Casele HL, Laifer SA. Factors Influencing Preconception Control of Glycemia in Diabetic Women. Arch Intern Med. 1998;158(12):1321-1324. doi:10.1001/archinte.158.12.1321
Although periconceptional glycemic control directly impacts perinatal outcome for pregestational diabetic women, these women still frequently enter pregnancy with suboptimal control of glycemia.
To determine how frequently diabetic women enter pregnancy with suboptimal glycemic control and to identify factors associated with not achieving optimal periconceptual control of glycemia.
Patients and Methods
Pregestational diabetic women (n=55) who attended the Maternal Fetal Medicine Faculty Practice or the Resident Medical Complications Obstetric Clinic, Magee Women's Hospital, Pittsburgh, Pa, were administered a brief questionnaire pertaining to periconceptional education and control of glycemia. Levels of periconceptional hemoglobin A1c were measured in all patients.
Sixty-one percent of pregestational diabetic women presented for prenatal care with suboptimal control of their blood glucose levels (hemoglobin A1c measurement, >0.08). Of the 55 diabetic women who participated in the questionnaire, 47 (86%) were under the care of a physician for their diabetes, 45 (82%) monitored their glucose level at least 3 times per day, 34 (60%) stated that they had been advised to plan a pregnancy, and 29 (53%) stated that they had planned their pregnancies. Women who had not been advised to achieve target hemoglobin A1c levels were more likely to enter pregnancy with suboptimal control of their blood glucose levels (P=.02). Women who experienced prior complications with pregnancy were significantly more likely to enter pregnancy with suboptimal control of their blood glucose levels (P=.02).
Diabetic women frequently enter pregnancy with suboptimal control of glycemia. Women not advised to achieve target glucose or hemoglobin A1c levels were more likely to enter pregnancy with suboptimal control of glycemia.
THE IMPORTANCE of periconceptual intensive glycemic control for the prevention of congenital anomalies and spontaneous abortions in women with diabetes has been clearly established.1- 5 Earlier studies demonstrated a direct correlation between glycemic control and hemoglobin A1c values and the incidence of congenital anomalies and spontaneous abortions.6,7 Later studies demonstrated that with intensification of glycemic control prior to conception and during the early weeks of pregnancy, the incidence of congenital anomalies and spontaneous abortions could be significantly reduced.8,9 Understanding the importance of periconceptual intensive glycemic control led to the strategy of preconception care of women with diabetes. The goals of preconception care include education, assessment of risks associated with pregnancy, modification (if needed) of diet, insulin therapy, frequency of self-monitoring of glucose levels, and improvement in glycemic control. Despite the universal acceptance in the obstetric community of the efficacy of preconception care, it was our observation that many patients with diabetes did not seek or were not referred for preconception care and maintained suboptimal glycemic control at the time of their first prenatal visit.
In 1993, results of the multicenter Diabetes Control and Complications Trial10 were published and provided important and convincing evidence that in contrast to conventional glycemic control, intensive glycemic control in patients with diabetes delays the onset and slows the progression of diabetic retinopathy, nephropathy, and neuropathy. By delaying the onset of end-organ damage and disability, intensive therapy is estimated to be cost-effective.11 The intensive therapy regimen was designed to achieve blood glucose levels as close to the reference range as possible.
It was our expectation that after dissemination of the results of the Diabetes Control and Complications Trial, more women of childbearing age would enter pregnancy with better glycemic control. The purpose of the present study was to determine what percentage of women with diabetes enter pregnancy with optimal glycemic control. We also sought to determine what factors were associated with achieving optimal periconceptional glycemic control.
All patients with pregestational diabetes attending the faculty Maternal Fetal Medicine Practice or Resident Medical Complications in Pregnancy Clinic, Magee Women's Hospital, Pittsburgh, Pa, between October 1994 and March 1997 were eligible to participate. During the first prenatal visit, all eligible patients were invited to participate in a research study, which was approved by the Research Review and Human Experimentation Committee at Magee Women's Hospital, and written informed consent was obtained. No patient participated in the study during more than 1 pregnancy, and no patient had participated in a formal preconceptional counseling program. Each patient was administered a questionnaire by a physician or nurse. The questionnaire consisted of 10 questions pertaining to preconception management of diabetes, pregnancy planning, diet, monitoring of blood glucose levels, preconception counseling, and preconception interventions. The hemoglobin A1c level was determined either before conception or during the first trimester. Measurements of preconception hemoglobin A1c, used in 2 patients, were obtained within 2 months of conception. Additional demographic information, such as maternal age, race, maternal weight, maternal height, presence of vascular complications, gravidity, prior poor outcome of pregnancy, and payer status, were taken from the patients' medical records. Poor outcome of pregnancy was defined as a congenital anomaly, 2 or more spontaneous abortions in the first trimester, or a fetal death in the second or third trimester.
Microsoft Excel (Microsoft, Seattle, Wash) and SPSS for Windows (SPSS Inc, Chicago, Ill) were used for all analyses. The data were analyzed using unconditional logistic regression to obtain point estimates of risk between the factors, such as prior poor outcome of pregnancy, gravidity, percentage of ideal body weight, age at diagnosis of diabetes, vascular complications, and the outcome variable of glycemic control. Unconditional logistic regression analysis was also used to evaluate the relationship between responses to the questionnaire and glycemic control. The logistic model provides odds ratios that are upwardly biased estimates of risk ratios when the outcome of interest (suboptimal glycemic control) is present in 60% of the cohort evaluated. The odds ratio is used as a measure of association independent of its estimate of a risk ratio.12 χ2 tests were used to compare proportions and Student t tests were used to compare means, as appropriate. A P value of less than .05 was considered significant.
Based on recent recommendations by the American Diabetes Association, we defined suboptimal glycemic control as a hemoglobin A1c value greater than 0.08.13 The normal nondiabetic hemoglobin A1c range at our institution is 0.04 to 0.06, and we defined optimal control as a periconceptional value in this range. Adequate control was defined as a hemoglobin A1c level between 0.06 and 0.08. Hemoglobin A1c measurements were done at our institution using cation exchange chromatography (DiaSTAT Hemoglobin A1C program, Bio-Rad Laboratories, Hercules, Calif).
Between October 1994 and March 1997, 94 pregestational diabetic women presented for prenatal care, and 55 participated in the study. The 39 women not included either declined to participate or were not brought to the investigators' attention. The mean hemoglobin A1c level and gestational age at first visit of both the entire pregnant diabetic population and the study cohort were not significantly different (P>.05). The proportion of patients with optimal, adequate, and suboptimal glycemic control was similar between the study patients and the entire population of pregestational diabetic women (Table 1).
Of the patients who participated in the study, 33 (60%) had suboptimal control while only 6 (11%) had hemoglobin A1c values in the normal range. Women with suboptimal glycemic control did not differ from women with good or adequate control with regard to payer status, maternal age, race, or maternal weight (P>.05). Multiparity, obesity (ideal body weight ≥120%), presence of vascular complications, and age at diagnosis of diabetes were not factors predictive of suboptimal glycemic control (Table 2). However, women with a prior poor outcome of pregnancy were significantly more likely to enter pregnancy with poor glycemic control (P=.02).
Unconditional logistic regression analysis revealed that not being advised to achieve target glucose or hemoglobin A1c values was significantly associated with entering pregnancy with poor glycemic control (P=.02) (Table 3). A total of 47 patients (86%) were followed up before conception by a family practitioner, internist, or endocrinologist for their diabetes, 45 (82%) monitored their blood glucose level at least 3 times per day, 33 (60%) stated that they had been advised to avoid an unplanned pregnancy, and 34 (62%) stated that they were aware of the complications of diabetes in pregnancy, but these were not significant factors associated with periconceptional glycemic control. Overall, 29 diabetic women (53%) stated that they had planned their pregnancies, but only 12 (22%) saw a physician before conception to modify their insulin intake or glycemic control.
Pregestational diabetes complicates 1 in 200 pregnancies.14 Congenital anomalies represent the leading cause of perinatal mortality in infants of diabetic mothers and occur early in pregnancy, by 5 to 8 weeks of gestation, usually before patients present for prenatal care.3 The likelihood of a congenital anomaly is directly related to the degree of glycemic control at the time of conception and during organogenesis. Data from the Joslin Clinic indicate that hemoglobin A1c values greater than 6 SDs above the mean (>0.09) increase the risk for the development of malformations.6 To improve periconceptional glycemic control and prevent malformations, diabetic women must be encouraged to seek preconception care; this strategy has been found to be effective in clinical trials. However, the findings of our study and a recent study by Rodgers and Rodgers15 support our impression that preconception care and optimization of periconceptual glycemic control is not occurring in clinical practice. Only 12% of patients in our cohort had normal hemoglobin A1c concentrations at their first prenatal visit, and almost 40% of patients had hemoglobin A1c values in the range associated with an increased risk of congenital malformations.
Why is the strategy of preconception care ineffective in clinical practice? We sought to determine what factors might influence whether patients with diabetes enter pregnancy with adequate or suboptimal glycemic control. We found that women who were not advised to achieve target hemoglobin A1c and glucose levels in preparation for pregnancy were significantly more likely to enter pregnancy with suboptimal glycemic control. It is our impression that patient education delivered by physicians may be insufficient to prepare patients for possible diabetic complications during pregnancy. Although most patients were followed up by a physician for their diabetes and most were counseled to plan their pregnancies, patients may not be properly educated about the specific target levels of blood glucose and hemoglobin A1c they should achieve. Achieving target blood glucose levels may require more frequent communication and testing prior to conception. We suspect that this does not happen with many patients. Interestingly, patients with a prior adverse outcome of pregnancy were more likely to have suboptimal control of their blood glucose levels. Our interpretation of this finding, based on clinical experience, is that there is a subgroup of patients with diabetes who do not adhere to recommendations regarding their care (diet, self-monitoring, and/or insulin doses). This may be due to attitude (ie, motivation) or organizational (ie, access to care) barriers.16 Presently, it is not clear which strategies may be effective in eliminating these barriers.
We recognize the limitations of a questionnaire study and the subjective interpretation of our findings. However, it is clear from the initial hemoglobin A1c values in the entire cohort of women with diabetes that optimal periconceptual control is not being achieved. We encourage other groups and investigators to determine how effectively periconceptional glycemic control is being achieved in populations of diabetic women of childbearing age. Since the completion of this study, the American Diabetes Association has published a new position statement on the standards of care for patients with diabetes.17 This latest position recommends a hemoglobin A1c goal of less than 0.07 for diabetic patients and specifically calls for preconception and contraceptive counseling in diabetic women of childbearing age.17 Innovative techniques to improve physician and patient education and patient motivation and compliance are also needed.
Another major obstacle to the success of any preconception strategy is that many patients do not plan their pregnancies. Rodgers and Rodgers15 found that 73% of diabetic women did not plan their pregnancies and speculated that this was possibly etiologic in suboptimal periconceptual glycemic control. Although in our study glycemic control was not significantly different between patients who stated that they planned or did not plan their pregnancies, we also observed 26 unplanned pregnancies (47%). We propose that in addition to preconceptional counseling, an ideal strategy for optimizing periconceptual glycemic control would be to implement strict monitoring of blood glucose levels for all diabetic women of childbearing age. The goal of intensive therapy, normalization of blood glucose levels, is the same as the goal of periconceptional glycemic control. Because the goals of these 2 strategies are similar, women who do not plan their pregnancies or who do not seek preconception care are still likely to enter pregnancy with optimal glycemic control. Patients who frequently monitor and adjust their diabetes regimen are more likely to maintain strict control of their blood glucose levels throughout pregnancy. It is our opinion that implementation of strict monitoring of blood glucose levels, in addition to its efficacy in reducing long-term complications in persons with diabetes, is likely to have a positive impact on preconception control and pregnancy outcome in diabetic women of childbearing age.
In conclusion, we have shown that diabetic women frequently enter pregnancy with poor glycemic control. Although many diabetic women are aware of the complications of diabetes in pregnancy, these patients may not be sufficiently educated about the target glucose levels recommended for periconceptual control. We propose that the implementation of an intensive effort to control blood glucose levels in all diabetic women of childbearing age, based on the findings of the Diabetes Control and Complications Trial, would obviate the need for preconception regimen changes and would result in patients entering pregnancy with better glycemic control.
Accepted for publication December 30, 1997.
Reprints: Holly L. Casele, MD, Division of Maternal Fetal Medicine, Evanston Hospital, 2650 Ridge Ave, Suite 1600 WH, Evanston, IL 60201.