Detection and Management of Diabetes during Pregnancy in Low Resource Settings: Insights into Past and Present Clinical Practices

Background. Timely and adequate treatment is important to limit complications of diabetes affecting pregnancy, but there is a lack of knowledge on how these women are managed in low resource settings. Objective. To identify modalities of gestational diabetes detection and management in low and lower middle income countries. Methods. We conducted a scoping review of published literature and searched the databases PubMed, Web of Science, Embase, and African Index Medicus. We included all articles published until April 24, 2016, containing information on clinical practices of detection and management of gestational diabetes irrespective of publication date or language. Results. We identified 23 articles mainly from Asia and sub-Saharan Africa. The majority of studies were conducted in large tertiary care centers and hospital admission was reported in a third of publications. Ambulatory follow-up was generally done by weekly to fortnightly visits, whereas self-monitoring of blood glucose was not the norm. The cesarean section rate for pregnancies affected by diabetes ranged between 20% and 89%. Referral of newborns to special care units was common. Conclusion. The variety of reported provider practices underlines the importance of promoting latest consensus guidelines on GDM screening and management and the dissemination of information regarding their implementation.


Introduction
The prevalence of diabetes in women of reproductive age is growing in developing countries, where changes in lifestyle are largely accountable for this increase [1]. However, diabetes affecting pregnancy is not yet regarded as a major problem in settings where mothers are dying of more obvious causes such as hemorrhage and hypertensive diseases [2]. Despite a higher risk of affected women to develop direct obstetric complications such as preeclampsia and prolonged labor due to macrosomic babies [3], the attention paid to gestational diabetes (GDM) in low resource countries remains negligible, and not much is yet known about the burden of GDM in these contexts.
Local studies show prevalence rates of GDM between 6% and 14% in East and West Africa [4,5] and between 13% and 18% in South Asia [6,7]. There is evidence that treating pregnant women affected by diabetes in pregnancy improves maternal and perinatal outcomes and leads to a fourfold decrease of severe perinatal complications [8,9]. In mothers who received treatment, excessive growth of a baby and the occurrence of shoulder dystocia and preeclampsia have been reduced [10,11].
Timely and adequate treatment of affected mothers is key for reducing potential complications particularly when considering the limited timeframe for management from detection until birth. Assuming that one out of ten pregnant women might already be affected by GDM in some low and lower middle income countries, where access to care and resource availability are often major challenges, it is important to understand how and when pregnant women are detected and subsequently managed. Therefore we conducted a review of the published literature to explore documented past and present practices regarding a health problem that will gain increasing attention in the years to come.

Search Strategy.
We carried out a scoping review of the literature to broadly map available literature and to identify existing gaps [12]. The electronic databases PubMed, Web of Science, Embase, and African Index Medicus as well as reference lists of identified studies were searched using the following search terms "gestational", "pregnancy", "diabetes", "manage * ", "screen * ", "developing countr * ", "low income countr * ", and "lower middle income countr * " without methodological, language, or date restrictions. The search strategies were designed to correspond to the specific requirements of each database. 24 April 2016 that contained any information on clinical practices of detection and management of gestational diabetes/diabetes in pregnancy in low and lower middle income countries as defined by the World Bank [36] were included. We included descriptive and observational studies but excluded intervention studies assessing the prevalence of GDM and publications providing only theoretical recommendations without specific information on observed clinical practices.

Data Extraction and Synthesis.
After the removal of duplicates, titles and abstracts were independently screened by two reviewers (BU and AD) regarding their relevance and conformity with the inclusion criteria followed by full text review of selected articles. Any discrepancies could be resolved after discussion between the two reviewers. For reference management we used Endnote version X7 (Thomson Reuters). Data of selected articles was entered into a summary table listing study design, country, year, and study population. Screening and management practices were grouped into the themes screening, hospital admission and inpatient treatment, ambulatory care and follow-up during pregnancy, medication, delivery, newborn, and postpartum care ( Table 1).
As management of gestational diabetes was not in all studies distinguishable from the management of preexisting diabetes, we refer to the combination of both as diabetes during pregnancy.

Results
Our initial search yielded 973 publications. After the removal of duplicates, 801 papers were retrieved. Through hand searching the bibliographies of screened articles, five additional articles were added. We identified in total 23 papers for inclusion ( Figure 1). Of these, eight papers originated from Sub-Saharan Africa with one publication from South Africa [20], three from Nigeria [15,23,34], and one from Kenya [14], Tanzania [16], Ivory Coast [18], and Sudan [21]. Two papers were from North Africa with one article from Morocco [24] and one from Libya [17]. There were eleven publications from Asia of which seven papers alone were from India [26-28, 31-33, 35], two publications were from Pakistan [19,22], one was from Sri Lanka [25], and one was from Vietnam [29]. One paper originated from Fiji [13] and one article included information on six countries in Asia, Africa, and the Caribbean [30].
Aspects of clinical management of pregnant women affected by diabetes were described in 19 articles, and information on screening was provided in 16 publications. The earliest publication dates back to 1977 and the latest article was published in 2015.
3.1. Screening. Screening practices varied and available screening tests either used alone or used in combination included in four settings the measurement of random glucose [25,30,33,35] or fasting glucose [24,26,30,33]; postprandial glucose was listed in two studies [24,25], screening by a glucose challenge test (GCT) was highlighted in five settings [19,22,24,26,30], and an oral glucose tolerance test (OGTT) was mentioned in 15 out of the 16 studies containing information on screening practices [19-26, 28-31, 33-35]. In one publication, rapid testing of urine for screening was noted [25]. The screening approach was described in 11 studies, with universal testing taking place in five settings [19,24,[29][30][31] or as being the preferred approach for the majority of providers in two publications [26,35]. Selective screening was described in three studies [23,25,34] and reported as being preferred by health care providers in one publication [33]. Early selective screening of women with risk factors at their first ANC booking in combination with universal screening after 24 weeks of gestational age was mentioned in two studies [24,30]. The timing of screening was described in 11 publications and took place between 16 and 24 weeks [35], between 20 and 28 weeks [19], after 24 weeks [25], between 24 and 28 weeks [24,26], or at 28 weeks [29,34]. Late screening between 28 and 32 weeks of gestational age was mentioned in one study from Pakistan [22] but was also reported in a publication from Nigeria, where the majority of women (78.7%) were diagnosed with GDM between gestational weeks 28 and 30, whereas 21.3% of women were only detected after 30 weeks of gestation [23]. Screening in every trimester was described in one study from India [31] and in one publication health care providers favored screening at the first ANC visit [33].

Hospital Admission and Inpatient Treatment.
Overall, five studies described initial routine hospital admission of pregnant mothers after they had been diagnosed with diabetes [14,15,18,20,21], whereas in a study from Vietnam, only patients whose glucose levels were not controllable by ambulatory care alone were hospitalized [29]. In a Tanzanian hospital, no specific policy on hospitalization existed for pregnant patients with diabetes and decisions were rather taken on an individual basis to achieve better glucose control and in the presence of obstetric or medical indications [16].
Specific diet advice was described in three studies and patients were instructed to follow a diet restricted in carbohydrates [14] or in calories [15,20]. Information on glucose monitoring during hospitalization was reported in five papers and glucose values of hospitalized patients were measured regularly from two to six times within 24 hours [14, 15, 20, 21],     Gestational OR pregnancy AND diabetes AND manage * OR screen * AND developing countr * OR low income countr * OR lower middle income countr * Figure 1: Flowchart of screening process.
either on a daily basis [20] or once or twice a week [14,15,18,21]. In a study from Ivory Coast, blood and urine tests, vaginal swabs, ophthalmological checks, and ultrasound investigations were performed during hospitalization [18]. Patient education included diet counselling in two papers [18,20] and the initiation to self-monitoring of blood glucose (SMBG) in one publication [20]. In five studies, self-monitoring was reported as not being performed [16,17,21,23,28]. Timing of discharge was specified in three studies and women were discharged after two weeks [18] or after glucose values were sufficiently controlled [14,21]. Admission or readmission in gestational weeks 32 to 34 was identified in four studies [13][14][15]21]. In one publication [17], only high risk pregnant mothers with diabetes during pregnancy were admitted in week 34/35, three to four weeks earlier than women with a normal or low risk, whereas in another publication patients were readmitted only around term [18]. Regarding the management of readmitted patients, reported practices ranged from bed rest [13], surveillance of caloric intake [13], blood glucose measurements [13,18,21], control of fetal wellbeing [18], and daily measurements of urine and blood pressure to weekly measurements of weight and uterine height [14]. In two publications dating back more than 30 years, amniotic fluid was examined once a week from 36 weeks onwards to assess fetal pulmonary maturity [13,14] with corticosteroids being administered a week before the planned delivery [13].

Medication.
According to 13 articles, insulin was the treatment of choice if diet alone was not sufficient to control hyperglycemia [13-18, 20, 21, 23, 28, 29, 31, 33]. Insulin was started as early as three days after diet modification [20] or after two weeks of diet [28] and was administered between once and three times daily [14,15,17,28]. Regarding diabetes control, poor control was reported for 6% of diabetic mothers in a setting in South Africa [20], for 20% in a study from Sudan [21], for 40% of patients in a study from Nigeria [23], and for 58% of pregnant diabetics in a publication from Pakistan [19].
3.5. Delivery, Newborn, and Postpartum Care. Specific information regarding the mode of delivery, intrapartum and newborn care, and maternal follow-up postpartum were mentioned in 14 studies [13-17, 20-23, 27, 31, 32, 34, 35]. Induction in week 37/38 was the preferred mode of delivering high risk pregnant patients with diabetes whereas medium and low risk patients were allowed to deliver vaginally at term in one publication [17]. In four papers, labor was induced at a gestational age of 38 to 39 weeks [21][22][23]31] corresponding to reported health care provider practices from India where the majority of obstetricians delivered women with GDM up to a gestational age of 38 weeks [27]. In a study from Pakistan, women who delivered vaginally were routinely assisted by vacuum or forceps in the second stage of labor [22]. In a publication from Tanzania [16], labor was not routinely induced and vaginal delivery attempted. This corresponds to findings from South Africa, where pregnancies were allowed to proceed to term given that glycemic control was good and no other obstetric complications were present [20]. In two studies, delivery took place electively after pulmonary maturity had been ascertained [13,14], either vaginally or by caesarean section in case of obstetric indications [13]. Caesarean section rates described in 14 studies ranged between 20% and 89% with a median rate of 42%. Glucose control in labor was mentioned in two articles reporting twohourly monitoring of parturients [21,22]. Regarding the care of the newborns of diabetic mothers, a study from Kenya mentioned the presence of pediatricians at each delivery and observation of the newborns for several days [14]. Referral of newborns was reported in five studies. In one publication all newborns were admitted to the special care unit [15] and in two publications nearly 50% of newborns were referred to special care [22,34]. In a hospital setting in India, all newborns of mothers receiving antidiabetic medication were admitted to special nursing care where they were regularly fed and referred back after achieving normal glucose levels [32]. Results of a provider practice survey in India revealed that one-third of clinicians refer every second newborn and 57% of providers refer one in 10 babies [27].
Retesting of women six weeks postpartum for a repeat 75 g OGTT was mentioned in one article [23] and in two papers between 54% and 93% of clinicians stated they would recommend a postpartum test [27,35].
3.6. Limitations. Our descriptive approach enabled us to shed some light on the diversity of screening and management procedures regarding diabetes during pregnancy, although the identified studies reflect only individual practices in specific hospitals or reported provider practices. National guidelines, where available, would have provided a more comprehensive picture of the national situation but as their availability alone does not necessarily mirror performed practice, we intentionally choose to identify only observed or reported clinical practices. As some of the publications date back almost 40 years, practices have changed over time and may not reflect the actual situation anymore. A specific focus on gestational diabetes practices was not always possible as some authors did not distinguish between preexisting diabetes and GDM when reporting on management practices.

Discussion
Our review revealed a limited number of publications describing detection and management practices of gestational diabetes in low resource settings. The majority of studies were conducted in large tertiary hospitals and although this may be partly due to publication bias with providers from referral centers more likely to publish, the findings indicate that the majority of patients with diagnosed diabetes during pregnancy are usually referred and treated through higher level services, where specialists and multidisciplinary teams are available.
Screening for GDM in developing countries is often not performed routinely and where applied, screening approaches are not uniform. Modes of screening vary not only between countries but also within countries, even within the same service [24], and highlight the importance of training health care providers on uniform practice standards that are applicable to low resource settings.
Therapeutic options for pregnant women diagnosed with GDM include nutritional therapy and antidiabetic medication. Although nutritional therapy has been highlighted in most of the publications as the foundation of treatment, overreliance on adequate nutrition coupled with a potential anxiety to administer insulin by health care providers based on a fear of hypoglycemia [37] might be one of the reasons for poorly controlled diabetes during pregnancy. Furthermore, information on nutrition may be misleading if recommendations are not adapted to local dietary habits and food availability [29].
In most publications, insulin was the drug of choice for pregnant women whose diabetes was not controlled by diet alone. Availability and storage issues of insulin are particular barriers to adequate management in low resource settings [38] and underline the need for a safe oral alternative. Recent international consensus guidelines of the International Federation of Gynecology and Obstetrics (FIGO) include Metformin as first-line treatment [39] which has proven to be a safe oral alternative to insulin [40] and more feasible to use in settings where regular monitoring of blood glucose cannot be assured.
Our results revealed that in several settings patients were initially admitted for nutritional education, close glucose monitoring, and initiation of treatment, although it seems that over time there was a trend to ambulatory management. In some of the earlier studies, bed rest was advised for pregnant women with diabetes, a practice for which current evidence is missing [41]. Hospitalization for treatment initiation and improved monitoring might be considered for settings where regular access to health services is not secured, Journal of Diabetes Research 13 although inpatient treatment might have substantial financial implications for affected women and their families [42].
Accessibility to regular follow-up as well as accessibility to self-testing of blood glucose for women with GDM are challenges that need to be addressed. In the majority of studies, ambulatory glucose controls were performed weekly to fortnightly and ranged from measuring urine glucose and random or fasting blood glucose to postprandial glucose checks. In several settings, self-monitoring of glucose was not considered a feasible option often due to difficulties for patients to purchase equipment. Other factors such as not feeling comfortable to self-monitor or to interpret blood sugar results at home made women return to the hospital for monitoring [29]. Contextualized local solutions to ensure regular monitoring in proximity to the women need to be developed, offering an alternative to self-monitoring and might include support through peers or care providers based within the communities.

Conclusion
In various settings, certain aspects of the above described management of GDM might still prevail and have possibly been adapted to local circumstances. However, the paucity of published information on clinical practices coupled with a lack of uniformity in the management of gestational diabetes requires a focus on the promotion of universal guidelines on GDM screening and management that are applicable to low resource settings. With the recent publication of the FIGO consensus recommendations grouped according to income setting [39], an important step has been taken. Information about their implementation and examples of best practice, particularly from low income settings, need to be disseminated.