Close Window

Pote Aimpun
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2000

Major Advisor: John H. Cross, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: A prevalence of intestinal parasites in southern Belize

Abstract

A biomedical survey of stool specimens from 82% of the population (n=672) of 5 villages in Toledo District, Belize were examined by the formalin-ethyl acetate concentration technique for the prevalence of intestinal parasitic infections. Seventy-six percent of the population was infected. The most common infection was hookworm (55%), followed by Ascaris lumbricoides (30%), Entamoeba coli (21%), Trichuris trichiura (19%), Giardia lamblia (12%), and Entamoeba histolytica (6%). The mean age of infected persons was 19 years. The frequency of infections was higher in younger age groups. Females had higher prevalence of hookworm infection than males. The living conditions of 111 surveyed households were characterized as 60% with dirt floor, 43% without toilets, 35% in overcrowded living condition, 10% using stream water and 16% drinking untreated water. A cross-tabulation and logistic regression analysis was used to identify risk and protective factors of the parasites. The risk factors for intestinal parasites were Mayan Ketchi [1.6(1,2.4)], houseworker [2.4(1.2,4.6)], and use of stream water [2.3(1.2,4.5)]. The protective factors were drinking treated water [0.4(0.2,0.9)], and wearing shoes [0.6(0.4,1)]. Prevention and control programs focusing on significant factors associated with parasite infections could save time and money by targeting populations by risk characteristics.


Joseph J. Amon
Department of Preventive Medicine and Biometrics

Doctor of Philosophy
September 2002

Major Advisor: Dr. John H. Cross, Department of Preventive Medicine and Biometrics

Dissertation Title: The Molecular Epidemiology of Malaria in Western Kenya

ABSTRACT

Malaria epidemiology reflects a complex web of inter-related factors: host, parasite, vector and environment. The nature, duration and severity of malaria infection depend upon these fixed and changing factors, and are complicated by varying levels of acquired immunity in individuals. The present research used molecular biology tools to examine three distinct topics related to malaria epidemiology: 1) the distribution of TNF-a, IFN-g, IL-6, TGF-b, and IL-10 gene polymorphisms; 2) the rate and pattern of mixed Plasmodium species infections; and 3) Plasmodium falciparum growth dynamics. The first two research topics were examined in a cohort of 248 males recruited from three highly endemic villages in western Kenya where severe malaria anemia is common. The third topic was investigated among 22 volunteers during the post-challenge phase of a malaria vaccine trial sponsored by Walter Reed Army Institute of Research. Individuals in the Kenya cohort were found to have a marked bias toward genotypes associated with low expression of IFN-g and IL-6, cytokines that, at high plasma levels, have been previously implicated in anemia and poor malaria outcomes. By contrast, the frequency of the TNFa -238A allele, which has been associated with severe malarial anemia, was found to be similar to frequencies reported in a number of diverse populations. Over the course of three malaria transmission seasons, non-falciparum malaria was consistently seen in a minority (5-10%) of volunteers, most commonly as a dual infection with P. falciparum. In contrast to previous reports of long-term stable parasite density among individuals with mixed species infections, in 80% of cases, multi-species infections were detected when individuals had their highest rate of parasite density. Among clinical trial volunteers, a highly sensitive real-time PCR assay found a ~48 hour periodicity in parasite density and a relatively wide range in parasite multiplication rate. Four individuals were identified as having some degree of resolution of infection, and an additional five volunteers were found to have a delayed pre-patent period. These results demonstrate the potential of molecular epidemiology and illustrate the subtle and complex relationship developed between humans and malaria parasites over millions of years of co-evolution.

Key words: malaria, epidemiology, molecular epidemiology, Kenya, cytokines, PCR, real-time PCR, molecular beacon probes, vaccine.


Joy Austin-Lane
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2003

Major Advisor: Galen L. Barbour, M.D., Department of Preventive Medicine and Biometrics

Dissertation Title: Influences on Tobacco Control Funding Decisions:
Explaining State Variation in Appropriations

ABSTRACT

This policy research on state tobacco control funding decisions provides a conceptual framework and quantitative model for explaining this outcome. Studying resource allocation for tobacco control is important because of the impact increased funding can have on health outcomes. One published study on state variation in tobacco control appropriations reported little explanatory value of tobacco burden of disease factors. The current research used both qualitative and quantitative methods to identify factors that were useful in explaining this outcome following the state settlements with the tobacco industry.

Key informant interviews with state policy experts yielded a comprehensive list of 26 influences on state tobacco control funding decisions. In a modified Delphi process, experts completed ratings of importance and the 11 factors that received the highest ratings were: budget situation, state priorities, tobacco industry economic and political activity, the role of the governor, the role of a legislator who champions tobacco control, public opinion, dedicated tobacco control funds, tobacco control advocacy, leadership, and coalitions. A conceptual diagram was developed to portray the relationships of these factors with state tobacco control funding.

To test the explanatory value of these factors, a multiple linear regression model was developed with existing cross-sectional data on all 50 states. Five measures were significantly associated with the outcome and explained 49% of the variance. The 26 states with Gross State Product from tobacco had lower allocations for tobacco control (Tobacco Economy). The 27 states with laws preempting stricter local ordinances were less likely to allocate funds for tobacco control (Tobacco Industry Political Activity). States that had Democratic governors with a high degree of executive authority allocated more funding for tobacco control (Role of Governor). Citizen liberalism was positively associated with higher allocations for tobacco control (Public Opinion).
These results imply that a strong scientific case for tobacco control is not sufficient to secure prevention funding at the state government level. Attention to political and economic aspects of the state budget process could lead to increased tobacco control funding. Further research is needed to determine whether these factors are predictive of appropriations in future years and how the factors can be used to influence future funding decisions.


Judith Chamberlin
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2001

Major Advisor: Larry W. Laughlin, M.D., Ph.D.
Department of Preventive Medicine and Biometrics

Dissertation Title: The Epidemiology of Bartonellosis in Peru

Abstract

Bartonellosis has caused debilitating illness and death since pre-Inca times. It is a vector-borne infectious disease found in the medically underserved communities of the high Andean mountain valleys of South America. The disease manifests in two clinical stages, a hematic stage followed by an eruptive stage. The etiologic agent is Bartonella bacilliformis, an alpha-2 proteobacteria that intracellularly infects erythrocytes and endothelial cells. Although recognized for centuries, relatively little is known about its epidemiology. This study examines the epidemiology and transmission dynamics of bartonellosis with the goal of developing a rationale control program.

In January 1997 a two year, population-based prospective cohort investigation was initiated in an area of Peru with endemic bartonellosis. Using house-to-house surveys, the population was censused, interviewed and then followed for evidence of bartonellosis. Active and passive case detection methods were used to identify cases and to determine prevalence, incidence, morbidity, mortality, and chronic carrier state. An indirect fluorescence antibody assay was developed to aid diagnosis. Individuals with clinical disease were enrolled in a cases-series investigation and re-evaluated every six months to determine the natural history of disease and the IgG antibody response over time. Case-houses were mapped using geographic information systems (GIS) to determine any possible spatial patterns of disease. Changes in incidence as a function of average monthly precipitation and temperature were analyzed using climatic data obtained from the National Centers for Environmental Prediction.

Of the 690 participants enrolled in the cohort investigation, 45% had IgG antibodies to B. bacilliformis. At enrollment, 0.5% of participants had asymptomatic bacteremia. After two years of follow-up, the incidence rate was 12.7 per 100 person-years. Significant risk factors for infection included young age and living in the household of another case; 70% of the cases were clustered in 18% of the households, and 75% of the cases occurred during the 1998 "El Nino", a period when average temperatures and precipitation levels increased. The 292 participants in the case-series investigation had a case fatality rate of 6%. Of those participants who were followed, 10% had persistent bacteremia after therapy, 41% experienced continual symptoms, and within 12 months, 44% of first stage clinical cases progressed to the second clinical stage with cutaneous lesions. IgG antibodies provided little protection from continual symptoms; participant's antibody titers slowly reverted to negative at a rate of 3% per year. The quantity and character of bartonellosis in this community represents a significant public health problem. Efforts to reduce disease prevalence should focus both on host and environmental factors and on methods to more effectively eradicate bacteremia in infected individuals.


LCDR David Merrell Claborn, USN
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2001

Major Advisor: Richard G. Andre, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: Remote Sensing and Geographic Information Systems as Decision Support Tools for Malaria Control in the Republic of Korea

Abstract

Cost-comparisons are a necessary component of decision-making in the control of vector-borne disease. Remote sensing and geographical information systems (GIS) were used to estimate the size of vector larval habitats to allow a cost comparison of chemoprophylaxis and larviciding, two malaria control methods currently being considered for use in the Republic of Korea (ROK).

Two U.S. Army camps (Cp Casey and Cp Greaves) were selected as research sites. The cost of chemoprophylaxis was estimated for each population assuming a 19-week treatment regimen consisting weekly chloroquine chemoprophylaxis with terminal primaquine treatment, a single pre-treatment G-6-PD deficiency test, and a 50% turnover in personnel during the malaria transmission season. Annual cost of chemoprophylaxis was $37.53/person.

Larval habitats were sampled from June through September, 2000. Anopheline larvae were reared to the adult stage and identified using adult and pupal morphological characteristics. Gene sequencing and random amplification of polymorphic DNA (RAPD) analyses were performed to confirm the identification of the mosquitoes. Both molecular work and pupal morphology indicated that most of the anophelines collected during the study were Anopheles sinensis Wiedemann and that the abundance of other anopheline species is generally overestimated by adult keys.

The size of vector larval habitats within the 1-km flight range of An. sinensis around the two U.S. Army camps was determined using satellite-acquired images, and the cost of treating those areas with an insect-growth regulator was estimated. At Cp Greaves, the cost of the requisite three larvicidal applications exceeded the cost of chemoprophylaxis by a factor of four, but at Cp Casey, chemoprophylaxis was about 21 times as expensive as larviciding due to the larger number of at-risk personnel and the smaller size of the vector habitats.

This study demonstrates the usefulness of remote sensing and GIS as decision support tools for estimating costs of vector control methods used in the control of vector-borne disease. Such information can then serve as one component of the decision-making process for the design and implementation of disease prevention strategies.

Key Words (Indexing): Anopheles sinensis, Anopheles lesteri, Anopheles yatsushiroensis, malaria, Plasmodium vivax, remote sensing, larviciding, chemoprophylaxis, cost- comparison.


John Paul Grieco
Department of Preventive Medicine and Biometrics

Doctor of Philosophy
2001

Major Advisor: Donald R. Roberts, Department of Preventive Medicine and Biometrics

Dissertation Title: The bionomics and vector competence of Anopheles albimanus Weidemann and Anopheles vestitipennis Dyar and Knab (Diptera: Culicidae) in southern Belize, Central America

ABSTRACT

During an 18 month study period (1997-1998) in the Toledo District, southern Belize, a series of studies were undertaken to evaluate the vector competence of the Anopheles mosquitoes. At both Golden Stream and Rancho, the population densities of An. vestitipennis showed positive correlation with human cases of P. Vivax. and tested positive for P. Vivax. 210 (MFIR of 0.01 and MFIR of 0.0004, respectively). Population densities of An. darlingi from Golden Stream showed a negative correlation with high numbers of malaria cases but also tested positive for P. Vivax. 210 (MFIR of 0.001). Population abundance of An. albimanus from Rancho showed a negative correlation with malaria transmission and was not found naturally infected with malaria parasites. The feeding preference study showed An. vestitipennis to prefer feeding on human blood while An. albimanus showed a strong tendency for feeding on cattle and non-human hosts.

Human landing collections showed An. albimanus exhibited a bimodal biting pattern with an 0:1 ratio of 1 :0.1. Anopheles vestitipennis , however, exhibited a high level of biting activity throughout the night and a strong endophagic behavior (0:1 ratio of I :0.9). Post spray, deltamethrin functioned primarily as an irritant, causing premature exiting behavior. DDT exerted a powerful repellency effect resulting in a 97% post-spray reduction of An. vestitipennis females entering the hut.

Results from the comparative susceptibility study showed An. darlingi maintains a high salivary gland infection rate (41.0%) with P. falciparum (NF -54). Anopheles vestitipennis had a low salivary gland infection rate of9.3% with moderate numbers of sporozoites (i.e. 85.7% containing 50-250 sporozoites). Anopheles albimanus was the least susceptible to salivary gland infection (Golden Stream, 0.0% and Buena Vista, 2.2%).

Although the rate of infection for An. vestitipennis is low, its strong endophagic and anthropophagic behaviors augment its overall vector potential. This species clearly demonstrates characteristics of an important vector of malaria in Belize. The opposite appears to be true of An. albimanus . A low infection rate and both exophagic/zoophagic behaviors reveal this species to be a poor vector of P. falciparum.

Key Words (Indexing): Anopheles albimanus, Anopheles vestitipennis , DDT, deltamethrin, comparative susceptibility, experimental huts, Belize.


Shilpa Hakre
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2003

Major Advisor: Donald Roberts, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: The Epidemiology of Malaria in Belize, 1989-1999

ABSTRACT

Retrospective analyses were conducted to describe the epidemiology of malaria in Belize over a 10-year period and to determine if environmental factors influenced the incidence of malaria at macro- and micro-scales. The purpose was to contribute to the current body of knowledge regarding malaria transmission in Belize, which may aid in region-specific malaria control efforts.

Malaria data were obtained from the National Malaria Control Program's National Malaria Database. Malaria rates were calculated using the 1991 population census of Belize (Central Statistics Office). Other data were obtained from the National Meteorological Service (Belize), National Center for Environmental Prediction (NCEP), National Weather Service (U.S.A.), Global Land One-Kilometer Base Elevation (GLOBE), National Geophysical Data Center, Land Information Centre (LIC, Belize), village and vector surveys, Ministry of Health (MOH), SPOT and Landsat multi-spectral images. Most of the data from these sources were collected for purposes other than the specific aims of the studies in this dissertation.

Over a 10-year period, malaria incidence rates varied temporally and spatially; southern and central areas of Belize had consistently higher rates of malaria than northern areas. Toledo District had the highest Plasmodium vivax incidence; whereas, Stann Creek District had the highest P. falciparum incidence. Malaria incidence was highest during 1993 through 1996. Plasmodium falciparum incidence was highest in the transitional months preceding the wet season in Stann Creek.
Vector surveys conducted in villages in the districts of Cayo, Stann Creek, and Toledo indicated that Anopheles darlingi was most common and abundant in Stann Creek District whereas, An. albimanus was most common and abundant in the other two districts. The epidemiology of P. falciparum incidence among very young children in Stann Creek, along with the common occurrence and seasonal abundance of An. darlingi, indicates that malaria transmission occurred locally and An. darlingi was the likely vector of P. falciparum in that district.

Preliminary results indicated malaria incidence differed geographically by season, type of vegetation, and proximity of villages to rivers or streams. Examination of associations between weather and malaria incidence indicated that precipitation was associated with malaria transmission. Higher total rainfall was associated with a higher malaria risk in villages. An assessment of the relationship between rainfall and malaria incidence in microenvironments, represented by districts, indicated that the relationship seen for the country was especially significant in Cayo and Toledo Districts, where higher rainfall increased malaria risk in villages; whereas, the opposite was seen for Corozal and Orange Walk. Examination of the relationship between vegetation and malaria incidence indicated that more forest cover was associated with higher risk of malaria in villages. This relationship was specifically seen in Belize, Cayo, and Stann Creek Districts.

Environmental risk factors and malaria incidence were assessed in households in San Martin, Cayo District and Red Bank, Stann Creek District. In San Martin, in 1997, proximity of a household to a stream, number of male occupants in a household, and having a history of malaria in a household were predictive of whether a household had malaria. In San Martin, malaria incidence was highest in males, especially in the 11 to 14 and 35 to 39 age groups. In Red Bank, in 1997, having a history of malaria in a household, construction of the outer walls, and the number of females in a household were predictive for malaria in a household. The 0 to 4 year-old age group had the highest malaria incidence in Red Bank.
Vector surveys conducted in 1997 and 1998 in both study villages showed An. albimanus was most common in San Martin; whereas, An. darlingi was most common in Red Bank. Anopheles darlingi and An. vestitipennis were also collected in San Martin, and An. albimanus, but not An. vestitipennis, was also collected in Red Bank in addition to An. darlingi. Malaria cases clustered by household in both villages. In San Martin during 1993 through 1998, only three to eight percent of households produced 50 percent or more of malaria cases. Similarly, in Red Bank during 1993 through 1998, only five to 12 percent of households produced 50 percent or more of malaria cases.

Malaria incidence varied by region, year, season, and populations in Belize during the study periods and these variations are linked to differences in environmental variables. Malaria control efforts might be more effective if environmental variables were accommodated in malaria control planning.


N. Neely Kazerouni
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2002

Major Advisor: Heidi Friedman, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: Family History of Breast Cancer as a Determinant of the Risk of Developing Endometrial and Ovarian Cancers: A Nationwide Cohort Study

ABSTRACT

Statement of the problem: Although endometrial and ovarian cancers share some of the same reproductive, hormonal, and genetic risk factors with breast cancer, it is not well established if a family history of breast cancer is associated with endometrial and ovarian cancer risk in a general population setting. We examined these associations in a prospective cohort study.

Methods: The women in the endometrial (n=37,583) and ovarian (n=49,975) cancer studies were former participants in a national breast cancer screening who were selected for additional follow-up (1979-1998). During follow-up, 648 and 362 women with endometrial and ovarian cancers, respectively, were identified. We examined information on the breast cancer history of mothers, sisters, daughters, aunts, and grandmothers of the study participants as well as the number of relatives affected with breast cancer, their age at diagnosis, and breast cancer laterality. We used Poisson regression to estimate rate ratios and 95% confidence intervals to characterize the precision of these point estimates.

Results: The presence of breast cancer in a first-degree (RR=0.96, 95% CI= 0.78-1.2) or a second-degree (RR=1.0, 95% CI=0.81-1.2) relative did not influence the risk of developing endometrial cancer. In addition, the risk of endometrial cancer did not vary by age of the relative at breast cancer diagnosis or by the number of affected relatives with breast cancer. However, there was a non-significant increase in the risk of endometrial cancer among women with a 1st degree relative with bilateral breast cancer (RR=1.4, 95% CI= 0.84-2.4) but not among women with a 1st degree relative with unilateral breast cancer (RR=0.83, 95% CI=0.62-1.1). Women with a personal history of prior breast cancer were more likely to develop endometrial cancer during the course of follow-up (RR=1.3; 95% CI=1.1-1.7), but even in this subgroup, family history of breast cancer did not confer additional risk of endometrial cancer.

On the other hand, breast cancer in a first- or second-degree (RR=1.4, 95% CI=1.1-1.7), and any second-degree (RR=1.3, 95% CI=1.0-1.7) relative, increased the risk of ovarian cancer. Participants with two or more first-degree relatives with breast cancer also had a significantly increased risk (RR=1.8, 95% CI=1.1-2.8). Risk was particularly high among women with 2 or more first-degree affected relatives, at least one of whom had bilateral breast cancer (RR=4.2, 95% CI=1.7-10) or younger age (<50) at breast cancer diagnosis (RR=2.6, 95% CI=1.4-4.8), and among women with a personal history of breast cancer who also had a first-degree relative with younger age at breast cancer diagnosis (RR=3.5, 95% CI: 1.7-7.4).

Conclusions: These results provide support for the hypothesis that a family history of breast cancer is a strong predictor of the risk of developing ovarian cancer, but is not a predictor of endometrial cancer risk.


Timothy A. Kluchinsky, Jr
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2001

Major Advisor: Philip A. Smith, Ph.D. Department of Preventive Medicine and Biometrics

Dissertation Title: Identification of CS-Derived Compounds Formed During Heat Dispersion of CS Riot Control Agent and the Temperature Ranges Associated with Their Formation

Abstract

Background: Military, correctional officer, and law enforcement personnel, are routinely exposed to 2-chlorobenzylidenemalononitrile (CS) during combat training, chemical protective mask confidence training, and riot control operations training. The public may be involuntarily exposed during rioting or civil disturbances. The potential exists for CS to form new compounds when dispersed by means of oxidizer-supported combustion of a chemical fuel from inside a canister.

Objectives: Research objectives were to identify CS-derived compounds that may be relevant to the preservation of health among those exposed, and to explore the effect of temperature on their formation. In addition to the need to identify these compounds and to understand their toxicity, knowledge of the temperature ranges associated with their formation is also important. If CS-derived compounds produced during heat-dispersion of CS canisters prove to be a public health hazard, alternative dispersion methods may be warranted..

Specific Aims: Research aims were to report on the objectives discussed above, publish the analytical methods and instrumental techniques necessary to characterize CS-derived compounds, and explore conditions leading to their formation.

Study Design: Air samples collected during heat-dispersion of CS were analyzed using open tubular gas chromatography coupled to mass spectrometry and compared to known standards using spectral interpretation techniques guided by spectral library comparison. Additional analytical methods were employed to explore formation of select inorganic analytes. A tube furnace was used to explore the effect of temperature on formation of CS-derived compounds.

Results: Compounds observed in air sample filter extracts, in addition to CS included 2- chlorobenzaldehyde, 2-chlorobenzonitrile, quinoline, 2-chlorobenzylcyanide, 1,2- dicyanobenzene, 3-(2-chlorophenyl)propynenitrile, cis and trans isomers of 2- chlorocinnamonitrile, 2 ,2-dicyano-3-(2-chlorophenyl)oxirane, 2-chlorohydrocinnamonitrile, benzylidenemalononitrile, cis and trans isomers of 2-cyanocinnamonitrile, 2- chlorobenzylmalononitrile, 3-quinoline carbonitrile, 3-isoquinoline carbonitrile, 4- Chlorobenzylidenemalononitrile, hydrogen cyanide, and hydrogen chloride. Experimentation using a tube furnace showed that formation of most of the CS-derived compounds observed was temperature dependent. The results of this research support the need for future evaluation of exposure to potentially harmful CS-derived compounds. Control of such exposures through implementation of appropriate personal protective equipment or alternative delivery methods may be warranted.


Elizabeth M. Maloney
Department of Preventive Medicine and Biometrics

Doctor in Public Health
2002

Major Advisor: Terry L. Thomas, Ph.D., National Cancer Institute

Dissertation Title: A Cohort Study of Health Effects of HTLV-I Infection in Jamaican Children and their Associations with Viral, Immunologic and Host Genetic Markers

Abstract

Statement of the problem: Human T-lymphotropic virus type I (HTLV-I) infection is associated with infective dermatitis of childhood. Early childhood infection is also thought to play a role in development of a rare malignancy associated with HTLV-I, adult T-cell leukemia/lymphoma (ATL). ATL develops in <5.0% of persons infected with HTLV-I in childhood. Identifying markers associated with increased risk of ATL among infected persons could be used to target persons for early clinical intervention. Several case reports have documented ATL patients with childhood histories of infective dermatitis. Infective dermatitis may be a cutaneous marker of risk for ATL. There may be additional health effects of HTLV-I infection in children that could be potential markers of risk for development of ATL, however this age group has not been well studied.

Methods: A cohort study of 28 HTLV-I infected and 280 uninfected children born to women who attended one of two antenatal clinics in Kingston, Jamaica between January, 1989 and August, 1990 were enrolled in this study. Children received physical examinations and phlebotomy at clinic visits scheduled from six weeks to ten years of age. The primary analysis compared incidence rates for targeted health outcomes between HTLV-I infected and uninfected children. Based on the results of the primary analysis, the secondary analysis examined associations of HTLV-I-associated health outcomes with pre-diagnostic levels of viral and immunologic markers, as well as host genetic markers among HTLV-I infected children. Additionally, levels of immunologic markers at the time of diagnosis were described in infected children with a specific health outcome.

Results: HTLV-I infected children had significantly increased incidence rates of seborrheic dermatitis, eczema and hyperreflexia of the lower limbs compared to HTLV-I uninfected children. Additionally, compared to uninfected children, HTLV-I infected children had elevated rates of lymphadenopathy, severe anemia, abnormal lymphocytes and a decreased rate of eosinophilia that were of borderline statistical. Among infected children, seborrheic dermatitis and severe anemia were associated with elevated HTLV-I proviral loads in pre-diagnostic specimens. Children with seborrheic dermatitis had elevated pro-inflammatory cytokines at the time of diagnosis, but not 12 months post-infection. Health outcomes among infected children were not associated with the HLA Class II alleles studied.

Conclusions: HTLV-I infection in children may be associated with seborrheic dermatitis and eczema. Further study is needed to confirm these associations and clinically define these HTLV-I associated diseases. Other abnormalities associated with HTLV-I have previously been reported in association with HTLV-I in adults. HTLV-I infected children with seborrheic dermatitis, eczema or severe anemia had elevated levels of proviral load 12 months post-infection that were of borderline statistical significance. At the time of diagnosis, infected children with seborrheic dermatitis had elevated levels of pro-inflammatory cytokines. HTLV-I associated seborrheic dermatitis may be an obvious marker of immune system dysregulation in children.


Lisa M. May
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2003

Major Advisor: Arthur Lee, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: The Military Deployment Human Exposure Assessment Study (MDHEXAS): Blood and urine exposure biomarkers as environmental
surveillance tools for assessing military personnel exposure
to chemicals during deployment to Camp McGovern, Bosnia

ABSTRACT

Currently the Department of Defense (DoD) does not use exposure biomarkers to measure environmental exposures to chemicals. Blood and urine exposure biomarkers for volatile organic compounds (VOC), selected heavy metals, depleted uranium (DU), and chemical warfare agents are currently available but have not been field tested or validated in military deployments as a tool to document exposures by the DoD. The Military Deployment Human Exposure Assessment Study, a prospective cohort of 46 soldiers deployed to Bosnia, was designed to validate blood and urine exposure biomarkers as a mechanism to document exposures to these chemicals during military deployments. Blood and urine were collected pre-, during, and post deployment. Standard questionnaire was administered, and environmental and occupational monitoring methods were conducted for comparison to the exposure biomarker results. The urine depleted uranium, blood VOC, urine heavy metals, and blood heavy metals results are compared pre-, during, and post deployment and against standard US reference ranges for the same compounds. The results of the study indicate that natural uranium and styrene environmental exposures increased during deployment. Therefore, exposure biomarkers may be a valuable tool in assessing exposures and risk from environmental and occupational chemicals and hence imperative to include in a comprehensive DoD preventive medicine program.


Barry A. Miller
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2000

Major Advisor: Terry L. Thomas, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: The Impact of Sociodemographic Factors on Racial/Ethnic Differences in Tumor Stage and Tumor Size for Cancer of the Female Breast

ABSTRACT

A population-based, case-control study was conducted to determine the importance of sociodemographic factors in explaining racial/ethnic differences in tumor stage and size at the time of diagnosis among women with invasive, primary breast cancer. The study group included 106,607 women newly diagnosed with breast cancer during the years 1992 through 1996 while residing in any of the eleven reporting areas in the United States that comprise the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute (NCI).

Descriptive tabulations of the study variables indicated that Japanese and White women tended to be diagnosed at an earlier stage, with smaller diameter tumors, and at a lower tumor grade than other groups. Black and Hispanic women were more likely than other groups to be diagnosed with metastatic disease, with tumors 2 cm or larger in diameter, and with poorly differentiated tumors. In the regression analysis, elevated odds ratios among Black and Hispanic patients for later stage and larger size tumors were reduced by 50% to 60% when sociodemographic factors were added to a model already containing age and geographic area. Tumor grade and hormone receptor status only explained a small amount of the excess odds for distant stage disease among Black and Hispanic women, and did not explain any of the racial/ethnic differences in regional stage disease or larger tumor size. In the analysis of tumor size, odds ratios for Black, Hispanic, Filipino, Chinese, and Korean women remained elevated relative to White women after adjustment for sociodemographic factors, tumor grade, and hormone receptor status. Japanese women, conversely, had consistently lower odds ratios (relative to White women) for every study outcome.

Results from this study suggest that sociodemographic factors account for a significant portion of the observed racial/ethnic differences in the stage of disease and tumor size at the time of diagnosis, but that unmeasured differences in socioeconomic or biological characteristics of breast tumors among some racial/ethnic groups may also exist. The special cancer data base created for this study may now be used to investigate the importance of sociodemographic factors in explaining population patterns for other types of cancer.



Sara B. Newman
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2002

Major Advisor: Heidi Friedman, Ph.D., Department of Preventive Medicine & Biometrics

Dissertation Title: An Epidemiologic Analysis of Chlamydia trachomatis and Neisseria gonorrhoeae Infections in Female Federal Prisoners

ABSTRACT

Statement of the problem: Infections caused by Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (GC) are the most commonly reported bacterial sexually transmitted infections in the U.S. No studies have been conducted to determine the prevalence of these two infections in women federal prisoners. Information about the prevalence of and risk factors for the infections may assist the Federal Bureau of Prisons to implement a rational screening approach for CT and GC in female inmates.

Methods: Two study phases were implemented as part of this protocol. The first study phase relied on qualitative techniques: focus groups and group-based cognitive interviews were implemented to assist in the design of the study instrument and study procedures to be used in the second phase of the study. For the second, quantitative phase, urine specimens and self-collected swabs were taken from 983 women incarcerated in two federal prison facilities in the U.S. Participants completed a self-administered questionnaire on sociodemographic characteristics, and sexual and clinical history. Another questionnaire was administered to participants after specimen collection on preferences for self-collected swab, urine collection and pelvic examination. Specimens were analyzed at the Johns Hopkins University Chlamydia laboratory using strand displacement amplification technology to detect the presence of CT and/or GC DNA. Prevalence of CT and GC was calculated for each of the prison sites. Potential risk markers associated with infection were assessed. Participants' preferences of self-collected swab and urine were also compared between the two sites.

Results: CT infection was found in 1.3% of the participants at one site where women are screened when they enter the prison, and in 2.3% at the other site, where women are not screened. One case of gonococcal infection was detected at the site where women are screened and no cases in the other site. Among women age 18-22, prevalence of CT infection was 8.5% in the prison with the highest prevalence of infection. Prevalence of CT infection among women age 30 and younger exceeded 3.5%. Screening women age 30 and younger would identify more than 60% of cases at an estimated cost of less than $60,000 per year. Approximately 83% of infections could be detected if women age 35 and younger were screened, but the cost for screening would approach $90,000.

More than half of the participants (57%) found no difference between giving urine or swab samples in terms of ease of collection. Approximately 30% of participants said they would prefer to give a swab specimen in the future as compared to urine (21%), but nearly half of women expressed no preference for one method over the other. Most participants (60%) expressed a preference for doing a self-collected swab rather than having a pelvic exam (23%) to test for the infections in the future, but nearly 17% had no preference for one over the other.
Conclusions: While prevalence among the study population was low, targeted screening in women age 30 and younger is recommended to detect more than half of the cases cost-effectively. This study provided evidence that to inmates noninvasive screening techniques are acceptable alternatives to pelvic examination.



Michael R. Sardelis
Department of Preventive Medicine and Biometrics

Doctor of Philosophy
2001

Major Advisor: Richard G. Andre, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: Potential for Aedes albopictus and Ochlerotatus j.japonicus to change the field ecology of arboviruses of human health importance in the mid-Atlantic region of the United States

ABSTRACT

Aedes albopictus and Ochlerotatus J. japonicus are mosquitoes that have been recently introduced into the United States. Since their introduction, they have been implicated in the transmission of one or more of the arboviruses of principal public health importance in the mid-Atlantic region. To more fully understand the potential of Ae. albopictus and Oc. J. japonicus to be vectors of endemic arboviruses, field and laboratory studies were conducted to assess their distribution in the region and to determine their vector competence for arboviruses for which data are lacking.

Aedes albopictus, a mosquito found in virtually all counties of states in the southeastern United States, was found to be established as far north as south-central Pennsylvania. Ochlerotatus J. japonicus, a mosquito that is generally found in more northern climates within its native range, was found to be established as far south as Maryland, and its relative abundance in Frederick County, Maryland, was found to be comparable to or greater than that of other container- inhabiting mosquito species.

Laboratory studies showed that Oc. j. japonicus is a competent vector of West Nile (WN), St. Louis encephalitis (SLE), eastern equine encephalitis (EEE), and La Crosse (LAC) viruses. Estimated transmission rates for these viruses by Oc. J. japonicus and by known principal or suspected vector mosquito species were determined concurrently for comparison. Based on these comparisons, Oc. j. japonicus was a more efficient laboratory vector of WN virus than Culex quinquefasciatus, as efficient as Culex pipiens in transmitting SLE virus, a less efficient laboratory vector of EEE virus than Ae. albopictus, and as efficient as Aedes triseriatus in transmitting LAC virus. Additional studies indicated that the extrinsic incubation period (EIP) for WN and EEE virus in Oc. j. japonicus held at 26C was between 7-11 days and around 5 days for, respectively.

Experimental transmission studies showed that North American (NA) strains of Ae. albopictus were competent vectors of WN virus. Transmission rates varied among the NA strains tested, ranging from 36-92%. The EIP for WN virus in Ae. albopictus held at 26oC was estimated to be 10 days. In a study to evaluate vertical transmission, no virus was recovered from >12,000 F1 progeny.
This study provided key information for implicating Ae. albopictus and Oc. j. japonicus as vectors of arboviruses of public health importance. These data, combined with the behavioral and other biological characteristics of Ae. albopictus and Oc. j. japonicus, indicate that these two introduced species could change the field ecology of arboviruses of human health importance in the mid-Atlantic region of the United States.


Diana L Schneider
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2000

Major Advisor: Heidi B. Friedman, Ph.D., Department of Preventive Medicine and Biometrics

Dissertation Title: Evaluation of Cervicography Screening for Cervical Cancer in a High-Risk Population

Abstract

Statement of the problem:, CervicographyTM was first described in 1981 as a visual screening system for early detection of cervical neoplasia and cancer. Early studies to assess the validity of cervicography showed the method to have an acceptable sensitivity but an unacceptably low specificity for mass screening. Following revision of the cervicography classification scheme, specificity improved, but at the expense of lowered sensitivity. Most previously published studies have had some methodologic inadequacies which may have affected the outcome.

Methods: Cervigrams were taken for 8460 women who enrolled into a population-based, natural history study of cervical neoplasia in Guanacaste Province, Costa Rica. Cervicography and three cytologic screening tests were the basis for referral for colposcopic examination and directed biopsy. Initial cervicography classification was compared with a referent diagnosis determined by histology and three cytologic tests, cytology, and presence of cancer-associated human papillomavirus types. Cervicography was submitted to additional review and arbitration to achieve an optimal classification. Interobserver agreement was assessed, and the performance of the optimal cervicography result was compared with the referent diagnosis. Sensitivity, specificity, and predictive values were estimated, and results were stratified by characteristics of the woman and visual characteristics of the cervigram image. Digital colposcopic images were interpreted to evaluate the perceived appropriateness of the decision to biopsy and biopsy placement, and the impact of these on sensitivity and specificity.

Results: Moderate agreement on cervigram classification was observed (kappa=0.47 when cervigram results were classified into seven categories and 0.54 when cervigram results were classified into dichotomous categories of referred for colposcopy versus not referred). For the detection of high grade squamous intraepithelial lesion or cancer, optimized cervicography yielded a sensitivity of 55.2% and a specificity of 94.3 %, which was only slightly improved over the initial estimates of 49.3% sensitivity and 95.0% specificity at enrollment. Higher sensitivity was associated with younger age, premenopausal status, the presence of metaplasia, the absence of cervicovaginal atrophy, and improved quality of the acetic acid effect.

Conclusions: Evaluator agreement with cervicography is moderate. The arbitrated cervigram classification improved the performance of cervicography only slightly over a single interpretation. Cytology performed better than cervicography for the detection of high grade squamous intraepithelial lesions, but the two methods performed similarly for the detection of invasive cervical cancer. Cervicography is not recommended for postmenopausal women and/or women ages 50 and older.


Stephanie L. Scoville
Department of Preventive Medicine and Biometrics

Doctor of Public Health
2002

Major Advisor: COL John Gardner, MC, USA, Department of Preventive Medicine and Biometrics

Dissertation Title: Mortality During U.S. Armed Forces Basic Training:
A 25-year Review (1977-2001)

ABSTRACT


Background: Efforts to understand and prevent the rare, but tragic, occurrence of death among healthy, young military recruits trying to serve their country depend upon medical surveillance data and accurate determination of mortality rates by specific cause. The purpose of this study was to create a Recruit Mortality Registry that includes deaths that have occurred during Air Force, Navy, Marine Corps, and Army basic military training from 1977 through 2001, and to describe the epidemiology of recruit mortality due to traumatic and nontraumatic deaths.

Methods: Recruit deaths were identified and confirmed through redundant sources. Attempts were made to obtain a complete file on each death including all available medical and personnel records. Demographic, circumstantial, and medical information was recorded on an abstraction form developed for this study. Mortality rates per 100,000 recruit-years were calculated using recruit accession data from the Defense Manpower Data Center.

Results: There were 276 recruit deaths from 1977 through 2001 identified through the Recruit Mortality Registry. Age-specific recruit mortality rates were less than half of same-age U.S. civilian mortality rates. Between Services, the age-adjusted mortality rates were highest in the Army and lowest in the Air Force and Navy, with the Marine Corps in between. The majority (71%) of recruit deaths were classified as nontraumatic and 69% (136/196) of these were exercise-related. Of the exercise-related deaths, 42% (57/136) were cardiac deaths and at least one-third (45/136) were related to heat stress. Infectious agents accounted for 24% (48/196) of the nontraumatic deaths. Only 29% (80/276) of recruit deaths were classified as traumatic. Of these, the majority (58%) were due to suicide, followed by unintentional injuries (37%), and homicide (5%).

Conclusions: Comparison of military recruit mortality rates with the same-age U. S. civilian population establishes the safety of the basic military training environment. This can be attributed to selection factors in inducting healthy recruits, a well-supervised training environment, the tremendous focus on safety during recruit training, and lack of access to alcohol and motor vehicles. Preventive measures focused on reducing heat stress during exercise may be effective in reducing high rates of exercise-related death.