Treatment of Gestational Diabetes Mellitus and Outcomes on Blood Glucose Levels:
A Meta Analysis
By Talia Sankari and Song Kim
Preliminary Abstract:
Obesity serves as a leading cause of gestation-related issues in women. Varying from infertility to eventual congenital birth defects, excessive weight gain prior to and during pregnancy can cause adverse gestational consequences and fetal abnormalities. In early gestation, obese women are more prone to insulin resistance, which can lead to gestational diabetes mellitus (GDM). Many treatment options are available for both high risk pregnant women and those with rapid weight gain during gestation. While lifestyle changes including diet and exercise are endorsed, common medical therapies also include glucose regulation through insulin injection and Metformin oral ingestion. This meta-analysis examines the effects of lifestyle changes in the form of regular exercise and Metformin medication intake in patients who develop gestational diabetes by comparing the fed two hour postprandial blood glucose levels across various screened studies.
Ten clinical studies involving a total of _____ individuals were included in the final data analysis. The dataset averages produced from each trial had statistically insignificant regression tests in comparison of 2-h-postprandial glucose levels measured in the second and third trimesters of gestation. Compared to the control group, exercise as a natural treatment option was related to lower overall blood glucose levels, and similar results were found when comparing Metformin to a control insulin group. All other outcomes compared were statistically insignificant.
Introduction:
Gestational Diabetes Mellitus (GDM) is a condition characterized by hyperglycemia and insulin intolerance during pregnancy which can ultimately cause fetal overgrowth, antagonistic cardiovascular health in offspring, and eventual adolescent obesity (American Diabetes Association, 2018). Worldwide obesity has tripled since 1975 and as levels continue to rise, cases of GDM have risen to a striking 9.2%, contributing to a vicious cycle of GDM-induced childhood obesity leading to future GDM, largely due to lack of available treatment and early diagnosis (American Diabetes Association, 2018). Women who experience GDM have been found to be six times more likely to develop type 2 diabetes later in life, with up to 50% of affected women doing so (Redden et. al., 2011); however, most women affected by GDM have never experienced high glucose levels prior to pregnancy (American Diabetes Association, 2018).
Among the various recognized factors which lead to development of GDM, the most common and harmful has been found to be maternal obesity prior to or during pregnancy, as well as excessive weight gain during gestation (Simmons et. al., 2017). One of the most apparent direct consequences of this condition is its influence on the birth weight of the child; GDM is known to cause macrosomia (American Diabetes Association, 2018). Babies born large at gestational age are at risk for injury at birth, low blood sugar, and future complications, including higher predisposition to develop obesity later in life (National Institutes of Health). Thus, the causes and effects of GDM—both long-term and immediate—can put the lives and welfare of both mother and child at great risk. Researching and understanding the mechanisms which contribute to the disease as well as best possible treatments may help us prevent these negative outcomes and hopefully alleviate some of the physical and mental stress placed on both individuals involved.
Despite its short lifespan within the body of the mother, untreated GDM has been known to greatly increase health risks for both the mother and child involved, including cardiovascular disease, macrosomia, defined as above average fetal weight at birth, and predisposition to develop type 2 diabetes (Reece, 2010). Healthy pregnancies proceed to prevent delivering excess glucose to the fetus by releasing extra insulin from the beta cells within the pancreas, bringing down blood sugar levels accordingly; in pregnancies affected by GDM, however, the pancreas fails to produce this added insulin and glucose levels rise to unhealthy levels, inducing diabetes and causing unhealthy weights in both the mother and fetus (Brawerman and Dolinsky, 2018).
In patients with type 2 diabetes, the implementation of regular exercise has been found to help combat the effects of hyperglycemia over time by managing blood glucose levels, blood pressure, body weight, and promoting a healthy lifestyle overall (Stanford et. al., 2014). In addition, exercise can aid in lowering risk of complications resulting from the disease later in life, such as kidney failure, nerve damage, and even amputations of limbs (Harrison et. al., 2016). People with diabetes suffer from insulin resistance, meaning that their muscles, liver, and tissue fails to uptake the insulin which is circulating through the body (Stanford). In response, the beta cells in the pancreas proceed to produce extra insulin to compensate, which can lead to the shutdown of this system all together (Stanford). This excess amount of insulin in the bloodstream impairs the process of glucose transport throughout the body to these same areas: the liver, adipose tissue, and—most importantly for our research here—the skeletal muscles.
Despite the failure of skeletal muscles to uptake glucose when induced by insulin in people with Type 2 diabetes, this same process is successful at nearly normal levels when induced by exercise instead (Stanford). Both systems activate glucose uptake by the muscles through stimulation of the GLUT4 glucose transporter, effectively translocating it into the plasma membrane by way of specific signaling mechanisms (Stanford). This glucose is thus successfully taken out of the blood and absorbed into the skeletal muscles, directly lowering blood glucose levels and making this measurement the most indicative metric by which to assess successful glucose uptake. By doing the job which insulin has ceased to perform in the body, exercise successfully increases glucose uptake and decreases insulin resistance in people with Type 2 Diabetes (Harrison).
However, in terms of gestational diabetes, results of this type of treatment tend to vary greatly, likely due to the high number of variations in exercise regimens set up for the patients in question. Some studies (reference these studies specifically) have found significant reductions in both fasting and fed blood glucose levels resulting from regiments of regular physical activity, indicating exercise as a possible natural alternative to medical treatment when combined with diet control (Redden et. al., 2011). Various other studies (reference these studies specifically), however, have reported either inconclusive or unsuccessful results. Research on the subject of natural exercise intervention as compared to medicinal metformin treatment has not been conducted to sufficiently discern the advantage of one over the other. Because background information is so inconclusive, this meta-analysis aims to compare both exercise and medical treatments options in regulating gestational blood glucose levels.
INSERT BIORATIONALE FOR METFORMIN HERE:
Medical treatment for and prevention of gestational weight gain typically follow one of two medicinal treatments: insulin injection therapy or oral hypoglycemic treatments such as Metformin intake (Brawerman and Dolinsky, 2018).
Early identification of GDM can be beneficial in implementing intervention plans in order to lessen neonatal consequences and possibly reduce the severity of cyclical family obesity. The end goal of treatment is to reduce hypoglycemia in the mother, in turn controlling gestational weight gain with the hopes of reducing the adverse effects of gestational diabetes on both parties involved (Ali et. al., 2018). Success of various treatments is measured by reduction of fed glucose levels, taken intermittently throughout and after treatment trials (Simmons et. al., 2017).
Studies that examine the effects of exercise therapy and Metformin intake on neonatal outcomes of mothers with GDM will be utilized and analyzed to compare the significance of the two medical interventions. This meta-analysis demonstrates the effects and impact of medical treatment interventions and implementations of exercise regiments on GDM, measured through fed glucose levels, and whether one has an advantage over the other.
We predict Metformin intake to have a clear benefit on patients of GDM, measured through a blatant decrease in blood glucose levels over time and reduction of neonatal outcomes. We expect lifestyle changes through exercise to also have a positive effect on individuals diagnosed with GDM, although the measurable difference in blood glucose may not be as great. As for consideration of a comparison between the intervention of medicine versus exercise, we expect to find that medicinal regulation of hypoglycemia has more tangible and reliable differences on the negative effects of GDM.
Methods:
Exercise as natural intervention
In order to assess the effects of exercise as intervention on GDM, we looked for randomized control trials consisting of women already diagnosed with GDM, ages 18-50, at 32 weeks of pregnancy or less, to leave sufficient amount of time for progress to be seen and measured prior to birth. To find these, we searched the databases Wiley Online Library, Ovid, and Science Direct, using the text words “gestational diabetes,” “exercise,” “blood glucose,” and “treatment.” Efficacy of treatment across trials was measured by changes in measured 1- or 2-hour postprandial (post-meal) glucose levels at the end of each trial period. Studies were excluded from consideration if they did not report post-trial postprandial glucose levels for both exercise and control groups or if they did not report the exact time and frequency of exercise treatment. Trials using women with prior exposure or diagnoses of diabetes (GDM or otherwise) were not considered.
According to the selection criteria previously described, the following five articles were chosen for data collection and assessment: “Resistance exercise and glycemic control in women with gestational diabetes mellitus,” a 2010 study by de Barros et. al.; Kokica et. al.,’s 2016 “Combination of a structured aerobic and resistance exercise improves glycaemic control in pregnant women diagnosed with gestational diabetes mellitus” ; “Simple lifestyle recommendations and the outcomes of gestational diabetes” from Bo et. al. in 2014; another 2014 study on “The effects of mindfulness eating and yoga exercise on blood sugar levels of pregnant women with gestational diabetes mellitus” done by Youngwanichsetha et. al.; and lastly, a classic 1997 study on “Effects of a partially home-based exercise program for women with gestational diabetes” by Avery et. al.
To assess the efficacy of increased exercise in reducing the effects of GDM, we compared the average observed difference in 2-h postprandial glucose levels between exercise groups and control groups measured after the trial period was over. This analysis allowed us to directly assess the impact of exercise intervention versus no exercise at all. In order to determine the significance of these differences across trials, we ran a regression test to obtain a p-value which will directly display the success or failure of exercise in regulating hyperglycemia on patients with GDM. In addition to this, we assessed the ability of increased exercise to further bring down glucose levels by comparing the minutes per week of exercise performed across trials with the respective measured postprandial glucose levels. We ran regression tests of these values to obtain p-values which will display the significances, or lack thereof, of increased exercise on further bringing down blood glucose levels. To achieve the raw numbers with which to run the regression tests, we went through the results of 10 total studies and subtracted the group’s postprandial glucose from each of those of the control groups’, respectively. All blood-glucose measurements will be calculated in mmol/L, and were converted from mg/dl if necessary.
B. Metformin as medical intervention
Irregular insulin levels in the body can be combated by hormonal maintenance through medication. Trials used in this meta-analysis regard Metformin oral tablet ingestion, the most common treatment option for GDM. Only studies with supplemental and pharmacologic tablet milligrams were considered. Five studies were analyzed in evaluate data surrounding the effects of different medical interventions on prevention of GDM. To determine the most relevant studies for analysis, the terms “gestational diabetes mellitus,” “randomized,” “ Metformin, ” and “fed and fasting blood glucose levels” were searched in various University of Wisconsin-Madison library databases, and the ones used in this analysis were found on Web of Science or Science Direct.
Each trial used in this analysis was a randomized, cohort study with a reference control group of insulin injection therapy and ethically practiced with guidance of the patients’ primary obstetricians and emergency personnel. The trials consisted of pregnant women and their offspring with minimum two interactions with researchers (prenatal and postnatal), but all studies considered were longitudinal cases spanning primary treatment for GDM: during the first and second trimesters of gestation. Studies performed on animals were eliminated for consistency.
In each study, fed and fasting glucose level data was obtained at least twice a day: at fasting and two hours after mealtime. In all studies, Metformin intake ranged from starting values of 500-750 mg per day (mg/d) with necessary dose adjustments made up to 2550 mg/d. Insulin intake varied from 0.2-1.0 units/kg per day. Patients already taking Metformin prior to gestation or who were previously diagnosed with diabetes were not provisioned in order to yield more accurate and consistent results. In order to evaluate the capacity of Metformin as a intervention method for GDM, this meta analysis will compare the average observed difference in 2-h postprandial blood glucose levels between Metformin group and a control insulin therapy group. Doing so allows us to directly compare the impact of medical intervention through Metformin versus insulin injection therapy as a control method. In order to determine the significance of this difference across trials, regression tests were run to obtain a p-value. This value will prove whether or not Metformin is a significant intervention method to combating the effects of GDM. The raw data was found and narrowed down into five different studies which were run through regression tests. The figures used were calculated by subtracting the group’s pre-treatment postprandial glucose levels from the post-treatment postprandial glucose levels.
Results:
Exercise as natural intervention
After compiling the data from the 5 selected studies, we sorted data into categories of average minutes per week of exercise, average 2-h postprandial glucose levels of both control and treatment (exercise) groups, and calculated difference between the two groups’ glucose levels at the end of treatment, a value we achieved in order to assess the actual difference exercise may have induced. We also organized non-numerical data pertaining to the studies, including the type of exercise performed by the treatment group (aerobic or resistance), whether a recommended diet plan was associated with the study, and if the reported adherence to at-home regiments was reported to be over the 70% cutoff to be deemed “satisfactory.”
Once we averaged out the data among these categories and converted all measurements to the same metric, we established set ranges found for each numerical dataset. Average minutes of exercise per week ranged from about 70-320 minutes. Blood glucose levels of the control group went from 5.30 ± .47 up to 6.51 ± .92 mmol/L, while the exercise group showed averages of 4.66 ± 0.46 to 5.75 ± .55 mmol/L. The calculated differences between these values for control and exercise groups ranged from .14 to .64 mmol/L.
Different types of exercise routines were assigned to participants: 20% performed resistance exercises consisting of circuit training using an elastic band to target main muscle groups; 40% did an aerobic exercise regimen aimed at cardiovascular activity such as walking or cycling; and 40% were assigned to do both aerobic and resistance exercises, either through an activity like yoga which combines the two practices or by performing a resistance workout a few times per week in addition to going on daily walks. 60% of participants in exercise groups were also advised to incorporate some kind of diet or general eating recommendations. 80% of those involved in exercise reported satisfactory adherence to at-home plans, making only 20% (one study) questionable in terms of commitment and efficacy.
In order to assess efficacy of exercise on decreasing blood glucose levels, we compared in a bar graph the average 2-h postprandial glucose levels taken at the end of the trial period of the control group vs. the exercise group (Figure 1). An ANOVA Single Factor test ran on this dataset provided an insignificant p-value of 0.1859, despite the control group having, on average, a blood glucose level .482 mmol/L higher than the exercise group.
An analysis of the relationship between minutes of exercise performed per week and the resulting average 2-h postprandial glucose level was done by comparing these two variables on a scatter plot, which yielded an negative slope (Figure. 2). A regression test was done on this relationship, which proved to be insignificant, with the p-value of the intercept at 0.1225. However, considering the unsatisfactory 66.3% average adherence of one of the data points (140 minutes, 5.89 mmol/L), this test was redone excluding this input value; the resulting regression test provided a significant p-value of .002957 (Figure 3). A similar assessment was then done on …
A bar graph relating type of exercise (aerobic, resistance, or both) and resulting average blood glucose level was produced and
B. Metformin as medical intervention
Data from the studies were classified into average Metformin dosages, average 2-h postprandial glucose levels of both control and treatment (exercise) groups, and the calculated differences when compared to the control group treatment, which was then statistically analyzed to determine impact of Metformin use to combat GDM. The averages of all data points were found, all measurements were converted into the same metric.
Average Metformin dosage was 1595 mg/day, with the range of every study being between 500-2500 mg/day. Average fed blood glucose levels of the Metformin was 7.057 mmol/L, while average fed blood glucose levels for the control insulin group was 7.186 mmol/L. The calculated differences between these values for control and exercise groups ranged from .01 mmol/L to .7 mmol/L.
Comparison to determine the success of Metformin in decreasing blood glucose levels was represented with a bar graph which contrasts average 2-h postprandial glucose levels taken at the end of the experimental period of the Metformin group group vs. the control insulin therapy group (Figure 2). By using ANOVA Single Factor to run a regression test, our dataset for Metformin versus control group insignificant p-value of 0.9026, despite the control group have an average of .1286 mmol/L higher than the Metformin group.
Discussion:
Exercise as Natural Intervention (by Talia Sankari)
Our analysis of the differences in both fed and fasting glucose levels across the 5 exercise trials results in p-values that were both around .08, indicating an insignificant relationship between the various trials’ differences in exercise vs. control groups, as it is higher than the cutoff value of 0.05. Based on this, we cannot definitively confirm that regulation of hyperglycemia through a disciplined exercise regimen is effective in bringing down glucose levels and managing gestational diabetes. This, however, is despite the fact that there was a positive difference between control groups and exercise groups in every study for the postprandial glucose levels, and in almost every trial for the fasting glucose levels; the only exceptions were in the case of 2 confounding studies which yielded results of no difference between the two groups in fasting glucose measurements.
Throughout the literature, there seems to be a general lack of consensus considering the efficacy of regular exercise as a confirmed method of treatment of GDM; (This lack of consistency is the reason we are doing the meta-analysis, and the issue your analysis should be attempting to resolve.)–should probably change this/take it out, mention it but show how I resolve it. however, there is agreement among researchers and patients alike concerning the general benefits of increased physical activity in the life of someone suffering from GDM, as well as consistent lack of adverse effects of exercise during pregnancy, as reported by the patients themselves. Various confounding factors may be heavily influencing these unreliable, (Why are they unreliable? Inconsistent doesn’t mean unreliable.)–either take this out or address why they are unreliable. and sometimes contradictory results, such as the extreme variety in exercise programs across trials, as well as variety in supplementary diet plans which were oftentimes added onto treatments and sometimes purposely excluded. In trials designed with at-home exercise regimens or amounts of exercise that were simply recommended to be performed, there was the factor of adherence to exercise plan by the patients to be considered, which was sometimes taken into account as a confounding variable and added into the significant differences in data, but often not considered at all.
B. Metformin as medical intervention
Analysis of all studies considered drew inconclusive results as some showed a statistically significant benefit of Metformin use in treating GDM in patients over insulin administration, while some showed no statistically significant correlation with total average P-value of 0.902. However, the insulin control group was met with a average blood glucose level .1286 mmol/L than the average blood glucose levels of the Metformin groups. The inconsistencies in results could be attributed to the varying lengths of the trials and the lack of strict diet and exercise regimens in the studies analyzed, which could act as confounding variables. Since many of the trials only lasted until birth, the positive effects of Metformin in terms of postnatal child development (i.e., reduced likelihood of childhood obesity or development of type 2 diabetes) seen in two of the trials may not have been accounted for, thus deducing inconclusively. None of the researchers in the studies analyzed implemented a strict diet and exercise plan in order to rule out nutritional confounding variability. To analyze this discrepancy, trials with greater duration and stricter lifestyle guidelines must to be conducted in order to analyze the effects of Metformin on its own. Metformin is still considered to be the leading medication for treating hyperglycemia perinatally, and continues to be used both as an alternative to insulin therapy and as a pre-insulin treatment option.
C. Comparative Discussion
All studies used concluded that early GDM diagnosis when combined with both insulin therapy and metformin could provide optimal neonatal results, reducing hyperglycemia in the mother as well as the risks for macrosomia, respiratory distress and birth trauma. This research is crucial to the understanding of preventative measures affecting perinatal health of mothers and their unborn children and how to reduce the likelihood of adverse neonatal outcomes. GDM affects the lives of millions of newborns and their mothers individuals across the world and understanding the importance of early diagnosis and treatment options may be able to provide these individuals with reduced effects and even prevention of future diseases.