|Year : 2021 | Volume
| Issue : 1 | Page : 25-30
Diagnostic accuracy of Diabetes in Pregnancy Study Group of India criteria for the screening of gestational diabetes mellitus in primary care setting
Nitya Balagopalan1, Rambha Pathak2, Farzana Islam1, Aruna Nigam3, Prem Kapur4, Sarita Agarwal5
1 Department of Community Medicine, Hamdard Institute of Medical Sciences and Research, Jamia Hamdard, New Delhi, India
2 Department of Community Medicine, Government Institute of Medical Sciences, Greater Noida, Uttar Pradesh, India
3 Department of Obstetrics and Gynaecology, Hamdard Institute of Medical Sciences and Research, Jamia Hamdard, New Delhi, India
4 Department of Medicine, Hamdard Institute of Medical Sciences and Research, Jamia Hamdard, New Delhi, India
5 Department of Biochemistry, Hamdard Institute of Medical Sciences and Research, Jamia Hamdard, New Delhi, India
|Date of Submission||03-Feb-2020|
|Date of Decision||03-Apr-2020|
|Date of Acceptance||21-May-2021|
|Date of Web Publication||29-Jun-2021|
Prof. Rambha Pathak
Department of Community Medicine, Government Institute of Medical Sciences, Greater Noida, Uttar Pradesh - 201 310
Source of Support: None, Conflict of Interest: None
Introduction: Although the Diabetes in Pregnancy Study Group of India (DIPSI) criterion is recommended by the Government of India guidelines, there is lack of consensus on a universal criterion for diagnosis of gestational diabetes. This has led to a wide variation of pregnant women being diagnosed with gestational diabetes mellitus (GDM). The WHO 1999 and International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria are widely used globally and in India as well. The objective of this study was to evaluate the diagnostic accuracy of DIPSI criteria in comparison to WHO 1999 and IADPSG criteria for diagnosis of GDM.
Materials and Methods: A community-based study was conducted for a period of 1 year. Oral glucose tolerance test was done on 506 pregnant women identified through house-to-house survey. The proportion of GDM cases by WHO, IADPSG, and DIPSI criteria was calculated. The diagnostic accuracy of DIPSI criteria was assessed by calculating sensitivity, specificity, and predictive values taking WHO and IADSPG criteria as gold standard.
Results: The prevalence of GDM was 14.2% by WHO criteria, 13% by DIPSI criteria, and 27.3% by IADPSG criteria; 10.3% were diagnosed by all the three criteria. The sensitivity and specificity of DIPSI criteria when the WHO criteria was taken as the gold standard was found to be 86.1% and 99.08%. The sensitivity and specificity of DIPSI criteria when the IADPSG criteria was taken as gold standard was found to be 44.93% and 98.91%, respectively.
Conclusion: The prevalence of GDM is found to be much higher by IADPSG criteria as compared to the WHO and DIPSI criteria. The single-step approach DIPSI criteria have good diagnostic accuracy and can be used in epidemiological studies and are feasible for diagnosis of GDM in primary care settings.
Keywords: Diabetes, Diabetes in Pregnancy Study Group of India, diagnosis of gestational diabetes mellitus, gestational diabetes
|How to cite this article:|
Balagopalan N, Pathak R, Islam F, Nigam A, Kapur P, Agarwal S. Diagnostic accuracy of Diabetes in Pregnancy Study Group of India criteria for the screening of gestational diabetes mellitus in primary care setting. Indian J Community Fam Med 2021;7:25-30
|How to cite this URL:|
Balagopalan N, Pathak R, Islam F, Nigam A, Kapur P, Agarwal S. Diagnostic accuracy of Diabetes in Pregnancy Study Group of India criteria for the screening of gestational diabetes mellitus in primary care setting. Indian J Community Fam Med [serial online] 2021 [cited 2022 Jan 23];7:25-30. Available from: https://www.ijcfm.org/text.asp?2021/7/1/25/319962
| Introduction|| |
Gestational diabetes mellitus (GDM) which is glucose intolerance first detected in pregnancy is emerging as a major public health problem. The prevalence of gestational diabetes ranges from 1% to 28%. It depends upon the population characteristics, screening, and diagnostic criteria used. Approximately 4 million pregnancies are complicated by GDM annually in India. Therefore, there is a need for universal screening of all pregnant women for GDM.
Globally, GDM is being diagnosed by using various criteria and there is lack of a universal accepted criterion. This may have an effect on the estimated prevalence of GDM and related health outcomes while also impacting on their health costs and quality of life.
The Diabetes in Pregnancy Study Group of India (DIPSI) criteria which is a modification of the older WHO criteria, is recommended by the Government of India guidelines and used in epidemiological studies across India. However, poor sensitivity and low positive predictive value of the test have also been reported.,
The International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria are based on the findings of the large-scale Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study and hence popular globally, but its drawback is argued to be the large number of false-positive cases due to lower fasting cutoffs and hence adding to the burden of GDM. Furthermore, it has a higher 2-h cutoff of 153 mg/dl and it is a matter of debate if the 2-h blood glucose levels between 140 and 153 mg/dl can be safely left untreated or not. In addition, diagnosing the Indian population by international studies can be inconclusive as the HAPO study lacked Indian representativeness in its findings.
Based on a large retrospective study comparing the IADPSG criteria and the WHO 1999 criteria, WHO criteria of 2-h plasma glucose of ≥140 mg/dl appeared to be sufficient to diagnose GDM. Since both these tests are internationally accepted and widely used, it was planned to evaluate the diagnostic accuracy of DIPSI criteria in comparison to WHO and IADPSG criteria for diagnosis of GDM.
| Materials and Methods|| |
This study is part of a prospective, community-based study to assess the maternal and fetal outcomes of GDM. However, in this article, only the diagnostic accuracy of the tests will be discussed. The study is set in the urban and rural field practice areas attached to the department of community medicine of a medical college in southeast district of Delhi and was conducted for a period of 1 year from December 2017 to December 2018. The study included consenting pregnant women between 18 and 28 weeks of gestation residing in the field practice area. The population of both the field practice areas is 1 lakh each. The number of pregnancies in both the areas together is 5000 approximately. Women with preexisting diabetes, renal disorders, pancreatic disorders, tuberculosis, or those who were on medications that are likely to cause dysglycemia were excluded from the study. The study was approved by the institutional ethical committee. Written informed consent was obtained from all the participants who agreed to participate in the study.
Adequate sample size was computed as in a diagnostic test study with calibrated outcome. The sensitivity and specificity of DIPSI and the candidate test were assumed to be 70% (with absolute precision of ± 10%) and 80% (with absolute precision of ± 10%), respectively, with WHO and IADPSG as gold standard separately. To detect the above sensitivity and specificity with 95% confidence level, we were required to enroll 61 GDM cases. In a multicentric study, the average prevalence of GDM was reported to be 15% from India. Thus, assuming prevalence of GDM of 15% in the screened population, we needed to screen at least 480 pregnant women in order to get about 61 women with GDM. Therefore, a total sample size of 500 was fixed.
The study used a two-stage cluster sampling technique to ensure the random selection of the study participants. The primary sampling units were the geographical clusters of the field practice area and the secondary sampling unit was households with pregnant women. Appropriate geographical clusters that were mutually exclusive were identified from the field practice areas. All the clusters in the population were listed, 20 clusters were selected by using simple random sampling strategy. In order to meet the required sample size, 10 clusters from rural and urban field practice area were selected and 25 women were taken from each cluster. Consecutive sampling was done in the clusters until the sample size for the cluster was reached. In the case of the required sample size not being met from that cluster, adjacent cluster was taken up to complete the sample. Since the complete enumeration of the pregnant women was not available, simple random sampling was not possible. Thus, to ensure representativeness of the study population, as well as keeping in mind the feasibility of the study, it was decided to use cluster sampling for the study.
The DIPSI criteria: It defines GDM as those cases who have a single blood glucose value of more than or equal to 140 mg/dl g taken 2 h after 75 g glucose load, given irrespective of fasting or nonfasting state.
- WHO criteria of GDM: Pregnant women having fasting blood glucose more than or equal to 126 mg/dl or those with a 2-h blood glucose after 75 g glucose load more than or equal to 140 mg/dl were labeled as GDM
- The IADPSG criteria: It defines GDM as fasting blood glucose more than or equal to 92 mg/dl or 1-h blood glucose following a 75 g oral glucose load more than or equal to 180 mg/dl or 2-h blood glucose following 75 g glucose load more than or equal to 153 mg/dl.
The pregnant women between 18 and 28 weeks of gestation residing in the geographical clusters were identified by house-to-house survey with the help of Medical Social Worker. The women were sensitized about GDM and the study being conducted and written informed consent was taken. A detailed history and thorough clinical examination was done. Body mass index (BMI) was calculated based on the prepregnancy weight and height. At the time of enrollment, all women were given 75-gm glucose load orally after dissolving in approximately 300 ml water within 10 min irrespective of their previous meal as recommended by DIPSI. Blood sugar was measured after 2 h. Considering the utility of capillary blood glucose assessment, plasma calibrated and standardized glucometer was used to evaluate blood sugar 2 h after the oral glucose load., If vomiting occurred within 30 min of oral glucose intake, the test was repeated the next day.
For the fasting oral glucose tolerance test (OGTT), the women were asked to report to the rural and urban health center, respectively, within 72 h of enrollment and after 8 h of overnight fasting and 48 h of regular diet. A fasting venous sample was taken for glucose estimation. The women were given a 75 g glucose in 300 ml water to be consumed in 10 min. After 2 h of the glucose load, sample was obtained of the venous blood for glucose estimation. The glucose estimation was done by hexokinase method. The machine used was Dimension RxL Max (Siemens). Calibration of the analyzers was done as per the directions of Dimension Rx Max. Internal Quality control was run once a day. The Quality control used was supplied by Biored. The laboratory also participates in the External Quality Assessment Scheme (EQAS) once a month. Samples are received from EQAS and analyzed in the laboratory every month and accuracy was compared with the standard. The flow of study is as shown in [Figure 1].
The data were entered in Microsoft Excel and analyzed in IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. (Armonk, NY: IBM Corp.). The proportion of GDM was computed based on WHO, IADPSG, and DIPSI criteria. The diagnostic accuracy of DIPSI criteria was assessed by calculating sensitivity, specificity, and predictive values taking WHO and IADSPG criteria as gold standard. Kappa statistic was also calculated to find out the level of agreement between different diagnostic criteria.
| Results|| |
A total of 506 pregnant women underwent both the fasting and nonfasting OGTTs. Out of the 506 women, 72 tested positive for GDM by WHO criteria, 138 tested positive by the IADPSG criteria, and 66 tested positive by DIPSI criteria. The women who were diagnosed by GDM by the three criteria were not always the same. The prevalence of GDM was found to be 14.2% by WHO criteria, 27.3% by IADPSG criteria, and 13% by DIPSI criteria. 11.06% of the women were diagnosed with GDM by both WHO and IADPSG criteria, 11.8% by both WHO and DIPSI criteria, 10.5% by both IADPSG and DIPSI criteria, whereas only 10.3% of the women were diagnosed with GDM by all the three criteria [Figure 2]. The age at conception and BMI of the pregnant women and its comparison by all the three criteria used in the study are shown in [Table 1].
|Table 1: Comparison of the study population characteristics according to the 3 diagnostic criteria|
Click here to view
|Figure 2: Venn diagram showing detailed breakup of patients diagnosed as gestational diabetes mellitus by different methods|
Click here to view
Out of the 72 women that were diagnosed with GDM using WHO criteria, 62 were also labeled as GDM using DIPSI criteria, whereas out of the 434 diagnosed as non-GDM using WHO criteria, 430 were also labeled as non-GDM using WHO criteria. Thus, the sensitivity of DIPSI criteria using the WHO criteria as gold standard was found to be 86.1%, whereas the specificity was found to be 99.08%. The positive predictive value and negative predictive value for DIPSI with WHO criteria as the gold standard were found to be 93.94% and 97.73%, respectively. The agreement kappa statistic for WHO and DIPSI was found to be 0.92 and this was found to be statistically significant with P < 0.001, indicating excellent agreement.
On the other hand, out of the 138 pregnant women who were diagnosed as GDM using IADPSG criteria, only 62 were also detected using DIPSI criteria, whereas out of the 368 women diagnosed as non-GDM using IADPSG criteria, 364 were also declared as non-GDM using DIPSI criteria. Thus, the sensitivity of DIPSI using IADPSG criteria as the gold standard was found to be 44.93%, whereas the specificity was found to be 98.91%. The positive predictive value and negative predictive value were found to be 93.9% and 82.7%, respectively. The agreement kappa for DIPSI and WHO was found to be 0.46 indicating good agreement, with a statistically significant P < 0.001.
The diagnostic accuracy parameters are shown in [Table 2].
|Table 2: Comparison of single step, non-fasting 2-hr DIPSI test with other oral glucose tolerance tests|
Click here to view
| Discussion|| |
In the present study, the prevalence of gestational diabetes was found to be 14.2% by WHO criteria and 27.3% by IADPSG criteria, 13% by DIPSI criteria, and 10.3% by all the three criteria. Similar findings were observed in a study done by Rajput et al. where the prevalence was found to be 13.9% by the WHO criteria and another study done by Agrawal et al., where the prevalence was observed to be 13.5% by DIPSI criteria.
In this study, the prevalence of gestational diabetes was found to be higher by IADPSG cutoffs (27.3%) as compared to WHO criteria (14.2%). A high prevalence of GDM (41.6%) by IADPSG in North India has also been found by Gopalakrishnan et al. The WINGS-6 study that compared the WHO and IADPSG criteria found a prevalence of 14.9% by WHO criteria and 18.9% by IADPSG criteria which was significantly higher. These results match with the findings of the present study as well and highlight the increased prevalence of gestational diabetes when estimated using the IADPSG criteria.
In our study, 77.8% of those diagnosed by WHO criteria also tested positive by IADPSG criteria whereas 40.6% who tested positive by IADPSG criteria also tested positive by WHO criteria. Similarly, in the WINGS-6 study, 65.1% of the women identified by WHO criteria were also picked up by the IADPSG criteria whereas 43.5% identified by the IADPSG criteria were picked up by the WHO criteria as well.
In the present study, we found the sensitivity, specificity, positive predictive value, and negative predictive value of DIPSI criterion to be 86.1%, 99.08%, 93.94%, and 97.73% respectively when compared to WHO criteria. Saxena et al. also found the sensitivity, specificity, positive predictive value, and negative predictive value of DIPSI criteria to be 96%, 98%, 79.3%, and 99.8%, respectively, whereas Sharma et al. found DIPSI to be 100% sensitive and 100% specific in diagnosing GDM. The agreement kappa statistic for DIPSI and WHO was found to be 0.92 (P = 0.01) indicating a 92% agreement. Likewise, Saxena et al. found the kappa statistic to be 0.868, indicating excellent agreement similar to the findings of this study.
The sensitivity and specificity for DIPSI criteria when compared to the IADPSG criteria were found to be 44.93% and 98.9%, respectively, in the present study. Similar results were observed by Srinivasan and Reddi who found the sensitivity and specificity to be 45.45% and 87.70%, respectively. The positive and negative predictive values of DIPSI in comparison to IADPSG were found to be 93.9% and 82.7%, respectively, in this study. In contrast, the study done by Srinivasan and Reddi found the positive predictive value to be 40.00% and negative predictive value 89.92%. Mohan et al. found a specificity of 97.8% for DIPSI criterion which was similar to our study results, but the sensitivity was 22.6% which was much lower than our study. The agreement kappa statistic between DIPSI and IADPSG criteria as observed in the present study was found to be 0.46 (P < 0.001) indicating a 46% agreement. However, kappa statistic by Mohan et al. was found to be 0.314 indicating disagreement between DIPSI and IADPSG.
Rationale for Nonfasting status OGTT: A nondiabetic pregnant woman would be able to maintain euglycemic state even after a meal due to adequate and brisk insulin response, whereas women with GDM will not be able to do so because of impaired glucose secretion. Therefore, in situations where a fasting test cannot be done, the nonfasting OGTT can be used, with lower cutoff points to increase sensitivity as a screening test, which forms the basis of DIPSI criterion for diagnosing GDM.
The DIPSI criterion has other advantages too. As the pregnant women need not be fasting, she will not experience morning sickness and will not have nausea or vomiting after glucose load. It also causes least disturbance in a pregnant woman's routine activities. It can serve as both screening and diagnostic procedure and in management.
In the Indian population, where there are challenges of accessibility to test centers, a test that requires a fasting state is often not feasible. Therefore, a one-step test with acceptable diagnostic accuracy is desirable, particularly in primary health-care settings. A one-step test that requires less training and which can be administered in the community using simple instruments such as a glucometer is beneficial to ensure that a larger population is covered for the screening of GDM.
Strengths of the study
This was a prospective community-based study where the study population was series of participants defined by the selection criteria. These participants were recruited by using cluster sampling technique. While carrying out the study, all the requirements needed as per the Standards for Reporting of Diagnostic Accuracy were strictly followed. The index test and reference test were done after a gap of 72 h and data were simultaneously collected.
| Conclusion|| |
The DIPSI criteria has good diagnostic accuracy and compares well with the WHO criteria. Therefore, it can be used in epidemiological studies and for diagnosis of GDM in primary care settings. The single-step approach of diagnosis makes it feasible and acceptable for use and therefore can ensure fewer noncompliance and dropouts and greater completion of the test. The findings of this study show that the prevalence of GDM is found to be much higher by IADPSG criteria as compared to the WHO and DIPSI criteria because of a lower fasting cutoff. The study recommends that universal screening which is already recommended must be scaled up and implemented at all peripheral levels. The medical officers and other health-care staff must be adequately trained in screening and diagnosis using the DIPSI criteria so as to ensure accurate diagnosis and minimal deviation from the protocol.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Jiwani A, Marseille E, Lohse N, Damm P, Hod M, Kahn JG. Gestational diabetes mellitus: Results from a survey of country prevalence and practices. J Matern Fetal Neonatal Med 2012;25:600-10.
Seshiah V, Balaji V, Balaji MS, Sanjeevi CB, Green A. Gestational diabetes mellitus in India. J Assoc Physicians India 2004;52:707-11.
Schmidt MI, Duncan BB, Reichelt AJ, Branchtein L, Matos MC, Costa e Forti A, et al
. Gestational diabetes mellitus diagnosed with a 2-h 75-g oral glucose tolerance test and adverse pregnancy outcomes. Diabetes Care 2001;24:1151-5.
Seshiah V, Das AK, Balaji V. Gestational diabetes mellitus – Guidelines. JAPI – DIPSI Guidelines. Gestational Diabetes Mellitus in India. J Assoc Physic of India 2004;52:707-115.
Koning SH, van Zanden JJ, Hoogenberg K, Lutgers HL, Klomp AW, Korteweg FJ, et al
. New diagnostic criteria for gestational diabetes mellitus and their impact on the number of diagnoses and pregnancy outcomes. Diabetologia 2018;61:800-9.
Rani PR, Begum J. Screening and diagnosis of gestational diabetes mellitus, where do we stand. J Clin Diagn Res 2016;10:E01-4.
Coustan DR, Lowe LP, Metzger BE, Dyer AR. The HAPO study: Paving the way for new diagnostic criteria for GDM. Am J Obstet Gynecol 2010;202:654.e1-654.e6.
Mohan V, Nallaperumal S, Bhavadharini B, Mahalakshmi M, Maheswari K, Jalaja R, et al
. Comparison of the world health organization and the International association of diabetes and pregnancy study groups criteria in diagnosing gestational diabetes mellitus in South Indians. Indian J Endocrinol Metab 2013;17:906-10.
Kotwal N, Pandit A. Variability of capillary blood glucose monitoring measured on home glucose monitoring devices. Indian J Endocrinol Metab 2012;16:S248-51.
Ignell C, Berntorp K. Evaluation of the relationship between capillary and venous plasma glucose concentrations obtained by the HemoCue Glucose 201+system during an oral glucose tolerance test. Scand J Clin Lab Invest 2011;71:670-5.
Rajput M, Bairwa M, Rajput R. Prevalence of gestational diabetes mellitus in rural Haryana: A community-based study. Indian J Endocrinol Metab 2014;18:350-4.
Agrawal S, Das V, Agarwal A, Pandey A, Namrata. Prevalence of gestational glucose intolerance and gestational diabetes in a tertiary care centre in northern India. J Clin Diagn Res 2018;12:QC04-6.
Yadav S, Gopalakrishnan V, Kapoor D, Bhatia E, Singh R, Pradeep Y, et al
. Evaluation of the prevalence of gestational diabetes mellitus in North Indians using the International Association of Diabetes and Pregnancy Study groups (IADPSG) criteria. J Postgrad Med 2015;61:155-8.
] [Full text]
Bhavadharini B, Mahalakshmi MM, Anjana RM, Maheswari K, Uma R, Deepa M, et al
. Prevalence of Gestational Diabetes Mellitus in urban and rural Tamil Nadu using IADPSG and WHO 1999 criteria (WINGS 6). Clin Diabetes Endocrinol 2016;2:8.
Saxena P, Verma P, Goswami B. Comparison of diagnostic accuracy of non-fasting DIPSI and HbA1c with fasting WHO criteria for diagnosis of gestational diabetes mellitus. J Obstet Gynaecol India 2017;67:337-42.
Sharma K, Wahi P, Gupta A, Jandial K, Bhagat R, Gupta R, et al
. Single glucose challenge test procedure for diagnosis of gestational diabetes mellitus: A Jammu cohort study. J Assoc Physicians India 2013;61:558-9.
Srinivasan S, Reddi RP. Comparative study of DIPSI and IADPSG criteria for diagnosis of GDM. Int J Reprod Contracept Obstet Gynecol 2018;7:932-7.
Mohan V, Mahalakshmi MM, Bhavadharini B, Maheswari K, Kalaiyarasi G, Anjana RM, et al
. Comparison of screening for gestational diabetes mellitus by oral glucose tolerance tests done in the non-fasting (random) and fasting states. Acta Diabetol 2014;51:1007-13.
Seshiah V, Balaji V, Balaji MS, Sekar A, Sanjeevi CB, Green A. One step procedure for screening and diagnosis of gestational diabetes mellitus. J Obstet Gynecol India 2005;55:525-9.
[Figure 1], [Figure 2]
[Table 1], [Table 2]