Jacky Galpin

Increased Prevalence of Severe Malaria in HIV-Infected Adults in South Africa

BACKGROUND Conflicting reports exist regarding the impact of human immunodeficiency virus (HIV) infection on the risk of severe malaria. We aimed to assess the effect of HIV infection status, advancing immunosuppression, and antimalarial immunity on the severity of malaria. METHODS A prospective cohort study was conducted. Consecutive hospitalized adult patients with falciparum malaria were tested for HIV antibodies and to determine CD4+ T cell count. Immunity to malaria was assessed by obtaining a history of childhood residence in an area where malaria is endemic. Patients were assessed for features of severe malaria. RESULTS Three hundred thirty-six patients were enrolled in the study, of whom 32 (10%) had severe malaria. The prevalence of HIV infection was 33%, and 111 patients (33%) were nonimmune to malaria. HIV-infected patients complained more frequently about respiratory and abdominal symptoms and less frequently about rigors and headache. Risk factors for severe malaria determined by multivariate analysis included being nonimmune to malaria, having a positive HIV serostatus, having an elevated parasite count, and having an increased white blood cell count. Risk of severe malaria was increased in HIV-infected patients with a CD4+ T cell count of < 200 x 10(6) cells/L (P < or = .001). Nonimmune HIV-infected patients were significantly more likely to have severe malaria (13 [36%] of 36 patients) than were nonimmune non-HIV-infected patients (9 [12%] of 75 patients; odds ratio, 4.15 [95% confidence interval, 1.57-10.97]; P = .003). HIV serostatus did not affect risk of severe malaria in the group from an area with endemicity (5 [7%] of 74 HIV-infected patients had severe malaria, and 5 [3%] of 151 non-HIV-infected patients had malaria; P = .248). CONCLUSIONS HIV-infected nonimmune adults are at increased risk of severe malaria. This risk is associated with a low CD4+ T cell count. This interaction is of great public health importance.

Multiple Method Comparison: Statistical Model Using Percentage Similarity

Method comparison typically determines how well two methods agree. This is usually performed using the difference plot model, which measures absolute differences between two methods. This is often not applicable to data with wide ranges of absolute values. An alternative model is introduced that simplifies comparisons specifically for multiple methods compared to a gold standard.

Staphylococcus aureus bacteraemia at two academic hospitals in Johannesburg

OBJECTIVES AND METHODS Staphylococcus aureus bacteraemia (SAB) remains a major problem worldwide. A retrospective study of patients with SAB seen from November 1999 to October 2002 was conducted at two academic hospitals in Johannesburg to determine mortality rates (death within 14 days of submission of blood culture) in patients bacteraemic with methicillin-sensitive (MSSA) and resistant S. aureus (MRSA) and to identify risk factors associated with mortality. RESULTS Of 449 patients with SAB, 104 (23.2%) died within 14 days of clinically suspected SAB. Of the 204 patients who acquired SAB in hospital, 6 patients died within 2 days, 39 between 2 and 14 days, and 41 more than 14 days after onset of SAB. One hundred and five patients (23.4%) had MRSA bacteraemia, 21 (20%) originating from the community. The MRSA bacteraemia rate among patients with hospital-acquired infection was 41.1%, significantly higher (p < 0.0001) than the 10.3% community-acquired MRSA bacteraemia. Thirty-five (33.3%) of the 105 patients with MRSA bacteraemia died within 14 days, compared with 69 (20.1%) of 344 MSSA patients (p = 0.0048). Admission to the intensive care unit (ICU) was significantly associated with mortality (p < 0.001)--30 of 79 patients admitted to ICU died (38%). Among 222 patients whose HIV status was known, 117 (52.7%) were positive, and of these 32 died (27.4%), a rate not significantly higher than that among HIV-seronegative patients (18 of 105 patients, p = 0.69). CONCLUSIONS Compared with MSSA, MRSA was shown to be significantly associated with mortality. Stay in ICU and infection with strains resistant to oxacillin, ofloxacin and rifampicin were highly significant predictors for mortality.

Evaluation of Linear Mixed Model Case Deletion Diagnostic Tools by Monte Carlo Simulation

A large number of statistics have been proposed to study the influence of individual observations in the linear mixed model. An extensive Monte Carlo simulation study is used to evaluate the appropriateness of these influence diagnostic measures. The sensitivity of the diagnostic measures to outliers and leverages is examined, and helpful results are obtained.

Clinical features and T-cell subsets in HIV-infected children with and without lymphocytic interstitial pneumonitis.

Summary Lymphocytic interstitial pneumonitis (LIP) is a non-infective lung condition common in untreated older children with vertically acquired HIV infection. Little is known about the prognosis in children with LIP, and diagnosis remains a problem where lung biopsy is not feasible. Our aim was to determine which clinical features aid the diagnosis of LIP in conjunction with the typical reticulonodular radiological picture, and whether the prognosis in children with LIP is different from that in HIV-infected children of the same age without LIP. We retrospectively compared the clinical features and T-cell subsets of 49 children with LIP with those of 56 children of similar age without LIP. Diagnosis of LIP was made radiologically. All children were apyrexial at the time of X-ray and acute intercurrent infections and tuberculosis had been excluded as far as possible. Ages ranged from 24 to 112 months in the non-LIP group and from 24 to 120 months in the LIP group. Digital clubbing and reticulo-endothelial hyperplasia were significantly more common in children with LIP than in those without. Children with LIP tended to have lower CD4+ counts and CD4% and higher CD8+ counts and CD8%, which resulted in significantly lower CD4/CD8 ratios in children under 5 years with LIP. It is possible in most cases to diagnose LIP using a combination of clinical and X-ray findings, as long as every effort is made to exclude tuberculosis. Lower CD4 + counts and CD4% as well as more frequent hospital admissions suggest that LIP adversely affects prognosis in children with HIV.

The association between environmental lead exposure with aggressive behavior, and dimensionality of direct and indirect aggression during mid-adolescence: Birth to Twenty Plus cohort

Chronic lead exposure is associated with neurological ill-health including anti-social behavior such as aggressive behavior. The main aim of this study was to examine the association between lead exposure at 13years old and dimensions of aggressive behavior during mid-adolescence. The study sample included 508 males and 578 females in mid-adolescence (age 14 to 15years) from the Birth to Twenty Plus cohort in Johannesburg, South Africa. Blood samples collected at age 13years were used to measure blood lead levels. Seventeen items characterizing aggression from the Youth Self Report questionnaire were used to examine aggressive behavior. Principal Component Analysis was used to derive composite variables from the original data for aggressive behavior; and data were examined for an association between blood lead levels and dimensionality of direct and indirect aggression and disobedience during mid-adolescence. We also examined the dimensions of aggression during mid-adolescence in relation to gender and socio-demographic factors. Blood lead levels ranged from 1 to 28.1μg/dL. Seventy two percent of males and 47.7% of females in the study had blood lead levels ≥5μg/dL. There was a positive association between elevated blood lead levels and direct aggression (p<0.05). Being male was positively associated with direct aggression (p<0.001) but, negatively associated with indirect aggression (p<0.001). Maternal education and age at birth were negatively associated with direct aggression during mid-adolescence. The significant association between elevated blood lead levels and direct aggressive behavior observed in this study may shed light on a possible environmental toxicological contribution to aggressive behavior in South African youth; and most importantly the type of aggressive behavior associated to lead exposure.

The association between elevated blood lead levels and violent behavior during late adolescence: The South African Birth to Twenty Plus cohort

Epidemiological studies have shown the adverse neuro-behavioral health effects of lead exposure among children, in particular. However, there is lack evidence in this regard from developing countries. The main aim of this study was to assess the association between blood lead levels (BLLs) during early adolescence and violent behavior in late adolescence. Our study sample from the Birth to Twenty Plus cohort in Soweto-Johannesburg, South Africa included 1332 study participants (684 females). BLLs were measured using blood samples collected at age 13years. Violent behavior was evaluated using data collected at ages 15 to 16years using the Youth Self Report questionnaire. First, bivariate analysis was used to examine data for an association between lead exposure in early adolescence and violent behavior items during late adolescence. Principal Component Analysis (PCA) was used for dimensionality reduction and six violent behavior components were derived. Data were further analyzed for an association between BLLs at age 13years and violent behavior using PCA derived components; to determine the specific type(s) of violent behavior associated with lead exposure. Median whole BLLs were 5.6μg/dL (p<0.001). Seventy five percent of males and 50% of females had BLLs≥5μg/dL. BLLs ranging from 5 to 9.99μg/dL were associated with physical violence (p=0.03) and BLLs≥10μg/dL were associated physical violence and fighting (p=0.02 and p=0.01, respectively). When data were analyzed using continuous BLLs physical violence was associated with lead exposure (p<0.0001). Furthermore, males were more likely to be involved in violence using a weapon (p=0.01), physical violence (p<0.0001), and robbing others (p<0.05) compared to females. The results from this study show the severe nature of violent behavior in late adolescence associated with childhood lead exposure. They highlight the urgent need for preventive measures against lead exposure among children in low or middle income countries such as South Africa.