M.Sc.(Epidemiology), Faculty of Health Sciences, University of the Witwatersrand, 2011 / Title: Correlates of delayed pulmonary tuberculosis diagnosis among HIVinfected pulmonary tuberculosis suspects in rural HIV clinic in South Africa Background South Africa is among the countries most heavily affected by Human Immunodeficiency Virus (HIV) and tuberculosis. Delay in pulmonary tuberculosis (PTB) diagnosis is more prevalent in HIV-positive patients and is one of the factors associated with the high prevalence of co-infection in this population. This study sought to determine factors associated with delayed PTB diagnosis among HIV-infected PTB suspects attending an HIV/AIDS clinic in rural South Africa. Methods This was a secondary analysis of the data collected in a retrospective cohort study conducted by Rural Aids and Development Action Research Programme (RADAR). Data were collected using record review of patients assessed as PTB suspects over 2 years from January 2006 to December 2007 at Rixile clinic. TB diagnosis delay was defined as PTB diagnosis after 8 weeks (56 days) from the date of first sputum for acid fast bacilli (AFB) collection, taking into account those diagnosed by culture as it takes up to 8 weeks to culture mycobacterium tuberculosis using Lowenstein Jensen method. Results PTB diagnosis delay was found in 78/162 (48.15%) of the participants with subsequent TB diagnosis. Median delay was 55 days (IQR = 20 – 302). The delay was between 1 to 30 days in 27/78 (34.62%) participants, between 31 to 180 days in 26/78 (33.33%) participants and 25/78 (32.05%) participants remained undiagnosed for more than 180 days. Factors predicting PTB diagnosis delay in multivariate analysis were older age > 40 years (adjusted OR 3.43 95%CI 1.38 – 8.55, p=0.008), high HIV viral-load (adjusted OR 3.13 95%CI 1.13 – 8.71, p=0.03) and being on Antiretroviraltherapy (ART) at the time of PTB diagnosis (adjusted OR 4.19 95%CI 1.66 – 10.58, p=0.002). Conclusion There is a considerable delay from PTB suspicion to diagnosis in these rural HIVinfected patients. Older patients, those with elevated viral load and those who are on ART at the time of PTB diagnosis are at higher risk of PTB diagnosis delay. Therefore active and collaborative efforts to reduce the PTB diagnosis delay are very essential.

A research report submitted to the School of Public Health, University of the Witwatersrand, Johannesburg, in partial fulfillment of the requirements for the Degree of Master of Public Health. / Introduction: Sero-discordant couples are a population at risk for new HIV infection in Swaziland due to heterosexual transmission from an HIV-infected partner to the other. Male circumcision and condom use have been key HIV prevention strategies for negative partners. The overall objective of this study was to determine sero-discordance rate and behavioural factors associated with HIV status among couples aged 18-49 years from Matsapha in Swaziland in 2013. Methods: This is a secondary analysis of the Matsapha voluntary counselling and testing (VCT) couple dataset of 2013 in Swaziland. The aim of the routine VCT couple data collection was to assess HIV prevalence in Swaziland amongst couples, aged 18-49. In the routine VCT data collection, a blood sample for HIV testing was taken and a Swaziland HIV testing and counselling (HTC) client record form was administered to all individuals accessing the service as couples in 2013 in Swaziland. For this study the sample size was 214 couples. The sample was limited to men and women aged 18-49 years documented as couples (in a relationship) where HIV status was known for both partners. Logistic regression analyses were conducted to detect associations between demographic factors, female and male sero-discordant characteristics and risk factors (i.e. condom use, sexual transmitted infections (STIs), number of sexual partners in the past 6 months, male circumcision). Results: Of 214 couples identified in the Matsapha VCT dataset, couple HIV prevalence in this study was found to be 13.6%. Couple HIV prevalence was as follows: 86.4% were concordant negative, 7.5% were HIV sero-discordant and 6.1% were concordant positive couples. Of the 16 sero-discordant couples, 11 of the HIV positive partners were women. The most significant independent predictor of at least one partner being HIV-infected (versus both partners HIV-uninfected) was the male partner having two or more sexual partners (AOR= 7.6; 95% CI 1.7 to 34.5; P=0.01). In addition there was some evidence that couples who reported only using condoms “sometimes” were at a greater risk of HIV-infection than couples who reported condom use “always” (AOR= 3.2; 95% CI 0.8 to 12.5 ; P=0.09). Conclusions: Overall, the findings in this study show positive associations between two or more sexual partners and the outcome variable of HIV infection. The study indicated that couple HTC programs should help HIV negative partners in sero-discordant relationships remain uninfected by promoting faithfulness to one sexual partner. / MT2017

A research report submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Medicine in the speciality of Psychiatry. Johannesburg, September 2015 / BACKGROUND Human Immunodeficiency virus infection remains a fearsome health condition and South Africa remains the epicentre of the pandemic. The relationship between HIV and psychiatric disorders is clearly established. Patients in mental institutions fit into the definition of key populations due to the higher incidence of the condition in relation to the general population. Whereas the prevalence in the general population has been thoroughly researched over the years, the magnitude and distribution of HIV/AIDS in psychiatric patients is not consistently studied and little is known about the sub-group of forensic patients in South Africa. The aim is to determine the prevalence of HIV and describe the socio-demographic characteristics of inpatients in a forensic unit in a psychiatric hospital in South Africa. OBJECTIVES The objectives of this research study are:  To describe the socio-demographic characteristics of the forensic inpatients.  To describe the psychiatric disorders of forensic inpatients.  To describe the types of crime necessitating admission to the forensic unit.  To ascertain the HIV seroprevalence of forensic inpatients. METHODS A cross-sectional study design was employed. Data was extracted retrospectively from inpatient records on socio-demographic characteristics and health status, including HIV status, using a standardised data collection tool. The study population consisted of all adult patients admitted at Sterkfontein Hospital from 01 January 2007 - 31 December 2011 on forensic grounds, in terms of Section 42 of the Mental Health Act 17 of 2002, with a concomitant psychiatric disorder and undergoing care at the forensic unit. Ethical clearance was obtained from the Human Research Ethics Committee of the University of the Witwatersrand and the Hospital Research Committee. RESULTS A total of 137 subjects were studied. The study subjects comprised of mainly males (92%), a large number of unmarried people (84.7%) and people who were unemployed (89.8%) at the time of admission. The mean age was 36 years with 75% of the inpatients aged less than 43 years on admission. Two diagnoses accounted for over 80% of the patients with an Axis l diagnoses, namely: schizophrenia (49.1%) and psychosis secondary to GMC (31.2%). Axis II diagnoses were made in 67 patients and the conditions were mainly Mental Retardation (MR) and Anti-social Personality Disorder (ASPD), both accounting for 94.2% of the diagnoses made. Multiple crimes were committed by 21% of study subjects and serious offences included: rape (45.3%); assault with grievous bodily harm (32.9%); theft (19.7%) and murder (19%). These accounted for almost 90% of the crimes that led to admission. Serological test for HIV were done on 124 (90%) of the patients and the HIV seroprevalence proportion was 21.8%. The highest prevalence of 37.5% was found in patients with bipolar mood disorder followed by mood/psychosis secondary to GMC and psychosis NOS at 35.3%. However, there was no statistically significant association between HIV status and Axis l diagnoses. Further analysis also revealed no statistically significant association, at p<0.05, between HIV status and Axis II diagnosis, gender, age, marital status, educational status, employment status, access to disability grants and substance use. However, previous crime was statistically significantly associated with HIV status. CONCLUSION AND RECOMMENDATION The seroprevalence of HIV in the Sterkfontein Hospital Forensic Unit is higher than the general population and justifies recognition of forensic psychiatric patients as a key population. With the exception of previous crime, there is no clinical (DSM Axis I and II diagnoses) nor socio-demographic factor associated with HIV status. Furthermore, HIV test was not done for 10% of the patients. Efforts should be made to make an early diagnosis among all patients, and manage the condition appropriately, with treatment offered to all eligible patients. Policies and procedures should be formalised to obtain 100% diagnostic testing of all patients. A comprehensive programme of action should be instituted and maintained to prevent transmission within the institution and break the cycle of transmission within the general society.

Lymphadenopathy is common in HIV-infected individuals as lymphoid tissue is a major target and reservoir of the Human immunodeficiency virus (HIV). Lymphadenopathy may occur at any stage of HIV infection. Lymphadenopathy is also a common clinical problem confronting the primary care physician. When lymphadenopathy occurs in the setting of underlying immunodeficiency, both benign and malignant aetiologies need to be considered. Indeed, in our study of 43 patients with HIV seropositivity, where a lymph node biopsy was performed (i.e. in 40 patients), TB was the most common cause of significant lymphadenopathy (16/40 - 60%), followed by malignancy (10/40 - 25%). However, 9/40 patients (22.5%) in this group, also had reactive lymphadenopathy, which may or may not be related to the HIV. The primary objectives of this study were: i) to define the causes and the clinical patterns of presentation of lymphadenopathy in an HIV sero-positive population. Secondary objectives were: i) To review the appropriateness of the investigations that may suggest or exclude a possible cause of the lymphadenopathy and ii) To correlate the results of a FNA and/or lymph node biopsy when this was performed. Furthermore, iii) it was questioned whether it was possible to identify criteria indicating a need for a FNA and/or a lymph node biopsy. A total of 43 Black African adult patients, 21 males (49%) and 22 females (51%), were prospectively studied during 2004 and 2005. The median age was 33 years (range 16-52 years). The median duration of lymphadenopathy was 16weeks (3-52 weeks). A history of constitutional symptoms was most common among the patients diagnosed with TB, but did not reach statistical significance. In our study, the finding of lymphadenopathy in the cervical and axillary regions did not differentiate between malignancy, TB and reactive nodes. However, the presence of these nodes was suggestive of significant pathology i.e. malignancy or TB. The presence of hard nodes was more in keeping with a diagnosis of malignancy and/or TB. The presence of matted nodes was classical of TB. In the ten patients diagnosed with malignancy, three had significant peripheral lymphadenopathy at only a single site, emphasizing the point that malignancy may present with localised adenopathy as well as generalised adenopathy. The lymph node biopsy in the three patients with localised adenopathy showed non-Hodgkin’s lymphoma. Lymphadenopathy in HIV positive patients does not seem to follow any specific clinical pattern. The nodes, in addition to the presence of malignancy, TB or other pathology may be distorted by the presence of PGL. This concept may still be true if one considers that some of the patients did indeed have more than one pathology in their lymph nodes. One patient was diagnosed as having a metastatic neuroendocrine tumour on the first FNA and TB on a second FNA. Lymph node biopsy was consistent with a neuroendocrine tumour. However, the patient also tested positive for acid fast bacilli on a sputum sample. Similarly, another patient had KS on one lymph node and reactive changes on another node from the same biopsy site. A third patient had TB and reactive lymphadenopathy from the same biopsy site. Even though there was no correlation between chest radiographs and lymph node biopsy pathology in our study, the finding of intra-thoracic lymphadenopathy suggested more significant pathology, such as TB or malignancy. A patient with unexplained cytopenias, especially with a pancytopenia should ideally have a bone marrow aspirate and trephine biopsy to clearly elucidate the underlying pathology. Interestingly in our study, a patient diagnosed with reactive lymphadenopathy on FNA and subsequently shown to have Castleman’s disease on lymph node biopsy, had a granuloma on bone marrow trephine consistent with TB. A relative monocytosis was suggestive of malignancy in our study, but may also be an indication of viral latency, as most of these patients with monocytosis also had a CD4+ lymphocyte count above 200 cells/mm3. In our study, thirty patients had both a FNA and a lymph node biopsy for comparison. Comparing FNA with lymph node biopsy as the true positive, eleven patients were diagnosed with TB, four with malignancy, three as reactive and two as having other aetiologies of their lymphadenopathy. In the discrepant group of ten patients comparing FNA to Biopsy, five patients were diagnosed as having reactive lymph nodes on FNA, but the biopsy confirmed two with Castleman’s disease, two with NHL and one with TB. One patient diagnosed as TB on FNA, had KS on biopsy. Another patient with a lymphoepithelial cyst on FNA had reactive lymphadenopathy on biopsy. A further patient diagnosed with TB on FNA had a sub-optimal lymph node biopsy. The differences in FNA and biopsy results could be explained as ‘errors’ of sampling. In other words the FNA and biopsy were done at different sites in the same patient, and therefore these patients potentially had more than one pathological cause for their lymphadenopathy. Also, less tissue was available for review from the FNA specimen and immunohistochemical stains were not possible on the paucity of FNA material available. In the patients who had a lymph node biopsy, sixteen had TB, ten were diagnosed with malignancy (seven with NHL, two with KS and one with a neuroendocrine tumour). The other group included two patients with Castleman’s disease, one with a lymphoepithelial cyst, one with fibrous tissue and one as having a suboptimal biopsy. One patient had a biopsy which showed KS on one lymph node and a second node as having reactive features. A second patient had a diagnosis of TB and reactive lymphadenopathy on the same biopsy. Nine patients had reactive lymphadenopathy on biopsy. Using chi-square analysis, the specificity and sensitivity of FNA was compared to the lymph node biopsy results. In general, the low specificity and sensitivity in the different groups with the wide confidence intervals may be due to the small sample size in our patient population. Nevertheless, when comparing FNA to lymph node biopsy for the diagnosis of malignancy, the specificity and sensitivity was 100% and 43% respectively, with wide confidence intervals. This indicates that although a positive result for malignancy on FNA is highly specific, FNA for the diagnosis of malignancy is much less sensitive and a significant number of patients could be missed if the diagnosis is based on the FNA only. The specificity and sensitivity for the diagnosis of TB was 86% and 95% respectively (FNA is less specific for TB, but more sensitive). Reactive lymphadenopathy had a lower specificity and sensitivity of 80% and 60% respectively. Similarly, the specificity and sensitivity for the diagnosis of other pathologies was 86% and 40% respectively. These results underline the importance of performing a lymph node biopsy (which remains the ‘gold standard’), more particularly where malignancy is suspected or where a reactive FNA result may not be sensitive enough to exclude a more definitive diagnosis for the lymphadenopathy. The clinical haematology department receives referrals of lymphoma patients from other hospitals and other departments within the hospital (such as surgery, ENT etc). Many of these patients are referred with a biopsy proven lymphoma. Morever, 60-70% of all patients with lymphoma(both NHL and HD) are HIV sero-positive. As a result of this pattern of referral, the reflection of the cause of lymphadenopathy at this hospital may not have been obtained in this study. This should be regarded as a limitation of the study. Other shortcomings of this study include the small sample size, making comparisons from a statistical point of view difficult to interpret. Nevertheless, the study emphasizes the importance of investigating the cause of significant lymphadenopathy in HIV seropositive individuals, as in 65% of such individuals a pathological cause such as TB or malignancy is found. The high discordance rates of 33,3% between the FNA and the lymph node biopsy, underlines the fact that a lymph node biopsy should remain the investigation of choice, where a definitive diagnosis needs to be established, particularly with respect to malignancy.

Proefschrift Universiteit van Amsterdam. / Auteursnaam op omslag: Jaap Maas. Met bibliogr., lit. opg. - Met samenvatting in het Nederlands.

Bochum, Univ., Diss., 2004. / Computerdatei im Fernzugriff.

Düsseldorf, Univ., Diss., 2004. / Erscheinungsjahr an der Haupttitelstelle: 2003. Computerdatei im Fernzugriff.

Düsseldorf, Univ., Diss., 2004. / Erscheinungsjahr an der Haupttitelstelle: 2003. Computerdatei im Fernzugriff.

Bochum, Universiẗat, Diss., 2004.

Düsseldorf, Universiẗat, Diss., 2004.