Characteristics of Youth-onset Noninsulin-dependent Diabetes Mellitus and Insulin-dependent Diabetes Mellitus at Diagnosis

DISCUSSION

The presentation of patients with NIDDM and IDDM in this study were similar. Classic signs and symptoms of diabetes such as polyuria, polydipsia, and polyphagia, are observed in all forms of diabetes and, therefore, unhelpful in discerning the etiology of diabetes. These symptoms and the others assessed including abdominal pain, headache, nocturia, and visual disturbance were shown to be of no clinical or statistical utility in differentiating between IDDM and NIDDM. Only weight loss demonstrated a statistically significant difference between NIDDM and IDDM patients. Weight loss was almost twice as likely to be reported by patients with IDDM. A history of weight loss may be a clinically useful symptom for distinguishing between these two groups, particularly when trying to assess obese youth with diabetes.

Physical characteristics, such as BMI, revealed significant differences between the two groups. NIDDM youths were markedly obese (96% BMI, >85th percentile) but not all patients in the IDDM group were of normal body weight. Twenty-four percent of the IDDM patients had a BMI ≥85th percentile for age. This percentage is consistent with the estimated 21% incidence of obesity in American youths aged 12 to 19 years²⁷ and the estimated 22% for children and adolescents.²⁸ Perhaps as significant as the severity of obesity in the youths with NIDDM was the finding that 86% of the youths were noted at the time of diagnosis to have acanthosis nigricans. Acanthosis nigricans has been linked to hyperinsulinemia and obesity in a randomized screening of schoolchildren²⁹ and Native-American children.³⁰ When insulin concentrations were measured, many of these obese individuals with acanthosis were also found to be hyperinsulinemic. It is interesting that obese patients, regardless of final diagnosis, were 4.6 times more likely to have had some measure of endogenous insulin production than nonobese patients. It is appropriate clinical management that more obese patients had endogenous insulin production assessed. Consequently fewer youths with IDDM had these studies obtained. Because of this collection bias, it is not possible to directly compare the endogenous insulin production between the two study groups.

The occurrence of hyperinsulinism and insulin resistance has been demonstrated to vary among ethnic populations at high risk for NIDDM, including youths of African-American,³¹Hispanic,³² Pima Indian,³³ and Japanese descent.³⁴ Recently, Arslanian and Suprasongsin³¹ studied a group of healthy adolescents with no family history of diabetes in first degree relatives and found an increase in insulin production and a decrease in insulin sensitivity in African-American participants. It is likely that diminished insulin sensitivity and increased insulin production play a role in the increased incidence of obesity and NIDDM in various ethnic populations. This likely occurs earlier than previously suspected, long before these problems become clinically apparent. It may also explain the childhood/adolescent presentation of obesity and NIDDM in our African-American population.

Another striking finding was the large percentage of NIDDM youths with hypertension at diagnosis. More than one third of these patients were already hypertensive, eight times as frequent as youths with IDDM, probably secondary to their obesity rather than as a complication from their diabetes. Youth-onset hypertension associated with obesity has been demonstrated in several national health surveys.³⁵

The gender and ethnic characteristics of our patient population are very similar to the patients reported from the population-based study conducted in Cincinnati.¹⁸ This finding is of particular interest as the study performed in Cincinnati was a large, population-based epidemiologic study designed to determine the incidence of NIDDM in the youths of that area. The population base for this study's referral center is a small urban area and a large rural area. Referral bias cannot be eliminated in this study based on the available data. Despite the differences in the studies population bases, demographics were similar. Female:male ratio between the study populations were almost identical and ethnic representations were similar. African-Americans were overrepresented in the youths with NIDDM from both study populations, whereas IDDM ethnic distribution in Arkansas mirrored the state's ethnic distribution. The youngest patient with NIDDM reported in our study was 8 years old, and 10 years old in the study by Pinas-Hamiel.¹⁸ Children as young as 5 years old have been reported among the Pima Indians³⁶ who have the highest reported incidence of NIDDM of any population. They are also considerably more obese than the general population.³⁶ This association between obesity and NIDDM holds true for all ages of Pima Indians presenting with diabetes.^36,37

Of the characteristics assessed, history of weight loss, BMI, acanthosis nigricans, and insulin or C-peptide concentration, seem to be the most useful in differentiating between types of diabetes. Interestingly, HbA1C was not useful in distinguishing between types of diabetes as the HbA1C at presentation was the same for both groups. Clearly, the majority of obese youths with diabetes had NIDDM which was supported by the presence of acanthosis nigricans and elevated measures of endogenous insulin production. NIDDM should be considered in the differential diagnosis of youths with new onset diabetes, especially in obese African-American youths with acanthosis nigricans and no history of weight loss. Perhaps with further study a history of no weight loss, which in this study was the only symptom statistically different between the two groups, can be shown to be a reliable predictor of NIDDM. Careful evaluation of these clinical and biochemical characteristics remains important to arrive at the proper diagnosis.

ICA and glutamic acid decarboxylase antibodies, markers of autoimmunity, have been linked to IDDM.^38,39 These markers have been detected before onset of clinical diabetes and may remain detectable for years after diagnosis.³⁸ These studies typically take 1 to 2 weeks before results are available, and thus are not useful for immediate decision making at the initial visit. They can confirm the etiology initially suspected, and guide long-term therapy. These markers also have implications for family members of the patient with diabetes.³⁸ Family members of patients with positive ICA and/or glutamic acid decarboxylase antibodies and clinical diabetes are at higher risk for developing IDDM.⁴⁰ Screening of these family members and preventive therapy is currently under investigation by the National Diabetes Prevention Trial.^38,41 Because of the implications for family members, the importance of making the correct diagnosis of a youth with diabetes should not be overlooked.

Initial therapy should be directed by the metabolic derangements at presentation. Patients with DKA require prompt institution of standard DKA therapy, including intravenous insulin, regardless of the suspected diagnosis. If a patient is subsequently found to have evidence of NIDDM, therapy can be tailored appropriately after resolution of DKA, as was seen in the earlier case presentation. Treatment of youths with NIDDM is not well studied and is presently based on experience with adult-onset NIDDM and MODY patients. Oral hypoglycemic agents are effective treatments for both MODY and adult-onset NIDDM.^14,42 New therapies directed at blood glucose control in adults may also be effective in youth-onset NIDDM but have not been studied.^43-45

Obesity has become a nationwide concern affecting both the adult and pediatric population. Numerous studies have documented the steady increase in adult body weight in this and other countries over the past 30 years.⁴⁶ National health and nutrition surveys have documented an increase in childhood and adolescent obesity in the United States (US) similar to that seen for adults.^9,28,47-51 There was an increase in the prevalence of obesity from 15 to 21% in adolescents 12 to 19 years of age between the years 1980 and 1990.⁹ Not only are more youths becoming more obese, but the upper centiles of all sexes and races assessed are becoming even more obese.³⁵ In one survey, based on skinfold thickness, the prevalence of obesity in the pediatric population was documented to have increased by 17% and for a subset of this population, the superobese children, 306%.³⁵ Obesity in adults and children has been linked to a number of health concerns, including diabetes.^1,3,7-9,35 This association has been established for many ethnic groups, including African-Americans and Native Americans.^24,25,52

It is known that the risk of developing NIDDM rises exponentially with increasing body weight in adults.⁸ This effect was seen even in the nonobese range of women studied in the Nurse's Health Study. From lowest to highest BMI ranges (BMI, 22 to 35 kg/m²) the incidence of NIDDM rose from 13 to 1190 per 100 000 person years.⁸ Similar findings have been reported in men.⁷ The mean BMI of the youths with NIDDM reported in this study, 35 ± 1.1 kg/m², corresponds to the adult group with the highest risk of NIDDM and the most extreme obesity.⁸ The mean age of these youths with NIDDM was 13.9 years.

Diabetes has become a significant burden on the national health care system. Of the US population, 4.5% had some form of diabetes in 1992 and were responsible for 14.6% (one in seven health-care dollars) of the total US health care expenditure.⁵³ This represents approximately 105 billion dollars annually and does not take into account loss of productivity directly attributable to diabetes. Ninety percent of the estimated 15 million Americans with diabetes have NIDDM (only one half of which have been diagnosed), the remaining 10% have IDDM.¹² Each year 160 000 individuals or more die from complications of diabetes making it the sixth leading cause of death in the US.⁵⁴ Arkansas has recently been ranked as one of 10 states with the highest incidence of obesity in the US.⁵⁵An estimated 6.8 to 7.5% of the population of Arkansas has some form of diabetes compared with 4.5% for US population.^26,53-56It is possible that this increased incidence of diabetes in Arkansas is related to the increased incidence of obesity. These data do not take into account the increase in childhood and adolescent obesity, nor do they include the impact of youth-onset NIDDM.

It seems logical that youths with NIDDM will be at risk for all of the health concerns associated with adult-onset NIDDM such as retinopathy, nephropathy, neuropathy, and macrovascular complications of diabetes. There also are treatment-related complications to be considered, such as hypoglycemia, which can lead to coma, seizures, and even death.⁵⁷ In addition, because of their extreme obesity, these children are at risk for hypertension, psychological dysfunction, respiratory disease, and several orthopedic conditions.^35,49 Childhood and adolescent obesity closely correlate to adult obesity.⁵⁸ Therefore, these very young patients are likely to have a lifelong battle with not only diabetes but obesity as well. A child developing NIDDM at age 13 years who lives an average lifespan, will have more than 60 years duration of disease.⁵⁹ Complications of diabetes increase with duration of disease and may develop after only 5 to 10 years of diabetes.⁶⁰ It is likely that youths with NIDDM will have a major impact on the US health care system as they will have much longer duration of diabetes than their adult counterparts.

Much remains unstudied in these youths. Prospective screening of all youths with diabetes is underway at our facility incorporating all characteristics presented in this paper as well as complete family history, assessment of weight loss, fatigue, measure of endogenous insulin production, markers of autoimmunity, and therapy. Therapy with existing oral hypoglycemic and antihyperglycemic drugs will be assessed. Psychosocial factors and family dynamics affecting the patient/family adherence to therapy and long-term metabolic control are also being assessed.

NIDDM, once believed to be a disease of predominantly obese adults, is on the increase in the pediatric population.¹⁸Pinas-Hamiel, et al recently reported that the incidence of adolescent NIDDM in the Greater Cincinnati area increased ten-fold, from .7/100 000 per year in 1982 to 7.2/100 000 per year in 1994.¹⁸ This increased incidence in youth-onset NIDDM has corresponded temporally to an unprecedented increase in body weight and prevalence of obesity in the US from 1980 to 1990.⁹

The increase in the number of youths with NIDDM noted at our facility has continued in 1996 with more than 16 youths diagnosed in the first six months alone. At present, it is unclear if this trend represents a true increase in incidence of NIDDM in our region or reflects increased physician awareness of this problem. All youths presenting with diabetes need a thoughtful evaluation to arrive at the correct diagnosis. Broad assumptions regarding age at presentation should not be the sole diagnostic criteria in the pediatric population to define IDDM, nor should obesity alone be used to diagnose patients with NIDDM. However, lack of weight loss, the presence of obesity, and/or acanthosis nigricans should raise the suspicion that the patient may have NIDDM and may require further evaluation for endogenous insulin production and markers of autoimmunity. It is believed that improved diagnosis will lead to improved medical therapy and better understanding of this challenging medical problem.