Search

Binary classification rules based on covariates typically depend on simple loss functions such as zero-one misclassification. Some cases may require more complex loss functions. For example, individual-level monitoring of HIV-infected individuals on antiretroviral therapy requires periodic assessment of treatment failure, defined as having a viral load (VL) value above a certain threshold. In some resource limited settings, VL tests may be limited by cost or technology, and diagnoses are based on other clinical markers. Depending on scenario, higher premium may be placed on avoiding false-positives, which brings greater cost and reduced treatment options. Here, the optimal rule is determined by minimizing a weighted misclassification loss/risk. We propose a method for finding and cross-validating optimal binary classification rules under weighted misclassification loss. We focus on rules comprising a prediction score and an associated threshold, where the score is derived using an ensemble learner. Simulations and examples show that our method, which derives the score and threshold jointly, more accurately estimates overall risk and has better operating characteristics compared with methods that derive the score first and the cutoff conditionally on the score especially for finite samples., (© 2019 John Wiley & Sons, Ltd.)

This retrospective study examined whether human papillomavirus (HPV) type-specific viral load changes measured in two or three serial cervical smears are predictive for the natural evolution of HPV infections and correlate with histological grades of cervical intraepithelial neoplasia (CIN), allowing triage of HPV-positive women. A cervical histology database was used to select consecutive women with biopsy-proven CIN in 2012 who had at least two liquid-based cytology samples before the diagnosis of CIN. Before performing cytology, 18 different quantitative PCRs allowed HPV type-specific viral load measurement. Changes in HPV-specific load between measurements were assessed by linear regression, with calculation of coefficient of determination (R) and slope. All infections could be classified into one of five categories: (i) clonal progressing process (R≥0.85; positive slope), (ii) simultaneously occurring clonal progressive and transient infection, (iii) clonal regressing process (R≥0.85; negative slope), (iv) serial transient infection with latency [R<0.85; slopes (two points) between 0.0010 and -0.0010 HPV copies/cell/day], and (v) transient productive infection (R<0.85; slope: ±0.0099 HPV copies/cell/day). Three hundred and seven women with CIN were included; 124 had single-type infections and 183 had multiple HPV types. Only with three consecutive measurements could a clonal process be identified in all CIN3 cases. We could clearly demonstrate clonal regressing lesions with a persistent linear decrease in viral load (R≥0.85; -0.003 HPV copies/cell/day) in all CIN categories. Type-specific viral load increase/decrease in three consecutive measurements enabled classification of CIN lesions in clonal HPV-driven transformation (progression/regression) and nonclonal virion-productive (serial transient/transient) processes.

Background: Up to half of all people with HIV infection have some degree of cognitive impairment. This impairment is typically mild, but nonetheless often disabling. Although early detection of cognitive impairment offers the greatest hope of effective intervention, there are important barriers to this goal in most clinical settings. These include uncertainty about how self-reported cognitive symptoms relate to objective impairments, and the paucity of bedside measurement tools suitable for mild deficits. Clinicians need guidance in interpreting cognitive symptoms in this population, and a brief cognitive measurement tool targeted to mild impairment. We addressed these two problems together here. The objective of this study was to determine the extent to which performance on cognitive tests and self-reported cognitive symptoms form a unidimensional construct., Methods: Two hundred three HIV+ individuals completed the Montreal Cognitive Assessment, computerized cognitive tasks and a questionnaire eliciting cognitive symptoms. Rasch measurement theory was applied to determine whether patient-reported and performance items could be combined to measure cognition as a unidimensional latent construct., Results: Performance-based items and cognitive symptoms are arranged hierarchically along the same continuum of cognitive ability, forming a measure with thresholds covering a broad spectrum of ability that has good internal reliability. The cognitive symptoms that fit the measurement model relate to important aspects of everyday life, providing evidence that the identified construct is meaningful., Conclusions: This finding lays the foundation for a rapid measure of cognitive ability in people with HIV infection that is feasible for routine clinical use, and shows that some cognitive symptoms are systematically related to performance in this population.

In this article, we examine the concept of HIV viral load and how it has evolved over time (1995-2013) in the field of HIV/AIDS. Although the term viral load is used extensively in this field, few efforts have been directed toward the conceptualization of HIV viral load, which is often left unquestioned, undertheorized, and portrayed as a neutral and objective laboratory value that has remained relatively stable over time--with the exception of progressive advancements in technology, techniques, and sensitivity. The purpose of this article is to apply the evolutionary concept analysis method developed by Rodgers (1989, 2000a) to the concept of HIV viral load. To set the stage, we establish the need for a concept analysis of HIV viral load and provide an overview of the evolutionary view. Then, drawing on the steps proposed by Rodgers (2000a), we outline the process of data collection, management, and analysis. We then offer an in-depth discussion of the findings (attributes, antecedents, and consequences) informed by Wuest's (2000) critical approach to concept analysis. We conclude by highlighting the implications of this analysis for clinical practice, research, and theory.

The aim was to evaluate the presence of Staphylococcus spp., Enterobacteriaceae and Pseudomonadaceae in the oral cavities of HIV-positive patients. Forty-five individuals diagnosed as HIV-positive by ELISA and Western-blot, and under anti-retroviral therapy for at least 1 year, were included in the study. The control group constituted 45 systemically healthy individuals matched to the HIV patients to gender, age and oral conditions. Oral rinses were collected and isolates were identified by API system. Counts of microorganisms from HIV and control groups were compared statistically by a Mann-Whitney test (α=5%). The percentages of individuals positive for staphylococci were similar between the groups (p=0.764), whereas for Gram-negative rods, a higher percentage was observed amongst HIV-positive (p=0.001). There was no difference in Staphylococcus counts between HIV and control groups (p=0.1008). Counts were lower in the oral cavities of patients with low viral load (p=0.021), and no difference was observed in relation to CD4 counts (p=0.929). Staphylococcus aureus was the most frequently isolated species in HIV group, and Staphylococcus epidermidis was the prevalent species in the control group. Significantly higher numbers of enteric bacteria and pseudomonas were detected in the oral cavities of the HIV group than in the control (p=0.0001). Enterobacter cloacae was the most frequently isolated species in both groups. Counts of enteric bacteria and pseudomonas were significantly lower in patients with low CD4 counts (p=0.011); however, there was no difference relating to viral load. It may be concluded that HIV group showed greater species diversity and a higher prevalence of Enterobacteriaceae/Pseudomonadaceae., (Copyright © 2011 Elsevier Ltd. All rights reserved.)

Aims: This was to determine the presence and types of oral mucosal lesions in a sample of HIV(+)/AIDS South African children taking antiretroviral therapy and to investigate the relationship between CD4(+) lymphocyte counts, viral load, duration of taking antiretroviral therapy (DART), and age on presence of oral lesions., Methods: The samples consisted of 56 South African children aged 0-4 years (mean age =7.09 years) with HIV(+)/AID, infected at birth. Subjects were divided into two groups according to the presence of oral lesions with test group (patients with oral lesions) and control group (patients with no oral lesions). Children were also divided into two groups, those <6 years and those >6 years old to study the effect of age on presence of lesions., Results: Oral Candidiasis was the most common lesion reported in 19/56 children, followed by Recurrent Herpetic Infection in 9 children. Other lesions such as Kaposi's sarcoma, Multifocal Epithelial Hyperplasia, Oral Hairy Leukoplakia, Linear Gingival Erythema, and oral ulceration were also present. A statistical significant difference in CD4(+) lymphocyte count (p value 0.005), and viral load (p value 0.002) was found between the oral lesion and no oral lesion groups, those with oral lesions having a significantly higher viral load and lower CD4+ count. No statisticaly significant difference between the two groups in terms of the DART effect (p value 0.811) was found. Furthermore, there was no effect of age groups on the presence of lesions in children with HIV(+)/AIDS., Conclusion: This study contributes to the relatively scant literature on the prevalence of oral lesions in children with HIV infection in South Africa and also the relationship of these lesions to the viral load and CD4(+) lymphocyte counts.

Objective: To explore the type and prevalence of oral mucosal lesions among adults with primary HIV infection (PHI) compared with HIV-negative adults at high risk for HIV disease, and in relation to HIV viral load., Methods: We conducted standardized oral examinations to identify specific oral mucosal lesions among adults with PHI, both pre-seroconversion and post- seroconversion-recently infected, compared with HIV-negative adults. We compared the group with oral lesions to those without oral lesions with respect to HIV-RNA load and CD4 + T-cell count., Results: Among 115 adults (predominantly men), pseudomembranous candidiasis was the most common oral lesion among those with PHI, and was found in 4% of the 23 participants in pre-seroconversion and in 9% of 69 participants with post-seroconversion recent infection, compared with none found among 23 HIV negatives. Among those with PHI, the median viral load was higher and the median CD4 + T-cell count lower among the 15 participants with an oral lesion of any type than among the 77 participants without oral lesions (P = 0.02 and 0.04, respectively)., Conclusion: This finding suggests that individuals with PHI who have oral lesions may be more likely to transmit HIV because of their higher viral load.

The HIV-specific central and effector CD4 and CD8 memory T cell populations disappear from the peripheral blood of infected individuals under highly active antiretroviral therapy (HAART) with a mean half-life of 6.0 and 7.7 months, respectively. By contrast, cytomegalovirus (CMV)-specific responses are stable or increase. The striking quantitative differences between T cell memory to two persistent viral infections are instructive as to how antigen dosage contributes to the maintenance of antigen-specific memory T cell responses in humans.

This study aims to determine the relationship among cerebral metabolite concentrations (on proton magnetic resonance spectroscopy or (1)H MRS), cognitive function, and clinical variables (CD4, plasma and CSF viral loads, and lipids) in antiretroviral medication-nai;ve HIV patients. We hypothesized that the probable glial markers myo-inositol [MI] and choline compounds [CHO] would correlate with cognitive function, CD4 count, and viral loads, but not with serum lipids. Forty-five antiretroviral-drug-nai;ve HIV patients and 25 control subjects were evaluated. Frontal lobe [MI], [CHO], and total creatine [CR] were elevated, while basal ganglia [CR] were decreased, with increasing dementia severity. As a group, HIV patients showed slowing on fine motor (Grooved Pegboard) and psychomotor function (Trails A & B), and deficits on executive function (Stroop tasks). Lower CD4 counts and elevated plasma viral loads were associated with elevated frontal white matter [MI], which in turn correlated with the Stroop tasks. These findings suggest that systemic factors (resulting from suppressed immune function and higher plasma viral load) may lead to glial proliferation (elevated [MI], [CHO], and [CR]) in the frontal white matter, which in turn may contribute to deficits on executive function in HIV. Studying antiretroviral-nai;ve patients minimized the confounding effects of antiretroviral treatment on the clinical, MRS, and neuropsychological variables, and allowed for a more accurate assessment of the relationships among these measurements. Metabolite concentrations, rather than metabolite ratios, should be measured since [CR], a commonly used reference for metabolite ratios, varies with disease severity in both frontal lobe and basal ganglia.

Objectives: This study was designed to assess sociodemographic and medical insurer factors associated with receipt of highly active antiretroviral therapy (HAART)., Methods: Patients included (n = 959) were enrolled in the Johns Hopkins HIV Clinic after April 1, 1996, received > or = 90 days of care, and had a CD4 count > or = 500 cells/mm3 or HIV-1 RNA > 20 000 copies/mL. We assessed the associations of sociodemographic factors and medical insurance with receipt of HAART, stratified by 2 time periods (April 1996 through March 1997 versus April 1997 through March 1999)., Results: HAART was more likely to be used in patients who were > 39 years, White, had CD4 counts < 350 cells/mm3, had fewer missed clinic visits, and did not have intravenous drug use as their risk factor for HIV transmission. In period 1 (April 1996 through March 1997), HAART was more likely to be used in patients who were commercially insured than in other payer groups; differences between payers narrowed in period 2 (April 1997 through March 1999), however, as did differences by race., Conclusions: Differences in use of HAART on the basis of payer have narrowed since 1996. This encouraging finding may demonstrate the importance of programs that lower economic barriers to medical care.

Objective: In vitro studies report that valproic acid causes an increase in HIV-1 replication. This retrospective study examines a sample of patients with HIV disease and behavioural disturbances, for which the treatment of choice is divalproex sodium (DVP), to determine whether DVP causes an increase in HIV-1 replication., Method: A chart and database review identified 15 patients with HIV disease presenting with either 1) mania or hypomania or 2) dementia with mania or with hypomania or behavioural disturbances. HIV-1 viral load was compared before and after mood stabilizer initiation., Results: Eleven patients started therapy with DVP, and 4 patients declined treatment with a mood stabilizer. Nine of the 11 patients taking DVP were also receiving antiretroviral therapy. HIV-1 viral load did not increase in 6 of the 9 patients who had measurements between 1 week and 3.5 months after DVP initiation. No follow-up was available for the other 3 patients. Of the 2 patients receiving DVP but not antiretroviral medication, 1 had an increase of 0.17 log in HIV-1 viral load at 4 months. No follow-up record was available for the second patient. The 4 patients not taking DVP were all on antiretroviral therapy; viral loads in 2 of them remained nondetectable over 3 to 4 months, and 1 had an increase of 0.32 log in HIV-1 viral load at 3 months. No follow-up record was available for the fourth patient., Conclusions: These preliminary data suggest that, in the presence of effective antiretroviral therapy, HIV-1 viral load appears not to be adversely affected by the administration of DVP. The results for the patients receiving DVP in the absence of antiretroviral medication remain indeterminate. Further prospective study is required.