Screening and management of gestational diabetes mellitus in Saudi Arabia

Screening for Gestational Diabetes

Gestational diabetes mellitus (GDM) is caused by the development of glucose intolerance during pregnancy (National Institutes of Health 2013). In the United States the National Institutes of Health (2013), U.S. Preventive Services Task Force (2008), and the American Diabetes Association (2013) are just a few agencies and organizations who have weighed in on this topic. Elsewhere, the Cochrane Systematic Reviews (Tieu et al. 2010) and the World Health Organization (WHO 2013a) have also published their recommendations. Despite the overwhelming number of recommendations concerning GDM, the best GDM screening protocols, diagnostic methods, and treatment approaches remain controversial.

To better understand recommended best practices for GDM care this proposal will first examine what is known about this disease and then review the evidence-based rationales underlying current recommendations. Particular attention will be paid to the health care challenges facing emerging developed nations, in particular Saudi Arabia.

GDM Epidemiology

The International Diabetes Federation (IDF 2013) estimates that 382 million people globally were suffering from diabetes in 2013, of which nearly half remain undiagnosed. This number is expected to increase to 592 million by 2035, with the vast majority (80%) residing in low- and middle-income countries. The economic burden is estimated to be over half a trillion dollars (U.S.), which represents close to 11% of all healthcare spending on adults.

A recent systematic review estimated the global prevalence of hyperglycemia during pregnancy to be 14.8% (Guariguata et al. 2013). The prevalence rate for North Africa and the Middle East was higher at 17.5%, but the range among all seven IDF regions reached a high of 25.0% for Southeast Asia and a low of 10.4% for the North American Continent. Of these women, close to 16% worldwide were diabetic before pregnancy or undiagnosed.

A recent WHO report (2013b: 13) reported a diabetes prevalence rate of 20% in Saudi Arabia, among the highest in the world. Al-Daghri and colleagues (2011) examined the prevalence of GDM and found 1.4% of women between 18 and 45 and 1.5% of women between 46 and 60 developed the disease during pregnancy. The higher rate among older women is not too surprising given that age is a risk factor for GDM. In addition, an estimated 36.4% of women between 18- and 45-years of age were obese compared to 61.9% for women between 46- and 60-years of age. Obesity also predicts the prevalence of type 2 diabetes among Saudi women, with 9.5% having the disease between the ages of 18 and 45 and 44.1% between 46 and 60-years of age.

GDM Etiology and Risk Factors

Hyperplasia of the pancreatic ?-cells leads to increased insulin production, which is fortunate because pregnancy causes increased insulin resistance after a short period of increased insulin sensitivity (Prutsky et al. 2013). The progressive development of insulin resistance is normal and caused by diabetogenic hormones, such as placental lactogen, estrogen, and prolactin. When insulin production is insufficient to control blood glucose levels, however, GDM develops. There are a number of GDM risk factors that have been identified, including non-European ancestry, overweight, obesity, age, poor diet, sedentary lifestyle, fertility problems, and a family history of diabetes (Zhang et al. 2013).

Adverse Outcomes Associated with GDM

The general view is that GDM should be treated during pregnancy due to the significant risks of adverse outcomes for both mother and child (National Institutes of Health 2013). The short-term adverse outcomes for the mother are gestational hypertension, proteinuria, and preeclampsia, while the long-term consequences include eventual development of type 2 diabetes, metabolic syndrome, and cardiovascular disease. The National Diabetes Education Program (n.d.) in the U.S. estimates that 35 to 60% of women who developed GDM will develop diabetes within 10 to 20 years. The risks to the fetus are hyperinsulinemia, macrosomia, shoulder dystocia, caesarean delivery, respiratory distress syndrome, and the emergence of metabolic problems during the perinatal period (National Institutes of Health 2013). Notably, of the children born during 2013 an estimated 21 million were exposed to maternal hyperglycemia (IDF 2013). Fetuses exposed to GDM are also believed to have an increased risk of developing diabetes later in life (National Diabetes Education Program n.d.).

Treatment of GDM

The first-line treatment approach for women diagnosed with GDM, according to the American College of Obstetricians and Gynecologists (2013: 409-410), is nutrition therapy to reduce hyperglycemia, prevent ketosis, and weight management. The recommended calorie distribution is 33 to 40% carbohydrates, 20% protein, and 40% fat. When possible, a low-glycemic index diet should be combined with regular, moderate exercise. Should diet modification and exercise fail to normalize plasma glucose levels then GDM should be treated pharmacologically. The primary recommended drug therapy is insulin, because it does not cross the placenta and achieves tight control over blood glucose levels. The initial dose ranges between 0.7 and 1.0 units/kg/d, divided evenly between several administrations during the day.

GDM Screening in Practice

OGTT, in its various forms, is the most widely-recognized laboratory screening method for GDM (Hanna et al. 2008). Screening for risk factors is also a critical method for identifying which patients need to be screened during the first trimester; however in countries which have not implemented national recommendations, such as the United Kingdom, the screening methods being utilized by clinicians can vary widely. A recent survey in the UK has found that routine screening was performed in 52% of the antenatal clinics who responded to a survey, while the others relied on a two-step protocol involving an evaluation of risk factors before conducting an OGTT (Hanna et al. 2008). The reasons given for not performing universal OGTT screening included no national recommendation, low GDM risk, and organizational obstacles.

In terms of screening methods, respondents to the survey predominantly relied on glycosuria (54%) and/or random plasma glucose (52%) for a first screen (Hanna et al. 2008). The second most common first screen was risk factors (34%), followed by fasting plasma glucose (27%). The cutoff values for fasting glucose used by respondents varied from 5.5 to 7.8 mmol/L. The secondary screening method used was 2-hour OGTT with a 75 g glucose challenge, but the cutoff values also varied from 7.0 to 11.1 mmol/L. These findings reveal a lack of consensus on screening strategies in some countries.

The use of reagent test strips to detect glycosuria when screening for GDM is a common practice in many countries, even the United States (Alto 2005). However, government and organizational guidelines do not recommend dipsticks for GDM screening. When William Alto reviewed the research literature on this topic in 2005, he found that four published studies revealed urinalysis unreliable across the board. The false-positive and -- negative rates are too high compared to oral glucose testing to justify its continued use as a GDM screening method. Mshelia and Buba (2005) examined how reproducible glycosuria testing was in a Nigerian clinic and discovered that only 32% were replicated on the day that OGTT was performed. Of these, only 22 and 11% were later diagnosed with GDM and pregestational diabetes, respectively.

Study Justification

I am a staff nurse in the Obstetrics and Gynecology ward of the King Fisal Specialist Hospital and Research Centre, Jeddah Branch, Saudi Arabia. My interest is in prenatal screening, especially evidence-based GDM screening methods. In my practice I encounter many pregnant women diagnosed with GDM, but the screening and diagnosis often depends solely on the detection of glycosuria. Given the findings from the above literature review the chances that a significant percentage of these patients represent false positives are high. In addition, many women who may have GDM are remaining undiagnosed. Both patient groups would obviously benefit from the use of fasting plasma glucose (FPG) and/or OGTT screening methods, given the greater sensitivity and specificity these methods offer.

Another common practice in my area of practice is forgoing urinalysis testing for glycosuria unless patients present with a history of diabetes or complain of symptoms consistent with diabetes, such as vaginal irritation or burning sensation. As discussed in the above literature review, when using urinalysis to test for hyperglycemia it is important to repeat the test often whether or not patients are symptomatic. In addition, positive tests should be replicated before giving patients a diagnosis of GDM. The obvious preference, however, would be to screen and diagnose GDM using FPG or OGTT. I have also observed clinicians using a finger stick to test for gestational diabetes, but based on research discussed above laboratory testing should be performed before a definitive diagnosis is given and treatment started. A Medline search for research articles on GDM and finger sticks retrieved only two citations and both discussed the value of finger sticks for effective management of GDM, not as a screening method or diagnostic tool.

There is thus a great need in my practice and country's healthcare system for a better understanding of best practice guidelines for GDM screening and diagnosis. Given the large number of recent systematic reviews that universally endorse the use of OGTT for GDM screening and diagnosis the following systematic review proposal will be limited to a cost-effectiveness comparison between the recommended screening methods and protocols. My hope is that this information will help clinicians dispense with the use of ineffective screening and diagnostic tools, especially in light of the potential harm suffered by undiagnosed patients or patients needlessly treated for GDM.

Part II: Systematic Review Proposal for GDM Screening Cost-Effectiveness

Based on the above analysis the most reliable GDM screening method for high-risk and low-risk patients is plasma glucose sampling before and after an oral glucose challenge. If the controversy over OGTT cutoff points can be ignored for now, these screening methods have been universally endorsed by all the international, national, and professional organizations concerned with diabetes screening. The main challenge facing health agencies with limited resources and developing nations is whether one screening method is more cost-effective than another. For this reason, a proposal for a systematic review of cost-benefit analyses of different screening protocols will be presented here.

Objectives

Regardless of whether low-risk or high risk pregnant women are being screened, the research literature and published recommendations by international and national health agencies agree that both should be screened by testing blood glucose levels. The various screening methods that have been employed include FPG, random plasma glucose, and OGTT. From a cost-benefit analysis, the primary question concerning these two groups of patients is whether low-risk women should be screened at all and under what circumstances.

The relative efficacy of FPG, random plasma glucose, and OGTT as screening methods for GDM should also be addressed; however, the U.S. Preventive Services Task Force recently conducted a systematic review on this question (Donovan et al. 2013). They discovered that using a cutoff of 85 mg/dL, both FPG and OGTT were good at identifying which women do not have GDM (specificity), but OGTT was a better predictor of which women have GDM (sensitivity). If a high cutoff (? 95 mg/dL) was used for FPG then the predictive power of the test increased to the point that it was equivalent to OGTT with respect to specificity and sensitivity. Random plasma glucose testing was not even evaluated, but testing HbA1c levels was examined and no conclusion could be drawn due to high inter-study variability. The use of HbA1c levels was recommended, however, as a quick and easy check of symptomatic women, but as a routine GDM screening tool for asymptomatic pregnant women it should not be used.

The objective of this systematic review will therefore be the evaluation of the research literature that will answer the following question: "Which GDM screening methods and protocols are the most cost-effective for a given population?" Only two patients groups will be considered here and these are low- and high-risk women for GDM who do not have undiagnosed diabetes. The problem being addressed through the PICO question format is as follows:

P = gestational diabetes

I = gestational diabetes screening methods

C = cost comparison between the GDM screening methods and protocols

O = the most cost-effective screening method for a given patient population: low- or high-risk

Literature Search Strategy

The systematic review will adhere as much as possible to the guidelines published by the Cochrane Collaboration (Higgins 2011). A search for relevant peer-reviewed research articles, which could answer the question stated in the previous section, will be conducted using online publication databases and resources including PubMed, Web of Science, ProQuest, EBSCO Host, Wiley Online Library, Ovid LWW Total Access Collection, Academic Search Premier, Cochrane Library, Science Direct, International Diabetes Federation, World Health Organization, American Diabetes Association, AHRQ National Guideline Clearinghouse, and Google. Additional resources may be accessed during the search due to opportunistic discoveries made when using the above resources and reading through retrieved publications (Figure).

All search strings used to search publication databases and other resources will be recorded, along with the date and time of the search, to ensure that the systematic review can be replicated by independent researchers. To ensure that no mistakes are made, search strings will be copied and pasted directly from search windows into the appendix section of the review. Any additional steps taken during the search will also be recorded. The search string strategy will begin with the most general terms, but if the search retrieves too many irrelevant citations the search strings will increase in stringency. The following is one such strategy proposed for the Medline database:

1. Gestational AND diabetes AND screening AND cost

2. "(Gestational OR GDM)"[title] AND diabetes AND screening AND (cost OR cost-effectiveness)

3. "(Gestational OR GDM)"[title] AND diabetes AND "screening"[title] AND (cost OR cost-effectiveness)

4. "(Gestational OR GDM)"[title] AND diabetes AND "screening"[title] AND "(cost OR cost-effectiveness)"[title]

Following retrieval of publications using search strings and opportunistic searches, the titles will be screened for relevance (Figure). A determination of relevance will be based on whether the focus of the article is on an empirical examination of the relative cost-effectiveness of gestational diabetes screening methods. The publications that pass this first screen will then be evaluated for relevance by reading through the abstracts and comparing the content to the eligibility criteria discussed below. The publications that survive this screen will be downloaded as full text and then evaluated comprehensively for eligibility criteria.

Eligibility Criteria

The retrieved publications will be screened to exclude non-English language, non-peer-reviewed, non-quantitative, and predigestional diabetes-only articles. Any research article that does not include low- and/or high-risk pregnant women without pregestational diabetes will be excluded from the systematic review. Although randomized controlled trials (RCTs) are the preferred study design, other quantitative studies will not be excluded from consideration. The reason for this is because RCTs are rare among papers examining GDM screening methods because not screening for GDM is unethical; however, the efficacy of the two screening methods under consideration here has already been established so the inclusion of lower quality studies is appropriate. The inclusion criteria will be an empirical evaluation of the cost-effectiveness of GDM screening methods and/or strategies using low- and/or high-risk patients without pregestational diabetes, the use of OGTT, FPG, and/or FPG plus OGTT, and a comparison of one-step or two-step screening protocols. Two-step screening typically involves a preliminary screen using FPG or risk factors, followed by OGTT (WHOa 2013: 14, 18, 22). Other screening tests, such as glycosuria or finger stick, are no longer considered best practice globally and will therefore be excluded from consideration. Only those studies that examine the cost effectiveness of GDM screening during the first trimester for high-risk women and between 24 and 28 weeks gestation for low-risk women will be included, because best practice guidelines generally address only these two patient groups.

Data Collection

Data will be collected from selected publications for further analysis and will include demographic data, such as patient age, ethnicity, low- or high-risk, geographic location, and socioeconomic status. What demographic data will be included in the analysis will depend on what has been included in the selected studies. The study design will be categorized into RCT, cohort, or case-control. Other information extracted from the selected publications will include interventions, randomization, blinding, sample size, study period, number of follow-ups, and outcome measures. Studies that took place in the same geographic area will be evaluated for the possibility of overlapping patient populations. If such studies are found a decision will be made about whether they represent duplications. Duplicate studies will eliminated by excluding the published article that provides the least amount of information about methods and data treatment.

A standard form will be created in Microsoft Excel to facilitate data collection. The headings for the columns in this form include exclusion study design, screening methods, cut-offs used, sample size, sample age range, Low/High risk patients, geographic location, outcome measures, gestational age, treatment, maternal outcomes, pregnancy outcomes, infant outcomes, and ICER (incremental cost-effectiveness ratio). The data entered into this form will be used to determine whether the study is appropriate for inclusion in the systematic review.

A Quality Rating Scale (QRS) will be created to evaluate the quality of the study in a number of ways. The QRS will be adapted from a published QRS (Churchill et al. 2001) to meet the needs of this systematic review. For example, the modified QRS will include evaluations of the quality of stated objectives, sample size, randomization, blinding, demographic information, statistical methods, and interpretations of findings. The QRS form will be created using Microsoft Excel.

Evaluation of the cost-effectiveness data will be modelled on the work of Pennington and colleagues (2013). Publications selected for analysis will be evaluated for the possibility of combining data, but if this is not possible an ICER will be calculated for each study to facilitate comparisons. The ICER formula is: ICER = ?C/?Q, where ?C will represent the change in direct healthcare costs and ?Q the change in outcome attributed to the intervention. Only direct care costs will be evaluated for the comparison, although interesting results concerning indirect costs will be mentioned in the text for the reader's benefit.

A limited search for systematic reviews and meta-analysis on this topic has already been performed and none were identified, but one or more could be published during the interval between the completion of this proposal and completion of the systematic review. If publication of such studies does occur then they will be included in the analysis when possible.

Figure. Flow chart depicting the search and selection strategy for systematic review of the cost-effectiveness of different GDM screening methods and strategies.

Conclusion

GDM is a condition that can have devastating consequences if left untreated and despite considerable global attention to this condition no 'gold standard' screening method has been identified. A large number of opinions and recommendations have been published by scientists, governments, and international and professional health organizations, but a consensus has not been reached. However, most stakeholders agree that checking for glycosuria is not only useless, but a waste of resources. The health of patients screened using urinalysis is risked when false negatives prevent the patient from receiving treatment. A similar adverse outcome is produced by a false positive result because unnecessary treatment consumes valuable resources and risks the patient's health. Random plasma screening is rarely mentioned as a viable screening method in the research literature, which leaves risk factor evaluation, FPG, and OGTT as the primary screening methods mentioned in published guidelines. There is also general agreement that when risk factor evaluation is used as a first screen, a two-step protocol involving a second screen using FPG or OGTT should be employed. The timing of the second screen will depend on whether the patient is high-risk or low-risk, with high-risk patients screened anytime during the first trimester. If a one-step screening protocol is used all pregnant women should be screen by 28 weeks gestation using either FPG or OGTT. Another controversy surrounding GDM screening is which cutoffs to use and the length of the OGTT test and retest interval. Any cost-effectiveness study of the screening methods and patient outcomes may help reveal which cutoffs produce the greatest benefit for a given cost; however, if a comparison reveals that two or more screening methods have similar efficacy but differ significantly in cost, this information will prove valuable to any clinician or health system concerned about the cost of health care.

In light of the remaining controversies concerning GDM screening and the costs associated with antenatal care, a systematic review of GDM screening cost-effectiveness studies is proposed. The primary goal of this review is to help healthcare organizations and agencies with limited resources make the best choices for their patients concerning GDM screening methods. The systematic review may also help clinicians decide whether FPG or OGTT would work best for their organization, whether a one- or two-step model makes the most sense, which cutoffs provide the most benefit for a given cost, and whether low-risk and high-risk patients should be screened differently. While cost will be the main focus of this systematic review, any findings and conclusions will be considered in relation to the efficacy of the different screening methods and strategies. The efficacy of a screening method and strategy will depend on the incidence rates for false positive and false negative results (reproducibility) and patient outcomes. Patient outcomes will also be broken down into pregnancy, maternal, and infant outcomes. In other words, the cost-effectiveness of a GDM screening method will depend on patient outcomes.