PREGESTATIONAL DIABETES MELLITUS AND PREGNANCY

Josip Djelmis1, Zeljko Metelko2, Ivana Pavlic-Renar2, Akkina Suresh Babu2

1Department of Obstetrics and Gynecology, School of Medicine, University of Zagreb,
Petrova 13, 10000 Zagreb, Croatia

2Vuk Vrhovac Institute, University Clinic for Diabetes, Endocrinology and Metabolic Diseases,
Dugi dol 4a, 10000 Zagreb, Croatia

Editorial

Received: January 8, 1999

Key words: insulin dependent diabetes mellitus, non-insulin dependent diabetes mellitus, pregnancy complications, early embryonic delay, fetal growth, fetal macrosomia

SUMMARY

At the Department of Obstetrics and Gynecology, Perinatal Unit for Diabetes and Fetal Growth, Zagreb University School of Medicine, perinatal care of diabetes-complicated pregnancies has been practiced for more than 36 years now. The purpose of this paper is to present results of a study of diabetic pregnancies and latest clinical advances in the perinatal care of such pregnancies. Pregnancy complicated with diabetes is at risk for a number of maternal, fetal, and neonatal complications. Recent advances in medicine, especially in diabetology and perinatology, help the clinician avoid or lessen antenatal or perinatal complications in diabetic pregnancies. The main result of improved perinatal care is that today, the fetal and neonatal mortality in diabetic pregnancy is almost equal to that in the healthy pregnant population. Intensive preconceptional care and optimal regulation of insulin dependent diabetes have resulted not only in decreased perinatal mortality but also in a decreased rate of congenital malformations. On the other hand, tight glycemia control during pregnancy has an impact on fetal growth. Intensive control of fetal growth, verification of lung maturation at term by amniocentesis, and surveillance of fetal oxygenation will result in giving birth to a mature, eutrophic newborn with the lowest rate of neonatal complications possible. Perinatal mortality of less than 2% in diabetic pregnancy can be achieved by planned delivery between week 38 and 39 of gestation, either by the vaginal route or by cesarean section, depending on the indication. After delivery, intensive care of the newborn is required.

INTRODUCTION

Diabetes in pregnancy poses numerous problems for both the mother and fetus. Poor metabolic control in pregestational diabetes (insulin dependent diabetes /IDDM/, and non-insulin dependent diabetes /NIDDM/) during pregnancy is associated with an increased risk of spontaneous abortion, congenital malformations, pre-eclampsia, fetal distress, stillbirths, macrosomia, and neonatal death. Tight metabolic control (defined as fasting blood glucose <5 mmol/L and peak postprandial glucose <7.8 mmol/L, with normal glycosylated hemoglobin, HbA1c) has been shown to decrease the rate of all these complications to an incidence approaching what is seen in the normal low risk population.
IDDM is a chronic autoimmune disease resulting from destruction of the pancreatic beta cells. Predisposition is genetically determined, with the relevant genes being located on chromosome six in association with the major histocompatibility complex. The major histocompatibility complex is responsible for encoding three HLA class II molecules known as HLA-DP, DQ, and DR, and 40% of women with IDDM possess DR3 or DR4 (1). The final expression of autoantibody production in IDDM is the result of genetic susceptibility, perhaps triggered by environmental factors such as certain viral infections, chemical toxins, or even milk proteins. The offspring of women with IDDM have a low risk of developing IDDM, with an incidence of 1%-3%. However, the risk is 6.1% if the father has IDDM (2). If both parents have IDDM, the offspring's risk is 20%. The cause of non-insulin dependent diabetes mellitus (NIDDM) lies in a combination of impaired B-cell function, with a markedly increased hepatic glucose output. It is subdivided into obese and non-obese. The prevalence varies from country to country and within country. Being a slow onset disease, NIDDM remains undiagnosed at onset, or even if diagnosed it is often ignored by the affected individuals. During pregnancy in women with NIDDM, the aims of their management are similar to those for women with IDDM. The women with poorly controlled NIDDM as well as their fetuses are subject to the same hazards as pregnant women with IDDM. Tight glycemic control is therefore indicated, and in virtually all cases this implies the use of insulin. In general, women with NIDDM should switch to insulin therapy, ideally before pregnancy or early in the first trimester. Oral hypoglycemic agents are unlikely to achieve adequate glycemic control. They cross the placenta, and are potentially teratogenic; sulfonylureas also stimulate the fetal B cells directly, aggravating fetal hyperinsulinemia and macrosomia.

PERINATAL MORTALITY

Improved fetal surveillance, neonatal intensive care, and maternal metabolic control have together reduced perinatal mortality in pregestational diabetic pregnancies to 2% in the last five years (Fig. 1).

Figure 1. Perinatal mortality from 1952 till 1997 at the Department of Obstetrics and Gynecology, School of Medicine, University of Zagreb


 Figure 1

Thirty years ago, it was more than 20% (3). Figure 2 shows relationship between gestational age and perinatal mortality.

Figure 2. Fetal, neonatal and perinatal mortality according to gestational age


 Figure 2

The lowest perinatal mortality rate has been recorded in the group of children born between the 38th and 40th week, increasing in newborns born between the 33th and 37th week, and reaching the highest rate in those born between the 29th and 32nd week of gestation. It is important to note that perinatal mortality increases in the newborns born after the 40th week of gestation. Fetal mortality is several fold in fetuses with intrauterine growth retardation (IUGR, 33.3%), while in macrosomic fetuses it does not differ significantly from that in fetuses of appropriate size for the respective gestational age (Fig. 3).

Figure 3. Fetal, neonatal and perinatal mortality according to fetal growth


 Figure 3

SGA=small for gestational age; AGA=appropriate for gestational age; LGA=large for gestational age


In diabetic pregnancies complicated by maternal hypertension (BP >140/90 mm Hg), perinatal mortality is significantly higher (13%) compared to women without hypertension (6%). Diabetic women have an increased incidence of spontaneous abortion if the first-trimester HbA1c hemoglobin concentrations are above 12% (Fig. 4), or if the median first-trimester preprandial blood glucose concentrations exceed 6.7 mmol/L. The rate of spontaneous abortion shows dependence on glycemia control, and is increased in uncontrolled diabetic patients (4).

Figure 4. Correlation of glycosylated hemoglobin A1c with spontaneous abortion and congenital anomalies in insulin dependent pregnant diabetics


 Figure 4

CONGENITAL MALFORMATIONS

If hyperglycemia is present during the first trimester of pregnancy when organogenesis is taking place, congenital malformations may occur. The incidence is said to be as high as 8% in uncontrolled diabetic pregnancies (uncontrolled during the first eight weeks of gestation), which is 2-3 times greater than in the general population. The malformations often involve the heart and central nervous system, and are potentially lethal.
Major congenital anomalies are 2-4 times more common in diabetic pregnancies than in normal pregnancies (4). Lethal congenical anomalies account for 26% of perinatal deaths. In IDDM pregnancies, the risk factor for major congenital anomalies is poor preconception diabetic control. Figure 4 shows relationship between the rate of congenital malformations and state of glycemia during the first trimester of IDDM pregnancy. The first trimester of pregnancy is the period of embryogenesis. HbA1c values are considered relevant for grading the regulation of maternal glycemia as follows: <8%, good metabolic control; 8%-12%, poor metabolic control; and >12% very poor metabolic control. During the first trimester of pregnancy in IDDM patients with poor metabolic control, the rate of congenital malformations increased from 5.9% to 20% in those with very poor metabolic control (2,4,5).
In the development of these congenital anomalies, organogenesis must be disrupted very early in pregnancy by the diabetic metabolic disturbance. Maternal hyperglycemia has been shown to be teratogenic in animal models. In humans, indirect evidence that hyperglycemia is the cause of major anomalies come from several reports linking the incidence of fetal abnormalities with maternal HbA1c concentrations (4). Although the site of action has not been identified, an attractive hypothesis is that the yolk sac is primarily affected, and that glucose-induced embryopathy is secondary to yolk sac impairment (6). If maternal hyperglycemia is the only factor causing birth defects, then the anomaly rate would be expected to correlate with the mean early first trimester maternal glucose levels. More recently, a support for the hyperglycemic cause has come from the evidence that tight control of maternal blood glucose concentrations in the early weeks of pregnancy is associated with a highly significant reduction in the risk of serious congenital anomalies (7). Diabetic embryopathy (birth defect and spontaneous abortion) results from metabolic abnormalities during 6-7 weeks of gestation. It is known that the incidence of congenital malformations in IDDM patients with poor glycemic control is 20% versus only 2% in IDDM patients with good glycemic control. The maternal metabolic milieu during the early and critical phase of organogenesis has a direct impact on the early embryonic development. The effect of various metabolites was studied in vitro, and elevated concentrations of glucose and beta-hydroxybutyrate were found to directly affect embryonic morphogenesis. Elevated concentrations of glucose induce hyperaccumulation of sorbitol on the one hand, and decrease the concentrations of arachidonic acid, prostaglandins and myoinositol on the other hand. The level of free radicals in the embryonic tissue is elevated in the state of hyperglycemia. All these changes can induce embryonic dismorphogenesis (7). Embryonic transvaginal ultrasonography is helpful in detecting the viable embryo and its appropriate size for gestational age as well as for noninvasive surveillance of early development (8). There is a relationship between the embryonic growth, congenital malformations, and inappropriate metabolic control, as clearly shown by the results of our studies (7-9). Obvious defects in the yolk sac structure suggest that hyperglycemia during organogenesis has a primary deleterious effect on the yolk sac function with resultant embryopathy. Damm and Molsted-Pedersen showed the relationship between abnormal embryogenesis and early embryonic retardation. They found major malformations in seven out of 53 fetuses with an early growth delay by ultrasonographic crown-rump length (CRL) measurement at 7 to 14 weeks of gestation. The most profound delay in embryonic size was found in cases with poor control of maternal diabetes, as judged by high HbA1c (10).
IDDM pregnancy carries a high risk of neural tube defect (16%). A rarer complication is the caudal regression syndrome (sacral agenesis, phocomelic diabetic embryopathy), which is a continuum of malformations ranging from agenesis of the lumbosacral spine to sirenomelia with fusion of the lower extremities. Although the syndrome has been considered "pathognomonic" of diabetes, only 11% of affected infants are born to diabetic mothers. Cardiac fetal abnormalities are found in 44% of IDDM pregnancies and include ventricular septal defects, transposition of the great vessels, and coarctation of the aorta. Renal anomalies (agenesis, ureteral duplication) and gastrointestinal anomalies (duodenal atresia, anorectal atresia) are also more common in these pregnancies. Investigations leading to prenatal diagnosis of fetal anomalies include the potential risk test (HbA1c levels); maternal serum alpha-fetoprotein estimation for neural tube defects (15-18 weeks); and fetal anatomic survey (ultrasound scan, 18-20 weeks). The accuracy of diagnosis is highest when all the three investigations are done in conjunction.

FETAL IMPLICATIONS OF MATERNAL HYPERGLYCEMIA

Traditionally, the hallmark of diabetic pregnancy has been excessive birth weight. The term "macrosomia" is commonly used for infants of birth weight greater than 4000 grams or the 90th percentile at any gestational age. Macrosomia is generally believed to be due to fetal hyperinsulinemia, with the increased insulin production being primarily of fetal origin. Despite improved methods of glycemic control in pregnant women with diabetes, the rates of macrosomia still vary between 8% and 43% (4,11). These high rates may simply reflect the difficulty encountered in achieving euglycemia, especially in long-standing IDDM. The incidence of macrosomia rises significantly when the mean maternal blood glucose concentrations exceed 7.2 mmol/L, and one of the best predictors of fetal birth weight in IDDM may be the third-trimester one-hour postprandial glucose concentration.
A constant glucose oversupply to the fetus stimulates the fetal pancreatic beta cells to increase insulin production and gradually induces hypertrophy and hyperplasia. Our own studies showed the newborns with diabetogenic fetopathy to produce higher insulin levels than healthy ones (Table 1) (12).

Table 1. Umbilical vein IGF-I, insulin, HGH and glucose values in normal, IUGR and IDDM pregnancies (12)


Growth factor IGF-I
µmol/L
Insulin
mU/L
HGH
µg/L
Glucose
mmol/L
Pregnancy group
Normal (N=20) 9.8±4.9a 9.6±7.0b 20.6±13.4 2.4±2.1c
IUGR (N=14) 8.0±5.4 8.0±1.8 19.4±14.0 1.9±0.9d
IDDM (N=23) 12.4±6.4a 43.3±37.7b 15.6±9.4 5.2±3.2c,d
a(IGF-I in IDDM vs IUGR pregnancy) F=3.92; p<0.05
b(insulin in IDDM vs IUGR pregnancy) F=8.68; p<0.05
c(glucose in normal vs IDDM pregnancy) F=4.84; p<0.05
d(glucose in IDDM vs IUGR pregnancy) F=6.88; p<0.05

Since the placental barrier prevents excess fetal insulin from leaving the fetoplacental unit, a disproportionate amount of the fetomaternal glucose pool is metabolized by the fetus resulting in overweight newborns, and in increased lipogenesis, visceromegaly, and typical cushingoid features of the newborns.
The phenomena of human fetal growth and development are complex processes involving interaction of the mother, placenta, and fetus. The growth and development of the fetus are intertwined with and dependent upon the amount of nutrients such as glucose, lipids, amino acids as well as on the genetic makeup and obviously on the maternal and fetal endocrine status. Glucose is a predominant source of metabolic energy in the human fetus, and insulin has been shown to be an essential factor for rapidly growing fetal tissues. Therefore, the presence of the maternal diabetic state increases the circulating blood concentration of glucose (maternal blood glucose is freely available to the fetoplacental unit and, in fact, is supplied to this compartment at higher rates in diabetic pregnancy). The higher amount of glucose may, in turn, increase the synthesis and release of insulin from the fetal pancreas, thereby stimulating fetal growth and development via an increase in glucose utilization and/or cell proliferation (12). The positive correlation of glucose and insulin levels in fetal circulation and amniotic fluid with birth weight points to their strong influence on fetal growth (13). A positive and statistically significant correlation (r=0.6; p<0.003) was recorded between IGF-I umbilical vein values and weeks of gestation. The circulating IGF-I levels correlated with fetal growth and duration of pregnancy.
Diabetic fetopathy, a specific syndrome, is the direct result of fetal hyperinsulinemia and its consequences. However, in diabetics with pre-eclampsia, placental insufficiency can lead to a fetal glucose deficit, hypoinsulinism, and prenatal dystrophy. At the other extreme, women with diabetic vascular involvement, or those who develop superimposed pre-eclampsia, are at an increased risk of developing a small-for-gestational-age infant. Other predisposing factors for IUGR include poor maternal renal function, chronic hypertension, and poor metabolic control during organogenesis. IUGR, with the associated risk of increased fetal morbidity and mortality, complicates 7% of diabetic pregnancies.

MATERNAL COMPLICATIONS ASSOCIATED WITH DIABETIC PREGNANCY

White's classification is now mainly used for presentation of epidemiologic data from clinical trials. Pedersen's classification suggests the use of four prognostically poor signs of pregnancy, related to maternal complications, i.e. diabetic ketoacidosis, pre-eclampsia, pyelonephritis, and maternal neglect. The occurrence of any of these complications in an IDDM pregnancy continues to be associated with poor pregnancy outcome but, apart from pre-eclampsia, their incidence can be reduced by tight metabolic control. However, other maternal complications, both diabetic and non-diabetic, may also contribute to perinatal morbidity and mortality.

Maternal infections
Maternal infections occur more frequently in pregestational diabetics than in normal pregnancies, and they are related in part to poor metabolic control. Diabetic pregnant women have significantly higher rates of urogenital, respiratory, wound, and endometrial infections (14). The rate of postpartum infection is higher than in the non-diabetic pregnant population. The risk of post-cesarean-section endometritis and wound infection is significantly greater among diabetic than among non-diabetic women. Pyelonephritis occurs in 4% of diabetic pregnancies, compared with 1% of the non-diabetic population.

Pre-eclampsia
Diabetic pregnancies complicated by pre-eclampsia are of great concern because of poor perinatal outcome (4,15). Pre-eclampsia complicates 20% of pregestational diabetic pregnancies. The proportion rises with increasing severity and duration of diabetes according to White's classification. The perinatal mortality rate is 13.2% for preeclamptic diabetic women versus 6% for normotensive diabetic women.

Diabetic nephropathy
Overt diabetic nephropathy is present in 5% of diabetic pregnancies, and a first-trimester finding of >0.3 g/24 h proteinuria is associated with an adverse pregnancy outcome (16). Hypertension is found in 30% of diabetic nephropathy pregnancies in the first trimester, and in 75% at delivery. Premature delivery rates are also high, because in many instances hypertension signals superimposed pre-eclampsia. In more than a half of diabetic pregnancies complicated by nephropathy, delivery occurs before 37 weeks of gestation, and the rate of cesarean section is 70%. Other maternal complications include severe edema related to hypoalbuminemia. The choice of an antihypertensive drug in pregnancy complicated by diabetic nephropathy may be difficult. Angiotensin-converting enzyme inhibitors, which are useful in reducing glomerular hyperfiltration in diabetic nephropathy, are contraindicated in pregnancy because of reports of embryopathy, i.e. neonatal renal failure. The use of beta-adrenergic blocking agents may lessen awareness of hypoglycemia. Methyldopa and calcium-channel-blocking agents such as nifedipine are more commonly used.
Despite the potential fetal and maternal complications, successful outcome is recorded in 90% of pregnancies complicated by maternal nephropathy. Prepregnancy counseling should emphasize two important issues. Poor perinatal outcome is predicted by proteinuria greater than 3.0 g/24 h and serum creatinine greater than 130 mmol/L. Even though pregnancy may be successful, maternal longterm morbidity and mortality are high and may severely compromise a diabetic mother's ability to raise her child.

Diabetic retinopathy
Proliferative retinopathy is characterized by neovascularization consisting of capillary growth over the surface of the retina. Pregnancy worsens proliferative retinopathy, when compared with progression of the disease over a 40-week period in non-pregnant diabetics (17). Macular edema, which is related to changes in macular capillary permeability, may also worsen in pregnancy, particularly in women with associated nephropathy and hypertension. Laser photocoagulation can be used either before or during pregnancy, and not only to treat proliferative retinopathy but also to prevent it. Ophthalmological evaluation should ideally be done before pregnancy, again early in the first trimester, and then periodically according to severity.

Diabetic neuropathy
Peripheral and cranial neuropathies are seldom prominent during pregnancy. Autonomic neuropathy may be very troublesome in pregnancy, with worsening postural hypotension, diminished catecholamine response to hypoglycemia, and deterioration in gastropathic symptoms. Diabetic gastropathy during pregnancy contributes to nausea and vomiting, nutritional problems, and difficulty with glucose control. Treatment with H2-receptor antagonists can be tried in pregnancy.

Diabetic control during pregnancy
All women with diabetes who are considering pregnancy should undergo prepregnancy counseling. Unfortunately, despite diabetic education and outreach programs, this advice is often not heeded, with serious implications for mother and fetus. Prepregnancy counseling should consist of pregnancy dietary advice, insulin regimen adjustment to attain euglycemia, and renal/retinal evaluation (18). Pregnancy should be deferred until an HbA1c concentration <8% is achieved, since this level is associated with the lowest incidence of both congenital anomalies and spontaneous abortion. Efforts to achieve maternal euglycemia are important in reducing perinatal morbidity. Ideal plasma glucose concentrations are less than 5.0 mmol/L in the fasting state and less than 7.8 mmol/L after a meal (19). The frequency and timing of glucose monitoring varies according to severity and difficulty of control. Monthly HbA1c measurements have been advocated, but they may be no better than mean glucose concentrations in predicting neonatal birth weight or neonatal hypoglycemia (19,20). A total calorie intake should range between 30-35 kcal per kg ideal body weight, given as three meals and three snacks daily. An ideal dietary composition is 55% carbohydrate, 20% protein, and 25% fat. Obese women may be managed on a lower calorie intake as long as weight loss and ketonuria are avoided.
Insulin therapy should be tailored for each individual. In early pregnancy, good control can often be obtained by a regimen of intermediate and regular insulin before breakfast, regular insulin before dinner and supper, and intermediate insulin at bedtime. Hypoglycemia episodes are more common and may not be as readily perceived during pregnancy as in the non-pregnant state, perhaps in part because the normal counter-regulatory hormone release may be suppressed.
Ketoacidosis now affects less than one percent of diabetic pregnancies, but it remains one of the most serious complications. Diabetic ketoacidosis may occur in association with hyperemesis gravidarum, tocolytic therapy (beta-sympathomimetic agents), infections, and use of corticosteroids. Since pregnancy is marked by a relative insulin resistance, enhanced lipolysis, and ketogenesis, diabetic ketoacidosis may occur with minimal hyperglycemia. If diabetic ketoacidosis complicates a diabetic pregnancy, fetal loss is 20%, and it occurs mainly during the first trimester. For the welfare of the fetus, maternal acidosis must be corrected rapidly, the management of diabetic ketoacidosis in pregnancy being similar to that in the non-pregnant state.

Pregnancy termination
Diabetic pregnancy can be terminated either vaginally or by cesarean section. In the last decade, the rate of cesarean section in diabetic pregnancy has been on a significant increase worldwide. At our Department of Gynecology and Obstetrics, the frequency of cesarean section in diabetic pregnancies has increased from 24.9% in 1962-1970 to 61% in 1998. The indications for cesarean section generally have been broadened, however, there should be no dilemma about what to do, when and how to terminate a diabetic pregnancy, when all parameters and possible maternal and fetal risks have been properly considered. Cesarean section is definitely the more appropriate route of delivery for a fetus of diabetic mother.

REFERENCES

  1. Atkinson MA, MacLaren NK. The pathogenesis of insulin-dependent diabetes mellitus. N Engl J Med 1994;331:1428-1436.
  2. Warram JH, Krolewski AS, Kahn CR. Determinants of IDDM and perinatal mortality in children of diabetic mothers. Diabetes 1988;37:1328-1334.
  3. Drazancic A. Diabetes and pregnancy. Gynaecol Perinatol 1994;3(Suppl 1):15-29.
  4. Drazancic A, Blajic J, Djelmis J. Early spontaneous abortions and congenital malformations in relation to diabetes regulation in very early pregnancy. 24th Annual Meeting of Diabetic Pregnancy Study Group, Silkeberg, Denmark, 1993; Abstract No. 3.
  5. Mills JL, Simpson JL, Driscoll SG, et al. Incidence of spontaneous abortion among normal women and insulin-dependent diabetic women whose pregnancies were identified within 21 days of conception. N Engl J Med 1988;319:1617-1623.
  6. Rosenn B, Miodovnik M, Combs CA, Khoury J, Siddiqi TA. Glycemic thresholds for spontaneous abortion and congenital malformations in insulin-dependent diabetes mellitus. Obstet Gynecol 1994;84:515-520.
  7. Reece EA, Hobbins JC. Diabetic embryopathy: pathogenesis, prenatal diagnosis, and prevention. Obstet Gynecol Surv 1986;41:325-335.
  8. Djelmis J. Early embryonic growth delay in IDDM pregnancies. 28th Annual Meeting of Diabetic Pregnancy Study Group, Aberdeen, Great Britain, 1997; Abstract No. 9.
  9. Starcevic V. Utjecaj glikemije na rani embrionalni razvoj u trudnica dijabeticarki. MS Thesis, Zagreb, 1998.
  10. Damm P, Molsted-Pedersen L. Significant decrease in congenital malformations in selected newborn infants of an unselected population of diabetic women. Am J Obstet Gynecol 1989;161:1163-1167.
  11. Cousins L. Pregnancy complications among diabetic women. Obstet Gynecol Surv 1987;42:140-149.
  12. Djelmis J, Drazancic A, Ivanisevic M, Suchanek E. Glucose, insulin, HGH and IGF-1 levels in maternal serum, amniotic fluid and umbilical venous serum: comparison between late normal pregnancy and pregnancies complicated with diabetes and fetal growth retardation. J Perinat Med 1992;20:47-56.
  13. Djelmis J, Ivanisevic M, Pfeiffer D. Fetal hyperinsulinemia in macrosomic newborns of healthy mothers. Acta Med Croat 1992;46:209-211
  14. Stamler EF, Cruz ML, Mimouni F, et al. High infectious morbidity in pregnant women with insulin-dependent diabetes: an understated complication. Am J Obstet Gynecol 1990;163:1217-1221.
  15. Drazancic A, Dejlmis J, Blajic J, Kuvacic I, Latin V. Bacteriuria in diabetic pregnancies. Diabetol Croat 1997;26:175-181.
  16. Garner PR, D'Alton ME, Dudley DK, Huard P, Hardie M. Pre-eclampsia in diabetic pregnancies. Am J Obstet Gynecol 1990;163:505-508.
  17. Hare JW. Diabetic complications of diabetic pregnancies. Semin Perinatol 1994;18:451-458.
  18. Klein BEK, Moss SE, Klein R. Effect of pregnancy on progression of diabetic retinopathy. Diabetes Care 1990;13:34-40.
  19. Drazancic A. Prekoncepcijska regulacija bolesti trudnica s dijabetesom. Diabetol Croat 1983;12:193-208.
  20. Djelmis J. Gestational diabetes mellitus. Diabetol Croat 1997;26:167-173.


 HOME PAGE