Rapid Progression From Pre-diabetes to Severely Ill Diabetes While Under“Expert Care”: Suggestions for Improved Screening for Disease Progression Free

B.B. is a 16.8-year-old African-American male who presented to the Lifestyle Clinic at Le Bonheur Children's Medical Center with a chief complaint of abnormal weight gain and random serum glucose levels of 175 and 164 mg/dl. He was the product of a full-term pregnancy complicated by excessive maternal weight gain, with a birth weight > 8 lb. He had lifelong excessive weight gain, for which he had seen his pediatrician 3.5 months earlier (at age 16.5 years) and was found to have hypertension(138–172/72–112 mmHg). Referral by his primary physician to the renal clinic 10 weeks earlier (at age 16.6 years) led to treatment with an 1,800-calorie diet and quinapril, 40 mg/day.

His symptoms at the time of the first Lifestyle Clinic visit included nocturnal enuresis, snoring, daytime somnolence, and chronic fatigue. Family history was positive for obesity, diabetes in parents, CVD, and hypertension. He was in regular classes in the 11th grade, lived with his mother, and had good family support.

Physical examination showed height 74.3 inches, weight 389 lb, BMI 49.6 kg/m², waist and hip circumferences each 52.8 inches, waist-to-hip ratio (WHR) 1.00, chest measurement 56.7 inches, heart rate (HR) 91 bpm, and blood pressure 182/77 mm Hg. He had central obesity, acanthosis nigricans grade 4, gynecomastia, hepatomegaly, and Tanner Stage 4 pubertal development.

Laboratory studies done at the time of diagnosis of hypertension (age 16.6 years) had shown normal concentrations of serum electrolytes, glucose, blood urea nitrogen (BUN), and creatinine (Cr). B.B. had increased levels of uric acid at 8.1 mg/dl (normal range 3.0–7.0 mg/dl), alanine aminotransferase(ALT) 160 IU/l (normal range 0–45 IU/l), hemoglobin A_1c (A1C)6.1% (normal range ≤ 6.0%), and an abnormal lipid profile(Table 1). Urinalysis was negative for glucose but positive for trace protein and blood and a slightly increased microalbumin-to-Cr ratio of 36 (normal < 25). Serum thyrotropin(TSH), plasma renin activity, and aldosterone levels were normal.

Studies obtained at the initial visit to the Lifestyle Clinic (week 0, age 16.8 years) showed normal serum levels of fibrinogen, C-reactive protein(CRP), free thyroxine (free T4) and cortisol, and testosterone (344 ng/dl,normal range 200–620). Levels of A1C, insulin, C-peptide, and proinsulin were elevated (Table 1). Subsequently, a sleep study showed primary snoring with only minimal episodes of obstructive apnea. An oral glucose tolerance test (OGTT) performed at week 9 (age 17.0 years) showed hyperinsulinism and hyperglycemia(Table 2).

The above findings of abnormal weight gain with acanthosis nigricans,hyperinsulinism, hyperglycemia, hypertension, and dyslipidemia categorized B.B. as having insulin resistance, pre-diabetes, and metabolic syndrome. Treatment recommendations focused on implementing dietary changes, instituting a program of aerobic exercise and strength training to reduce weight and improve fitness, and initiating daily home monitoring of blood pressure and fasting blood glucose. Frequent clinic visits were scheduled to provide further education and motivation. The addition of metformin therapy was considered but deferred in hopes that the patient would succeed with lifestyle changes. B.B.'s dose of quinapril was increased to 40 mg twice daily, and hydrochlorothiazide, 12.5 mg/day, was added.

Five months later, at the fourth clinic visit (week 21, age 17.2 years),the patient's blood pressure was lower, but his weight, BMI, WHR, and A1C were unchanged (Table 1). He was exercising and lifting weights. Measurements (n = 36) taken at home over the preceding 7 weeks showed an average blood pressure of 135/85 mmHg(range 103–168/58–113). His fasting blood glucose levels averaged 104 mg/dl (range 92–114 with four outliers excluded). At previous visits, B.B. had complained of pain and numbness in his feet and lower legs,which had been attributed to his high blood glucose levels. The patient had not mentioned that he had also been experiencing polyuria, polydipsia, and polyphagia, because he did not realize the significance of these symptoms.

Just 7.5 weeks later (week 28.6, age 17.4 years), B.B. developed nausea and vomiting for 2 days, with marked weakness and tiredness. During the previous 2 months, his symptoms of polyuria, polydipsia, and polyphagia had worsened. During the past month, he had become weaker, felt more tired, and needed more sleep. The pain and numbness in his feet and lower legs had also worsened.(The patient does not recall being told to contact anyone if these symptoms developed or if glucose levels above a certain range would indicate a need for medical attention.) Despite severe illness, he went to school to take an important test, intending to go home after the test. During the test, he had to leave the room to use the restroom. On the way, his legs became weak and wobbly and he reported that he “passed out.” He remembered people talking to him but could not respond to them.

B.B. was taken to a local hospital emergency department. He complained of pain in his legs and arms and of a sore throat. His vital signs were: heart rate 120 bpm, respiratory rate 22/minute, blood pressure 133/55 mmHg,temperature 100.3° F, and weight 361 lb (decreased 32 lb or 8.2% from his weight at the clinic visit in week 21). He was variably oriented or confused,his mucous membranes were dry, his abdomen was soft, and he had tenderness in his arms and legs. Laboratory tests showed sodium 118 mm/l (normal 135–145), potassium 7.0 mm/l (normal 3.5–5.3), blood glucose 1,560 mg/dl (normal 70–115), BUN 32 mg/dl (normal range 9–21), Cr 2.4 mg/dl (normal range 0.5–1.2), calculated serum osmolality 334 mOsm/kg(normal range 270–290), acetone negative, venous pH 7.33, carbon dioxide 48 mmHg, bicarbonate 25.3 mm/l, urine glucose 4+, and ketones trace. He was diagnosed as having new-onset diabetes with dehydration, hyperglycemic hyperosmolality, and viral syndrome. Intravenous saline and insulin therapies were given.

The patient was transferred to Le Bonheur Children's Medical Center and admitted to the ICU. His serum glucose had decreased to 753 mg/dl and acetone was positive at a titer of 1:8. His sensorium cleared over the next hours, and an ultrasound study was negative for deep vein thrombosis. His A1C result was 11.2%, markedly higher than the 6.8% measured just 7.5 weeks earlier, and his levels of insulin and C-peptide were strikingly lower than those measured with the OGTT 5 months earlier (Tables 1 and 2). However, before discharge on hospital day 6 and with insulin therapy withheld, a glucagon stimulation test performed to assess residual β-cell function showed recovery with a surprisingly excellent basal and peak C-peptide response(Table 3), supporting a diagnosis of type 2 diabetes.

B.B.'s hospital course was additionally complicated by the development on day 5 of a dark red urine that showed 3+ blood but only 3–5 red blood cells per high power field on microscopic exam, with negative serum myoglobin but elevated creatine kinase at 601 IU/l (normal range 30–210),diagnostic of rhabdomyolysis. Hydration and alkalinization therapy produced clearing of the urine 12 hours later. At discharge, B.B. was treated with insulin, 0.9 units · kg^–1 ·day^–1.

Five weeks after discharge (week 35, age 17.5 years), B.B. had metformin therapy, 500 mg three times a day, added. His blood pressure had improved to 110/60 mmHg, permitting his quinapril to be reduced to 20 mg/day. His diabetes control worsened in the following months, however, with insulin requirements increasing to 1.35 units · kg^–1 ·day^–1 and his A1C increasing to 14.0% (normal range ≤6.0). At week 52, age 17.8 years. B.B.'s compliance in taking insulin was not good, and a 4-month trial with an insulin pump brought no improvement. His weight decreased because of poor diabetes control along with increased exercise and strength training.

At his 12th clinic visit (week 95, age 18.6 years, 22 months since the first Lifestyle Clinic visit), B.B.'s blood pressure remained well controlled,but he continued to be poorly compliant in taking his insulin, and his weight and BMI were at their lowest. Laboratory studies showed normal serum levels of electrolytes, ALT, fibrinogen, CRP, TSH, free T4, and cortisol but abnormal serum elevations of BUN at 31 mg/dl and Cr at 1.6 mg/dl. He now had marked elevations in A1C and serum triglycerides, with lower levels of insulin,C-peptide, and proinsulin (Table 1). Efforts were made to improve his insulin therapy, and combination rosiglitazone/metformin, 2/500 mg/day, was added with metformin,500 mg, given at night. Because of his age, he was referred to the adult service for future care.

The patient, at age 20.2 years (3.4 years after his initial clinic visit),has received his medical care at a general health clinic and has only recently been followed by one physician. He has experienced occasional slurred speech,left-sided weakness, and foot drop. His doctor believes he may have had a stroke. His blood pressure is stable without medication. His weight fluctuates between 250 and 260 lb, and he cannot gain weight. His acanthosis nigricans is diminished or gone since he lost weight. He was recently hospitalized for high blood glucose levels and restarted on insulin therapy twice daily.

B.B. entered culinary school after high school. The program involved 1 year of class work and 4 months of fieldwork. He left school after 4 months because he felt sick all of the time, which worried him because he was far from home. He he has been living on his own since, but he has been unable to hold a job because of his illness and has filed for disability support. He is planning to move back to his mother's home. Nevertheless, he is upbeat and says he still has goals and believes he will meet them someday.

The rapid progression of this patient from pre-diabetes to full-blown type 2 diabetes with acute complications, while under our care, raised the following challenges: 1) to identify which patients are at very high risk for developing diabetes, and 2) to recognize early on, before acute complications develop, when a patient has progressed to diabetes.

The identification of patients at very high risk for developing type 2 diabetes centers on understanding its causative factors and natural history. Type 2 diabetes is defined as resulting from a progressive insulin secretory defect on a background of increased insulin resistance.^4,5 Genetic and environmental/behavioral factors play important roles. While most studies have been done in adults, it is likely that the same explanations apply to the pediatric age group.

There is a strong association of development of type 2 diabetes with a family history of the disease; 74–100% of patients have a first- or second-degree relative with diabetes, and 45–80% have at least one parent with diabetes.⁶  This association extends to first-degree relatives with no known history of diabetes, in whom OGTT has revealed 19% to have unrecognized diabetes and 36%to have pre-diabetes, with only 45% having normal glucose tolerance.⁷ 

Periodic testing of normal offspring (ages 16–60 years) of patients with type 2 diabetes found that 16% went on to develop diabetes, a rate eight times the general population risk. The average period of follow-up testing was 13 years. At initial evaluation (1–2 decades before they developed diabetes), the affected offspring already had reduced glucose clearance with accompanying hyperinsulinemia on intravenous glucose tolerance testing(IVGTT). This suggests that the initial defect is in the peripheral tissue response to insulin and glucose, not in the pancreaticβ-cell.⁸ 

A study of 54 healthy white children and adolescents, 28 with a positive family history of type 2 diabetes and 26 with a negative family history, found in those with a family history of diabetes that insulin sensitivity and insulin clearance were lower, whereas first- and second-phase insulin levels were not different. The glucose disposition index (GDI; insulin sensitivity× first-phase insulin secretion) was also lower. These results demonstrate that family history of type 2 diabetes in white children is associated with an impaired relationship between insulin action andβ-cell compensation. Positive family history suggests that at least some of the determinants of GDI are genetic/heritable.⁹ However, a significant association with family history of diabetes may not always be seen.¹⁰ 

The occurrence of type 2 diabetes shows racial/ethnic differences as well. The disease occurs in all racial groups, but its prevalence is increased in certain groups of non-European ancestry, including American Indians, African Americans, Hispanics, Asian/Pacific Islanders, and Asian Amer-icans.^6,7,11–14 One explanation is that there appear to be racial differences in insulin sensitivity in childhood. African-American children and adolescents have higher insulin levels and higher insulin responses to the OGTT than white children and adolescents.^15,16 African Americans may also have 30% lower insulin sensitivity than whites,¹⁶  although a second study failed to confirm this and found no difference.¹⁷ Normal, young, nonobese Mexican Americans have decreased insulin sensitivity and greater insulin secretion on IVGTT compared to whites, indicating increased insulin resistance.¹⁸ 

Analysis of data from the Third National Health and Nutrition Examination Survey, 1988–1994, also revealed racial and ethnic differences. The rates of diabetes for adults ≥ 20 years of age in the United States were similar between the sexes, but the rates for non-Hispanic blacks and Mexican Americans were 1.6 and 1.9 times the rate for non-Hispanic whites.¹⁹  For 7,968 children and young adults aged 5–24 years and not treated for diabetes,mean A1C levels, when adjusted for confounders, were higher in African Americans (5.15%) and Mexican Americans (5.01%) than in whites (4.93%). These differences persisted within age subgroups, among both males and females and both overweight and nonoverweight individuals, and at all levels of education.²⁰ 

The key environmental/behavioral factors associated with the occurrence of type 2 diabetes are those that increase insulin resistance. These include obesity, sedentary lifestyle, and puberty. Smoking and sleep apnea or lack of sleep are also associated with development of diabetes

Obesity. Obesity is strongly associated with insulin resistance and the occurrence of type 2 diabetes. Increasing severity of overweight and obesity in adults is associated with a graded increase in prevalence of diabetes (also high blood pressure, gallbladder disease, and osteoarthritis)in both sexes and across all racial and ethnic subgroups.²¹  Almost all pediatric patients with type 2 diabetes are overweight or obese. Obesity is not a sine qua non, however; ∼ 35% of adults and a small number of pediatric patients with type 2 diabetes are not overweight or obese.¹ 

Visceral (intra-abdominal) adiposity is more directly associated with insulin resistance and abnormal glucose and lipid metabolism than is total(visceral plus subcutaneous) adiposity. A study of obese adults found that those with a visceral-to-subcutaneous adipose tissue ratio (VAT-to-SAT ratio)≥ 0.4 had a significantly higher fasting plasma glucose (FPG) level, area under the plasma glucose concentration curve after oral glucose loading(AUC-Glu), and fasting serum triglyceride and total cholesterol levels than did those with a lower VAT-to-SAT ratio. The VAT-to-SAT ratio was significantly correlated with the AUC-Glu and with serum triglyceride and cholesterol levels.²²  Possible explanations given to explain a link with visceral fat are a key role of elevated nonesterified fatty acid concentrations; the deposition of triglycerides in muscle, liver, and pancreatic cells that interfere with insulin sensitivity; and the adipose tissue as endocrine organ, with secretion of various adipocytokins (e.g., leptin, tumor necrosis factor-α[TNF-α], resistin, and adiponectin), which are implicated in insulin sensitivity and possibly β-cell function.²³ 

Sedentary lifestyle. A sedentary lifestyle or lack of physical activity is an independent risk factor for insulin resistance and is significantly associated with pre-diabetes and type 2 diabetes. Regular physical exercise improves insulin action in skeletal muscle in insulin-resistant individuals and improves glucose tolerance and insulin action in individuals predisposed to develop diabetes.²⁴  Studies in adults have shown that increased exercise or activity reduces the incidence of type 2 diabetes by 25–40%.²⁵ 

A 6-year study of 577 adult Chinese men and women with pre-diabetes examined whether diet and exercise interventions would reduce the incidence of type 2 diabetes. When adjusted for differences in baseline BMI and fasting glucose, the diet, exercise, and diet-plus-exercise interventions were associated with 31% (P < 0.03), 46% (P < 0.0005), and 42% (P < 0.005) reductions in risk of developing diabetes,respectively. The reductions were similar when subjects were stratified as lean or overweight. Thus, the implementation of an exercise or diet intervention produces a significant decrease in the incidence of type 2 diabetes over a 6-year period in those with pre-diabetes.²⁶ 

Increased fitness is associated with a lower incidence of type 2 diabetes. In the Physicians' Health Study, the age-adjusted relative risk of type 2 diabetes decreased with increasing frequency of exercise: 0.77 for once weekly, 0.62 for two to four times per week, and 0.58 for five or more times per week (P for trend =0.0002).²⁷  In children and adolescents, a sedentary lifestyle is associated with obesity,and brief (8- to 15-week) exercise interventions have been shown to improve glycemic status,²⁸ although longer-term studies are lacking.

Puberty. Insulin resistance transiently increases during puberty Tanner Stages 2–4. Children develop a ∼ 30% increased insulin resistance during puberty. Euglycemic clamp studies of 357 normal children(159 girls) aged 10–14 years showed that insulin resistance increased immediately at the onset of puberty (Stage 2), peaked at Tanner Stage 3, and returned to near-prepubertal levels by the end of puberty (Stage 5). For the BMI categories < 27 kg/m², girls were more insulin resistant than boys at all stages, which is partially explained by differences in adiposity. This difference disappeared at higher BMI categories.¹⁷ 

It is believed that the insulin resistance of puberty is explained by higher levels of growth hormone and IGF-1 that occur during puberty. The insulin resistance correlates with mean 24-hour serum growth hormone levels and serum IGF-1 levels,^29–31 and the peak insulin resistance is temporally related to the pubertal growth spurt in each sex.¹⁷ 

Smoking and sleep apnea disorders. Smoking is positively related to central obesity and to the development of type 2 diabetes, particularly in men with normal BMI, for whom heavy smoking may double the risk of developing diabetes.²⁵  Smokers have increased insulin levels and insulin resistance compared to nonsmokers,and cessation of smoking results in improvements in insulin sensitivity and HDL cholesterol levels.^32–34 

Obesity leads to a decrease in residual lung volume and ventilatory capacity because of increased abdominal pressure on the diaphragm and effects of visceral fat. Sleep apnea and increased snoring may be caused by increased pharyngeal fat.¹ Excess leptin levels may also play a role.²  However,obstructive sleep apnea syndrome is also an independent risk factor for insulin resistance. Continuous positive airway pressure therapy produces a prompt improvement in insulin sensitivity.³⁵ 

Acanthosis nigricans and skin tags. Acanthosis nigricans is a thickening of the stratum corneum that becomes pigmented in a racially dependent manner. Insulin and IGF-1 receptors are present and respond to high levels of insulin, which promote keratinocyte proliferation. Acanthosis nigricans or skin tags are useful markers of the presence of hyperinsulinism. Its severity is proportional to the degree of insulin responsiveness in the midst of insulin resistance. Its appearance precedes the development of hyperglycemia and pre-diabetes. Its severity may lessen with weight loss and improvement of hyperinsulinism.² 

Metabolic syndrome. The metabolic (insulin resistance) syndrome represents a link of insulin resistance with hypertension, dyslipidemia, type 2 diabetes, and other metabolic abnormalities. Together these result in an increased risk for atherosclerotic CVD.

The most common cause of insulin resistance in the pediatric age group is obesity. Evaluation of a large, multiethnic, multiracial cohort of children and adolescents showed the prevalence of the metabolic syndrome increased with severity of obesity and reached 50% in severely obese youngsters. In turn,after adjustment for race or ethnic group and degree of obesity, the prevalence of the metabolic syndrome increased directly with insulin resistance (P for trend < 0.001). In affected youngsters,biomarkers for an increased risk of adverse cardiovascular outcomes were also already present, with CRP levels increased and adiponectin levels decreased with increasing obesity.³⁶ 

The linkage of metabolic syndrome with insulin resistance is briefly explained as follows. Insulin resistance is defined as the impairment of insulin at usual concentrations to adequately promote peripheral glucose disposal, suppress hepatic glucose production, and inhibit very low density lipoprotein output. It is diagnosed by finding hyperinsulinism at fasting or on OGTT (Table 4) or by determination of insulin sensitivity on IVGTT. A diminished first-phase insulin response on IVGTT with high proinsulin levels is strongly predictive for the future development of type 2 diabetes.² 

The pathogenesis of insulin resistance is multifactorial and polygenic and has a strong familial association and racial and ethnic propensity as described above. Small-for-gestational-age birth weight and excessive weight gain during childhood are important risk factors. Environmental and behavioral factors that promote obesity and insulin resistance increase expression of the metabolic syndrome.² Obesity further causes a proinflammatory and prothrombotic state that potentiates atherosclerosis. Fat-derived cytokines, (e.g., TNF-β and adiponectin) may serve as pathogenic contributors or protective factors. Lipid accumulation in certain organs and increased free fatty acids are mediators of insulin resistance.³⁷ Hypertension and dyslipidemia (high triglyceride and low HDL cholesterol levels) are seen in the metabolic syndrome, and their molecular pathogenicities are reviewed.^2,37 

The majority of patients with insulin resistance will not develop diabetes,their genetic background influencing the adequacy of β-cell compensatory insulin secretion. Nevertheless, progression of insulin resistance to pre-diabetes and diabetes and to metabolic syndrome, and progression of atherosclerotic disease to coronary heart disease and stroke are seen with increasing frequency from early puberty onward.² 

Polycystic ovary syndrome. Women with polycystic ovary syndrome(PCOS) are insulin resistant, have insulin secretory defects, and are at high risk for pre-diabetes and type 2 diabetes. A study of 254 women aged 14–44 years with PCOS found the overall prevalence of pre-diabetes was 31.1% and of diabetes was 7.5%. Nonobese women with PCOS had prevalence rates of 10.3 and 1.5%, respectively. The prevalence of glucose intolerance was significantly higher in women with PCOS than in control subjects. Variables most associated with 2-hour glucose levels on OGTT were FPG levels (P< 0.0001), PCOS status (P = 0.002), WHR (P = 0.01), and BMI (P = 0.021). In summary, women with PCOS have both insulin resistance and abnormal β-cell function and are at significantly increased risk for pre-diabetes and type 2 diabetes at all weights and at a young age.³⁸ 

The development of type 2 diabetes is thought to occur in this sequence: 1) onset of insulin resistance (impaired insulin sensitivity) with compensatory insulin hypersecretion, 2) development of pre-diabetes because of impaired pancreatic β-cell secretion of insulin, and 3) progression to diabetes because of further impairment ofβ-cell insulin secretion and insulin sensitivity with increased hepatic glucose production.^3,5,39–42 

In adults, the development of type 2 diabetes is preceded by the presence of pre-diabetes.^{26,41,43–45} Pre-diabetes is an intermediate stage defined by an elevated FPG (IFG) or elevated 2-hour glucose on OGTT (IGT) as outlined in Table 5. As with adults, it is likely that pediatric patients with type 2 diabetes all have pre-diabetes as their preceding condition.^42,46 

A study of a multi-ethnic cohort of 167 severely obese children (aged 4–10 years) and adolescents (aged 11–18 years), a group at high risk for the eventual development of type 2 diabetes, found pre-diabetes to be already present in 25% of the children and 21% of the adolescents. Their insulin and C-peptide levels were markedly elevated both at fasting and in response to the OGTT. In addition, the children with normal glucose tolerance(without pre-diabetes) as a group already showed elevated levels (mean± SE) of fasting insulin of 20 ± 5 mU/l (normal range 0–13) and C-peptide 2.6 ± 0.3 ng/ml (normal range 0.4–2.2). After controlling for BMI, insulin resistance (estimated by homeostatic model assessment) was greater in the pre-diabetes cohort and was the best predictor of pre-diabetes. Silent type 2 diabetes, interestingly, was identified in 4 of the 112 adolescents. Unlike those with pre-diabetes, their insulin and C-peptide levels were not elevated in response to the OGTT, and theirβ-cell function (estimated from the insulinogenic index [the ratio between the changes in the insulin level and the glucose level during the first 30 minutes after the ingestion of glucose]) was reduced. In summary,insulin resistance and pre-diabetes are highly prevalent among severely obese children and adolescents irrespective of ethnic group, and pre-diabetes is associated with insulin resistance, while β-cell function is still relatively preserved. Overt type 2 diabetes is linked to β-cell failure.¹⁰ 

In another study, an evaluation of insulin sensitivity, pancreaticβ-cell function, and the balance between the two was done in 14 adolescents with type 2 diabetes and 20 obese control subjects. In the adolescents with diabetes, insulin sensitivity was lower, fasting insulin was higher, and fasting glucose rate of appearance was higher. More significantly,both the first- and second-phase insulin secretions were lower, and GDI was∼ 86% lower. A1C correlated with first-phase insulin secretion (r= –0.61, P = 0.025) with no relationship to insulin sensitivity. While both insulin sensitivity and β-cell function were impaired in adolescents with type 2 diabetes, the magnitude of the derangement was greater in β-cell function than in insulin sensitivity compared to that in obese control subjects. The inverse relationship between β-cell function and A1C may either reflect the impact of deteriorating β-cell function on glycemic control or be a manifestation of a glucotoxic phenomenon on β-cell function. Understanding the natural course of β-cell failure may lead to treatment to prevent it.⁴² 

Patients with type 2 diabetes come to diagnosis through a range of presentations:

• Serendipitously, by casual finger-stick blood glucose testing at home,random urinalysis or blood screen on a school or sports physical, or during evaluation of an unrelated problem that led to urine or blood testing.

• After development of vaginal candidiasis, urinary tract infection,furunculosis, or abscess that may raise suspicion for diabetes.

• Based on presence of one or more risk factors(Table 4) that instigates a work-up leading to discovery of diabetes.

• After development of the classic symptoms of polyuria, polydipsia, and nocturnal enuresis or of acute complications of dehydration, hyperosmolar syndrome, severe diabetic ketoacidosis, or neurological change that require hospitalization.

The fourth category above is no longer considered surprising as the initial presentation of a patient with type 2 diabetes. In the past, essentially all patients presenting with classic diabetic ketoacidosis or acute severe complications were assumed to have type 1 diabetes; we now know this is not true. Our review of all patients admitted to the Le Bonheur Children's Medical Center ICU during the period 2000–2003 with severely ill new-onset diabetes (n = 79) revealed that 17 (22%) had type 2 diabetes (R.K.D.,unpublished observations). This compares with a ∼ 34% prevalence of type 2 diabetes among all cases of diabetes diagnosed during this same period. Complications of severe dehydration, mental status changes, rhabdomyolysis,and deep vein thrombosis were seen in some of these ICU patients with type 2 diabetes. It is therefore important to recognize early when a patient has developed diabetes to prevent its further progression to acute complications that may lead to morbidity or mortality and to initiate treatment to reverse its clinical course.

Fasting glucose versus 2-hour glucose. Detecting the presence of pre-diabetes requires laboratory testing. The tests accepted as being diagnostic of pre-diabetes are fasting plasma glucose (FPG) and 2-hour glucose on the OGTT (Table 5). FPG(performed after no caloric intake ≥ 8 hours) has been recommended over the OGTT, which is more labor-intensive and costly to perform.⁴⁷  The FPG,unfortunately, has proven to be far less sensitive than the 2-hour glucose on the OGTT, with resultant failure in diagnosing patients that have pre-diabetes or diabetes.^25,48 

Data on healthy individuals who underwent an OGTT between 1986 and 2002,when the lower limit for IFG was 110 mg/dl instead of the current 100 mg/dl,were analyzed to determine the relationships between FPG and 2-hour OGTT results over the normal range of A1C levels. Of 404 individuals with normal A1C levels (< 6.0%), 60% had normal glucose tolerance, 33% had IGT, 1% had isolated impaired FPG (IFG), and 6% had type 2 diabetes. Of 161 individuals with IGT, 80% had normal FPG levels. Both FPG and 2-hour OGTT levels increased as A1C increased and were significantly correlated (r = 0.63, P < 0.001), but the 2-hour OGTT level increased at a rate four times greater than the FPG and accounted for a greater proportion of A1C. Most individuals with A1C values between 6.0 and 7.0% had normal FPG levels but abnormal 2-hour OGTT levels.⁴⁸ 

In the cohort of 167 severely obese children and adolescents described above, FBG levels were similar between normal children and those with pre-diabetes. In adolescents, FBG levels averaged 82 mg/dl in normal and 90 mg/dl in pre-diabetic subjects and only 118 mg/dl in the four subjects with type 2 diabetes, well below the diagnostic levels of 100 and 126 mg/dl,respectively. The 2-hour OGTT levels, in contrast, were higher than normal and diagnostic in the children and adolescents with pre-diabetes, and they were highest in those with diabetes (P < 0.001 for both comparisons). In this study, the 2-hour OGTT result additionally showed excellent reproducibility on OGTT repeated 3 months later in a subset of 10 subjects.¹⁰ 

In summary, fasting hyperglycemia is an indicator of a more advanced stage of pre-diabetes or diabetes but is a very insensitive method for detecting pre-diabetes or diabetes. In contrast, the 2-hour OGTT result is a reliable and more sensitive tool for the early detection and diagnosis of pre-diabetes and diabetes, at least in the obese.¹⁰  This finding is innately commonsensical, as a defect in a system is far more likely to be uncovered when a load is imposed than under a basal, unloaded state.

Other predictors of disease progression. To identify youth at highest risk for developing diabetes and the factors that have the strongest impact on glucose tolerance, a group of 117 obese children and adolescents had OGTT performed at baseline and after ∼ 2 years. At initial testing, 84 subjects had normal glucose tolerance, and 33 had IGT based on the 2-hour OGTT result. Eight subjects, all of whom had IGT, developed diabetes, whereas 15 subjects with IGT reverted to normal glucose tolerance. In this cohort, severe obesity, IGT, and African-American ethnicity emerged as the best predictors of developing diabetes, whereas FBG, insulin, and C-peptide levels were nonpredictive. Changes in insulin sensitivity, strongly related to weight increase (BMI > 40 kg/m²), had a significant impact on the 2-hour OGTT level on the follow-up study. Severely obese children and adolescents with IGT, particularly those of African-American descent, are at very high risk for developing diabetes over a short period of time. Parameters derived from an OGTT, but not fasting samples, can serve as predictors of changes in glucose tolerance.⁴⁶ 

There are suggestions that other markers of glycemic status be considered to screen or diagnose at-risk patients for pre-diabetes or diabetes. FPG,2-hour OGTT result, and A1C are each useful as markers of glycemic status and to predict complications of diabetes, and their clinical attributes are reviewed.⁴⁹ 

It has been proposed that A1C be used as a screening or diagnostic test. It correlates well with FPG, mean postprandial glucose levels, and 2-hour OGTT results and is predictive of future diabetes and of diabetes-related complications. It is a very reproducible measurement, having an intraindividual coefficient of variation of 1.9–4.2% (compared to 6.4–11.4% for FPG and 14.3–16.7% for 2-hour OGTT). A random plasma glucose ≥ 200 mg/dl and an A1C level > 2 SD above the laboratory mean may be sufficient to diagnose diabetes. If only one test is positive, then measurement of FPG is used to determine whether a patient has pre-diabetes or diabetes.⁴⁹  A meta-analysis of 10 studies that employed this method showed it to have a sensitivity of 66% and specificity of 98%.⁵⁰ 

Alternatively, in a patient found initially to have an elevated FPG, it has been proposed that A1C be measured as the second test rather than a repeat measurement of FPG.⁵¹  In the clinic setting, the addition of A1C with a random glucose or FPG may offer a useful screen for pre-diabetes or type 2 diabetes in high-risk patients and may help improve their clinical management.

Finally, another finding¹⁰  was that fasting proinsulin levels were nearly twice as high in children and adolescents with pre-diabetes than normal and 2.5 times higher in adolescents with type 2 diabetes (P = 0.002 for both comparisons). On the other hand, the proinsulin-to-insulin ratio, a marker of β-cell (dys)function in adults, was not significantly different in children and adolescents with pre-diabetes or diabetes because of an accompanying higher insulin secretion. In this study, the best predictors of pre-diabetes in obese children and adolescents were insulin resistance with hyperinsulinemia, and fasting hyperproinsulinemia.¹⁰ 

B.B. was diagnosed by OGTT as having insulin resistance with pre-diabetes,based on the elevated insulin and C-peptide levels at fasting and at 2 hours and his elevated 2-hour OGTT result (Table 2). The elevated fasting proinsulin level was also predictive of possible future development of diabetes. His elevated BMI (> 40 kg/m²) and A1C ≥ 6.0% (Table 1) placed him at high risk for the development of diabetes.

In light of this, B.B. was educated about the significance of polyuria,polydipsia, and other diabetes symptoms. However, he did not remember their importance at the time of their occurrence. Thus, critical aspects of patients' education require greater reinforcement.

B.B. was asked to monitor his FBG and did so, providing excellent glucose logs at his clinic visits. His FBG levels, however, averaged only 104 mg/dl(range 92–114) and lacked the rise suggestive of progression to diabetes, indicating their inadequacy. Monitoring postprandial glucose levels might have offered a more sensitive screen, but this had not been requested. The marked weight loss that he suffered before his acute presentation to the hospital emergency department was not yet evident at his fourth clinic visit.

Suggestions for how to improve the screening of patients for possible progression from pre-diabetes to diabetes are offered below and in Table 5. A summary of suggestions for screening patients for progression from normal glucose tolerance to pre-diabetes and to type 2 diabetes is provided in Table 6. These suggestions are based on our experience and a review of the recent literature but remain to be proven. They are offered with our view that the American Diabetes Association(ADA) recommendation to use the FPG (and abandon the use of the 2-hour OGTT result except in pregnant women) to screen and diagnose patients for the development of type 2 diabetes is flawed because the FPG suffers from poor sensitivity. Like the ADA, we make efforts to keep costs down. The screening process suggested below is low-cost and expected to limit the use of the formal OGTT to patients identified as likely to have diabetes.

At a minimum, these patients should have insulin resistance(Table 4) and pre-diabetes(Table 5). Factors placing patients at very high risk for progression to diabetes include a BMI > 40 kg/m² and an A1C ≥ 6.0%. Whether an elevated fasting proinsulin level has additional usefulness in predicting very high risk for progression to diabetes is not known and is presently not recommended.

First and most important, patients should screen themselves at home with a 1-hour postprandial blood glucose measurement once every 2 weeks or twice monthly using individually wrapped glucose test strips. A bottle of 25 strips will last 1 year, making this a simple, low-cost effort. Alternate methods,such as screening FBG or urine for ≥ 1+ glucosuria, are too insensitive and not recommended.

Second, institute a strong education program regarding the recognition of symptoms of diabetes decompensation, particularly polyuria, nocturia/enuresis,polydipsia, polyphagia, acute unintended weight loss, excessive fatigue, and acute nausea and vomiting. Provide the criteria for when patients should make emergency phone contact if any of the above factors become positive.

First, document whether patients have developed symptoms of diabetes decompensation. Second, evaluate for acute or excessive weight loss (a prompt for diabetes decompensation) and, conversely, for a large increase in weight(a cue that insulin sensitivity is deteriorating). Third and most important,introduce a modified OGTT to measure the 2-hour OGTT level.

A modified OGTT involves giving the standard dose of glucose (1.75 g glucose/kg, maximum 75 g, by mouth) on a patient's arrival at the clinic (or earlier at home). The patient need not be fasting because this is a screen and not a formal OGTT used for diagnosis. Only the 2-hour glucose result is obtained, thereby minimizing technician workload and cost. Consider obtaining a reflex measurement of the 2-hour levels of insulin and C-peptide if the 2-hour glucose is ≥ 200 mg/dl to determine whether the patient has type 1 or type 2 diabetes.

If the modified OGTT is not performed, then use a casual glucose ≥ 200 mg/dl to screen for progression to diabetes. A limitation of this method would be if the patient has not eaten or has taken a negligible carbohydrate load before the screening.

Measurement of A1C is a useful adjunctive test. A sharp rise strongly suggests progression to diabetes. An A1C ≥ 6.0% coupled with a high 2-hour OGTT result or a high random glucose or high FBG strongly supports the likelihood that a patient has developed diabetes. Measurement of FBG as the initial screen is too insensitive and not recommended, although it may have usefulness as a second screen.

Recognize that the above evaluations and tests are used to screen patients for the possibility that they have developed diabetes, but not to diagnose them as having diabetes. A definitive diagnosis requires either that a formal OGTT be performed or that a patient has presented with symptoms of diabetes and a laboratory-measured glucose level ≥ 200 mg/dl(Table 5).

This article presented a case in which limitations in screening and inpatient education methods delayed the recognition that a patient had progressed from pre-diabetes to type 2 diabetes. The delay in diagnosis resulted in the patient experiencing severe acute complications of diabetes that required ICU management. It also delayed implementation of more aggressive therapies and efforts at lifestyle change to reverse the clinical course of his disease. To avoid such problems, clinicians should be aware of the advantages and disadvantages of various screening tests and may wish to implement the low-cost methods outlined above to identify and treat at earlier stages patients who are progressing from pre-diabetes to diabetes.