Includes bibliographical references. / [Background] Research in adults and older children has shown an association between Mycobacterium tuberculosis and helminth infection, with those infected with helminths at greater risk of tuberculosis. This association is believed to be on the basis that chronic helminth infection can result in a functional impairment of the immune response that is necessary to clear or control infection by Mycobacterium tuberculosis (Elias et al. 2001; Rook et al. 2006; Fincham 2001). It is thus possible that the introduction of regular deworming programmes in a vulnerable population of children under the age of five years could assist their immune systems to ward off tuberculosis infection and reduce the risk of tuberculosis disease in such a population. A randomised controlled trial to compare two methods of administering bacille Camlette-Guerin (BCG) vaccination to newborns from a high tuberculosis risk population provided an opportunity to test this hypothesis in a sub-study. [Objective] The objective of this study is to determine if young children in a high-risk tuberculosis population who have been vaccinated with BCG at birth and have been treated for helminth infection are at lower risk of tuberculosis disease than children who have been vaccinated with BCG at birth but not treated for helminth infection. [Method] A case control study nested within a cohort recruited for a separate randomised control trial to compare two methods of administering BCG vaccination was carried out. Children who presented to their local clinic or hospital with symptoms of tuberculosis or a history of exposure to tuberculosis were admitted to a case verification (CV) ward for investigation of tuberculosis. Investigation of tuberculosis included a detailed history, including past helminth treatment, physical examination, tuberculin skin test, chest radiograph, gastric washing and induced sputum for culture of tuberculosis and clinical examination. A diagnostic algorithm was developed by specialist physicians and biostatisticians to classify the children into one of five tuberculosis categories. A total of 510 children (median age 18.13 months) were included in the primary analysis of this case control study. Those defined as cases were the 328 classified as "definite or probable TB" and 182, classified as "not TB", comprised the control group. Those classified as "possible TB" or "unlikely TB" were excluded. A secondary analysis was performed that included the 337 children who had been classified as "unlikely TB" with the controls resulting in a total of 847 children (median age 18.37 months). The 328 children classified as "definite or probable TB" were defined as cases and the 519 classified as "unlikely or not TB" comprised the control group. Univariate analysis was used to explore a possible relationship between tuberculosis and helminth treatment using all the variables in the sub-study (n=510 primary analysis; n=847 secondary analysis). For both the primary and secondary analysis a multivariate logistic regression model was built using a reduced sample that had a complete set of data for all the variables: primary analysis (n=435); secondary analysis (n=724). This final model was then fitted on a more complete sample as the final variables selected had fewer missing data for the observations: primary analysis (n=493); secondary analysis (n=822). [Result] A total of 35.69% of the study sample in the primary analysis had been treated for helminth infection. The proportion of children who had been treated for helminth infection was similar in the cases and controls (35.98% and 35.16% respectively). Univariate logistic regression showed no association between tuberculosis and treatment for helminth infection: [odds ratio (OR) 1.04; 95% confidence interval (CI) 0.71 - 1.51]. Multivariate analysis adjusted for the effect of nutritional status, recorded as height for age z score (haz), number of occupants sharing the same dwelling as the child, gender and birth site showed a similar result: (OR 1.03; 95% CI 0.69 " 1.53). The OR is very close to 1 with a 95% CI that includes 1, which indicates that there is not a statistically significant association between tuberculosis and helminth treatment. In the secondary analysis, a total of 38.61% of the study sample had been treated for helminth infection. In this analysis the proportion of children who had been treated for helminth infection showed a difference between the cases and controls (35.98% and 40.27% respectively). Univariate logistic regression showed a 17% relative reduction in tuberculosis odds but this was not a statistically significant result: (OR 0.83; 95% CI 0.63 " 1.11). Multivariate analysis adjusted for the effect of haz, number of children sharing the same dwelling as the child and gender, showed a similar result: (OR 0.85; 95% CI 0.63 " 1.15). [Conclusion] The primary analysis of this observational study does not support the hypothesis that helminth treatment reduces the risk of tuberculosis disease in young children in a high-risk tuberculosis population. Although the secondary analysis showed a 15% relative reduction in tuberculosis odds after adjusting for the effect of haz, number of occupants sharing the same dwelling as the child and gender, this was not a statistically significant result. [Final Conclusion] This study does not support the hypothesis that helminth treatment reduces the risk of tuberculosis disease in young children in a high-risk tuberculosis population.
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:uct/oai:localhost:11427/9315 |
| Date | January 2009 |
| Creators | Workman, Lesley |
| Contributors | Ehrlich, Rodney, Little, Francesca, Hatherill, Mark |
| Publisher | University of Cape Town, Faculty of Health Sciences, Department of Public Health and Family Medicine |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Master Thesis, Masters, MPH |
| Format | application/pdf |
Page generated in 0.0023 seconds