Screening for Type 1 and Type 2 Diabetes

Abstract

Key Messages•In the absence of evidence for interventions to prevent or delay type 1 diabetes, screening for type 1 diabetes is not recommended.•Screening for type 2 diabetes using a fasting plasma glucose (FPG) and/or glycated hemoglobin (A1C) should be performed every 3 years in individuals ≥40 years of age or in individuals at high risk using a risk calculator.•Diabetes will be diagnosed if A1C is ≥6.5% (see Definition, Classification and Diagnosis chapter, p. S8).•Testing with a 2-hour plasma glucose (2hPG) in a 75 g oral glucose tolerance test (OGTT) should be undertaken in individuals with an FPG of 6.1–6.9 mmol/L and/or an A1C of 6.0%–6.4% in order to identify individuals with impaired glucose tolerance (IGT) or diabetes.•Testing with a 2hPG in a 75 g OGTT may be undertaken in individuals with an FPG 5.6–6.0 mmol/L and/or A1C 5.5%–5.9% and ≥1 risk factor in order to identify individuals with IGT or diabetes. •In the absence of evidence for interventions to prevent or delay type 1 diabetes, screening for type 1 diabetes is not recommended.•Screening for type 2 diabetes using a fasting plasma glucose (FPG) and/or glycated hemoglobin (A1C) should be performed every 3 years in individuals ≥40 years of age or in individuals at high risk using a risk calculator.•Diabetes will be diagnosed if A1C is ≥6.5% (see Definition, Classification and Diagnosis chapter, p. S8).•Testing with a 2-hour plasma glucose (2hPG) in a 75 g oral glucose tolerance test (OGTT) should be undertaken in individuals with an FPG of 6.1–6.9 mmol/L and/or an A1C of 6.0%–6.4% in order to identify individuals with impaired glucose tolerance (IGT) or diabetes.•Testing with a 2hPG in a 75 g OGTT may be undertaken in individuals with an FPG 5.6–6.0 mmol/L and/or A1C 5.5%–5.9% and ≥1 risk factor in order to identify individuals with IGT or diabetes. The clinical spectrum of diabetes ranges from a low-risk to a higher-risk individual or to the symptomatic patient who needs immediate treatment. Screening for diabetes implies testing for diabetes in individuals without symptoms who are unaware of their condition. Screening for diabetes will also detect individuals at increased risk for diabetes (prediabetes) or individuals with less severe states of dysglycemia who may still be at risk for type 2 diabetes. Screening strategies vary according to the type of diabetes and evidence of effective interventions to prevent progression of prediabetes to diabetes and/or reduce the risk of complications associated with diabetes. The growing importance of diabetes screening is undeniable (1Gilmer T.P. O’Connor P.J. The growing importance of diabetes screening.Diabetes Care. 2010; 33: 1695-1697Crossref PubMed Scopus (17) Google Scholar). In contrast to other diseases, there is no distinction between screening and diagnostic testing. Therefore, to screen for diabetes and prediabetes, the same tests would be used as for diagnosis of both medical conditions (see Definition, Classification and Diagnosis chapter, p. S8). Type 1 diabetes mellitus is primarily a result of pancreatic beta cell destruction due to an immune-mediated process that is likely incited by environmental factors in genetically predisposed individuals. An individual’s risk of developing type 1 diabetes can be estimated by considering family history of type 1 diabetes with attention to age of onset and sex of the affected family members (2Harjutsalo V. Reunanen A. Tuomilehto J. Differential transmission of type 1 diabetes from diabetic fathers and mothers to their offspring.Diabetes. 2006; 55: 1517-1524Crossref PubMed Scopus (58) Google Scholar) and profiling immunity and genetic markers (3Decochez K. Truyen I. van der Auwera B. et al.Belgian Diabetes RegistryCombined positivity for HLA DQ2/DQ8 and IA-2 antibodies defines population at high risk of developing type 1 diabetes.Diabetologia. 2005; 48: 687-694Crossref PubMed Scopus (39) Google Scholar). The loss of pancreatic beta cells in the development of type 1 diabetes passes through a subclinical prodrome that can be detected reliably in first- and second-degree relatives of persons with type 1 diabetes by the presence of pancreatic islet autoantibodies in their sera (4Bingley P.J. Interactions of age, islet cell antibodies, insulin autoantibodies, and first-phase insulin response in predicting risk of progression to IDDM in ICA+ relatives: the ICARUS data set. Islet Cell Antibody Register Users Study.Diabetes. 1996; 45: 1720-1728Crossref PubMed Google Scholar). However, in a recent large study, one-time screening for glutamic acid decarboxylase antibodies (GADAs) and islet antigen-2 antibodies (IA-2As) in the general childhood population in Finland would identify 60% of those individuals who will develop type 1 diabetes over the next 27 years. Initial positivity for GADAs and/or IA-2As had a sensitivity of 61% (95%confidence interval [CI] 36–83%) for type 1 diabetes. The combined positivity for GADAs and IA-2As had both a specificity and a positive predictive value of 100% (95% CI 59–100%) (5Knip M. Korhonen S. Kulmala P. et al.Prediction of type 1 diabetes in the general population.Diabetes Care. 2010; 33: 1206-1212Crossref PubMed Scopus (116) Google Scholar). Ongoing clinical studies are testing different strategies for preventing or reversing early type 1 diabetes in the presence of positive autoimmunity. Given that the various serological markers are not universally available and in the absence of evidence for interventions to prevent or delay type 1 diabetes, no widespread recommendations for screening for type 1 diabetes can be made. Undiagnosed type 2 diabetes may occur in >2.8% of the general adult population (6Cowie C.C. Rust K.F. Byrd-Holt D.D. et al.Prevalence of diabetes and impaired fasting glucose in adults in the U.S. population: National Health And Nutrition Examination Survey 1999-2002.Diabetes Care. 2006; 29: 1263-1268Crossref PubMed Scopus (1037) Google Scholar), and the number increases to >10% in some populations (7Rolka D.B. Narayan K.M. Thompson T.J. et al.Performance of recommended screening tests for undiagnosed diabetes and dysglycemia.Diabetes Care. 2001; 24: 1899-1903Crossref PubMed Scopus (120) Google Scholar, 8Rathmann W. Haastert B. Icks A. et al.High prevalence of undiagnosed diabetes mellitus in Southern Germany: target populations for efficient screening. The KORA survey 2000.Diabetologia. 2003; 46: 182-189PubMed Google Scholar). Tests for hyperglycemia can identify these individuals, many of whom will have, or will be at risk for, preventable diabetes complications (5Knip M. Korhonen S. Kulmala P. et al.Prediction of type 1 diabetes in the general population.Diabetes Care. 2010; 33: 1206-1212Crossref PubMed Scopus (116) Google Scholar, 6Cowie C.C. Rust K.F. Byrd-Holt D.D. et al.Prevalence of diabetes and impaired fasting glucose in adults in the U.S. population: National Health And Nutrition Examination Survey 1999-2002.Diabetes Care. 2006; 29: 1263-1268Crossref PubMed Scopus (1037) Google Scholar). To be effective, population-based screening would have to involve wide coverage and would have the goal of early identification and subsequent intervention to reduce morbidity and mortality. Using various multistaged screening strategies, the ADDITION-Europe study showed that 20% to 94% of eligible people in primary care practices attended the first blood glucose test of the screening process, and diabetes was detected in 0.33% to 1.09% of the target populations, which was lower than expected (9Van den Donk M. Sandbaek A. Borch-Johnsen et al.Screening for type 2 diabetes: lessons from the ADDITION-Europe study.Diabet Med. 2011; 28: 1416-1426Crossref PubMed Scopus (48) Google Scholar). In the subsequent ADDITION-Europe cluster randomized trial of intensive multifaceted cardiovascular risk factor management vs. routine diabetes care among screening-identified type 2 diabetes patients, intensive management did not reduce cardiovascular events (hazard ratio 0.83; 95% CI 0.65–1.05) or all-cause mortality (hazard ratio 0.91; 95% CI 0.69–1.21) (10Griffin S.J. Borch-Johnsen K. Davies M.J. et al.Effect of early intensive multifactorial therapy on 5-year cardiovascular outcomes in individuals with type 2 diabetes detected by screening (ADDITION-Europe): a cluster-randomised trial.Lancet. 2011; 378: 156-167Abstract Full Text Full Text PDF PubMed Scopus (357) Google Scholar). Of note, a very high proportion of the routine care group also received optimal cardiovascular risk factor management, which may have diluted any potential benefits. In ADDITION-Cambridge, population-based screening for type 2 diabetes was not associated with a reduction in all-cause, cardiovascular or diabetes-related mortality within 10 years compared to a no-screening control group. However, the low rate of type 2 diabetes in the screened population (3%) was likely too small to affect overall population mortality (11Simmons R.K. Echouffo-Tcheugui J.B. Sharp S.J. et al.Screening for type 2 diabetes and population mortality over 10 years (ADDITION-Cambridge): a cluster-randomised controlled trial.Lancet. 2012; 380: 1741-1748Abstract Full Text Full Text PDF PubMed Scopus (178) Google Scholar). Nonetheless, there is no current evidence of clinical benefit to support a strategy of population-based screening for type 2 diabetes. Although the relatively low prevalence of diabetes in the general population makes it unlikely that mass screening will be cost effective, testing for diabetes in people with risk factors for type 2 diabetes or with diabetes-associated conditions is likely to result in more benefit than harm and will lead to overall cost savings (12Raikou M. McGuire A. The economics of screening and treatment in type 2 diabetes mellitus.Pharmacoeconomics. 2003; 21: 543-564Crossref PubMed Scopus (31) Google Scholar, 13The cost-effectiveness of screening for type 2 diabetes. CDC Diabetes Cost-Effectiveness Study Group, Centers for Disease Control and Prevention.JAMA. 1998; 280: 1757-1763Crossref PubMed Scopus (202) Google Scholar, 14Kahn R. Alperin P. Eddy D. et al.Age at initiation of screening to detect type 2 diabetes: a cost-effectiveness analysis.Lancet. 2010; 375: 1365-1374Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar, 15Simmons R.K. Rahman M. Jakes R.W. et al.Effect of population screening for type 2 diabetes on mortality: long term follow-up of the Ely cohort.Diabetologia. 2011; 54: 312-319Crossref PubMed Scopus (45) Google Scholar, 16Gilles C.L. Lambert P.C. Abrahms K.R. et al.Different strategies for screening and prevention of type 2 diabetes in adults: cost-effectiveness analysis.BMJ. 2008; 336: 1180-1184Crossref PubMed Scopus (217) Google Scholar, 17Hoerger T.J. Hicks K.A. Sorensen S.W. et al.Cost-effectiveness of screening for prediabetes among overweight and obese U.S adults.Diabetes Care. 2007; 30: 2874-2879Crossref PubMed Scopus (85) Google Scholar). Routine testing for type 2 diabetes is, therefore, justifiable in some but not all settings (18Knowler W.C. Screening for NIDDM. Opportunities for detection, treatment, and prevention.Diabetes Care. 1994; 17: 445-450Crossref PubMed Scopus (62) Google Scholar, 19Simmons R.K. Echouffo-Tcheugui J.B. Griffin S.J. Screening for type 2 diabetes: an update of evidence.Diabetes Obes Metab. 2010; 12: 838-844Crossref PubMed Scopus (36) Google Scholar). Screening individuals as early as age 40 years in family physicians’ offices has proved to be useful in detecting unrecognized diabetes (20Leiter L.A. Barr A. Bélanger A. et al.Diabetes Screening in Canada (DIASCAN) StudyDiabetes Screening in Canada (DIASCAN) Study: prevalence of undiagnosed diabetes and glucose intolerance in family physician offices.Diabetes Care. 2001; 24: 1038-1043Crossref PubMed Scopus (77) Google Scholar). While fasting plasma glucose (FPG) and/or glycated hemoglobin (A1C) are the recommended screening tests, a 75 g oral glucose tolerance test (OGTT) is indicated when the FPG is 6.1 to 6.9 mmol/L (14Kahn R. Alperin P. Eddy D. et al.Age at initiation of screening to detect type 2 diabetes: a cost-effectiveness analysis.Lancet. 2010; 375: 1365-1374Abstract Full Text Full Text PDF PubMed Scopus (194) Google Scholar) and/or A1C is 6.0% to 6.4%. It may be indicated when the FPG is 5.6 to 6.0 mmol/L and/or A1C is 5.5% to 5.9% and suspicion of type 2 diabetes or impaired glucose tolerance (IGT) is high (e.g. for individuals with risk factors listed in Table 1) (Figure 1).Table 1Risk factors for type 2 diabetes•Age ≥40 years•First-degree relative with type 2 diabetes•Member of high-risk population (e.g. Aboriginal, African, Asian, Hispanic or South Asian descent)•History of prediabetes (IGT, IFG or A1C 6.0%–6.4%)∗Associated with insulin resistance.•History of gestational diabetes mellitus•History of delivery of a macrosomic infant•Presence of end organ damage associated with diabetes:○Microvascular (retinopathy, neuropathy, nephropathy)○Macrovascular (coronary, cerebrovascular, peripheral)•Presence of vascular risk factors:○HDL cholesterol level <1.0 mmol/L in males, <1.3 mmol/L in females∗Associated with insulin resistance.○Triglycerides ≥1.7 mmol/L∗Associated with insulin resistance.○Hypertension∗Associated with insulin resistance.○Overweight∗Associated with insulin resistance.○Abdominal obesity∗Associated with insulin resistance.•Presence of associated diseases:○Polycystic ovary syndrome∗Associated with insulin resistance.○Acanthosis nigricans∗Associated with insulin resistance.○Psychiatric disorders (bipolar disorder, depression, schizophrenia†The incidence of type 2 diabetes is at least 3 times higher in people with schizophrenia than in the general population (25,26). Using data collected in 1991, the prevalence of diabetes was assessed in >20,000 individuals diagnosed with schizophrenia. The rate of diagnosed diabetes was 9% to 14%, exceeding rates for the general population prior to the widespread use of new antipsychotic drugs (27).)○HIV infection‡HIV and HAART increase the risk of prediabetes (IGT) and type 2 diabetes by 1.5- to 4-fold compared to the general population (28).○OSA§OSA is an independent risk factor for diabetes (hazard ratio 1.43) (29).•Use of drugs associated with diabetes:○Glucocorticoids○Atypical antipsychotics○HAART‡HIV and HAART increase the risk of prediabetes (IGT) and type 2 diabetes by 1.5- to 4-fold compared to the general population (28).○Other (see Appendix 1)•Other secondary causes (see Appendix 1)A1C, glycated hemoglobin; HAART, highly active antiretroviral therapy; HDL, high-density lipoprotein; HIV, human immunodeficiency virus-1; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; OSA, obstructive sleep apnea.∗ Associated with insulin resistance.† The incidence of type 2 diabetes is at least 3 times higher in people with schizophrenia than in the general population 25McKee H.A. D’Arcy P.F. Wilson P.J. Diabetes and schizophrenia: a preliminary study.J Clin Hosp Pharm. 1986; 11: 297-299PubMed Google Scholar, 26Mukherjee S. Decina P. Bocola V. et al.Diabetes mellitus in schizophrenic patients.Compr Psychiatry. 1996; 37: 68-73Abstract Full Text PDF PubMed Scopus (392) Google Scholar. Using data collected in 1991, the prevalence of diabetes was assessed in >20,000 individuals diagnosed with schizophrenia. The rate of diagnosed diabetes was 9% to 14%, exceeding rates for the general population prior to the widespread use of new antipsychotic drugs 27Dixon L. Weiden P. Delahanty J. et al.Prevalence and correlates of diabetes in national schizophrenia samples.Schizophr Bull. 2000; 26: 903-912Crossref PubMed Scopus (523) Google Scholar.‡ HIV and HAART increase the risk of prediabetes (IGT) and type 2 diabetes by 1.5- to 4-fold compared to the general population 28Samaras K. Chisholm D.J. Diabetes, insulin resistance and glucose metabolism in HIV infection and its treatment.in: Ekoe J.M. Rewers M. Williams R. The Epidemiology of Diabetes Mellitus. Wiley Blackwell, Chichester, UK2008: 665-675Crossref Scopus (3) Google Scholar.§ OSA is an independent risk factor for diabetes (hazard ratio 1.43) 29Botros N. Concato J. Mohsenin V. et al.Obstructive sleep apnea as a risk factor for type 2 diabetes.Am J Med. 2009; 122: 1122-1127Abstract Full Text Full Text PDF PubMed Scopus (278) Google Scholar. Open table in a new tab A1C, glycated hemoglobin; HAART, highly active antiretroviral therapy; HDL, high-density lipoprotein; HIV, human immunodeficiency virus-1; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; OSA, obstructive sleep apnea. People with prediabetes, especially those with IGT or an A1C of 6.0% to 6.4%, not only are at increased risk of developing type 2 diabetes, but they also have an increased risk of macrovascular complications, particularly in the context of the metabolic syndrome (21Hu G. Qiao Q. Tuomilehto J. et al.DECODE Study GroupPrevalence of the metabolic syndrome and its relation to all- cause and cardiovascular mortality in nondiabetic European men and women.Arch Intern Med. 2004; 164: 1066-1076Crossref PubMed Scopus (907) Google Scholar). These individuals would benefit from cardiovascular risk factor reduction strategies (1Gilmer T.P. O’Connor P.J. The growing importance of diabetes screening.Diabetes Care. 2010; 33: 1695-1697Crossref PubMed Scopus (17) Google Scholar). Members of high-risk ethnic populations (Table 1) should be screened for prediabetes and type 2 diabetes using the recommended screening tests, such as FPG, OGTT and A1C. However, the high prevalence of hemoglobinopathies among these populations may considerably reduce the accuracy of A1C as a reliable screening tool in these populations. Furthermore, high-risk ethnic groups may have A1C levels that are slightly higher than those of Caucasians at the same level of glycemia, and more studies may help determine ethnic-specific A1C thresholds for diabetes diagnosis (see Definition, Classification and Diagnosis chapter, p. S8). A number of risk scores based on clinical characteristics have been developed to identify individuals at high risk of having undiagnosed diabetes. However, the impact of known risk factors on having undiagnosed type 2 diabetes differs between populations of different ethnic origins, and risk scores developed in Caucasian populations cannot be applied to populations of other ethnic groups (22Glumer C. Vistisen D. Borch-Johnsen K. et al.Detect 2 CollaborationRisk scores for type 2 diabetes can be applied in some populations but not all.Diabetes Care. 2006; 29: 410-414Crossref PubMed Scopus (110) Google Scholar). Furthermore, the prevalence of individuals at risk for developing type 2 diabetes varies considerably according to the scoring system. Risk scoring systems must, therefore, be validated for each considered population in order to adequately detect individuals at risk and eventually implement efficacious preventative strategies (23Schmid R. Vollenweider P. Waeber G. et al.Estimating the risk of developing type 2 diabetes: a comparison of several risk scores. The Cohorte Lausannoise Study.Diabetes Care. 2011; 34: 1863-1868Crossref PubMed Scopus (18) Google Scholar). The Canadian Diabetes Risk Assessment Questionnaire (CANRISK) is a statistically valid tool that may be suitable for diabetes risk assessment in the Canadian population and is available on the Internet at www.phac-aspc.gc.ca/cd-mc/diabetes-diabete/canrisk/index-eng.php (24Robinson C.A. Agarwal G. Validating the CANRISK prognostic model for assessing diabetes risk in Canada’s multi-ethnic population.Chron Dis Inj Can. 2011; 32: 19-31PubMed Google Scholar).Recommendations1.All individuals should be evaluated annually for type 2 diabetes risk on the basis of demographic and clinical criteria [Grade D, Consensus].2.Screening for diabetes using FPG and/or A1C should be performed every 3 years in individuals ≥40 years of age or at high risk using a risk calculator [Grade D, Consensus]. More frequent and/or earlier testing with either FPG and/or A1C or 2hPG in a 75 g OGTT should be considered in those at very high risk using a risk calculator or in people with additional risk factors for diabetes [Grade D, Consensus]. These risk factors include:•First-degree relative with type 2 diabetes•Member of high-risk population (e.g. Aboriginal, African, Asian, Hispanic or South Asian descent)•History of prediabetes (IGT, IFG, or A1C 6.0%–6.4%)•History of gestational diabetes mellitus•History of delivery of a macrosomic infant•Presence of end organ damage complications associated with diabetes:○Microvascular (retinopathy, neuropathy, nephropathy)○Macrovascular (coronary, cerebrovascular, peripheral)•Presence of vascular risk factors:○HDL cholesterol <1.0 mmol/L in males, <1.3 mmol/L in females○Triglycerides ≥1.7 mmol/L○Hypertension○Overweight○Abdominal obesity•Presence of associated diseases:○Polycystic ovary syndrome○Acanthosis nigricans○Obstructive sleep apnea○Psychiatric disorders (bipolar disorder, depression, schizophrenia)○HIV infection•Use of drugs associated with diabetes:○Glucocorticoids○Atypical antipsychotics○HAART○Other (see Appendix 1)•Other secondary causes (see Appendix 1)3.Testing with 2hPG in a 75 g OGTT should be undertaken in individuals with FPG 6.1–6.9 mmol/L and/or A1C 6.0%–6.4% in order to identify individuals with IGT or diabetes [Grade D, Consensus].4.Testing with 2hPG in a 75 g OGTT may be undertaken in individuals with FPG 5.6–6.0 mmol/L and/or A1C 5.5%–5.9% and ≥1 risk factor(s) in order to identify individuals with IGT or diabetes [Grade D, Consensus].Abbreviations:2hPG, 2-hour plasma glucose; A1C, glycated hemoglobin; FPG, fasting plasma glucose; HAART, highly active antiretroviral therapy; HDL, high-density lipoprotein; HIV, human immunodeficiency virus-1; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; OGTT, oral glucose tolerance test. 1.All individuals should be evaluated annually for type 2 diabetes risk on the basis of demographic and clinical criteria [Grade D, Consensus].2.Screening for diabetes using FPG and/or A1C should be performed every 3 years in individuals ≥40 years of age or at high risk using a risk calculator [Grade D, Consensus]. More frequent and/or earlier testing with either FPG and/or A1C or 2hPG in a 75 g OGTT should be considered in those at very high risk using a risk calculator or in people with additional risk factors for diabetes [Grade D, Consensus]. These risk factors include:•First-degree relative with type 2 diabetes•Member of high-risk population (e.g. Aboriginal, African, Asian, Hispanic or South Asian descent)•History of prediabetes (IGT, IFG, or A1C 6.0%–6.4%)•History of gestational diabetes mellitus•History of delivery of a macrosomic infant•Presence of end organ damage complications associated with diabetes:○Microvascular (retinopathy, neuropathy, nephropathy)○Macrovascular (coronary, cerebrovascular, peripheral)•Presence of vascular risk factors:○HDL cholesterol <1.0 mmol/L in males, <1.3 mmol/L in females○Triglycerides ≥1.7 mmol/L○Hypertension○Overweight○Abdominal obesity•Presence of associated diseases:○Polycystic ovary syndrome○Acanthosis nigricans○Obstructive sleep apnea○Psychiatric disorders (bipolar disorder, depression, schizophrenia)○HIV infection•Use of drugs associated with diabetes:○Glucocorticoids○Atypical antipsychotics○HAART○Other (see Appendix 1)•Other secondary causes (see Appendix 1)3.Testing with 2hPG in a 75 g OGTT should be undertaken in individuals with FPG 6.1–6.9 mmol/L and/or A1C 6.0%–6.4% in order to identify individuals with IGT or diabetes [Grade D, Consensus].4.Testing with 2hPG in a 75 g OGTT may be undertaken in individuals with FPG 5.6–6.0 mmol/L and/or A1C 5.5%–5.9% and ≥1 risk factor(s) in order to identify individuals with IGT or diabetes [Grade D, Consensus].Abbreviations:2hPG, 2-hour plasma glucose; A1C, glycated hemoglobin; FPG, fasting plasma glucose; HAART, highly active antiretroviral therapy; HDL, high-density lipoprotein; HIV, human immunodeficiency virus-1; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; OGTT, oral glucose tolerance test. Definition, Classification and Diagnosis of Diabetes, Prediabetes and Metabolic Syndrome, p. S8 Reducing the Risk of Developing Diabetes, p. S16 Type 1 Diabetes in Children and Adolescents, p. S153 Type 2 Diabetes in Children and Adolescents, p. S163 Appendix 1. Etiologic Classification of Diabetes Mellitus Appendix 1: Etiologic Classification of Diabetes MellitusCanadian Journal of DiabetesVol. 37Preview Full-Text PDF