Windows of opportunity to improve diabetes care when patients with diabetes are hospitalized for other conditions.

OBJECTIVE: The overwhelming majority of hospitalizations for patients with diabetes occur for treatment of comorbid conditions. This study assessed broad-based interventions to improve diabetes care for patients hospitalized with cardiac conditions. RESEARCH DESIGN AND METHODS: A pre-post quasi-experimental study design was used to evaluate the implementation of two quality improvement interventions: 1) revision of the hospital's capillary blood glucose monitoring form into a color-coded process control chart and 2) a clinical path for type 2 diabetes as a secondary diagnosis. Interventions were implemented on the medical and surgical cardiac care units (not including the intensive care units on these services) of a tertiary academic medical center. A chart abstraction sample included 328 subjects with no exposure to the interventions and 336 subjects hospitalized after both interventions were implemented. Telephone surveys were conducted after discharge on 446 patients. RESULTS: The frequency of patients with severe hyperglycemia (at least one glucose level >400 mg/dl) and prolonged hyperglycemia (at least three consecutive glucose levels >250 mg/dl) decreased from 12 and 17% preintervention to 6.6 and 10% postintervention (P = 0.017 and P = 0.013, respectively). We found that 9% of the patients preintervention and 5% of the patients postintervention (P = 0.05) had nosocomial infections. Patient-reported receipt of self-care instruction varied from 44 to 69% on nine survey items preintervention. Postintervention linear regression slopes for receipt of self-care instruction were all greater than preintervention slopes, but the differences did not achieve statistical significance. We found that 40% of the patients had important diabetes knowledge deficits. CONCLUSIONS: Our broad-based interventions were associated with a decreased frequency of prolonged and severe hyperglycemia and a decreased frequency of nosocomial infections. We also identified opportunities to improve diabetes self-care instruction before discharge and to address important knowledge deficits of patients.

Grading the evidence for diabetes performance measures.

CONTEXT: Grading scientific evidence is a critical step in developing practice guidelines and quality performance measures. GENERAL QUESTION: What is the most useful way to grade evidence? SPECIFIC QUESTION: How should we grade the recommended clinical practices for patients with diabetes? STANDARD APPROACH: Hierarchical grading systems (e.g., grades I, II, and III), such as that used by the U.S. Preventive Services Task Force, have traditionally been used to rank the research designs of studies that support a particular clinical practice. POTENTIAL DIFFICULTIES: Many studies that support the clinical practices of diabetes care do not clearly conform to the categories traditionally used in hierarchical grading systems. As a result, there is a tendency to inaccurately characterize the level of evidence, leading to the phenomenon of evidence inflation or evidence deflation. To avoid exaggerating the evidence, important sources of information may be excluded, resulting in an understatement of the available supporting evidence. ALTERNATE APPROACH: This paper offers a more descriptive typologic system that uses the study design and an explanatory modifier to grade the evidence supporting the clinical practices of diabetes care. The study grades are randomized, controlled trial (RCT); RCT-embedded component; RCT-treatment only; RCT-different population; observational study-risk factor; and expert opinion. Using this grading system, the authors were able to more accurately describe the best available evidence supporting the clinical practices of diabetes care.

Higher incidence of diabetes in liver transplant recipients with hepatitis C.

We assessed the clinical and biochemical parameters associated with the development of posttransplantation diabetes (PTDM) in 52 liver transplant recipients followed up for 1 year. Diabetes was present before transplantation in 9.6% (5 of 52) of patients, and PTDM occurred in 23% (11 of 47) of the remaining liver transplant recipients. Of the 13 patients who had hepatitis C as the cause of their liver failure (HC-LD), 8 (62%) developed PTDM; of the 34 patients with other causes of liver failure, 3 (9%) developed PTDM (p < 0.001). Posttransplantation diabetes was also associated with the development of early posttransplantation hyperglycemia, a higher number of liver rejection episodes, and lower serum albumin levels at 6 months. The association of PTDM with HC-LD remained significant in a logistic regression model after adjustment for potential confounding variables. We conclude that PTDM is common in liver transplant recipients. Associated clinical parameters predictive of PTDM include a diagnosis of HC-LD before transplantation, the development of early hyperglycemia after transplantation, multiple episodes of posttransplantation liver rejection and low serum albumin levels at 6 months. The fact that HC-LD remained an independent risk factor for the development of PTDM may suggest a direct or immune-mediated pancreatic effect of the virus.

Influence of an interdisciplinary diabetes specialist team on short-term outcomes of diabetes at a community health center.

OBJECTIVE: To assess the effect on short-term diabetes outcomes of intervention with an interdisciplinary diabetes specialist team at a primary-care community health center in East Harlem, New York City. METHODS: An interdisciplinary diabetes specialist team, consisting of a diabetologist, a bicultural certified diabetes nurse-educator, and a nutritionist, attended weekly clinics at a primary-care community health center in East Harlem. Emphasis was placed on communicating in the patient's primary language and providing nutritional counseling, diabetes education, and diabetes management. After 1 year, a retrospective review of medical records for patients seen by the diabetes team was performed to assess the influence of this intervention on performance of home glucose monitoring (HGM), frequency of hypoglycemia, and changes in diabetes treatment regimens. Of 70 patients referred to the diabetes team by their primary-care providers, 50 underwent follow-up for at least 6 months and were included in the statistical analysis. RESULTS: Of the 50 study patients, 94% had type 2 diabetes, with a mean duration of 11.2 years. Eighty-two percent were Hispanic, and 18% were Afro-American. The mean age was 54.6 years. Microvascular complications were present in 44%, and macrovascular complications were present in 22%. HGM was done by 13 patients (26%) before and 33 patients (66%) after diabetes team intervention (P<0.001). Before intervention by the diabetes team, 14 patients (28%) were having episodes of unrecognized hypoglycemia. Unrecognized hypoglycemia resolved after intervention in all but two patients with type 1 diabetes (P<0.001). Before intervention, diabetes treatment was dietary in 6 patients, a sulfonylurea in 19, and insulin in 25; after intervention, 5 patients had dietary management of their diabetes, 14 were taking a sulfonylurea, and 31 were receiving insulin (P<0.05). CONCLUSION: Providing access to an interdisciplinary diabetes specialist team at the site of primary care had a beneficial effect on short-term diabetes outcomes in inner-city Hispanic and Afro-American patients with diabetes. Providing specialty care in the primary-care setting may be one model for improving the quality and long-term outcomes of diabetes care, particularly in high-risk populations.

Management of diabetes mellitus from a public health perspective.

This article details the scope and the impact of diabetes in the United States including the prevalence, demographics, morbidity and mortality, and costs. The current status of medical care for diabetes is used to illustrate why diabetes should be approached from a proactive public health perspective, rather than a reactive, traditional medical perspective. The importance of early diagnosis and secondary intervention programs for prevention and early detection of diabetes complications are emphasized. The specialty-based intensive management model, the primary care-based co-management model, the systems management model, and the mini-clinic models are presented as potential paradigms for approaching diabetes care from a public health perspective. Requirements for implementing a public health approach to diabetes care including long-term planning, targeting patients for improved care, and goal setting for outcomes of care, are discussed.

Effect of a bicultural community health worker on completion of diabetes education in a Hispanic population.

OBJECTIVE: To determine the effect of a bicultural community health worker (CHW) on completion of diabetes education in an inner-city Hispanic patient population and to evaluate the impact of completion of the education program on patient knowledge, self-care behaviors, and glycemic control. RESEARCH DESIGN AND METHODS: Patients were randomized into CHW intervention and non-CHW intervention groups. All patients received individualized, comprehensive diabetes education from a certified diabetes nurse educator after baseline demographic information, diabetes knowledge, diabetes self-care practices, and glycohemoglobin levels were assessed. Rates of education program completion were determined. Diabetes knowledge, self-care practices, and glycohemoglobin levels were reassessed at program completion and at a later postprogram follow-up medical appointment and compared to baseline. Logistic regression analysis and the Mantel-Haenszel chi 2 statistic were used to determine the effect of the CHW assignment on program completion. Analyses of covariance were performed with end-of-treatment behavior scores, knowledge scores, and glycohemoglobin levels as outcome variables, controlling for baseline values and testing for the effect of CHW assignment. RESULTS: Of 64 patients enrolled in the study, 40 (63%) completed and 24 (37%) dropped out before completing the diabetes education program. Of the patients having CHW intervention, 80% completed the education program, compared with 47% of patients without CHW intervention (P = 0.01). "Dropouts" were younger (age 47.5 +/- 12.5 years [mean +/- SD]) compared with patients who completed the program (55.9 +/- 9.9 years) (P = 0.004). Dropout status showed no significant relationship to educational level achieved or literacy level. For the program "completers," knowledge levels and selected self-care practices significantly improved, and glycohemoglobin levels improved from a baseline level of 11.7% to 9.9% at program completion (P = 0.004) and 9.5% at the postprogram follow-up (P < 0.001). The effect of the CHW assignment on program completion, controlling for financial status and language spoken, was extremely robust (P = 0.007). The effect of the CHW on knowledge, self-care behavior, or glycohemoglobin outcome variables was not statistically significant. CONCLUSIONS: These findings suggest that intervention with a bicultural CHW improved rates of completion of a diabetes education program in an inner-city Hispanic patient population irrespective of literacy or educational levels attained. Our data further suggests that completion of individualized diabetes educational strategies leads to improved patient knowledge, self-care behaviors, and glycemic control.

An inpatient diabetes QI program.

OBJECTIVE: The goal of this study was to establish a continuous quality improvement (CQI) program for diabetes which would identify patterns in the problems of care encountered by hospitalized patients with diabetes and improve the in-hospital process of diabetes care delivery. RESEARCH DESIGN AND METHODS: The laboratory information system in an acute and tertiary care 1,000-bed urban teaching hospital provided us on a daily basis with a list of patients on the medical service having blood glucose (BG) levels < 40mg/dl or > 450mg/dl and positive serum acetones. We performed concurrent implicit chart review when BG levels were hypoglycemic (< 40mg/dl) or hyperglycemic (> 450mg/dl on two occasions) or when diabetic ketoacidosis (DKA) was present (acetones were > 1+) using preset indicators for documentation and appropriate medical management. Data were expressed as the ratio of number of cases in compliance with the indicator over total number of cases identified. A test for trend in proportions was used to assess compliance with the indicators over time. RESULTS: Documentation of nursing unit-based capillary blood glucose (CBGM) and insulin infusion monitoring improved significantly over time (p < 0.001 for both). The medical management of hypoglycemia, hyperglycemia and DKA improved (p = 0.1) over the three-year period. Identification of recurrent multidisciplinary process problems in the management of DKA, intravenous insulin infusion constitution and delivery, CBGM determination in the setting of anemia, and recognition of clinical settings conducive to the development of hypo- and hyperglycemia were identified and addressed with standardization in documentation, an insulin infusion protocol, administrative rules, and staff education. CONCLUSIONS: Efforts to standardize specific clinical and documentation processes had a positive impact on the care of hospitalized patients with diabetes and resulted in an institutional effort to improve inpatient diabetes care with a CQI team.

Mortality in hospitalized patients with hypoglycemia and severe hyperglycemia.

BACKGROUND: We designed a study to determine the incidence, cause, and implications of hypoglycemia (< or = 2.7 mmol/L, 49 mg/dL) and severe hyperglycemia (> or = 22.2 mmol/L, 400 mg/dL) in in-patients at an urban tertiary medical center. METHODS: A daily computer search of the Laboratory Information System identified all hospitalized patients with hypoglycemia and severe hyperglycemia during a 49-day period. Chart review was used to assess demographic information, risk factors, and epidemiologic variables. The eventual outcome of the hospitalization was obtained by follow-up. RESULTS: The incidence of hypoglycemia was 1.5%, and of hyperglycemia, 1.9%. Seventy-six percent of the hypoglycemic patients and 16% of the hyperglycemic patients had no prior history of diabetes. The mortality rate for hypoglycemic patients was 22.2%; for hyperglycemic patients it was 11.1%. For all other hospitalized patients it was 2.3% (p < 0.0001). Mortality rates for the black and Hispanic patients who were hypoglycemic (30% and 46%) were significantly higher than for white patients (6%, p < 0.01). CONCLUSIONS: Hypoglycemia and severe hyperglycemia are not uncommon in hospitalized patients and serve as metabolic markers for patients at increased risk for inhospital mortality. Early identification of at-risk patients and the impact of aggressive treatment of their underlying disease processes should be evaluated in future studies.

Long-term prospective study of postpartum thyroid dysfunction in women with insulin dependent diabetes mellitus.

Postpartum thyroid dysfunction (PPTD) is a common autoimmune disorder. Type I diabetes mellitus (IDDM) is an autoimmune disease with a high incidence of concomitant autoimmune thyroid failure. We hypothesized that women with IDDM would have an increased incidence of PPTD. Women with IDDM in New York City, were followed prospectively during the second and third trimester of pregnancy and at 6 weeks, 3 months, 6 months, 9 months, and 1 yr postpartum. A long-term follow-up was performed at 31 months postpartum. Forty-one women with IDDM were recruited at their initial prenatal visit. Two women (4.8%) had thyroid function test abnormalities observed at screening, three (7.3%) had a spontaneous miscarriage, and eight (19.5%) women were noncompliant with follow-up. Twenty-eight women (68.2%) completed the study. Thyroid function tests and thyroid autoantibody determinations were obtained at all visits. PPTD was defined as a TSH greater than 5.0 or less than 0.2 mU/L in the postpartum period with documented normal thyroid function tests during pregnancy. The incidence of PPTD in women with IDDM was 25%. This is a 3-fold increase compared to a similar study by our group in a nondiabetic population. Forty-three percent of the women (3/7) who developed PPTD required treatment in the immediate postpartum period and at long-term follow-up. The remainder of the women with PPTD, as well as all women who did not develop PPTD were euthyroid at 31 months postpartum. Women with IDDM are at high risk for PPTD. We recommend that all women with IDDM be screened for thyroid hormonal abnormalities during pregnancy and at 3 months postpartum for postpartum thyroid dysfunction. Long-term follow-up did not reveal an increased incidence of hypothyroidism in women who did not require treatment in the first postpartum year.

Lack of disease associated HLA-DQ restriction fragment length polymorphisms in families with autoimmune thyroid disease.

Autoimmune thyroid disease (AITD) is often familial and serological HLA disease associations have been described in many different populations. However, such HLA disease associations are weak and the precise molecular contribution of HLA antigens to thyroid disease susceptibility remains unknown. Much of the data available are cross-sectional and few studies have explored familial inheritance of AITD at the molecular level. We have, therefore, examined the inheritance of AITD in multiplex and multi-generational families using restriction fragment length polymorphism (RFLP) analysis of DNA digested with the restriction enzyme BamH1 and probed with a full length human HLA-DQ beta cDNA probe. Thirty seven subjects in 7 informative families were available for study. Eleven subjects had Graves' disease and 4 were diagnosed as having Hashimoto's thyroiditis. Segregation of polymorphic fragments enabled genotyping of each individual to produce fully informative families. LOD scores were computed, using the LIPED program, for dominant and recessive models of inheritance, for recombination fractions of 0.01 to 0.5 for each sex, and for penetrances of 0.1 to 1.0. The results showed that maximum LOD scores were negative for all of the inheritance models tested. If the primary locus for AITD were in the HLA region, LOD scores would be highly positive. These data, therefore, provide strong evidence against a disease locus for AITD in linkage disequilibrium with the HLA-DQ beta locus.

Leukocyte scanning with 111In is superior to magnetic resonance imaging in diagnosis of clinically unsuspected osteomyelitis in diabetic foot ulcers.

OBJECTIVE: To compare the accuracies of MRI and leukocyte scanning in diagnosing clinically unsuspected osteomyelitis in diabetic foot ulcers. RESEARCH DESIGN AND METHODS: A prospective study of 16 diabetic foot ulcers in 12 patients, including both ambulatory and hospitalized patients, was performed at a university medical center. Pedal images were obtained by leukocyte scanning with [111In]oxyquinoline and MRI. Definitive diagnosis of osteomyelitis then was determined by bone biopsy for culture and histology. RESULTS: Biopsy-proven osteomyelitis was present in 7 (44%) of the 16 foot ulcers. The diagnosis was suspected clinically in 0%. Leukocyte scanning was 100% sensitive, whereas MRI was only 29% sensitive in diagnosing osteomyelitis in diabetic foot ulcers. Specificities were 67 and 78%, respectively. The positive and negative predictive values (70 and 100%, respectively) for the leukocyte scan also were greater than those of MRI (50 and 58%, respectively). CONCLUSIONS: Leukocyte scanning is superior to MRI in detecting clinically unsuspected osteomyelitis in diabetic foot ulcers.

Maternal characteristics, neonatal outcome, and the time of diagnosis of gestational diabetes.

OBJECTIVE: A study was undertaken to evaluate the yield of early, routine screening for gestational diabetes and to determine whether maternal characteristics and neonatal outcome differ according to the time of diagnosis. STUDY DESIGN: A total of 2776 women were screened before 24 weeks of gestation, and each was delivered of a singleton infant on the clinic service of Mount Sinai Hospital in New York City between January 1986 and January 1991. RESULTS: An abnormal glucose tolerance test was diagnosed in 102 women < 24 weeks and in 252 patients at > or = 24 weeks of gestation. Logistic regression analysis showed that the group diagnosed early was significantly older and more likely to have hypertensive disorders and low maternal weight gain and to require insulin treatment, compared with the group diagnosed late. No significant differences were evident in neonatal outcome. CONCLUSIONS: These data indicate that a sizable proportion of patients with gestational diabetes can be diagnosed early in pregnancy. The differences in maternal characteristics and insulin requirements between the early- and late-diagnosis groups also suggest heterogeneity of gestational diabetes or the possibility of preexisting impaired glucose intolerance in the early-diagnosis group.

Genetics of autoimmune thyroid disease: lack of evidence for linkage to HLA within families.

Clinical and epidemiologic observations, including the association of Graves' disease (GD) and Hashimoto's thyroiditis (HT) with the HLA gene complex, support a role for specific disease-related genes in the development of autoimmune thyroid disease (AITD). The combination of HLA and immunoglobulin heavy chain allotypes (Gm) has previously been reported to be predictive of AITD in multiply affected Japanese families. We have investigated the immunogenetics of AITD in families in the United States. Twenty-seven pedigrees including 15 with GD, 8 with HT, and 4 with both HT and GD were immunogenetically typed and analyzed for population and within family disease associations. The majority of families (63%) were multiplex for AITD. HLA-DR3 was increased in affected family members with GD and HLA-DR5 was increased in affected family members with HT. Formal linkage analysis was applied to test for coinheritance of disease with the HLA locus within families. The LIPED computer program was used to calculate the probability of linkage in terms of the lod score. Evidence from linkage analysis was consistently against linkage of either GD or HT to the HLA region under various penetrances and different modes of inheritance. The combination of HLA and Gm was not found to be predictive of disease in 7 selected multiplex families with multigenerational instances of AITD. T cell function was also examined in 3 pairs of siblings genetically identical for HLA and Gm but discordant for disease expression. We found no evidence of a global T cell defect in the small number of patients examined. We conclude that whereas there is an association of AITD with the HLA region, our linkage analysis demonstrates that alleles of the HLA region are not cosegregating with either GD or HT within these families. Thus, whereas HLA may increase susceptibility to AITD, as shown by the existence of an HLA association, the major genetic influence on the inheritance of AITD must be at another locus.

A prospective study of lymphocyte-initiated immunosuppression in normal pregnancy: evidence of a T-cell etiology for postpartum thyroid dysfunction.

Immune function in normal pregnancy and the postpartum period remains poorly defined. We hypothesized that a comparative study between pregnant women with normal and abnormal immune function would further our understanding of the immune mechanisms of pregnancy. We chose to study a cohort of pregnant women at risk for the development of postpartum thyroid dysfunction (PPTD) as well as a group of normal controls. We chose PPTD as the model for abnormal immune function because of the relative ease of monitoring disease development and the relatively high prevalence for PPTD reported in earlier studies. Five hundred and fifty-two women were screened for the presence of thyroid autoantibodies in the first trimester of pregnancy. Thirty-three thyroid autoantibody-positive women and 28 thyroid autoantibody-negative women were followed prospectively throughout pregnancy and 6 months into the postpartum period. Lymphocyte subset analyses, thyroid function tests, and thyroid autoantibodies (antihuman thyroglobulin and antithyroid peroxidase) were performed at defined intervals. All patients were HLA serotyped. Normal pregnancy was principally characterized by decreased CD4+ T-cells and increasing CD8+ T-cells, causing a significant fall in the CD4+/CD8+ ratio in late pregnancy and into the postpartum period. Women who developed PPTD had 1) a higher CD4+/CD8+ ratio (P = 0.04), 2) activation of T-cells in the postpartum period (P = 0.02), and 3) significantly higher thyroid autoantibody titers (antihuman thyroglobulin, P = 0.02; antithyroid peroxidase, P = 0.0018). We found an overall incidence for PPTD of 8.8%. These data demonstrated that women who were thyroid autoantibody positive in the first trimester of pregnancy had a one in three chance of developing PPTD. We observed a significant fall in the T-cell helper/suppressor ratio in normal pregnant women, which was associated with distinct T-cell subset changes. This pregnancy-initiated T-cell regulation reflected an overall suppression of immune function. The development of PPTD was a frequent postpartum event in our population and was associated with a triad of immune markers: a reduction in the normal immune suppression of pregnancy (as indicated by higher T-cell helper/suppressor ratios), enhanced postpartum T-cell activation, and elevated thyroid autoantibodies. The reduction in the degree of immune suppression was, therefore, a major factor in the development of PPTD. Our results define immunological changes that occur in normal pregnancy and distinct immunological abnormalities necessary for the development of PPTD.

An approach to the heterogeneity of diabetes in adults.

Diabetes mellitus in the adult population consists of a group of etiologically heterogeneous disease processes, the common outcome of which is derangement of carbohydrate metabolism. New knowledge in the areas of insulin receptors, insulin action, glucose transport, pancreatic amylin, and pancreatic autoimmunity is likely to lead to the development of disease-process-specific treatment modalities.

Intensive insulin therapy: euglycemia with a price tag.

The benefit of lowering blood glucose levels with intensive insulin therapy, especially with regard to the development and progression of microvascular complications, is still being investigated. Although many studies have demonstrated a correlation between the degree of hyperglycemia and the risk of development of long-term diabetic complications, we still do not know at what level of glycemia that risk occurs. It is hoped the Diabetes Control and Complications Trial will answer the question. Until that time, patients who are candidates for intensive insulin regimen therapies should be educated and trained to regard the prevention of hypoglycemia as the highest priority of their self-care.

Unsuspected osteomyelitis in diabetic foot ulcers. Diagnosis and monitoring by leukocyte scanning with indium in 111 oxyquinoline.

OBJECTIVE: The prevalence of osteomyelitis in diabetic foot ulcers is unknown. Early diagnosis of this infection is critical, as prompt antibiotic treatment decreases the rate of amputation. We therefore assessed the prevalence of osteomyelitis in 35 diabetic patients with 41 foot ulcers. We compared results of roentgenograms, leukocyte scans with indium In 111 oxyquinoline, and bone scans with the diagnostic criterion standards of bone histologic and culture findings. Leukocyte scans were repeated at 2- to 3-week intervals during antibiotic treatment. DESIGN: Cohort study. SETTING: Institutional and private, ambulatory and hospitalized patients. PATIENTS: Consecutive sample of 54 diabetic patients. Thirty-five patients with 41 foot ulcers were included. RESULTS: As determined by bone biopsy and culture, osteomyelitis was found to underlie 28 (68%) of 41 diabetic foot ulcers. Only nine (32%) of the 28 cases were diagnosed clinically by the referring physician. Underscoring the clinically silent nature of osteomyelitis in these ulcers, 19 (68%) of 28 occurred in outpatients, 19 (68%) of 28 occurred in ulcers not exposing bone, and 18 (64%) of 28 had no evidence of inflammation on physical examination. All patients with ulcers that exposed bone had osteomyelitis. Of the imaging tests, the leukocyte scan had the highest sensitivity, 89%. In patients with osteomyelitis, the leukocyte scan image intensity decreased by 16 to 34 days of antibiotic treatment and normalized by 36 to 54 days. CONCLUSION: The majority of diabetic foot ulcers have an underlying osteomyelitis that is clinically unsuspected. Leukocyte scans are highly sensitive for diagnosing osteomyelitis in diabetic foot ulcers and may be useful for monitoring the efficacy of antibiotic treatment. We recommend that diabetic patients with foot ulcers that expose bone should be treated for osteomyelitis. Diabetic patients with foot ulcers that do not expose bone should undergo leukocyte scanning, which eliminates the risk of bone biopsy in diagnosing osteomyelitis and allows for the diagnosis and treatment of this well-known but often silent precursor of lower extremity amputation.

Detection of at-risk pregnancy by means of highly sensitive assays for thyroid autoantibodies.

We screened 552 women who presented to their obstetrician in the first trimester of pregnancy using highly sensitive enzyme-linked immunosorbent assays for the presence of thyroglobulin and thyroid peroxidase autoantibodies and found an incidence of positivity of 19.6%. The tendency to secrete detectable levels of thyroid autoantibodies was significantly correlated with an increased rate of miscarriage. Thyroid autoantibody-positive women miscarried at a rate of 17%, compared with 8.4% for the autoantibody-negative women. Individual levels of thyroglobulin and thyroid peroxidase autoantibodies were similarly related to this increased miscarriage rate, with no evidence of autoantibody specificity in the relationship. Furthermore, the increase in miscarriages could not be explained by differences in thyroid hormone levels, the presence of cardiolipin autoantibodies, maternal age, gestational age at the time of maternal entry into the study, or previous obstetric history. We conclude that thyroid autoantibodies are an independent marker of "at-risk" pregnancy.

Intrathyroidal accumulation of T cell phenotypes in autoimmune thyroid disease.

In this study we have correlated peripheral T cell subset phenotypes with intrathyroidal lymphocyte accumulation in patients with autoimmune thyroid disease (Graves' and Hashimoto's disease). Our study utilized euthyroid family members for one of our control groups (n = 48) thus significantly limiting familial, but not disease-specific, influences on these T cell phenotypes. Our principal new observations were found only in patients with Graves' disease. As previously reported, there was a decrease in CD8+ (suppressor/cytotoxic) T cells in the peripheral blood of patients with untreated hyperthyroid Graves' disease (n = 27) (mean +/- SEM, 19 +/- 1.1% in patients compared with 25 +/- 1.2% in controls, p = 0.03), a finding not observed in treated, euthyroid Graves' disease patients or their relatives. However, the relative number of CD8+ T cells, assessed by CD4:CD8 ratios, was increased in the intrathyroidal T cell populations (n = 10), when compared to normal and patient peripheral blood. There were no consistent changes in total CD4+ (helper) T cells in the peripheral blood of patients with treated and untreated Graves' disease but a reduction in CD4+2H4+ (suppressor-inducer) T cells was seen in patients undergoing surgery for Graves' disease (13 +/- 6.9% compared with 39 +/- 3.4%). Again, however, this T cell subset was increased within the target organ of the same patients (41 +/- 5.9%). These data point to either a selective accumulation, or a specific "homing", of certain T cell subsets within the thyroid gland of patients with Graves' disease where T cell differentiation may be strongly influenced by antithyroid drug treatment and the local immune environment.