- Email: [email protected]
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes Jenny T. Wang, PhD, Deborah J. Wiebe, PhD, MPH, and Perrin C. White, MD Objective To examine race/ethnicity and neighborhood income differences in longitudinal patterns of deterioration in hemoglobin A1c (HbA1c) values among youth (age 10 to 18 years) with type 1 diabetes.
Study design A sample of 225 youth (50.2% female), including 81 White, 81 Black, and 63 Hispanic youth with type 1 diabetes, was matched initially on age and sex. Neighborhood median family income was obtained through public census databases. Self-identified race/ethnicity and all HbA1c values (grand mean, 9.09% 2.02%) available in patients’ medical records between age 10 and 18 years were recorded and analyzed. Results Hierarchical linear modeling revealed age-related deterioration in HbA1c values that differed by race/ ethnicity and income. Controlling for income, White and Hispanic youth had similar HbA1c values at the start of adolescence (age 10) and demonstrated similar rates of deterioration across adolescence. Blacks had higher initial HbA1c values compared with Whites and Hispanics, but a similar rate of deterioration. Higher neighborhood income was associated with slower deterioration in HbA1c value among White teens, but not among Hispanic or Black teens. Conclusions Longitudinally, Black youth appear to experience disproportionate risks compared with White and Hispanic youth when income is statistically controlled. Neither Black nor Hispanic youth appear to benefit from living in higher-income neighborhoods. (J Pediatr 2011;159:571-6).
I
t is well documented that metabolic control deteriorates across the adolescent years in non-Hispanic White youth with type 1 diabetes,1,2 but to date no study has compared the longitudinal trajectories of hemoglobin A1c (HbA1c) values across White and minority youth. Cross-sectional studies suggest that Black and Hispanic youth have poorer metabolic control than Whites,3-5 and that such disparities extend into adulthood with grave consequences.6,7 Understanding when such disparities emerge may provide insight into the developmental processes involved and the critical developmental periods during which to implement interventions. We examined whether White, Black, and Hispanic youth entered adolescence with differing levels of metabolic control (suggesting differential risks beginning in childhood) and/or displayed differing rates of deterioration across adolescence (suggesting differential risks during the adolescent years). We hypothesized that racial/ethnic minority youth would display higher HbA1c values at the beginning of adolescence and different rates of deterioration in HbA1c values across the adolescent years.
Methods The Institutional Review Board of the University of Texas Southwestern Medical Center approved this study. Medical records at a large hospital-based pediatric endocrinology clinic at Children’s Medical Center Dallas were reviewed. Children’s Medical Center Dallas is the only academic pediatric healthcare facility in North Texas and has a catchment radius of approximately 150 miles north, east, and south, serving a diverse population. The vast majority of patients with type 1 diabetes within this catchment area are initially admitted to this hospital. The racial/ethnic breakdown for the entire hospital is 27% White, 19% Black, 45% Hispanic, and 9% Asian, other, or unknown. Insurance coverage of the endocrinology clinic population is 52% commercial/private insurance and 43% Medicaid/public insurance. In our sample (n = 225), 51.1% had commercial/private insurance coverage and 39.1% used public insurance, with the remaining classified as missing or unspecified. Our sample was fairly representative of the overall hospital as well as the endocrinology clinic. We oversampled minority youth to achieve our goal of understanding racial differences in metabolic control. A total of 737 self-identified White, Black, and Hispanic patients aged 14 to 17 years when seen in the clinic between January 1 and December 31, 2007, had From the Department of Psychiatry and Behavioral type 1 diabetes for >1 year, and had at least 3 clinic visits with an HbA1c record Sciences, Duke University Medical Center, Durham, NC (J.W.) and Departments of Psychiatry (J.W., D.W.); and within the last 5 years (ie, as necessary to analyze longitudinal trajectories) were Pediatrics (P.W.), University of Texas Southwestern Medical Center, Dallas, TX initially identified. Eighty-one Black youth met the inclusion criteria and were The authors declare no conflicts of interest.
HbA1c HLM M SES
Hemoglobin A1c Hierarchical linear modeling Grand mean Socioeconomic status
This manuscript is based on J.W.’s doctoral dissertation under the supervision of D.W. Portions of these data were presented at the annual convention of the Society of Behavioral Medicine, Seattle, WA, April 8, 2010. 0022-3476/$ - see front matter. Copyright ª 2011 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2011.03.053
571
THE JOURNAL OF PEDIATRICS
www.jpeds.com
matched to a sample of 81 White youth first on sex and then on age based on the closest proximity of birth date; all birth dates were matched within 2 months. Sixty-three Hispanic youth were identified but could not be completely matched on sex and age with Black and White counterparts, because fewer Hispanic patients met the inclusion criteria. This yielded a total sample of 225 youth for primary analyses. After the target sample was identified, age, time since diagnosis, and HbA1c values recorded at each routine clinic visit back to at least 10 years of age or as far as possible were obtained retrospectively from each participant’s medical records. Therefore, data for the full sample included HbA1c values for youth across ages 10 to 18. These repeated measures resulted in 4029 total HbA1c values across the full sample (1562 White, 1338 Black, and 1129 Hispanic). HbA1c values were measured at routine diabetes clinic visits using a point-of-care instrument (DCA2000 or DCA Vantage; Siemens, Deerfield, Illinois). The average number of clinic visits for all youth was 4.1 per year and was not significantly related to median family income or race/ethnicity. Type of insulin therapy was obtained both as a between-subjects variable (pump or no pump therapy at any point during the study period) and a within-subjects variable (pump or no pump at each clinic visit to capture changes in pump status over time). Because few youth changed pump status across time, only the categorical between-subjects variable was analyzed. Individually reported socioeconomic status (SES) could not be obtained given the retrospective review. Therefore, neighborhood SES data were collected from 2000 United States census tract data8 based on the subjects’ zip codes at the time of data collection in 2007. Fourteen participants had missing addresses, preventing us from deriving census tract–based income predictions (2 Black, 8 White, and 4 Hispanic); thus, the sample size was decreased to 211 when income was included in the analyses. Statistical Analysis The data were preliminarily evaluated for violations of statistical assumptions and for outliers. Primary analyses were conducted with hierarchical linear modeling (HLM) techniques.9 HLM analyzes systematic variance on 2 levels. The first level analyzes variables that account for within-subject changes across time (ie, within-subject age-related changes in HbA1c value), and the second level considers how these age-related trajectories differ as a function of individual differences, such as race/ethnicity or income. Age was centered at 10 years to capture the beginning of adolescence for all subjects. These analyses produced coefficients for the intercept, which provided information about average HbA1c value at age 10, and the linear slope, which provided information about the rate of change in HbA1c value across age. Finally, these analyses identified any significant variability in the intercepts or slopes, pointing toward individual differences that allowed us to examine differences across race/ethnicity and income. The HLM analyses included 4 models: (1) the basic model (model 1), in which we determined the basic associations 572
Vol. 159, No. 4 between age and HbA1c, the presence of individual differences in intercept and slope, and the need for covariates (sex, age at diagnosis, and pump status); (2) the race/ethnicity model (model 2), which examined whether race/ethnicity explained individual differences in intercept and slope; (3) the income model (model 3), in which income was included as a covariate to determine whether race effects were independent of income; and (4) the race-by-income interaction model (model 4) to evaluate whether these variables interacted to predict HbA1c value. In models 2, 3, and 4, race/ ethnicity variables were dummy coded to create two pairwise comparisons with White participants as the comparison group in each (White = 0, Black = 1 to test the effects for Black youth; White = 0, Hispanic = 1 to test the effects for Hispanic youth). In models that included income, median family income was centered at the grand mean (M) across all participants. As such, regardless of race/ethnicity, the effects of income were evaluated as a function of deviations from the M SD, $60 812 $28 662/year. Predicted means for significant interactions between race/ethnicity and income were computed by entering values for income that were 1 SD from this grand mean (low income, $32 150/ year; high income, $89 474).
Results Descriptive analyses indicated that White subjects were more likely to use insulin pump therapy, which was included as a covariate for primary analyses (Table I). There were no significant racial/ethnic differences in sex, age at diagnosis, age across all clinic visits, or number of available clinic visits across the data collection period; however, White youth were more likely to live in neighborhoods with higher median family income. We initially examined the linear relationship between HbA1c value and age at each clinic visit ignoring all other variables (model 1). The coefficients for both the intercept [b = 8.89; t(224) = 81.34; P < .001; model 1, Table II] and slope [b = .24; t(224) = 8.51; P < .001; model 1, Table II] were significant. The coefficients indicated that at age 10 years, all youth had an average HbA1c value of 8.89%, and that there was a 0.24% absolute increase in HbA1c value for every yearly increase in age. The significant variance components suggested variability in the intercept (P < .001) and slope (P < .001) not fully accounted for by model 1, which warranted the inclusion of other variables (eg, race/ethnicity, income) in the model (Table II). These results confirmed age-related deterioration in HbA1c values across adolescence in a racially/ethnically diverse sample. Sex, age at diagnosis, and pump status predicted variability in HbA1c across time and were thus entered as covariates in all subsequent analyses. The race/ethnicity model (model 2, Table II) evaluated differences between racial/ethnic groups in average HbA1c value at age 10 and the rate of change in HbA1c values across age 10 to 18. At age 10, Black youth had a 1.4% higher HbA1c value and Hispanic youth a 0.55% higher Wang, Wiebe, and White
ORIGINAL ARTICLES
October 2011
Table I. Descriptive demographic data by race/ethnicity Source
White
Black
Hispanic
N Female sex, % Age at diagnosis, years, M (SD) Average number of clinic visits per participant, M (SD) Average age across all clinic visits, years, M (SD) Average age at first and last clinic visit, years, M (SD) Pump therapy, %* Median family income, $*
81 50.6 9.76 (3.34) 19.28 (8.19) 13.30 (3.00) 18.5 76 289 (36 290)
81 50.6 10.16 (3.75) 16.49 (10.02) 13.34 (3.09) 3.7 55 081 (20 501)
63 49.2 10.13 (3.47) 17.92 (9.13) 13.54 (2.93) 7.9 49 337 (17 150)
Total range
11.09-17.01
*Differences between Whites and minority samples are significant at P < .001.
HbA1c value than White youth. However, rates of change in HbA1c across time were not significantly different among the 3 groups. This indicated that racial/ethnic minority youth had higher HbA1c values than their White counterparts at age 10 (intercept), but showed parallel trajectories in rates of change in HbA1c values over time (slope). When income was added as a covariate in the model (model 3, Table II), the intercept difference between White versus Hispanic teens was no longer significant (P > .10), whereas the difference between White and Black youth remained significant (P < .001). Specifically, Black youth had a 1.3% higher HbA1c value than White youth at age 10 years. All youth experienced similar rates of change in HbA1c values across adolescence regardless of racial/ethnic group. Thus, Black youth appeared to be at disproportionately greater risk independent of neighborhood income. It is notable that risks appeared to begin in childhood, resulting in poorer metabolic control before adolescence. In addition to these childhood risks, Black youth continued to accumulate
additional adolescent risks and evidenced similar rates of deterioration in HbA1c values as White and Hispanic teens across the adolescent years. However, White and Hispanic youth appeared to encounter similar levels of risk in childhood and adolescence when neighborhood income was statistically controlled. Race interacted with income to predict the rate of HbA1c deterioration with age. Specifically, higher neighborhood income was associated with smaller age-related HbA1c deteriorations in White youth, such that every $10 000 increase in income resulted in a slower deterioration in HbA1c value, at a rate of 0.02% per year (model 4, Table II). In contrast, higher neighborhood income was associated with more rapid deterioration in metabolic control in Black and Hispanic youth. Every $10 000 increase in income was associated with more rapid rates of deterioration in HbA1c value of 0.05% in Black youth and 0.08% in Hispanic youth. The Figure depicts pairwise comparisons of income interaction effects between high-income and low-income
Table II. Coefficients for each model from HLM Model 1 Intercept coefficients Intercept Sex Pump status Age at diagnosis Black race Hispanic ethnicity Income Black versus White income interaction Hispanic versus White income interaction Slope coefficients Slope Sex Pump status Age at diagnosis Black race Hispanic ethnicity Income Black versus White income interaction Hispanic versus White income interaction Variance components Intercept Slope
Model 2
Model 3
Model 4
b
P
b
P
b
P
b
P
8.89 -
<.001 -
7.53 .37 .61 .20 1.40 .55 -
<.001 .050 .045 <.001 <.001 .016 -
7.68 .37 .58 .20 1.28 .39 .000005 -
<.001 .048 .056 <.001 <.001 .126 .125 -
7.70 .40 .61 .20 1.31 .24 .000008 .000015 .000006
<.001 .029 .042 <.001 <.001 .321 .002 .092 .459
b
P
b
P
.24 -
<.001 -
.24 .08 .07 .02 .01 .02 -
.098 .154 .353 .045 .849 .748 -
Variance
P
Variance
P
2.42 .13
<.001 <.001
1.60 .13
<.001 <.001
b .23 .08 .07 .02 .02 .04 .000000 Variance 1.58 .13
P .237 .157 .438 .032 .736 .662 .673 P <.001 <.001
b .25 .08 .08 .02 .01 .07 .000002 .000005 .000008 Variance 1.56 .13
P .088 .136 .234 .047 .930 .397 .067 .029 .029 P <.001 <.001
*Sex was coded as 0 = female, 1 = male; pump status was coded as 0 = no pump, 1 = pump therapy; Black (B) race was coded as B = 1; White (W) race = 0 and Hispanic (H) ethnicity was coded as H = 1, W = 0; Black versus White income interaction was coded as B = 1, W and H = 0; and the Hispanic versus White income interaction was coded as H = 1, B and W = 0.
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes
573
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 159, No. 4
Black, Hispanic, and White youth. Further pairwise comparisons revealed that income associations with HbA1c value did not differ between Hispanic and Black youth when Hispanic status was dummy coded as the reference group [b = –.00; t(202) = –.652; P > .10]. This suggests that rising income was similarly detrimental to the rate of deterioration in metabolic control across Black and Hispanic youth.
Discussion Longitudinal patterns of deterioration in metabolic control across White, Black, and Hispanic youth with type 1 diabetes have not been studied previously. Our findings add to the literature documenting poorer metabolic control in minority youth1,3,10 in several important ways. First, our findings suggest that differences in metabolic control between White and Hispanic youth may be more related to neighborhood economic disparities than to racial/ethnic differences, whereas differences between White and Black youth may reflect variables not captured by neighborhood income disparities. Second, our findings suggest that there may be childhood risk factors that contribute to the poorer metabolic control before adolescence in Black youth, which need to be better understood and prevented; indeed, differences in metabolic control between White and Black youth appear to emerge shortly after diagnosis and are maintained for up to 5 years.11 Third, the finding of similar rates of HbA1c deterioration across the adolescent years in all 3 groups suggests that there might be similar barriers to effective diabetes management across all youth (eg, conflicting parent–child relationships, peer conformity, negotiation of autonomy in diabetes management),12-14 which may be useful targets for developing culturally sensitive interventions for minority youth. Finally, income was differentially related to metabolic control in all 3 groups. Although White youth tended to benefit from living in neighborhoods with higher income levels, Black and Hispanic youth actually experienced more rapid HbA1c deteriorations when they lived in higher-income neighborhoods. All of these findings occurred independent of pump status or number of clinic visits, suggesting that these differences are not obviously due to racial/ethnic differences in access to care. Racial/ethnic differences in income effects on metabolic control need to be replicated. Regardless, our results add to a growing set of findings suggesting that low-income environments are not always associated with poor health outcomes. These findings lend support to the ‘‘Hispanic paradox,’’ in which lower-income Hispanic individuals appear to have better health than would be expected.15 There may be protective factors associated with living in these lower-income neighborhoods; for example, ‘‘familism’’ or ‘‘allocentrism,’’ which reflects the tendency to place the welfare of the community or family above the needs of the individual, appear to be important in Hispanic communities16 and may be emphasized in communities with fewer monetary resources. Alternatively, factors related to an ‘‘ethnic density effect’’ may suggest that 574
Figure. High SES of $89 474/year (1 SD above M, $60 812) versus low SES of $32 150/year (1 SD below M) shows different age-related trajectories in HbA1c values among Black, Hispanic, and White adolescents. Higher SES is associated with more rapid deterioration in HbA1c values across the teenage years in Black and Hispanic youth compared with White youth.
minority children experience heightened risk with increases in family income as they relocate or live in areas in which they are more overtly racial minorities. Minority individuals living in neighborhoods that are more majority-race dense may experience poorer health due to greater exposure to racial tension, acculturative stressors, and discrimination and prejudice.17,18 Minority individuals living in mostly White neighborhoods also may perceive less social/community support and more social isolation, which may put them at greater risk for poorer health outcomes. Several methodological considerations should be taken into account when interpreting our findings. Our modest sample size in each race/ethnic group was countered somewhat by the increased power obtained with repeated measures of each individual across time (4029 total HbA1c measures in the entire sample). The data were collected and recorded during routine clinic visits for standard care at a single site, limiting the generalizability of our results to other medical centers or cities. Data were limited to information Wang, Wiebe, and White
October 2011 available in chart review; we were unable to consistently obtain information on such variables as pubertal status or timing of adoption of diabetes responsibilities, which might have influenced the trajectories and should be collected in future research. We cannot rule out the possibility that racial/ethnic differences in HbA1c values reflected artifacts of measurement such as might be caused by hemoglobinopathies; however, we believe that this was unlikely, because we are not aware of such cases in our sample and because hemoglobin S (HbS), the most common hemoglobin variant in the Black population, tends to decrease red cell survival and thus lead to underestimation of HbA1c. The HbA1c measurement method used in this study does not overestimate HbA1c in the presence of HbS.26 Neighborhood SES has been associated with trajectories of child health,20 which we used as a proxy for individual SES. Although this technique has often been used when individual SES is not available21,22 and is a recommended approach to overcoming the lack of socioeconomic data in health databases, it is not without limitations.23,24 In particular, because we collected each subject’s address at a single point in time, our data do not account for residential mobility of individuals to different geographical regions over several years. In addition, neighborhood estimates of individual income may overestimate or underestimate actual individual income, resulting in misclassification of income status.25 For example, Black individuals tend to live in neighborhoods of greater economic disadvantage compared with White counterparts of similar income levels.24 Future studies should prospectively collect both neighborhood and individual SES variables to determine their common and unique associations with metabolic control. Future longitudinal studies also should obtain measures of family structure to evaluate its impact on metabolic control in low-income families, given that single-parent households are more common in lowerincome households and are associated with poorer metabolic control.11,26 Although implications for interventions may be premature at this time, our data suggest the importance of early intervention, especially in minority populations. Based on intervention studies in White youth, parent–child relationships and peer influences may be potential areas of focus for such interventions with minority youth.12,13 If supported by future research, interventions that leverage resources unique to lower-SES minority communities or that address risks present in higher-SES minority communities may be useful. Clarifying the needs of minority families coping with chronic health conditions will aid the development of such interventions. This study represents a step toward a better understanding of how minority youth cope with diabetes management across the adolescent years. Childhood risk factors may be particularly important in Black youth, whereas adolescent risk factors appear to be equally important in Black, White, and Hispanic youth. Minority youth do not appear to benefit as much from living in more-affluent neighborhoods compared with White youth. Clarifying this finding may
ORIGINAL ARTICLES point to detrimental and beneficial neighborhood factors for intervention to reduce health disparities and improve the quality of life for diverse children, adolescents, and their families. Since the acceptance of this report, we have become aware that differences between Blacks and Whites in HbA1c measurements have been reported,19 but these racial differences in HbA1c measurements are smaller than those observed in the present study. n We thank the dissertation committee members, including Drs Sunita Stewart, Margaret Caughy, and Crista Wetherington at the University of Texas Southwestern Medical Center, Dallas, Texas for their support and guidance. Submitted for publication Oct 8, 2010; last revision received Feb 16, 2011; accepted Mar 28, 2011.
References 1. Anderson B, Ho J, Brackett J, Finkelstein D, Laffel L. Parental involvement in diabetes management tasks: relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulin-dependent diabetes mellitus. J Pediatr 1997; 130:257-65. 2. Weissberg-Benchell J, Glasgow AM, Tynan WD, Wirtz P, Turek J, Ward J. Adolescent diabetes management and mismanagement. Diabetes Care 1995;18:77-82. 3. Delamater AM, Albrecht DR, Postellon DC, Gutai JP. Racial differences in metabolic control of children and adolescents with type I diabetes mellitus. Diabetes Care 1991;14:20-5. 4. Delamater AM, Shaw KH, Applegate EB, Pratt IA, Eidson M, Lancelotta GX, et al. Risk for metabolic control problems in minority youth with diabetes. Diabetes Care 1999;22:700-5. 5. Gallegos-Macias AR, Macias SR, Kaufman E, Skipper B, Kalishman N. Relationship between glycemic control, ethnicity and socioeconomic status in Hispanic and white non-Hispanic youths with type 1 diabetes mellitus. Pediatr Diabetes 2003;4:19-23. 6. Summerson JH, Konen JC, Dignan MB. Race-related differences in metabolic control among adults with diabetes. South Med J 1992; 85:953-6. 7. Tull ES, Barinas E. A twofold excess mortality among black compared with white IDDM patients in Allegheny County, Pennsylvania. Pittsburgh DERI Mortality Study Group. Diabetes Care 1996;19:1344-7. 8. United States Federal Financial Institutions Examination Council. FFIEC Geocoding System based on data collected in 2007. Available from: http://www.ffiec.gov/geocode/default.htm. Accessed June 29, 2009. 9. Bryk AS, Raudenbush SW. Hierarchical linear models: applications and data analysis methods. Newbury Park, CA: Sage; 1992. 10. Auslander WF, Thompson S, Dreitzer D, White NH, Santiago JV. Disparity in glycemic control and adherence between African-American and Caucasian youths with diabetes: family and community contexts. Diabetes Care 1997;20:1569-75. 11. Frey MA, Templin T, Ellis D, Gutai J, Podolski CL. Predicting metabolic control in the first 5 yr after diagnosis for youths with type 1 diabetes: the role of ethnicity and family structure. Pediatr Diabetes 2007;8:220-7. 12. Palmer DL, Berg CA, Wiebe DJ, Beveridge RM, Korbel CD, Upchurch R, et al. The role of autonomy and pubertal status in understanding age differences in maternal involvement in diabetes responsibility across adolescence. J Pediatr Psychol 2004;29:35-46. 13. Hains AA, Berlin KS, Davies WH, Smothers MK, Sato AF, Alemzadeh R. Attributions of adolescents with type 1 diabetes related to performing diabetes care around friends and peers: the moderating role of friend support. J Pediatr Psychol 2007;32:561-70.
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes
575
THE JOURNAL OF PEDIATRICS
www.jpeds.com
14. Wiebe DJ, Berg CA, Korbel C, Palmer DL, Beveridge RM, Upchurch R, et al. Children’s appraisals of maternal involvement in coping with diabetes: enhancing our understanding of adherence, metabolic control, and quality of life across adolescence. J Pediatr Psychol 2005;30:167-78. 15. Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis 2001;11:496-518. 16. Gallo LC, Penedo FJ, Espinosa de los Monteros K, Arguelles W. Resiliency in the face of disadvantage: do Hispanic cultural characteristics protect health outcomes? J Pers 2009;77:1707-46. 17. Becares L, Nazroo J, Stafford M. The buffering effects of ethnic density on experienced racism and health. Health Place 2009;15:700-8. 18. Pickett KE, Wilkinson RG. People like us: ethnic group density effects on health. Ethn Health 2008;13:321-34. 19. Kamps JL, Hempe JM, Chalew SA. Racial disparity in A1C independent of mean blood glucose in children with type 1 diabetes. Diabetes Care 2010;33:1025-7. 20. McGrath JJ, Matthews KA, Brady SS. Individual versus neighborhood socioeconomic status and race as predictors of adolescent ambulatory blood pressure and heart rate. Soc Sci Med 2006;63:1442-53.
Vol. 159, No. 4 21. Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health 1992;82:703-10. 22. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Race/ ethnicity, gender, and monitoring socioeconomic gradients in health: a comparison of area-based socioeconomic measures. The Public Health Disparities Geocoding Project. Am J Public Health 2003;93:1655-71. 23. Braveman PA, Cubbin C, Egerter S, Chideya S, Marchi KS, Metzler M, et al. Socioeconomic status in health research: one size does not fit all. JAMA 2005;294:2879-88. 24. Diez-Roux AV, Kiefe CI, Jacobs DR, Jr., Haan M, Jackson SA, Nieto FJ, et al. Area characteristics and individual-level socioeconomic position indicators in three population-based epidemiologic studies. Ann Epidemiol 2001;11:395-405. 25. Hyndman JC, Holman CD, Hockey RL, Donovan RJ, Corti B, Rivera J. Misclassification of social disadvantage based on geographical areas: comparision of postcode and collector’s district analyses. Int J Epidemiol 1995;24:165-76. 26. Thompson SJ, Auslander WF, White NH. Comparison of single-mother and two-parent families on metabolic control of children with diabetes. Diabetes Care 2001;24:234-8.
50 Years Ago in THE JOURNAL OF PEDIATRICS Anemia Associated with Protein Deficiency: A Study of Two Cases with Cystic Fibrosis Shahidi NT, Diamond LK, Shwachman H. J Pediatr 1961;59:533-42.
t was as late as 1938 when cystic fibrosis (CF) was first recognized as a separate entity from celiac disease.1 For decades after this milestone, the intestinal manifestations continued to be dominant. Fifty years ago Shahidi and Diamond, with the giant of exocrine pancreas research Harry Shwachman, published a report in The Journal describing two infants who were admitted for diagnosis and treatment of pitting edema (albumin 1.50 and 1.64 gram percent) who were found to be profoundly anemic (hemoglobin 7.2 and 7.7 gram percent).2 Shwachman had introduced the sweat test 5 years earlier, so these infants received a quick diagnosis, but the article also addresses the extensive evaluation of these infants, including bone marrow examination and radio-labeled albumin studies. Fecal fat studies were not mentioned, and we now know, on the back of pioneering studies such as this one, that profound maldigestion often aggravated by inadequate caloric intake causes this clinical picture. Currently, the triad of hypoalbuminemia, edema, and anemia should strongly suggest the diagnosis of CF, and a sweat chloride concentration should be performed. A note of caution is indicated, however, because edema may cause a false-negative sweat test result, and the test should then be repeated when the child is in a better nutritional state. As we move into an era of neonatal screening for CF,3 this clinical picture will become rarer because infants are diagnosed before clinical symptoms develop.
I
Michael Wilschanski, MBBS Pediatric Gastroenterology Hadassah University Hospitals Jerusalem, Israel 10.1016/j.jpeds.2011.05.016
References 1. Anderson DH. Cystic fibrosis of the pancreas and its relationship to celiac disease. Am J Dis Child 1938;56:344-99. 2. Shahidi NT, Diamond LK, Shwachman H. Anemia associated with protein deficiency: a study of two cases with cystic fibrosis. J Pediatr 1961;53342. 3. Castellani C, Massie J. Emerging issues in cystic fibrosis newborn screening. Curr Opin Pulm Med 2010;16:584-90.
576
Wang, Wiebe, and White
Study design A sample of 225 youth (50.2% female), including 81 White, 81 Black, and 63 Hispanic youth with type 1 diabetes, was matched initially on age and sex. Neighborhood median family income was obtained through public census databases. Self-identified race/ethnicity and all HbA1c values (grand mean, 9.09% 2.02%) available in patients’ medical records between age 10 and 18 years were recorded and analyzed. Results Hierarchical linear modeling revealed age-related deterioration in HbA1c values that differed by race/ ethnicity and income. Controlling for income, White and Hispanic youth had similar HbA1c values at the start of adolescence (age 10) and demonstrated similar rates of deterioration across adolescence. Blacks had higher initial HbA1c values compared with Whites and Hispanics, but a similar rate of deterioration. Higher neighborhood income was associated with slower deterioration in HbA1c value among White teens, but not among Hispanic or Black teens. Conclusions Longitudinally, Black youth appear to experience disproportionate risks compared with White and Hispanic youth when income is statistically controlled. Neither Black nor Hispanic youth appear to benefit from living in higher-income neighborhoods. (J Pediatr 2011;159:571-6).
I
t is well documented that metabolic control deteriorates across the adolescent years in non-Hispanic White youth with type 1 diabetes,1,2 but to date no study has compared the longitudinal trajectories of hemoglobin A1c (HbA1c) values across White and minority youth. Cross-sectional studies suggest that Black and Hispanic youth have poorer metabolic control than Whites,3-5 and that such disparities extend into adulthood with grave consequences.6,7 Understanding when such disparities emerge may provide insight into the developmental processes involved and the critical developmental periods during which to implement interventions. We examined whether White, Black, and Hispanic youth entered adolescence with differing levels of metabolic control (suggesting differential risks beginning in childhood) and/or displayed differing rates of deterioration across adolescence (suggesting differential risks during the adolescent years). We hypothesized that racial/ethnic minority youth would display higher HbA1c values at the beginning of adolescence and different rates of deterioration in HbA1c values across the adolescent years.
Methods The Institutional Review Board of the University of Texas Southwestern Medical Center approved this study. Medical records at a large hospital-based pediatric endocrinology clinic at Children’s Medical Center Dallas were reviewed. Children’s Medical Center Dallas is the only academic pediatric healthcare facility in North Texas and has a catchment radius of approximately 150 miles north, east, and south, serving a diverse population. The vast majority of patients with type 1 diabetes within this catchment area are initially admitted to this hospital. The racial/ethnic breakdown for the entire hospital is 27% White, 19% Black, 45% Hispanic, and 9% Asian, other, or unknown. Insurance coverage of the endocrinology clinic population is 52% commercial/private insurance and 43% Medicaid/public insurance. In our sample (n = 225), 51.1% had commercial/private insurance coverage and 39.1% used public insurance, with the remaining classified as missing or unspecified. Our sample was fairly representative of the overall hospital as well as the endocrinology clinic. We oversampled minority youth to achieve our goal of understanding racial differences in metabolic control. A total of 737 self-identified White, Black, and Hispanic patients aged 14 to 17 years when seen in the clinic between January 1 and December 31, 2007, had From the Department of Psychiatry and Behavioral type 1 diabetes for >1 year, and had at least 3 clinic visits with an HbA1c record Sciences, Duke University Medical Center, Durham, NC (J.W.) and Departments of Psychiatry (J.W., D.W.); and within the last 5 years (ie, as necessary to analyze longitudinal trajectories) were Pediatrics (P.W.), University of Texas Southwestern Medical Center, Dallas, TX initially identified. Eighty-one Black youth met the inclusion criteria and were The authors declare no conflicts of interest.
HbA1c HLM M SES
Hemoglobin A1c Hierarchical linear modeling Grand mean Socioeconomic status
This manuscript is based on J.W.’s doctoral dissertation under the supervision of D.W. Portions of these data were presented at the annual convention of the Society of Behavioral Medicine, Seattle, WA, April 8, 2010. 0022-3476/$ - see front matter. Copyright ª 2011 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2011.03.053
571
THE JOURNAL OF PEDIATRICS
www.jpeds.com
matched to a sample of 81 White youth first on sex and then on age based on the closest proximity of birth date; all birth dates were matched within 2 months. Sixty-three Hispanic youth were identified but could not be completely matched on sex and age with Black and White counterparts, because fewer Hispanic patients met the inclusion criteria. This yielded a total sample of 225 youth for primary analyses. After the target sample was identified, age, time since diagnosis, and HbA1c values recorded at each routine clinic visit back to at least 10 years of age or as far as possible were obtained retrospectively from each participant’s medical records. Therefore, data for the full sample included HbA1c values for youth across ages 10 to 18. These repeated measures resulted in 4029 total HbA1c values across the full sample (1562 White, 1338 Black, and 1129 Hispanic). HbA1c values were measured at routine diabetes clinic visits using a point-of-care instrument (DCA2000 or DCA Vantage; Siemens, Deerfield, Illinois). The average number of clinic visits for all youth was 4.1 per year and was not significantly related to median family income or race/ethnicity. Type of insulin therapy was obtained both as a between-subjects variable (pump or no pump therapy at any point during the study period) and a within-subjects variable (pump or no pump at each clinic visit to capture changes in pump status over time). Because few youth changed pump status across time, only the categorical between-subjects variable was analyzed. Individually reported socioeconomic status (SES) could not be obtained given the retrospective review. Therefore, neighborhood SES data were collected from 2000 United States census tract data8 based on the subjects’ zip codes at the time of data collection in 2007. Fourteen participants had missing addresses, preventing us from deriving census tract–based income predictions (2 Black, 8 White, and 4 Hispanic); thus, the sample size was decreased to 211 when income was included in the analyses. Statistical Analysis The data were preliminarily evaluated for violations of statistical assumptions and for outliers. Primary analyses were conducted with hierarchical linear modeling (HLM) techniques.9 HLM analyzes systematic variance on 2 levels. The first level analyzes variables that account for within-subject changes across time (ie, within-subject age-related changes in HbA1c value), and the second level considers how these age-related trajectories differ as a function of individual differences, such as race/ethnicity or income. Age was centered at 10 years to capture the beginning of adolescence for all subjects. These analyses produced coefficients for the intercept, which provided information about average HbA1c value at age 10, and the linear slope, which provided information about the rate of change in HbA1c value across age. Finally, these analyses identified any significant variability in the intercepts or slopes, pointing toward individual differences that allowed us to examine differences across race/ethnicity and income. The HLM analyses included 4 models: (1) the basic model (model 1), in which we determined the basic associations 572
Vol. 159, No. 4 between age and HbA1c, the presence of individual differences in intercept and slope, and the need for covariates (sex, age at diagnosis, and pump status); (2) the race/ethnicity model (model 2), which examined whether race/ethnicity explained individual differences in intercept and slope; (3) the income model (model 3), in which income was included as a covariate to determine whether race effects were independent of income; and (4) the race-by-income interaction model (model 4) to evaluate whether these variables interacted to predict HbA1c value. In models 2, 3, and 4, race/ ethnicity variables were dummy coded to create two pairwise comparisons with White participants as the comparison group in each (White = 0, Black = 1 to test the effects for Black youth; White = 0, Hispanic = 1 to test the effects for Hispanic youth). In models that included income, median family income was centered at the grand mean (M) across all participants. As such, regardless of race/ethnicity, the effects of income were evaluated as a function of deviations from the M SD, $60 812 $28 662/year. Predicted means for significant interactions between race/ethnicity and income were computed by entering values for income that were 1 SD from this grand mean (low income, $32 150/ year; high income, $89 474).
Results Descriptive analyses indicated that White subjects were more likely to use insulin pump therapy, which was included as a covariate for primary analyses (Table I). There were no significant racial/ethnic differences in sex, age at diagnosis, age across all clinic visits, or number of available clinic visits across the data collection period; however, White youth were more likely to live in neighborhoods with higher median family income. We initially examined the linear relationship between HbA1c value and age at each clinic visit ignoring all other variables (model 1). The coefficients for both the intercept [b = 8.89; t(224) = 81.34; P < .001; model 1, Table II] and slope [b = .24; t(224) = 8.51; P < .001; model 1, Table II] were significant. The coefficients indicated that at age 10 years, all youth had an average HbA1c value of 8.89%, and that there was a 0.24% absolute increase in HbA1c value for every yearly increase in age. The significant variance components suggested variability in the intercept (P < .001) and slope (P < .001) not fully accounted for by model 1, which warranted the inclusion of other variables (eg, race/ethnicity, income) in the model (Table II). These results confirmed age-related deterioration in HbA1c values across adolescence in a racially/ethnically diverse sample. Sex, age at diagnosis, and pump status predicted variability in HbA1c across time and were thus entered as covariates in all subsequent analyses. The race/ethnicity model (model 2, Table II) evaluated differences between racial/ethnic groups in average HbA1c value at age 10 and the rate of change in HbA1c values across age 10 to 18. At age 10, Black youth had a 1.4% higher HbA1c value and Hispanic youth a 0.55% higher Wang, Wiebe, and White
ORIGINAL ARTICLES
October 2011
Table I. Descriptive demographic data by race/ethnicity Source
White
Black
Hispanic
N Female sex, % Age at diagnosis, years, M (SD) Average number of clinic visits per participant, M (SD) Average age across all clinic visits, years, M (SD) Average age at first and last clinic visit, years, M (SD) Pump therapy, %* Median family income, $*
81 50.6 9.76 (3.34) 19.28 (8.19) 13.30 (3.00) 18.5 76 289 (36 290)
81 50.6 10.16 (3.75) 16.49 (10.02) 13.34 (3.09) 3.7 55 081 (20 501)
63 49.2 10.13 (3.47) 17.92 (9.13) 13.54 (2.93) 7.9 49 337 (17 150)
Total range
11.09-17.01
*Differences between Whites and minority samples are significant at P < .001.
HbA1c value than White youth. However, rates of change in HbA1c across time were not significantly different among the 3 groups. This indicated that racial/ethnic minority youth had higher HbA1c values than their White counterparts at age 10 (intercept), but showed parallel trajectories in rates of change in HbA1c values over time (slope). When income was added as a covariate in the model (model 3, Table II), the intercept difference between White versus Hispanic teens was no longer significant (P > .10), whereas the difference between White and Black youth remained significant (P < .001). Specifically, Black youth had a 1.3% higher HbA1c value than White youth at age 10 years. All youth experienced similar rates of change in HbA1c values across adolescence regardless of racial/ethnic group. Thus, Black youth appeared to be at disproportionately greater risk independent of neighborhood income. It is notable that risks appeared to begin in childhood, resulting in poorer metabolic control before adolescence. In addition to these childhood risks, Black youth continued to accumulate
additional adolescent risks and evidenced similar rates of deterioration in HbA1c values as White and Hispanic teens across the adolescent years. However, White and Hispanic youth appeared to encounter similar levels of risk in childhood and adolescence when neighborhood income was statistically controlled. Race interacted with income to predict the rate of HbA1c deterioration with age. Specifically, higher neighborhood income was associated with smaller age-related HbA1c deteriorations in White youth, such that every $10 000 increase in income resulted in a slower deterioration in HbA1c value, at a rate of 0.02% per year (model 4, Table II). In contrast, higher neighborhood income was associated with more rapid deterioration in metabolic control in Black and Hispanic youth. Every $10 000 increase in income was associated with more rapid rates of deterioration in HbA1c value of 0.05% in Black youth and 0.08% in Hispanic youth. The Figure depicts pairwise comparisons of income interaction effects between high-income and low-income
Table II. Coefficients for each model from HLM Model 1 Intercept coefficients Intercept Sex Pump status Age at diagnosis Black race Hispanic ethnicity Income Black versus White income interaction Hispanic versus White income interaction Slope coefficients Slope Sex Pump status Age at diagnosis Black race Hispanic ethnicity Income Black versus White income interaction Hispanic versus White income interaction Variance components Intercept Slope
Model 2
Model 3
Model 4
b
P
b
P
b
P
b
P
8.89 -
<.001 -
7.53 .37 .61 .20 1.40 .55 -
<.001 .050 .045 <.001 <.001 .016 -
7.68 .37 .58 .20 1.28 .39 .000005 -
<.001 .048 .056 <.001 <.001 .126 .125 -
7.70 .40 .61 .20 1.31 .24 .000008 .000015 .000006
<.001 .029 .042 <.001 <.001 .321 .002 .092 .459
b
P
b
P
.24 -
<.001 -
.24 .08 .07 .02 .01 .02 -
.098 .154 .353 .045 .849 .748 -
Variance
P
Variance
P
2.42 .13
<.001 <.001
1.60 .13
<.001 <.001
b .23 .08 .07 .02 .02 .04 .000000 Variance 1.58 .13
P .237 .157 .438 .032 .736 .662 .673 P <.001 <.001
b .25 .08 .08 .02 .01 .07 .000002 .000005 .000008 Variance 1.56 .13
P .088 .136 .234 .047 .930 .397 .067 .029 .029 P <.001 <.001
*Sex was coded as 0 = female, 1 = male; pump status was coded as 0 = no pump, 1 = pump therapy; Black (B) race was coded as B = 1; White (W) race = 0 and Hispanic (H) ethnicity was coded as H = 1, W = 0; Black versus White income interaction was coded as B = 1, W and H = 0; and the Hispanic versus White income interaction was coded as H = 1, B and W = 0.
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes
573
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 159, No. 4
Black, Hispanic, and White youth. Further pairwise comparisons revealed that income associations with HbA1c value did not differ between Hispanic and Black youth when Hispanic status was dummy coded as the reference group [b = –.00; t(202) = –.652; P > .10]. This suggests that rising income was similarly detrimental to the rate of deterioration in metabolic control across Black and Hispanic youth.
Discussion Longitudinal patterns of deterioration in metabolic control across White, Black, and Hispanic youth with type 1 diabetes have not been studied previously. Our findings add to the literature documenting poorer metabolic control in minority youth1,3,10 in several important ways. First, our findings suggest that differences in metabolic control between White and Hispanic youth may be more related to neighborhood economic disparities than to racial/ethnic differences, whereas differences between White and Black youth may reflect variables not captured by neighborhood income disparities. Second, our findings suggest that there may be childhood risk factors that contribute to the poorer metabolic control before adolescence in Black youth, which need to be better understood and prevented; indeed, differences in metabolic control between White and Black youth appear to emerge shortly after diagnosis and are maintained for up to 5 years.11 Third, the finding of similar rates of HbA1c deterioration across the adolescent years in all 3 groups suggests that there might be similar barriers to effective diabetes management across all youth (eg, conflicting parent–child relationships, peer conformity, negotiation of autonomy in diabetes management),12-14 which may be useful targets for developing culturally sensitive interventions for minority youth. Finally, income was differentially related to metabolic control in all 3 groups. Although White youth tended to benefit from living in neighborhoods with higher income levels, Black and Hispanic youth actually experienced more rapid HbA1c deteriorations when they lived in higher-income neighborhoods. All of these findings occurred independent of pump status or number of clinic visits, suggesting that these differences are not obviously due to racial/ethnic differences in access to care. Racial/ethnic differences in income effects on metabolic control need to be replicated. Regardless, our results add to a growing set of findings suggesting that low-income environments are not always associated with poor health outcomes. These findings lend support to the ‘‘Hispanic paradox,’’ in which lower-income Hispanic individuals appear to have better health than would be expected.15 There may be protective factors associated with living in these lower-income neighborhoods; for example, ‘‘familism’’ or ‘‘allocentrism,’’ which reflects the tendency to place the welfare of the community or family above the needs of the individual, appear to be important in Hispanic communities16 and may be emphasized in communities with fewer monetary resources. Alternatively, factors related to an ‘‘ethnic density effect’’ may suggest that 574
Figure. High SES of $89 474/year (1 SD above M, $60 812) versus low SES of $32 150/year (1 SD below M) shows different age-related trajectories in HbA1c values among Black, Hispanic, and White adolescents. Higher SES is associated with more rapid deterioration in HbA1c values across the teenage years in Black and Hispanic youth compared with White youth.
minority children experience heightened risk with increases in family income as they relocate or live in areas in which they are more overtly racial minorities. Minority individuals living in neighborhoods that are more majority-race dense may experience poorer health due to greater exposure to racial tension, acculturative stressors, and discrimination and prejudice.17,18 Minority individuals living in mostly White neighborhoods also may perceive less social/community support and more social isolation, which may put them at greater risk for poorer health outcomes. Several methodological considerations should be taken into account when interpreting our findings. Our modest sample size in each race/ethnic group was countered somewhat by the increased power obtained with repeated measures of each individual across time (4029 total HbA1c measures in the entire sample). The data were collected and recorded during routine clinic visits for standard care at a single site, limiting the generalizability of our results to other medical centers or cities. Data were limited to information Wang, Wiebe, and White
October 2011 available in chart review; we were unable to consistently obtain information on such variables as pubertal status or timing of adoption of diabetes responsibilities, which might have influenced the trajectories and should be collected in future research. We cannot rule out the possibility that racial/ethnic differences in HbA1c values reflected artifacts of measurement such as might be caused by hemoglobinopathies; however, we believe that this was unlikely, because we are not aware of such cases in our sample and because hemoglobin S (HbS), the most common hemoglobin variant in the Black population, tends to decrease red cell survival and thus lead to underestimation of HbA1c. The HbA1c measurement method used in this study does not overestimate HbA1c in the presence of HbS.26 Neighborhood SES has been associated with trajectories of child health,20 which we used as a proxy for individual SES. Although this technique has often been used when individual SES is not available21,22 and is a recommended approach to overcoming the lack of socioeconomic data in health databases, it is not without limitations.23,24 In particular, because we collected each subject’s address at a single point in time, our data do not account for residential mobility of individuals to different geographical regions over several years. In addition, neighborhood estimates of individual income may overestimate or underestimate actual individual income, resulting in misclassification of income status.25 For example, Black individuals tend to live in neighborhoods of greater economic disadvantage compared with White counterparts of similar income levels.24 Future studies should prospectively collect both neighborhood and individual SES variables to determine their common and unique associations with metabolic control. Future longitudinal studies also should obtain measures of family structure to evaluate its impact on metabolic control in low-income families, given that single-parent households are more common in lowerincome households and are associated with poorer metabolic control.11,26 Although implications for interventions may be premature at this time, our data suggest the importance of early intervention, especially in minority populations. Based on intervention studies in White youth, parent–child relationships and peer influences may be potential areas of focus for such interventions with minority youth.12,13 If supported by future research, interventions that leverage resources unique to lower-SES minority communities or that address risks present in higher-SES minority communities may be useful. Clarifying the needs of minority families coping with chronic health conditions will aid the development of such interventions. This study represents a step toward a better understanding of how minority youth cope with diabetes management across the adolescent years. Childhood risk factors may be particularly important in Black youth, whereas adolescent risk factors appear to be equally important in Black, White, and Hispanic youth. Minority youth do not appear to benefit as much from living in more-affluent neighborhoods compared with White youth. Clarifying this finding may
ORIGINAL ARTICLES point to detrimental and beneficial neighborhood factors for intervention to reduce health disparities and improve the quality of life for diverse children, adolescents, and their families. Since the acceptance of this report, we have become aware that differences between Blacks and Whites in HbA1c measurements have been reported,19 but these racial differences in HbA1c measurements are smaller than those observed in the present study. n We thank the dissertation committee members, including Drs Sunita Stewart, Margaret Caughy, and Crista Wetherington at the University of Texas Southwestern Medical Center, Dallas, Texas for their support and guidance. Submitted for publication Oct 8, 2010; last revision received Feb 16, 2011; accepted Mar 28, 2011.
References 1. Anderson B, Ho J, Brackett J, Finkelstein D, Laffel L. Parental involvement in diabetes management tasks: relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulin-dependent diabetes mellitus. J Pediatr 1997; 130:257-65. 2. Weissberg-Benchell J, Glasgow AM, Tynan WD, Wirtz P, Turek J, Ward J. Adolescent diabetes management and mismanagement. Diabetes Care 1995;18:77-82. 3. Delamater AM, Albrecht DR, Postellon DC, Gutai JP. Racial differences in metabolic control of children and adolescents with type I diabetes mellitus. Diabetes Care 1991;14:20-5. 4. Delamater AM, Shaw KH, Applegate EB, Pratt IA, Eidson M, Lancelotta GX, et al. Risk for metabolic control problems in minority youth with diabetes. Diabetes Care 1999;22:700-5. 5. Gallegos-Macias AR, Macias SR, Kaufman E, Skipper B, Kalishman N. Relationship between glycemic control, ethnicity and socioeconomic status in Hispanic and white non-Hispanic youths with type 1 diabetes mellitus. Pediatr Diabetes 2003;4:19-23. 6. Summerson JH, Konen JC, Dignan MB. Race-related differences in metabolic control among adults with diabetes. South Med J 1992; 85:953-6. 7. Tull ES, Barinas E. A twofold excess mortality among black compared with white IDDM patients in Allegheny County, Pennsylvania. Pittsburgh DERI Mortality Study Group. Diabetes Care 1996;19:1344-7. 8. United States Federal Financial Institutions Examination Council. FFIEC Geocoding System based on data collected in 2007. Available from: http://www.ffiec.gov/geocode/default.htm. Accessed June 29, 2009. 9. Bryk AS, Raudenbush SW. Hierarchical linear models: applications and data analysis methods. Newbury Park, CA: Sage; 1992. 10. Auslander WF, Thompson S, Dreitzer D, White NH, Santiago JV. Disparity in glycemic control and adherence between African-American and Caucasian youths with diabetes: family and community contexts. Diabetes Care 1997;20:1569-75. 11. Frey MA, Templin T, Ellis D, Gutai J, Podolski CL. Predicting metabolic control in the first 5 yr after diagnosis for youths with type 1 diabetes: the role of ethnicity and family structure. Pediatr Diabetes 2007;8:220-7. 12. Palmer DL, Berg CA, Wiebe DJ, Beveridge RM, Korbel CD, Upchurch R, et al. The role of autonomy and pubertal status in understanding age differences in maternal involvement in diabetes responsibility across adolescence. J Pediatr Psychol 2004;29:35-46. 13. Hains AA, Berlin KS, Davies WH, Smothers MK, Sato AF, Alemzadeh R. Attributions of adolescents with type 1 diabetes related to performing diabetes care around friends and peers: the moderating role of friend support. J Pediatr Psychol 2007;32:561-70.
Developmental Trajectories of Metabolic Control among White, Black, and Hispanic Youth with Type 1 Diabetes
575
THE JOURNAL OF PEDIATRICS
www.jpeds.com
14. Wiebe DJ, Berg CA, Korbel C, Palmer DL, Beveridge RM, Upchurch R, et al. Children’s appraisals of maternal involvement in coping with diabetes: enhancing our understanding of adherence, metabolic control, and quality of life across adolescence. J Pediatr Psychol 2005;30:167-78. 15. Franzini L, Ribble JC, Keddie AM. Understanding the Hispanic paradox. Ethn Dis 2001;11:496-518. 16. Gallo LC, Penedo FJ, Espinosa de los Monteros K, Arguelles W. Resiliency in the face of disadvantage: do Hispanic cultural characteristics protect health outcomes? J Pers 2009;77:1707-46. 17. Becares L, Nazroo J, Stafford M. The buffering effects of ethnic density on experienced racism and health. Health Place 2009;15:700-8. 18. Pickett KE, Wilkinson RG. People like us: ethnic group density effects on health. Ethn Health 2008;13:321-34. 19. Kamps JL, Hempe JM, Chalew SA. Racial disparity in A1C independent of mean blood glucose in children with type 1 diabetes. Diabetes Care 2010;33:1025-7. 20. McGrath JJ, Matthews KA, Brady SS. Individual versus neighborhood socioeconomic status and race as predictors of adolescent ambulatory blood pressure and heart rate. Soc Sci Med 2006;63:1442-53.
Vol. 159, No. 4 21. Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health 1992;82:703-10. 22. Krieger N, Chen JT, Waterman PD, Rehkopf DH, Subramanian SV. Race/ ethnicity, gender, and monitoring socioeconomic gradients in health: a comparison of area-based socioeconomic measures. The Public Health Disparities Geocoding Project. Am J Public Health 2003;93:1655-71. 23. Braveman PA, Cubbin C, Egerter S, Chideya S, Marchi KS, Metzler M, et al. Socioeconomic status in health research: one size does not fit all. JAMA 2005;294:2879-88. 24. Diez-Roux AV, Kiefe CI, Jacobs DR, Jr., Haan M, Jackson SA, Nieto FJ, et al. Area characteristics and individual-level socioeconomic position indicators in three population-based epidemiologic studies. Ann Epidemiol 2001;11:395-405. 25. Hyndman JC, Holman CD, Hockey RL, Donovan RJ, Corti B, Rivera J. Misclassification of social disadvantage based on geographical areas: comparision of postcode and collector’s district analyses. Int J Epidemiol 1995;24:165-76. 26. Thompson SJ, Auslander WF, White NH. Comparison of single-mother and two-parent families on metabolic control of children with diabetes. Diabetes Care 2001;24:234-8.
50 Years Ago in THE JOURNAL OF PEDIATRICS Anemia Associated with Protein Deficiency: A Study of Two Cases with Cystic Fibrosis Shahidi NT, Diamond LK, Shwachman H. J Pediatr 1961;59:533-42.
t was as late as 1938 when cystic fibrosis (CF) was first recognized as a separate entity from celiac disease.1 For decades after this milestone, the intestinal manifestations continued to be dominant. Fifty years ago Shahidi and Diamond, with the giant of exocrine pancreas research Harry Shwachman, published a report in The Journal describing two infants who were admitted for diagnosis and treatment of pitting edema (albumin 1.50 and 1.64 gram percent) who were found to be profoundly anemic (hemoglobin 7.2 and 7.7 gram percent).2 Shwachman had introduced the sweat test 5 years earlier, so these infants received a quick diagnosis, but the article also addresses the extensive evaluation of these infants, including bone marrow examination and radio-labeled albumin studies. Fecal fat studies were not mentioned, and we now know, on the back of pioneering studies such as this one, that profound maldigestion often aggravated by inadequate caloric intake causes this clinical picture. Currently, the triad of hypoalbuminemia, edema, and anemia should strongly suggest the diagnosis of CF, and a sweat chloride concentration should be performed. A note of caution is indicated, however, because edema may cause a false-negative sweat test result, and the test should then be repeated when the child is in a better nutritional state. As we move into an era of neonatal screening for CF,3 this clinical picture will become rarer because infants are diagnosed before clinical symptoms develop.
I
Michael Wilschanski, MBBS Pediatric Gastroenterology Hadassah University Hospitals Jerusalem, Israel 10.1016/j.jpeds.2011.05.016
References 1. Anderson DH. Cystic fibrosis of the pancreas and its relationship to celiac disease. Am J Dis Child 1938;56:344-99. 2. Shahidi NT, Diamond LK, Shwachman H. Anemia associated with protein deficiency: a study of two cases with cystic fibrosis. J Pediatr 1961;53342. 3. Castellani C, Massie J. Emerging issues in cystic fibrosis newborn screening. Curr Opin Pulm Med 2010;16:584-90.
576
Wang, Wiebe, and White