, and for most out of province deliveries among Nova Scotia residents. It also includes follow-up information on death and cause of death in the first year of life. Standardized data collection forms (which are also used as clinical care tools) are used throughout the province. After discharge from hospital after delivery, data are abstracted from the medical records by health records personnel and entered in an Oracle database (Oracle, Redwood Shores, CA). Data has been shown to be reliable by validation studies.22 An ongoing data quality assurance program, which includes periodic abstraction studies, shows that the information in the database is reliable.
Maternal and fetal characteristics extracted from the database include age, parity, diabetes by White's classification,23 prepregnancy weight and weight gain during pregnancy, smoking status at first perinatal visit, chronic and pregnancy-induced hypertension (includes preeclampsia), gestational age and weight at birth, and method of delivery. Diabetes status was abstracted from the standardized antenatal record completed at the first antenatal visit and thus documented as present before pregnancy (White's classes B, C, D, F, and R). Thus, women with insulin-dependent diabetes mellitus, non–insulin-dependent diabetes mellitus requiring insulin before pregnancy, and non–insulin-dependent diabetes mellitus not requiring insulin before pregnancy were included. Pregnancies complicated by gestational diabetes (White's class A) were excluded from the analysis (n=4,067). Infants whose gestational ages or birth weights were unknown were excluded from the analysis. Only singleton pregnancies were included.
The primary neonatal outcomes of interest were perinatal mortality, major congenital anomalies, macrosomia, and fetal growth restriction. Perinatal mortality was defined as death between 20 weeks of gestation and day 7 after birth.24 Stillbirth was defined as death between 20 weeks of gestation and before birth.24 Infant mortality was defined as death after birth and before 1 year of age, these data being available for the population studied.24 Major congenital anomaly was defined in the database as one which is any of the following: lethal, life-shortening, life-threatening, requires major surgery, or affecting in a significant way the quality of life. Chromosomal abnormalities are included in this definition. Macrosomia was defined as birthweight more than the 90th percentile (large for gestational age, LGA) and fetal growth restriction as birthweight less than the 10th percentile (small for gestational age, SGA) for Canadian infants.25 Ethics approval (#3012) for the study was obtained from the Research Ethics Boards of the Izaak Walton Killam Health Centre and Dalhousie University, both in Halifax, Nova Scotia.
Data were analyzed using descriptive statistics as means±standard deviations unless otherwise specified. Continuous variables were analyzed using Student t test, and categorical variables by means of χ2 analysis, or Fisher exact test when appropriate. Relative risk (RR) with 95% confidence intervals (CIs) and attributable risk were calculated using standard methods. P values less than .05 were considered significant. Outcomes reaching statistical significance on univariable analysis were entered into a backward conditional regression to obtain adjusted RR. For the statistical analysis, SPSS for Windows 11.5 (SPSS Inc., Chicago, IL) was used.
There were 516 singleton infants of mothers with preexisting diabetes and 150,589 singleton infants of nondiabetic mothers born between January 1988 and December 2002. Characteristics of mothers and newborns are shown in Table 1. Women with preexisting diabetes were on average older and heavier at conception, and more likely to have chronic and pregnancy-induced hypertension compared with nondiabetic mothers. The majority of women had diabetes of Whites Class B (53.7%) or C (31.4%). Fewer women had diabetes of Whites class D (10.7%), F (1.2%), or R (3.1%).
Of the 516 diabetic pregnancies, there were 13 (2.5%) deaths in total. Five of the infants were stillborn. Four deaths occurred between birth and 7 days of age, three between 7 and 28 days, and one between 28 days and 1 year of age (Table 2). Major congenital anomaly was stated as the cause of five of the deaths (38.5%) among infants of diabetic mothers: one in the perinatal period, and the remainder by 1 year of age.
Perinatal mortality and infant mortality rates for infants of diabetic mothers were significantly higher than in the nondiabetic population (Table 3). After the rate of perinatal mortality was adjusted for major congenital anomaly, preterm birth, and preexisting maternal hypertension, maternal diabetes was not a statistically significant contributor to perinatal mortality (P=.760). The rate of premature birth (defined as birth before 37 weeks gestation) was significantly higher in infants of diabetic mothers, with most of these being near-term births (between 34 and 36 weeks) (Table 1).
There were 47 infants (9.1%) with one or more major congenital anomaly (Table 4). Of these, the majority of anomalies were cardiac. The most common cardiac anomaly was ventricular septal defect. Musculoskeletal, central nervous system, and ear, nose, and throat anomalies, as well as hypospadias complex were also relatively overrepresented. Maternal diabetes remained a significant contributor to infant congenital anomaly after adjusting for maternal smoking and age (P<.001).
The rate of cesarean delivery in our entire population was high, but was particularly high in infants born to mothers with preexisting diabetes (49.0% compared with 19.5%, P<.001). Infants of diabetic mothers who were LGA were not more likely to be born by cesarean delivery than infants of diabetic mothers who were not LGA (P=.251). Diabetic women who delivered vaginally were more likely than nondiabetic women to require assistance with forceps or vacuum (25.1% compared with 13.8%, P<.001).
The rate of LGA among infants of diabetic mothers in our population was remarkably high at 45.2%, compared with 12.6% in infants of nondiabetic mothers (Table 1). Maternal diabetes remained an independent risk factor for macrosomia, even after adjusting for prepregnancy weight and weight gain during pregnancy (P<.001). We found that infants of diabetic mothers were less likely to be born small for gestational age (P<.001).
A similar number of infants were born to diabetic mothers between 1988 and 1995 (256) and 1996 and 2002 (260), whereas the number of births in nondiabetic women decreased. In infants of diabetic mothers, there was no change over time in perinatal mortality, stillbirth, rate of congenital anomalies, or incidence of LGA infants (Table 5).
Infants of diabetic mothers in our population continue to suffer from a disproportionately high incidence of adverse fetal and neonatal outcomes compared with infants born to nondiabetic mothers. Infants born to diabetic mothers in Nova Scotia have a three times greater rate of dying in the perinatal period, nine times greater rate of death in the first year of life, three times greater rate of severe congenital anomaly, and 3.6 times greater rate of LGA birth compared with infants of nondiabetic women. Infants of diabetic mothers have more than five-fold greater risk of preterm birth, a major source of morbidity. It is of concern that despite improvements in organization of diabetes care, there was not a statistically significant improvement in these outcomes over the period studied. The goals of the St. Vincent Declaration have not been achieved.
This report is one of the largest recent population-based studies available and spans a period during which diabetes care has changed (MEDLINE; January 1966–November 2005; English language; search terms: “diabetes,” “pregnancy,” and “population-based”). These data are likely to be representative of other populations with similar ethnic representation and access to universal health care but may not be applicable to significantly different populations. The only recent study with larger numbers of infants is a population-based study from Denmark of 1,215 infants born to mothers with insulin-dependent diabetes mellitus between 1993 and 1999.13 Keeping in mind that it is difficult to compare results of studies that use different inclusion criteria and outcome definitions, our results in infants of diabetic women seem similar to those of previously published population-based studies from Europe.3–14
The perinatal mortality rate of 17.4 per 1,000 in infants of diabetic mothers was lower than the rates of 24.6 to 48.0 per thousand reported in other population-based studies published in the last 10 years.5,6 The RR of death in the perinatal period in Nova Scotia infants of diabetic mothers is also at the lower end of the range of 2.5–5.4 previously reported. The reasons for this are unclear; most previous studies are from Scandinavia, the United Kingdom, or Western Europe, where there is access to a similar system of socialized medicine; however, it is possible that the ethnic homogeneity of the Nova Scotia population is a factor. The only study reporting a lower RR at 2.5 included only infants of mothers with non–insulin-dependent diabetes mellitus and may thus reflect a less severely affected population.6 After adjusting for major congenital anomaly, maternal hypertension, and preterm birth, maternal diabetes was no longer a significant factor in perinatal mortality. This suggests that these three factors are important contributors to the excess mortality in infants of diabetic mothers and must be targeted to reduce perinatal mortality. Congenital anomalies are known to be related to poor maternal glycemic control in the early weeks of pregnancy, the period of organogenesis.26 Therefore, if improvements in congenital anomaly rates are to occur, efforts to ensure that women are referred to specialized diabetes in pregnancy clinics for preconceptional counseling and early pregnancy care must be targeted.16 In our center, most women are referred to the clinic when they are already several weeks into their pregnancy. For these women, targeting tight glycemic control and management of hypertension during the child bearing years is important. Further study to determine factors associated with preterm birth in women with diabetes is warranted.
The rate of congenital anomaly in infants of diabetic mothers in Nova Scotia at 91 per 1,000 is similar to many recently reported population-based studies that range from 83 to 99 per 1,000.3,5–7,10 However, two reports from Scandinavia9,13 and one from Scotland4 including only infants of women with insulin-dependent diabetes mellitus, have found lower rates of anomalies, ranging from 39 to 60 per 1,000. This wide variation in congenital anomaly rates may be due in part to differing definitions of major congenital anomaly and to differences in detection and pregnancy termination rates. Because we did not have access to data concerning pregnancies resulting in termination or pregnancies ending before 20 weeks of gestation or before the infant weight had reached 500 g, it is likely that our study underestimates the rates of congenital anomaly in our population, especially in the more recent years when improved detection methods may have led to a greater number of early terminations. These data were available in some other population-based studies, even those reporting lower anomaly rates.4,13 Therefore, our population has a relatively high rate of congenital anomaly compared with other recent reports, and the reasons for this must be sought.
We report a very high rate of LGA infants at almost one half (45.2%) of all births to diabetic women. It is difficult to compare our numbers with those of studies including only women with insulin-dependent diabetes mellitus who are more likely to have low birth weight babies; however, this rate of LGA is high even compared with the 28% found by Dunne et al6 in women with non–insulin-dependent diabetes mellitus, but similar to a recent report from Scotland.4 Infants of diabetic mothers in our population also had higher rates of complications associated with LGA, including shoulder dystocia, fractures, and brachial plexus injury compared with other infants. The rate of cesarean delivery in our population of diabetic women was also high at 49.0%; however, we found that nonmacrosomic infants were just as likely as macrosomic infants to be born by cesarean birth (P=.251). Despite improvements in the organization of management of diabetic pregnancy care in Nova Scotia, this population-based study revealed trends toward, but statistically insignificant improvements in perinatal mortality and LGA rates for infants of diabetic mothers between 1988 and 2002 and no change in rates of congenital anomalies. The reason for the lack of improvement is unclear. The population of women with diabetes experiencing pregnancy may have changed during this time, because there has been an increase in obesity in pregnancy,27 perhaps leading to more non–insulin-dependent diabetes mellitus among the latter group. It is also possible that the number of deaths was too small to show a statistically significant difference in mortality over time, because there were only nine perinatal deaths among infants of diabetic mothers in the entire 15-year period. During this same period, perinatal mortality rates decreased significantly in nondiabetic women.
The strengths of this study include the large number of births, the population-based design, with the availability of complete data regarding outcomes on all births to women with and without diabetes in the same period, and the high quality of the data in the Nova Scotia Atlee Perinatal Database. The quality of the Nova Scotia Atlee Perinatal Database was demonstrated when it was used to validate stillbirth and live birth infant death files by Statistics Canada.22 In this validation study, data from Statistics Canada were shown to be 92% to 99% complete for stillbirths and infant deaths respectively when evaluated against the Nova Scotia Atlee Perinatal Database. Our current study has a few limitations. The data available did not allow differentiation between type 1 and type 2 diabetes, but because diabetes was classified using White's Classification, gestational diabetes was easily excluded. Clinical information regarding glycemic control and attendance at a specialized diabetes clinic during pregnancy was not available. Finally, the database does not contain information regarding terminations or pregnancy losses before 20 weeks gestation or 500 g fetal weight. Despite these limitations, this study does provide valuable information in that it represents one of only a few Canadian studies on the subject, and is the first population-based study in North America to report neonatal outcomes in pregnancies in women with type 1 and type 2 diabetes.
Putting the findings of this study together, we see a very high rate of LGA, preterm birth, and cesarean delivery, and a relatively high rate of congenital anomalies compared with other reports of outcomes of diabetic pregnancy. All of these outcomes may be related to suboptimal glycemic control during pregnancy. The study spans a period of change in the organization of diabetes care, and it was hypothesized that outcomes would improve over time. We, however, failed to demonstrate a significant improvement in the major outcomes studied. This points out the importance of reliably documenting outcomes of diabetic pregnancy in populations and avoiding the assumption that changes in care will automatically translate into changes in outcomes. Careful examination of maternal hemoglobin A1c (preconception and during pregnancy) and attendance or nonattendance at the diabetes and pregnancy clinic is an essential next step in understanding the reasons for the persistence of increased rates of adverse outcomes of diabetic pregnancy in Nova Scotia.
1. Walkinshaw SA. Pregnancy in women with pre-existing diabetes: management issues. Semin Fetal Neonatal Med 2005;10:307–15.
2. Diabetes care and research in Europe: the Saint Vincent declaration. Diabet Med 1990;7:360.
3. Platt MJ, Stanisstreet M, Casson IF, Howard CV, Walkinshaw S, Pennycook S, et al. St. Vincent’s Declaration 10 years on: outcomes of diabetic pregnancies. Diabet Med 2002;19:216–20.
4. Penney GC, Mair G, Pearson, DW, Scottish Diabetes in Pregnancy Group. Outcomes of pregnancies in women with type 1 diabetes in Scotland: a national population-based study. BJOG 2003;110:315–18.
5. Hawthorne G, Robson S, Ryall EA, Sen D, Roberts SH, Ward Platt MP. Prospective population based study of outcome of pregnancy in diabetic women: results of the Northern Diabetic Pregnancy Audit, 1994. BMJ 1997;315:279–81.
6. Dunne F, Brydon P, Smith K, Gee H. Pregnancy in women with Type 2 diabetes: 12 years outcome data 1990–2002. Diabet Med 2003;20:743–8.
7. Casson IF, Clarke CA, Howard CV, McKendrick O, Pennycook S, Pharoah PO, et al. Outcomes of pregnancy in insulin dependent diabetic women: results of a five year population cohort study. BMJ 1997;315:275–8.
8. Vaarasmaki MS, Hartikainen A, Anttila M, Pramila S, Koivisto M. Factors predicting peri- and neonatal outcome in diabetic pregnancy. Early Hum Dev 2000;59:61–70.
9. Nielsen GL, Nielsen PH. Outcome of 328 pregnancies in 205 women with insulin-dependent diabetes mellitus in the county of Northern Jutland from 1976 to 1990. Eur J Obstet Gynecol Reprod Biol 1993;50:33–8.
10. Evers IM, de Valk, HW, Visser GH. Risk of complications of pregnancy in women with type 1 diabetes: nationwide prospective study in the Netherlands. BMJ 2004;328:915.
11. Sharpe PB, Chan A, Haan EA, Hiller JE. Maternal diabetes and congenital anomalies in South Australia 1986–2000: a population-based cohort study. Birth Defects Res A Clin Mol Teratol 2005;73:605–11.
12. von Kries R, Kimmerle R, Schmidt JE, Hachmeister A, Bohm O, Wolf HG. Pregnancy outcomes in mothers with pregestational diabetes: a population-based study in North Rhine (Germany) from 1988 to 1993. Eur J Pediatr 1997;156:963–7.
13. Jensen DM, Damm P, Moelsted-Pedersen L, Ovesen P, Westergaard JG, Moeller M, et al. Outcomes in type 1 diabetic pregnancies: a nationwide, population-based study. Diabetes Care 2004;27:2819–23.
14. Silva Idos S, Higgins C, Swerdlow AJ, Laing SP, Slater SD, Pearson DW, et al. Birthweight and other pregnancy outcomes in a cohort of women with pre-gestational insulin-treated diabetes mellitus, Scotland, 1979 – 95. Diabet Med 2005;22:440–7.
15. Pregnancy outcomes in the Diabetes Control and Complications Trial. Am J Obstet Gynecol 1996;174:1343–53.
16. McElvy SS, Miodovnik M, Rosenn B, Khoury JC, Siddiqi T, Dignan PS, et al. A focused preconceptional and early pregnancy program in women with type 1 diabetes reduces perinatal mortality and malformation rates to general population levels. J Matern Fetal Med 2000;9:14–20.
17. Cnattingius S, Berne C, Nordstrom ML. Pregnancy outcome and infant mortality in diabetic patients in Sweden. Diabet Med 1994;11:696–700.
18. Nordstrom L, Spetz E, Wallstrom K, Walinder O. Metabolic control and pregnancy outcome among women with insulin-dependent diabetes mellitus. A twelve-year follow-up in the country of Jamtland, Sweden. Acta Obstet Gynecol Scand 1998;77:284–9.
20. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group. N Engl J Med 1993;329:977–86.
21. Diabetes Care Program of Nova Scotia. In: Dunbar M, Feguson L, editors. The pregnancy & diabetes management guidelines. 1st ed. Halifax (Nova Scotia): Nova Scotia Department of Health; 1993.
22. Fair M, Cyr M, Allen AC, Wen SW, Guyon G, MacDonald RC, et al. Validation study for a record linkage of births and infant deaths in Canada (Statistics Canada, Catalogue 84F0013XIE) 1999. Available at: http://www.statcan.ca/english/research/84F0013XIE/84F0013XIE1992001.pdf
. Retrieved June 8, 2006.
23. White P. Classification of obstetric diabetes. Am J Obstet Gynecol 1978;130:228–30.
24. Statistics Canada. Health Indicators December 2000. Available at: http://www.statcan.ca/english/freepub/82-221-XIE/defin.htm
. Retrieved May 2, 2006.
25. Kramer MS, Platt RW, Wen SW, Joseph KS, Allen A, Abrahamowicz M, et al. A new improved population-based Canadian reference for birth weight for gestational age. Pediatrics 2001;108:E35.
26. Kitzmiller JL, Gavin LA, Gin GD, Jovanovic-Peterson L, Main EK, Zigrang WD. Preconception care of diabetes: glycemic control prevents congenital anomalies. JAMA 1991;265:731-6.
27. Robinson HE, O'Connell CM, Joseph KS, McLeod NL. Maternal outcomes in pregnancies complicated by obesity. Obstet Gynecol 2005;106:1357–64.