Original Article
Transplacental Passage of Insulin in Pregnant Women with Insulin-Dependent Diabetes Mellitus — Its Role in Fetal Macrosomia
N Engl J Med 1990; 323:309-315August 2, 1990
- Abstract
- Abstract
Background and Methods.
Fetal macrosomia occurs despite nearly normal maternal blood glucose levels in women with diabetes treated with insulin. We examined the hypothesis that it may be caused by insulin transferred as an insulin-antibody complex from the mother to her fetus. We adapted and validated a method based on high-performance liquid chromatography and used it to quantitate insulin in small volumes (0.5 to 1.0 ml) of cord serum from 51 infants born to mothers with insulin-dependent diabetes mellitus.
Results.
In mothers receiving only human insulin (n = 6), only human insulin was detected in cord serum. Of the remaining 45 infants, whose mothers received animal insulin during pregnancy, 28 (group 1) had levels of animal (bovine or porcine) insulin (mean [±SE], 707±163 pmol per liter) that constituted 27.4±2.5 percent of the total insulin concentration (2393±500 pmol per liter) measured in the cord serum. The cord-serum insulin concentration in the remaining 17 infants (group 2), in whom only human insulin was detected (381±56 pmol per liter), was only 15 percent of that in group 1 (P<0.001). There was a significant correlation between the maternal and the cord-serum concentrations of anti-insulin antibody and the concentration of animal insulin in the baby (r = 0.77, P<0.01, and r = 0.76, P<0.001, respectively), suggesting that the animal insulin was transferred as an insulin-antibody complex. In group 1 the mean concentration of animal insulin in cord serum was higher in the 12 infants with macrosomia than in the 16 infants without the condition (1113±321 vs. 402±110 pmol per liter; P<0.05), and the concentration of animal insulin in cord serum correlated with birth weight (r = 0.39, P<0.05). The maternal glycosylated hemoglobin values and the incidence of respiratory distress syndrome were similar in groups 1 and 2.
Conclusions.
Considerable amounts of antibody-bound insulin are transferred from mother to fetus during pregnancy in some women with insulin-dependent diabetes mellitus; the extent of transfer correlates with the maternal concentration of anti-insulin antibody. The correlation between macrosomia and the concentrations of animal insulin in cord serum indicates that the transferred insulin has biologic activity and suggests that the formation of antibody to insulin in the mother is a determinant of fetal outcome independent of maternal blood glucose levels. (N Engl J Med 1990; 323:309–15.)
Media in This Article
Figure 1
High-Performance Liquid Chromatography Elution Profiles.Figure 2
Correlation between Insulin-Antibody Activity and Concentrations of Insulin in Cord Serum from 45 Neonates Born to Mothers with IDDM Treated with Animal Insulin during Pregnancy.Article Activity
- Article
The human placenta is impermeable to free insulin, so the insulin in the fetus and in the amniotic fluid is presumed to be entirely fetal in origin.1 , 2 Furthermore, the inability to detect Radio-labeled insulin in the cord serum of neonates born to mothers with insulin-dependent diabetes mellitus (IDDM) who had received Radio-labeled insulin shortly before delivery has been interpreted as evidence against the placental transfer of insulin-antibody complexes.3 Thus, insulin administered to a pregnant woman with IDDM is not thought to cross the placenta to her fetus and has not been considered an etiologic factor in diabetic fetopathy. However, the plasma half-life of intravenously administered insulin is at most 30 minutes, and a slow transfer of insulin may therefore remain undetected by this technique.4 Indeed, Bauman and Yalow detected animal insulin in the cord serum of two infants whose mothers, who had IDDM, had been treated with animal insulin and had anti-insulin antibodies in their serum,5 but the clinical relevance of this finding, which was based on the use of an antibody that distinguished porcine and bovine from human insulin, has not been established.
The availability of a cohort of well-defined patients treated with animal insulin allowed us to examine in greater detail the possible placental transfer of insulin in pregnant women with IDDM and its biologic effect on the fetus. To do so we adapted a method based on reverse-phase, high-performance liquid chromatography to measure animal and human insulin separately in as little as 0.5 ml of serum or amniotic fluid. Using this technique, we tested the hypotheses that there is placental transfer of insulin from the mother to her fetus and that complications of fetal outcome, such as macrosomia, correlate with the amount of insulin that is transferred from the mother, suggesting a biologic effect of the transferred insulin. We provide evidence for the validity of both hypotheses by documenting a substantial placental transfer of insulin from the mother to the fetus and a correlation between the amount of insulin transferred and the presence of fetal macrosomia.
Methods
Subjects
Mixed arteriovenous cord serum was collected at delivery from 51 infants born to women with IDDM (mean [±SD] age, 26.0±5.3 years) from 1979 to 1987, and the levels of insulin and anti-insulin antibody were measured. In 8 of the 51 pregnancies, amniotic fluid was also collected during the third trimester. Serum collected at delivery from 21 of the mothers was available for the determination of insulin-antibody concentrations. The 51 women were participants in a prospective study of the effects of diabetes on pregnancy and fetal well-being at the University of Cincinnati. The condition of 8 women was classified as White Class B (age at onset ≥20 years and duration <10 years), of 20 women as Class C (age at onset or duration, 10 to 19 years), of 16 women as Class D (age at onset <10 years, duration >20 years, with background retinopathy or hypertension), of 4 women as Class R (proliferative retinopathy), of 1 woman as Class RT (previous renal transplantation), and of 2 women as Class F (nephropathy).6 The average (±SD) duration from the onset of IDDM to conception was 10.7±6.4 years. All the women received treatment for IDDM according to a standard protocol, the details of which have been described elsewhere,7 in which glycemic control was obtained with the use of split-dosage regimens of short-acting and intermediate-acting insulins. Forty-five of the 51 women received bovine or porcine insulin before and throughout pregnancy, whereas 6 women previously treated with bovine or porcine insulin received human insulin starting in the sixth week of pregnancy. The goals of glycemic control were a fasting blood glucose level of less than 5.5 mmol per liter and a blood glucose level of less than 7.8 mmol per liter 90 minutes after eating. The mean (±SD) daily doses of insulin during the second and third trimesters were 0.85±0.3 and 1.06±0.5 U per kilogram of body weight, respectively. Maternal blood glycosylated hemoglobin values were measured every month to assess the adequacy of glycemic control8; the mean value at the time of delivery was 8.2±1.7 percent (normal range, <8.5 percent). Ultrasonography was performed at the initial visit and repeated at 20 weeks of gestation to verify pregnancy dates. Gestational age was also confirmed by clinical assessment of the newborn. One infant was delivered at 32 weeks, 22 infants at 34 to 37 weeks, and 28 at more than 37 weeks of gestation; 80 percent were delivered by cesarean section. The cord-serum and amniotic-fluid samples were stored at -70°C from the time of collection. The study protocol was approved by the institutional review board of the University of Cincinnati Medical Center, and each woman gave informed consent for the study.
Macrosomia was defined as birth weight more than 2 SD above the mean weight for gestational age as described by Usher and McLean.9 For the purposes of statistical analysis birth weight was also expressed as the number of standard deviations from the mean for gestational age. The diagnosis of respiratory distress syndrome was based on clinical signs (grunting, retraction, and respiratory rate >60 per minute), typical radiologic findings, and increased oxygen requirements.
Extraction and Concentration of Insulin from Serum and Amniotic Fluid
In order to release insulin bound to anti-insulin antibodies in serum and amniotic fluid, 0.5 to 1 ml of serum or amniotic fluid was incubated with an equal volume of 1 percent trifluoracetic acid for one hour at room temperature. The mixtures of serum or amniotic fluid and trifluoracetic acid were then passed through octadecyl silica cartridges (Sep-Pak C-18, Waters Associates, Milford, Mass.). Insulin was eluted from the cartridge with use of a modification of the method of Cohen et al.10
To assess the yield of unbound insulin extracted from serum, 125I-labeled or unlabeled porcine insulin was added to serum from fasting normal subjects, and the extraction was carried out as detailed above. After elution, the dried eluate was either counted in a gamma counter for 125I radioactivity or subjected to radioimmunoassay for insulin. The mean (±SD) recovery of 125I-labeled insulin and unlabeled insulin from these samples that contained no insulin antibody was 72.6±1.7 percent (n = 11) and 93.7±14.2 percent (n = 4), respectively.
High-Performance Liquid Chromatography of Insulin
The method of Shoelson et al. was used with minor modifications.11 A mixture of insulin standards (9 pmol each of bovine, porcine, and human insulin) dissolved in 5 μl of 0.1 M acetic acid containing 0.5 percent bovine serum albumin was injected onto the high-performance liquid chromatography column, eluted with 30 percent (vol/vol) acetonitrile, and their elution volumes determined by absorbance at 214 nm. Knowledge of this elution profile permitted the detection of a smaller mass of each insulin by radioimmunoassay. The three insulins (bovine, porcine, and human) could be clearly separated by high-performance liquid chromatography (Fig. 1Figure 1
High-Performance Liquid Chromatography Elution Profiles.A), and the elution volumes for the three insulin analogues were constant (within 0.5 minute) during a single day. However, the system of high-performance liquid chromatography was very sensitive to small changes in the concentration of acetonitrile in the mobile phase (a 0.1 percent change in the acetonitrile concentration changed the elution volume by 3 to 4 ml), resulting in day-to-day variation in elution volumes. The elution volumes of serum extracts were therefore analyzed in comparison with those of insulin standards subjected to high-performance liquid chromatography on the same day.The vacuum-dried sample eluates were reconstituted in 300 μl of 0.1 M acetic acid containing 0.5 percent bovine serum albumin just belbre injection onto the high-performance liquid chromatography column. Fractions of 0.3 ml were collected in tubes containing 50 μl of borate buffer (0.5 mol of boric acid per liter containing 1 percent bovine serum albumin, pH 9.3), immediately frozen, and lyophilized. After lyophilization all fractions were reconstituted in radioimmunoassay buffer and underwent radioimmunoassay for insulin.
Internal Standard
To determine the efficacy of the procedure used to measure insulin, 0.43 pmol of a semisynthetic human insulin analogue (Hum [SerB24], a generous gift from Dr. H. Tager, University of Chicago) was added to the serum aliquot before incubation with trifluoracetic acid. The elution volume for this insulin preparation was significantly different from that of the native human, porcine, and bovine insulin, so that it did not interfere with the detection and quantitation of the natural insulins on high-performance liquid chromatography. Determining the amount of the semisynthetic analogue recovered after high-performance liquid chromatography permitted us to calculate the recovery of insulin in each serum sample. The mean (±SD) recovery of the internal standard (Hum[SerB24]) from serum samples averaged 50.1±10.9 percent. All values for insulin immunoreactivity after high-performance liquid chromatography were corrected for this recovery.
Insulin Radioimmunoassay and Insulin-Antibody Assay
An equilibrium double-antibody radioimmunoassay was used to measure insulin.10 The residue from the lyophilized elution buffer in each high-performance liquid chromatography fraction (0.3 ml) did not significantly inhibit the binding of 125I-labeled insulin to the guinea pig anti-porcine insulin serum. The natural insulins (bovine, human, and porcine) were equipotent in inhibiting the binding of 125I-labeled insulin to the anti-insulin antibody used in the assay; hence, the fractions predicted to contain these analogues were measured against a standard curve for porcine insulin. The inhibition of the binding of 125I-labeled insulin to the insulin antibody used in the assay by the semisynthetic analogue Hum[SerB24] differed from that of the binding of native bovine, human, and porcine insulin, and so the fractions predicted to contain the semisynthetic analogue were quantitated with use of a homologous standard curve. The sensitivity of the assay for the native insulin was 21 to 36 pmol per liter.
Insulin antibodies in maternal and cord serum and amniotic fluid were measured in duplicate with the method of Palmer et al. with minor modifications.12 The results are expressed as the percentage of 125I-labeled insulin (125I-monoiodinated A14 human insulin, Amersham, Arlington Heights, Ill.; 6.0 fmol per tube, 1900 Ci per millimole) added per tube. The mean (±SD) level of 125I-labeled insulin binding determined by this method was 0.14±0.02 percent in the serum of 19 normal subjects, and the mean insulin-antibody activity in amniotic fluid from 15 normal pregnant women was 0.06±0.08 percent.
Scatchard analysis of the binding of anti-insulin antibody in cord serum was performed with the use of deinsulinized cord serum.13 Deinsulinization was performed by incubating 1 ml of serum with 1.25 ml of 0.12 M hydrochloric acid and 0.5 ml of dextran-coated charcoal for five minutes at room temperature, followed by neutralization with 1.25 ml of 0.12 M sodium hydroxide. The charcoal was separated from the serum by two sequential centrifugations of 30 minutes each at 2000Xg. The binding of human insulin was assessed by incubating 40 μl of deinsulinized serum with 70 μl of 125I-labeled monoiodinated A14 human insulin and graded amounts of unlabeled insulin. The results were analyzed with the computer program EBDA/LIGAND (Biomedical Computing Technology Information Center, Nashville). The binding capacity is expressed in moles per liter.
Statistical Analysis
The data were analyzed with use of SAS software (SAS Institute, Cary, N.C.). Correlation analysis (Spearman's rank correlation) was used to examine the relation between the continuous variables. Group mean values were compared with use of Student's t-test for log-transformed data or the Wilcoxon rank-sum test for untransformed data. The relations between dichotomous variables were tested with Fisher's exact test. All results are expressed as means ±SE unless otherwise stated. P values below 0.05 were considered statistically significant.
Results
Analysis of Cord Serum
Insulin
Only one peak of immunoreactive insulin was detected in the cord-serum samples from the infants of the six women treated with human insulin during pregnancy (Fig. 1B). The elution volume of this peak corresponded to the elution volume of the human-insulin standard. The concentrations of human insulin and the insulin-antibody activity in these six samples ranged from 35 to 11,700 pmol per liter and from 0.9 to 13.7 percent, respectively. There was a significant correlation (r = 0.83, P<0.05) between the insulin concentration and the insulin-antibody activity in these six samples.
For purposes of analysis, the remaining 45 cord-serum samples from infants whose mothers had received animal (bovine or porcine) insulin during pregnancy were classified in two groups (Table 1Table 1
Cord-Serum Insulin and Anti-insulin Antibody Measured in 45 Newborn Infants of Women with IDDM Treated with Animal Insulin during Pregnancy.*). Group 1 included the 28 cord-serum samples in which animal insulin was detected on high-performance liquid chromatography (Fig. 1C), and group 2 included the 17 samples in which animal insulin was undetectable (<35 pmol per liter). In group 1, animal insulin (707±163 pmol per liter) formed 27.4±2.5 percent of the total insulin (2393±500 pmol per liter) measured. The elution profile of the insulin in the samples from group 2 was similar to that of the human-insulin standard. The total insulin measured in these 17 cord samples was 381±56 pmol per liter, only about 15 percent of that in group 1 (P<0.001). The maternal serum estriol concentrations at around 38 weeks of gestation were not significantly different in groups 1 and 2 (4.3±0.4 vs. 4.6±0.9 pmol per liter, respectively).Insulin Antibody
Insulin-antibody binding was measured in all 51 cord-serum samples. The mean binding of 125I-labeled insulin in the women treated with human insulin and in the two groups (groups 1 and 2) of women treated with animal insulin was 5.1±2.1, 19.3±2.3, and 5.0±1.0 percent, respectively (Table 1). Because of the limited volume of serum available, Scatchard analysis of insulin antibodies was restricted to the cord serum from 14 of the 45 infants whose mothers had received animal insulin (10 in group 1 and 4 in group 2). The Scatchard plots from all the samples were curvilinear, indicating heterogeneity of insulin-binding sites. When the results were analyzed with use of a two-site model, the binding capacities of the lowaffinity (Kd = 55.78±13.0 nmol) and high-affinity (Kd = 5.29±0.94 nmol) binding sites were 0.614±0.17 and 0.16±0.043 mmol per liter, respectively. In these 14 serum samples there was a strong correlation (r = 0.91, P<0.001) between the binding of l25I-labeled insulin measured with use of deinsulinized serum and that measured with use of nondeinsulinizcd scrum.
In the 14 infants whose cord-serum insulin antibody was subjected to Scatchard analysis, there were significant correlations between the total insulin-binding capacity of the serum and the total insulin concentration (r = 0.83, P<0.001), the concentration of animal insulin (r = 0.92, P<0.001), and the proportion of the total insulin that was animal insulin (r = 0.82, P<0.001). Likewise, the binding capacity of the highand low-affinity antibody-binding sites individually correlated with the same three indexes. The binding of 125I-labeled insulin determined in nondeinsulinized cord serum from all 45 infants whose mothers had received animal insulin correlated significantly with the total insulin concentration (r = 0.67, P<0.001), the concentration of animal insulin (r = 0.76, P<0.001), and the proportion of the insulin that was animal insulin (r = 0.69, P<0.05) (Fig. 2Figure 2
Correlation between Insulin-Antibody Activity and Concentrations of Insulin in Cord Serum from 45 Neonates Born to Mothers with IDDM Treated with Animal Insulin during Pregnancy.).For 21 infants (12 in group 1 and 9 in group 2), maternal serum collected at delivery was available for studies of 125I-labeled insulin binding. In these samples there was a close correlation (r = 0.93, P<0.01) between the insulin-binding activity in the maternal serum and that in the corresponding cord serum. The insulin-antibody activity in the samples of maternal scrum correlated with the corresponding total insulin (r = 0.83, P<0.01) and animal insulin (r = 0.77, P<0.01) concentrations.
Biologic Effects
Of the 45 infants born to mothers who had received animal insulin during pregnancy, 21 had macrosomia and 24 were of appropriate birth weight. The mean cord-serum concentration of human insulin was higher in the infants with macrosomia than in those without (1762±419 vs. 696±126 pmol per liter, P = 0.03), and the concentrations of animal insulin and total insulin were slightly but not significantly higher in the infants with macrosomia (animal insulin, 636±218 vs. 267±83 pmol per liter, P = 0.08; and total insulin, 2396±601 vs. 965±199 pmol per liter, P = 0.08). Of the 28 infants in whom animal insulin was detected in cord serum, 12 had macrosomia, whereas 16 were of appropriate birth weight. The 12 infants with macrosomia had significantly higher cord-serum concentrations of animal, human, and total insulin than the infants without macrosomia (1113±321 vs. 402±110 pmol per liter, P<0.05; 2726±599 vs. 908±163 pmol per liter, P<0.02; and 3839±840 vs. 1309±259 pmol per liter, P<0.02, respectively) (Table 2Table 2
Relation between Birth Weight and Cord-Serum Concentrations of Animal, Human, and Total Insulin in 28 Infants with Detectable Animal Insulin in Cord Serum.*). There were also significant correlations between birth weight and cord-serum concentrations of animal insulin (r = 0.39, P<0.05), human insulin (r = 0.43, P<0.05), and total insulin (r = 0.47, P<0.02). There was no significant difference between the infants with macrosomia and those without the condition in cord-serum insulin-antibody activity, years of maternal diabetes before conception, maternal insulin dosage (units per kilogram) during the second or third trimester of pregnancy, maternal blood glycosylated hemoglobin concentration, incidence of respiratory distress syndrome, or lowest blood glucose concentration in the infant during the first four hours of life.Analysis of Amniotic Fluid
The recovery of the internal standard (Hum[SerB24]) from the samples of amniotic fluid averaged only 21.0±4.2 percent. In six of the eight samples no insulin was detected by high-performance liquid chromatography followed by radioimmunoassay, and in one sample only human insulin (69 pmol per liter) was detected. In the remaining amniotic-fluid sample the concentration of total immunoreactive insulin was 93 pmol per liter, of which 67 percent was animal insulin. The insulin-antibody activity of this sample (1.7 percent) was clearly higher than that of the seven samples with no detectable animal insulin (0.2 to 0.7 percent). The cord-serum total insulin concentration of this infant was 8200 pmol per liter, of which 27.0 percent was animal insulin. In the eight infants who had both measurements performed, there was a significant correlation between the insulin-antibody activity in amniotic fluid and in serum (r = 0.88, P<0.01).
Discussion
Since endogenous insulin secreted by the fetal pancreas must be the human analogue, any animal insulin detected in cord serum must be insulin that was transferred from the mother to the fetus. Therefore, the identification of significant amounts of animal insulin in the cord serum of infants whose mothers received bovine or porcine insulin (group 1) demonstrates that animal insulin administered to the mother crossed over to the fetus. The concentration of animal insulin in our subjects ranged from 36 to 3474 pmol per liter, on average constituting about 25 percent of the total cord-serum insulin concentration (Table 1). The demonstration of transplacental passage of substantial quantities of insulin also provides an additional explanation for the reported lack of correlation between C-peptide and insulin concentrations in the cord serum of infants born to mothers with diabetes.14 , 15
The concentration of insulin antibody in the cord serum strongly correlated with that present in the mother at the time of delivery, suggesting that these antibodies were transferred from the mother.14 The total amount of animal insulin in cord serum correlated with the insulin-binding capacity of the maternal and cord serum, indicating that the transfer of insulin took place as an insulin-antibody complex. Maternal serum estriol concentrations, an index of placental function, were similar in groups 1 and 2, but the possibility of abnormal placental permeability for insulin in pregnant women with IDDM cannot be entirely ruled out.
One established role for insulin in the fetus is the promotion of growth — insulin is a growth factor implicated in fetal macrosomia.16 , 17 Among the 28 infants in whom animal insulin was detected in cord serum, the 12 who had macrosomia had a significantly higher mean cord-serum concentration of animal insulin than the other 16; the concentration of animal insulin in cord serum was significantly correlated with birth weight. Since endogenous human fetal insulin is equally a growth factor, the associations and correlations with birth weight would be expected to hold true for human insulin, particularly for the sum of the transferred (animal) and endogenous (human) insulin concentrations. As predicted, cord-serum human and total insulin concentrations were higher in the infants with macrosomia than in those without, and human and, to a greater degree, total insulin concentrations correlated with the degree of macrosomia. This association between cord-serum insulin concentration and birth weight could not be attributed to differences in the duration of maternal diabetes before conception, class of maternal diabetes, amount of insulin administered to the mother during the second or third trimester of pregnancy, or antenatal glycemic control as assessed by maternal glycosylated hemoglobin values, all of which were comparable in the mothers of the two groups of infants.
In contrast to previous studies,17 , 18 ours found no correlation between the cord-serum insulin concentration (animal, human, or total) and either the incidence of respiratory distress syndrome or the infant's blood glucose nadir during the first four hours of life. The relatively low incidence of respiratory distress syndrome reflected a policy of inducing delivery only after antenatal confirmation of the maturity of the fetal lungs, and we instituted intravenous glucose therapy in every infant with blood glucose concentrations ≤2.2 mmol per liter. Our management thus precluded the demonstration of a relation between either respiratory distress syndrome or blood glucose nadir and cordserum insulin concentrations.
The cause of fetal macrosomia in the offspring of women with IDDM has not been fully elucidated, but the most widely accepted hypothesis is that originally proposed by Pedersen.19 , 20 According to this hypothesis the serum concentration of glucose and other metabolic substrates in the fetus is directly proportional to the maternal concentration; an excess of fuels such as glucose and amino acids in a mother with poorly controlled IDDM will therefore lead to a similar excess in her fetus. Excesses of these nutrients stimulate the early maturation of fetal pancreatic beta cells and the excessive secretion of insulin, resulting in fetal hyperinsulinism and hence macrosomia. However, about 20 percent of mothers in whom nearly normal glycemia is maintained during pregnancy deliver infants who have macrosomia.21 Our finding of the transfer of insulin from mother to fetus as insulin-antibody complex offers a possible explanation for this apparent paradox. The transfer to the fetus of insulin injected into the mother would depend on the maternal concentration of anti-insulin antibody, and the total insulin concentration in the fetus would reflect not only the degree of maternal metabolic control, but also the maternal concentration of anti-insulin antibody. Since glycosylated hemoglobin measurements in these women were on average within the normal range, the correlation between the cord-serum concentration of animal insulin and birth weight in the 28 infants whose mothers received animal insulin supports the hypothesis that fetal growth was stimulated, at least in part, by the transferred insulin. Because women with IDDM treated with human insulin also acquire insulin antibodies, albeit to a lesser degree than those treated with porcine or bovine insulin,22 we postulate that the incidence of macrosomia in infants born to such mothers would Still be higher than that in the normal population. Indeed, our findings suggest that the complication of macrosomia will occur until a completely nonimmunogenic insulin becomes available for the treatment of IDDM.
Amniotic-fluid insulin is considered to be entirely fetal in origin and has been used as an index of the activity of fetal pancreatic beta cells.20 , 23 The finding of animal insulin in one sample of amniotic fluid we analyzed challenges this concept of exclusively fetal origin. Insulin in amniotic fluid may instead represent insulin administered to the mother, transferred to the fetal circulation, and thence into amniotic fluid.
In summary, we have documented the transfer to the fetus of insulin administered to women with IDDM. The magnitude of this placental transfer was directly related to the amount of anti-insulin antibody in the mother at the time of delivery, strongly suggesting that the transfer occurs as insulin-anti-insulin antibody complexes. The transferred insulin may appear in amniotic fluid and has biologic effects on the fetus, as reflected by the correlation with fetal macrosomia. These data also provide evidence for a deleterious effect of insulin-anti-insulin antibody complexes on neonatal morbidity in women with IDDM. Insulins of the lowest immunogenicity should be used in treating women with IDDM before and during their childbearing years.
Supported by a fellowship from the Jergens Foundation to Dr. Menon, a Research and Development Award from the American Diabetes Association and a FIRST award (DK 38541) from the National Institutes of Health to Dr. Cohen, grants (HD 12613 and HD 11725) from the Public Health Service to Dr. Sperling, and grants from the Diabetes Research and Education Foundation and Lilly Research Laboratories to Dr. Menon.
Presented at the Annual Scientific Sessions of the Society for Pediatric Research, May 7 to 11, 1990, Anaheim, Calif.
We are indebted to Mr. Walter Banach, Mr. Michal Rymaszewski, Mr. Michael Masnyk, Jr., and Ms. Victoria Neumann for their excellent technical help; to Dr. Steven Shoelson for helpful discussions regarding the high-performance liquid chromatography; and to Ms. Shirley Courtney for secretarial assistance.
Source Information
From the Divisions of Endocrinology (R.K.M., M.A.S., W.S.C.) and Neonatology (F.M.) and the Departments of Pediatrics and Medicine (R.M.C., J.C.K.), University of Cincinnati Medical Center, Cincinnati. Address reprint requests to Dr. Sperling at the Department of Pediatrics, Children's Hospital of Pittsburgh, 1 Children's Pl., Fifth Ave., Pittsburgh, PA 15213.
- References
References
1
Adam
Web of Science | Medline2
Buse
CrossRef | Web of Science | Medline3
Kalhan
CrossRef | Web of Science | Medline4
Berson
CrossRef | Web of Science | Medline5
Bauman
CrossRef | Web of Science | Medline6
White
Web of Science | Medline7
Mimouni
Web of Science | Medline8
Kaplan
Web of Science | Medline9
Usher
CrossRef | Web of Science | Medline10
Cohen
CrossRef | Web of Science | Medline11
Shoelson
CrossRef | Web of Science | Medline12
Palmer
CrossRef | Web of Science | Medline13
Scatchard
CrossRef | Web of Science14
Heding
CrossRef | Web of Science | Medline15
Knip
CrossRef | Web of Science | Medline16
Hill
CrossRef | Web of Science | Medline17
Sosenko
Full Text | Web of Science | Medline18
Robert
Full Text | Web of Science | Medline19
Pedersen
Medline20
Freinkel
CrossRef | Web of Science | Medline21
Small
CrossRef | Web of Science | Medline22
Deckert
Medline23
Greco
CrossRef | Web of Science | Medline
- Citing Articles (41)
Citing Articles
1
Elisabeth R. Mathiesen, Peter Damm, Lois Jovanovic, David R. McCance, Camilla Thyregod, Anders Boisen Jensen, Moshe Hod. (2011) Basal insulin analogues in diabetic pregnancy: a literature review and baseline results of a randomised, controlled trial in type 1 diabetes. Diabetes/Metabolism Research and Reviews 27:6, 543-551
CrossRef2
Kevin M. Pantalone, Charles Faiman, Leann Olansky. (2011) Insulin Glargine Use During Pregnancy. Endocrine Practice 17:3, 448-455
CrossRef3
Patrice Darmon. (2011) Des arguments expérimentaux rassurants pour l’utilisation de la glargine pendant la grossesse. Médecine des Maladies Métaboliques 5:1, H45
CrossRef4
Seok Hong Lee, Jihyun Ahn, Jaetaek Kim. (2011) Medical Therapy in Pregnant Women with Diabetes. Journal of Korean Diabetes 12:4, 201
CrossRef5
Kellie L.K. Tamashiro, Timothy H. Moran. (2010) Perinatal environment and its influences on metabolic programming of offspring. Physiology & Behavior 100:5, 560-566
CrossRef6
Michael F. Greene, Caren G. Solomon, Stephanie L. Lee, Robert A. Peterfreund. 2010. Diabetes Mellitus in Pregnancy. , 293-321.
CrossRef7
Chagit Klieger, Erika Pollex, Aleksey Kazmin, Gideon Koren. (2009) Hypoglycemics: Pharmacokinetic Considerations During Pregnancy. Therapeutic Drug Monitoring 31:5, 533-541
CrossRef8
SCOTT M NELSON, ROBERT S LINDSAY. (2009) TYPE 1 DIABETES IN PREGNANCY; INFLUENCES ON MOTHER AND FETUS. Fetal and Maternal Medicine Review 20:01, 17
CrossRef9
Gernot Desoye, Eleazar Shafrir, Sylvie Hauguel-de Mouzon. 2008. The placenta in diabetic pregnancy: Placental transfer of nutrients. , 47-56.
CrossRef10
Oded Langer. 2008. Oral anti-diabetic agents in pregnancy: Their time has come. , 217-227.
CrossRef11
Lois Jovanovic, John L. Kitzmiller. 2008. Insulin therapy in pregnancy. , 205-216.
CrossRef12
D. J. Pettitt, P. Ospina, C. Howard, H. Zisser, L. Jovanovic. (2007) Efficacy, safety and lack of immunogenicity of insulin aspart compared with regular human insulin for women with gestational diabetes mellitus. Diabetic Medicine 24:10, 1129-1135
CrossRef13
Charanpal Singh, Lois Jovanovic. (2007) Insulin Analogues in the Treatment of Diabetes in Pregnancy. Obstetrics and Gynecology Clinics of North America 34:2, 275-291
CrossRef14
B. E Levin. (2006) Metabolic imprinting: critical impact of the perinatal environment on the regulation of energy homeostasis. Philosophical Transactions of the Royal Society B: Biological Sciences 361:1471, 1107-1121
CrossRef15
Lois Jovanovic. (2005) Turning the Tide: Type 2 Diabetes Trends in Offspring of Mothers with Gestational Diabetes Mellitus. Metabolic Syndrome and Related Disorders 3:3, 233-243
CrossRef16
A. Lapolla, M. G. Dalfrà, D. Fedele. (2005) Insulin therapy in pregnancy complicated by diabetes: are insulin analogs a new tool?. Diabetes/Metabolism Research and Reviews 21:3, 241-252
CrossRef17
Lorin Lakasing, Catherine Williamson. (2005) Obstetric complications due to autoantibodies. Best Practice & Research Clinical Endocrinology & Metabolism 19:1, 149-175
CrossRef18
K. Gamson, S. Chia, L. Jovanovic. (2004) The safety and efficacy of insulin analogs in pregnancy. Journal of Maternal-Fetal and Neonatal Medicine 15:1, 26-34
CrossRef19
Alex C. Vidaeff, Edward R. Yeomans, Susan M. Ramin. (2003) Gestational Diabetes:. Obstetrical & Gynecological Survey 58:11, 759-769
CrossRef20
Samuel Edwin Fineberg, James H. Anderson. 2003. Complications of Insulin Therapy. .
CrossRef21
E. A. Masson, J. E. Patmore, P. D. Brash, M. Baxter, G. Caldwell, I. W. Gallen, P. A. Price, P. A. Vice, J. D. Walker, S. W. Lindow. (2003) Pregnancy outcome in Type 1 diabetes mellitus treated with insulin lispro (Humalog). Diabetic Medicine 20:1, 46-50
CrossRef22
M. R.G. Carrapato. (2003) The offspring of gestational diabetes. Journal of Perinatal Medicine 31:1, 5-11
CrossRef23
David Simmons. (2002) The utility and efficacy of the new insulins in the management of diabetes and pregnancy. Current Diabetes Reports 2:4, 331-336
CrossRef24
Oded Langer. (2002) Oral hypoglycemic agents and the pregnant diabetic: “From bench to bedside”. Seminars in Perinatology 26:3, 215-224
CrossRef25
LOIS JOVANOVIC. (2000) Role of Diet and Insulin Treatment of Diabetes in Pregnancy. Clinical Obstetrics and Gynecology 43:1, 46-55
CrossRef26
B.V.Rama Sastry. (1999) Techniques to study human placental transport. Advanced Drug Delivery Reviews 38:1, 17-39
CrossRef27
Peter A.M Weiss, Franz Kainer, Peter Pürstner, Gerda Zehetleitner, Ulla Hüttner, Josef Haas. (1998) Anti-insulin antibodies and birth weight in pregnancies complicated by diabetes. Early Human Development 53:2, 145-154
CrossRef28
Antoine Malek, Ruth Sager, Henning Schneider. (1998) Transport of Proteins Across the Human Placenta. American Journal of Reproductive Immunology 40:5, 347-351
CrossRef29
Annunziata Lapolla, Claudio Cardone, Paolo Negrin, Edoardo Midena, Stefania Marini, Carlo Gardellin, Daniela Bruttomesso, Domenico Fedele. (1998) Pregnancy Does Not Induce or Worsen Retinal and Peripheral Nerve Dysfunction in Insulin-Dependent Diabetic Women. Journal of Diabetes and its Complications 12:2, 74-80
CrossRef30
Antoine Malek, Ruth Sager, Alois B. Lang, Henning Schneider. (1997) Protein Transport Across the In Vitro Perfused Human Placenta. American Journal of Reproductive Immunology 38:4, 263-271
CrossRef31
Gernot Desoye, Michaele Hartmann, Carolyn J.P. Jones, Hans J. Wolf, Gabriele Kohnen, Georg Kosakke, Peter Kaufmann. (1997) Location of insulin receptors in the placenta and its progenitor tissues. Microscopy Research and Technique 38:1-2, 63-75
CrossRef32
Lois Jovanovic-Peterson, Charles M. Peterson. (1996) Review of Gestational Diabetes Mellitus and Low-calorie Diet and Physical Exercise as Therapy. Diabetes/Metabolism Reviews 12:4, 287-308
CrossRef33
Naomi Weintrob, Moshe Karp, Moshe Hod. (1996) Short- and long-range complications in offspring of diabetic mothers. Journal of Diabetes and its Complications 10:5, 294-301
CrossRef34
B. Persson, U. Hanson. (1996) Fetal size at birth in relation to quality of blood glucose control in pregnancies complicated by pregestational diabetes mellitus. BJOG: An International Journal of Obstetrics and Gynaecology 103:5, 427-433
CrossRef35
BERNADINE BRENNAN MOGLIA, DAVID S. PHELPS. (1996) Changes in Surfactant Protein A mRNA Levels in a Rat Model of Insulin-Treated Diabetic Pregnancy. Pediatric Research 39:2, 241-247
CrossRef36
Sarah R Wellik, Margarita de Veciana, Mark A Morgan, Kathleen M Berkowitz, Edward R Arquilla. (1995) Naturally occurring insulin autoantibodies in neonates of normal pregnancies and their relationship to insulinemia and birth weight. American Journal of Obstetrics and Gynecology 173:6, 1878-1884
CrossRef37
André J. Nahmias, Maurice Panigel, David A. Schwartz. (1994) Hematogenous infections of the placenta—An interdisciplinary and evolutionary perspective. Placenta 15, 107-136
CrossRef38
Gernot Desoye, Eleazar Shafrir. (1994) Placental metabolism and its regulation in health and diabetes. Molecular Aspects of Medicine 15:6, 505-682
CrossRef39
David Granot, Michael Snyder. (1993) Carbon source induces growth of stationary phase yeast cells, independent of carbon source metabolism. Yeast 9:5, 465-479
CrossRef40
(1991) Transplacental Passage of Insulin. New England Journal of Medicine 324:3, 198-199
Full Text41
Schwartz, Robert, . (1990) Hyperinsulinemia and Macrosomia. New England Journal of Medicine 323:5, 340-342
Full Text
- Letters
