Multiple Imputation for Partial Recording Periodontal Examination Protocols.

AIM: Partial-mouth recording protocols often result in underestimation of population prevalence and extent of periodontitis. We posit that multiple imputation of measures such as clinical attachment loss for nonselected tooth sites in partial-mouth samples can reduce bias in periodontitis estimates. METHODS: Multiple imputation for correlated site-level dichotomous outcomes in a generalized estimating equations framework is used to impute site-level binary indicators for clinical attachment loss exceeding a fixed threshold in partial-mouth samples. Periodontitis case definitions are applied to the imputed "complete" dentitions, enabling estimation of prevalence and other summaries of periodontitis for partial-mouth samples as if for full-mouth examinations. A multiple imputation-bootstrap procedure is described and applied for point and variance estimation of these periodontitis measures. The procedure is evaluated with pseudo-partial-mouth samples based on random site selection protocols of 28 to 84 periodontal sites repeatedly generated from full-mouth periodontal examinations of 3,621 participants in the 2013 to 2014 National Health and Nutrition Examination Survey (NHANES) survey. RESULTS: Multiple imputation applied to partial-mouth samples overestimated periodontitis mean extent, defined as the number of sites with clinical attachment loss 3 mm or greater, by 9.5% in random site selection protocols with 84 sites and overestimated prevalence by 5% to 10% in all the evaluated protocols. CONCLUSIONS: In the 2013 to 2014 NHANES data, multiple imputation of site-level periodontal indicators provides less biased estimates of periodontitis prevalence and extent than has been reported from estimates based on the direct application of full-mouth case definitions to partial-mouth samples. Multiple imputation provides a promising solution to the longstanding, vexing problem of estimation bias in partial-mouth recording, with potential application to a wide array of case definitions, periodontitis measures, and partial recording protocols. KNOWLEDGE TRANSFER STATEMENT: Partial-mouth sampling, while a resource-efficient strategy for obtaining oral disease estimates, often results in underestimation of periodontitis metrics. Multiple imputation for nonselected periodontal sites produces pseudo-full-mouth data sets that may be analyzed and combined to produce estimates with small bias.

E-cadherin autoantibody profile in patients with pemphigus vulgaris.

BACKGROUND: Pemphigus vulgaris (PV) is an autoimmune skin blistering disease. The main targets of autoantibodies are the desmosomal proteins desmoglein (Dsg)3 and Dsg1. Anti-E-cadherin antibody is the second most frequent antibody found in pemphigus foliaceus (fogo selvagem), but the frequency in PV is unknown. OBJECTIVES: To determine the anti-E-cadherin antibody profile in the two major subtypes of PV: mucosal PV (mPV) and mucocutaneous PV (mcPV). METHODS: Sera from 80 patients with PV and 80 controls were tested. Patients with PV were subdivided into mPV (n = 18) and mcPV (n = 62). Samples were tested by E-cadherin, Dsg1 and Dsg3 enzyme-linked immunosorbent assays (ELISAs), and immunoprecipitation coupled with Western blotting (IP-WB). RESULTS: Both mPV and mcPV sera have antibodies against E-cadherin as demonstrated by ELISA and IP-WB. Both subtypes of PV have low levels of anti-E-cadherin antibodies, but significantly higher levels than healthy controls by ELISA (P < 0·0001). No difference exists in antibody levels between subgroups (P = 0·82). By IP-WB, 78% of mcPV sera reacted to E-cadherin, vs. 33% of mPV sera tested. Correlation analysis suggests a moderate correlation between anti-E-cadherin antibodies and Dsg1 antibodies (average r = 0·61), but no correlation with Dsg3 antibodies (average r = 0·19). Patients with mPV can have lower levels of Dsg1 antibodies compared with controls by ELISA (P < 0·0001). A few mPV sera also reacted to Dsg1 protein by IP-WB (17%). CONCLUSIONS: Anti-E-cadherin antibodies are present in both major subtypes of PV. A moderate correlation exists between E-cadherin and Dsg1 antibodies. Patients with mPV can have low levels of both E-cadherin and Dsg1 antibodies.

Concurrent radiation therapy and chemotherapy followed by esophagectomy for localized esophageal carcinoma.

PURPOSE: A prospective study was performed to determine the outcome of patients with esophageal cancer who received preoperative radiation therapy and chemotherapy followed by esophagectomy, and to determine the role of preresection esophagogastroduodenoscopy (EGD) in predicting the patients in whom surgery could possibly be omitted, and the impact of surgery on survival. MATERIALS AND METHODS: Thirty-five patients with localized carcinoma of the esophagus received concurrent external-beam radiotherapy and chemotherapy followed by esophagectomy. Patients received 45 Gy in 25 fractions. Chemotherapy consisted of continuous infusion fluorouracil (5-FU; 1,000 mg/m2/d) on days 1 through 4 and 29 through 32 and cisplatin (100 mg/m2) on day 1. Patients underwent an Ivor-Lewis esophagectomy 18 to 33 days after completion of radiotherapy. RESULTS: Eighty percent of the patients had squamous cell carcinoma and 20% had adenocarcinoma. In addition, 51% had a pathologic complete response (CR). Twenty-two of the 35 underwent a preresection EGD before resection. Seventeen of the 22 (77%) had negative pathology from the preresection EGD, but seven of the 17 (41%) had residual tumor at surgery. The median survival and disease-free survival rates for all patients were 25.8 months and 32.8 months, respectively. Eighteen patients (51%) had no tumor at resection. The median survival for these patients was 36.8 months; the median disease-free survival time has not been reached. The median survival and disease-free survival rate for the patients with residual tumor in the surgical specimen were 12.9 months and 10.8 months, respectively. CONCLUSION: Preresection EGD is not reliable for determining the presence of residual disease or the patients in whom surgery could be omitted. Twenty-five percent of the patients with residual tumor in the resected surgical specimen were long-term survivors; this suggests a benefit from esophagectomy after concurrent radiotherapy and chemotherapy.

The role of subclass switching in the pathogenesis of endemic pemphigus foliaceus.

Endemic pemphigus foliaceus, like the sporadic form seen in the developed world, is mediated by IgG antibodies to desmoglein-1. We studied an endemic focus in Limao Verde, Brazil, where disease prevalence is 3.4%. We previously detected IgG antibodies to desmoglein-1 in 97% of patients, but also in 55% of normal subjects in the endemic focus, with progressively lower levels in normal subjects in surrounding areas. An environmental trigger is hypothesized to explain these and other findings. In this study we sought to determine if patients and enzyme-linked-immunosorbent-assay-positive normal subjects in Limao Verde differ in IgG subclass response to desmoglein-1. We developed a sensitive and specific subclass enzyme-linked immunosorbent assay using recombinant desmoglein-1 and standardized the assay to enable comparability between the four subclasses. We found that normal subjects have an IgG1 and IgG4 response, whereas patients have similar levels of IgG1 but a mean 19.3-fold higher IgG4 response. Patients in remission have a weak IgG4 response, and a 74.3-fold higher IgG4 response is associated with active disease. Finally, in five patients in whom we had blood samples from both before and after the onset of clinical disease, a mean 103.08-fold rise in IgG4 was associated with onset of clinical disease, but only a mean 3.45-fold rise in IgG1. These results suggest that the early antibody response in normal subjects living in the endemic area and in patients before the onset of clinical disease is mainly IgG1. Acquisition of an IgG4 response is a key step in the development of clinical disease.

Perceptions and opinions on the performance of Pap smears: a survey of clinicians using a commercial laboratory.

Office-based clinicians play a major role in cervical cancer screening in the United States, but little is known about their specific knowledge and understanding of the technical aspects of obtaining an adequate Pap smear. We surveyed 186 office-based clinicians using a major commercial laboratory for cytology services to obtain their opinions regarding the technical adequacy of Pap smears, collection procedures, and methods of communicating test results to patients. There was an 80.1% response rate. Two thirds of the respondents were men. Cytobrush use was reported significantly more by female clinicians. Respondents varied in reported rotation of the cytobrush, from 90 degrees to greater than 360 degrees. More female than male clinicians appropriately indicated that inflammation and heavy vaginal discharge caused sampling difficulties. Approximately half of the respondents believed that errors in cervical screening most often resulted from sampling the cervix or preparation of the Pap smear. Nineteen percent reported that ectocervical cells were not necessary for an adequate smear to be reported. Just over 25% of respondents indicated they did not report normal Pap smear results to the patient. The results indicate differences between clinicians by gender and specialty in reported knowledge, understanding, and technique in cervical cancer screening. The differences suggest that targeted education in this population may be important to improve the quality of cervical cancer screening. Medical Subject Headings (MeSH): screening; neoplasms, cervical; Papanicolaou smear; diagnosis, laboratory.

Recall and treatment decisions of primary care providers in response to Pap smear reports.

INTRODUCTION: Although the frequency of cervical cancer screening has been extensively studied, little is known about how clinicians decide to screen or recall patients for Pap smears. This study reports the management decisions made by office-based clinicians for 10 different Pap smear reports describing adequacy limitations and cytological diagnoses. METHODS: We surveyed 186 clinicians using a commercial laboratory in the southeastern United States and analyzed results by frequency and comparison statistics. RESULTS: Our respondents were 148 clinicians (79.6% response rate) from different specialties. There was variation in reported management of inflammation, atypia, and low-grade abnormalities (LGSIL), in regard to recall for repeat or routine testing as well as arranging colposcopy. In only 3 of 10 Pap smear results did more than 50% of respondents agree on a specific test recall interval. CONCLUSIONS: The variation in responses from office-based clinicians suggests either uncertainty or different opinions in making recall and treatment decisions for smears of limited quality even when associated with cytologic abnormalities. These differences may have relevance to outcomes, clinician workload, and costs of care in cervical cancer screening.

Simultaneous chemoradiation in the treatment of advanced head and neck cancer.

Fifty-eight patients with either advanced or unresectable squamous cell carcinoma of the head and neck were randomly selected to receive either twice daily radiation alone or twice daily radiation plus concomitant chemotherapy with cisplatin and fluorouracil (5-fluorouracil). There was no advantage in survival or time to progression with the addition of chemotherapy to twice daily radiation for patients with advanced resectable cancers. In the group of patients with unresectable cancers, however, there was a statistically significant advantage to the addition of chemotherapy, both in terms of disease-free survival and date to progression.

The effect of incarceration during pregnancy on birth outcomes.

OBJECTIVE: This study examined the effect of incarceration during pregnancy on infant birth weight. METHODS: North Carolina prison records were matched to North Carolina birth certificates and health services records to identify 94 women who were incarcerated during one pregnancy but not incarcerated during another pregnancy. Paired analyses examined differences between the pregnancies in terms of the women's characteristics and use of health services. A generalized estimating equations analysis modeled infant birth weight as a function of the number of days that a woman was incarcerated during pregnancy, accounting for the correlation between the birth weights of two infants born to the same mother and several potentially confounding variables. RESULTS: Since the women were significantly more likely to have been incarcerated during the second of the pair of pregnancies, as a group they were significantly older and had more children at the time of the incarcerated pregnancy than at the time of the home pregnancy. After controlling for important covariates, a higher number of pregnancy days spent incarcerated was found to be associated with higher infant birth weight. CONCLUSION: These findings suggest that aspects of the prison environment such as shelter and regular meals may enhance pregnancy outcomes among very high risk women. Health professionals should join others in efforts to assure that health-promoting resources such as adequate shelter, nutritional support, and substance abuse treatment programs are available to all pregnant women.

Marginal models for correlated binary responses with multiple classes and multiple levels of nesting.

A model for correlated binary data is presented. Marginal probabilities and odds ratios are allowed to have general regression structures that include multiple classes and multiple levels of nesting. Estimation is done through the generalized estimating equations approach of Liang and Zeger (1986, Biometrika 73, 13-22). They are contrasted with conditional models and recommendations for choosing between the two are given. Examples from genetic epidemiology are presented.

Robust regression for clustered data with application to binary responses.

Generalized estimating equations (GEE) can be highly influenced by the presence of unusual data points. A generalization of the GEE procedure, which yields parameter estimates and fitted values that are resistant to influential data, is introduced. Resistant generalized estimating equations (REGEE) include weights in the estimating equations to downweight influential observations or clusters. Influential observations are downweighted according to their leverage or residual in an example of correlated binary regression applied to 137 urinary incontinent elderly patients from 38 medical practices.

Characteristics and quality of Papanicolaou smears obtained by primary care clinicians using a single commercial laboratory.

BACKGROUND: Few data are available on factors associated with the quality of Papanicolaou smears performed in primary care. OBJECTIVE: To identify the patterns and proficiency of cervical cancer screening among different primary care specialties. MATERIALS AND METHODS: Clinical and cytologic data from 21,833 Papanicolaou smears, submitted to a single large commercial laboratory by 176 clinicians during a 7-month period, were correlated with individual clinician and specialty characteristics according to indexes of specimen quality. RESULTS: Obstetrician-gynecologists, nurse practitioners, and physician assistants provided screening to a younger population of women compared with family physicians, internists, and general practitioners. Factors positively associated with a greater probability of a "satisfactory" smear or the presence of endocervical cells (as a marker of adequate sampling) were increasing patient age, use of the cytobrush, and the specialty of the obstetrician-gynecologist. Satisfactory smears were not associated with any increased identification of cytologic abnormalities compared with "limited" smears. In contrast, smears with endocervical cells showed a higher proportion of abnormalities compared with specimens without such cells. CONCLUSIONS: Differences in the performance of obtaining Papanicolaou smears exist between primary care specialties, but need further clarification. The use of the cytobrush and the presence of endocervical cells are criteria that reflect clinician proficiency more realistically than the laboratory criterion of satisfactory smear.

The Carolina Breast Cancer Study: integrating population-based epidemiology and molecular biology.

The integration of epidemiology and molecular biology provides a new strategy to identify additional risk factors for breast cancer and to better understand the role played by traditionally recognized risk factors. The Carolina Breast Cancer Study (CBCS) is a population-based, case-control study designed to identify causes of breast cancer among Caucasian and African-American women who are residents of a 24-county area of central and eastern North Carolina. Information on established and potential breast cancer risk factors is obtained by personal interviews. Blood samples are collected from all consenting participants. Medical record documentation and paraffin-embedded tumor specimens are obtained for all breast cancer patients. DNA from tumor tissue is tested for a variety of molecular alterations characteristic of breast cancer. Germline DNA from blood lymphocytes is evaluated for presence of alleles increasing susceptibility to breast cancer. Statistical analyses evaluate gene-environment interaction by exploring the associations between environmental/behavioral factors and breast cancer in relation to specific molecular alterations (germline and tumor). Results will help identify high-risk women, clarify causal pathways, and hopefully contribute to the prevention of breast cancer.

Proliferation of carcinogen-damaged hepatocytes during cell-cycle-dependent initiation of hepatocarcinogenesis in the rat.

Hepatocyte proliferation and damage to DNA were characterized during the initiation phase of carcinogenesis in livers of rats that had received a single administration of the methylating agent methyl(acetoxymethyl)nitrosamine (DMN-OAc). Quiescent non-proliferating hepatocytes in intact livers did not appear to be susceptible to initiation by DMN-OAc, whereas proliferating hepatocytes in the S phase appeared to have greatest risk. To characterize the phenomenology of S-phase-dependent initiation further, the fractions of hepatocytes in the S and M phases of the cell cycle were enumerated at various times after treatment with DMN-OAc. Hepatocytes treated when in G1 experienced a delay of up to 20 h in the onset of S phase and a reduced rate of entry into the S and M cycle phases. Hepatocytes treated when in S phase experienced considerable delay in progression to mitosis due to part to inhibition of DNA replication. Hepatocytes treated when in late S/G2 also demonstrated a delay in progression into mitosis. The levels of 7-methylguanine and O6-methyldeoxyguanosine were quantified in the nuclear DNA of proliferating hepatocytes. The kinetics of removal of these lesions appeared to be first-order (half-life = 24 h). Hepatocyte risk of initiation was modeled by a function which summed over time the product of the fraction of hepatocytes in the S phase and the fraction of residual, unrepaired damage to DNA. For hepatocytes treated when in early G1, the time-weighted frequency of premutagenic DNA damage that was present during DNA replication was estimated to be less than half of that for hepatocytes treated when in early S. The results suggest that cell-cycle-dependent variation in sensitivity to initiation of hepatocarcinogenesis may be, in part, due to efficient removal of potentially carcinogenic lesions from DNA during an extended G1. The apparent high sensitivity of hepatocytes in late S/G2 suggests the contribution of additional factors.

Hepatitis B and hepatitis C in emergency department patients.

BACKGROUND: Infections with hepatitis B virus (HBV), hepatitis C virus (HCV), and the human immunodeficiency virus type 1 (HIV-1) are common in inner-city populations, but their frequency and interrelations are not well established. METHODS: During a six-week period, excess serum samples were collected, along with information on risk factors, from all adult patients presenting to an inner-city emergency department. The samples were assayed for hepatitis B surface antigen (HBsAg) and antibodies to HCV and HIV-1. RESULTS: Of the 2523 patients tested, 612 (24 percent) were infected with at least one of the three viruses. Five percent were seropositive for HBV, 18 percent for HCV, and 6 percent for HIV-1. HCV was found in 145 of the 175 intravenous drug users (83 percent), 36 of the 171 transfusion recipients (21 percent), and 5 of the 24 homosexual men (21 percent). Among black men 35 to 44 years of age, the seroprevalence of HCV was 51 percent. HBsAg was present in 9 percent of those whose only identifiable risk was possible heterosexual exposure. At least one viral marker was found in about 30 percent of the patients who were actively bleeding or in whom procedures were performed. Testing for HIV-1 alone would have failed to identify 87 percent of the patients infected with HBV and 80 percent of those infected with HCV. CONCLUSIONS: In a population of patients in an inner-city emergency room, HBV, HCV, and HIV-1 are all highly prevalent. However, routine screening for HIV-1 alone would identify only a small fraction of the patients who pose risks of severe viral infections, including HBV and HCV, to providers.