Trends in Kaposi sarcoma prevalence at HIV care initiation in adults at a large HIV treatment program in Jos, Nigeria


 Background: Improved access to antiretroviral therapy (ART) significantly reduced the prevalence of Kaposi Sarcoma (KS) among patients initiating care for Human Immunodeficiency Virus (HIV) care in western nations. A similar pattern has not been universally documented across sub-Saharan Africa where the majority of people with HIV reside. This study examined trends in KS prevalence among adults initiating HIV care in Nigeria from 2006 to 2017.Methods: Data of patients 18 years or older, who initiated HIV care at the Jos University Teaching Hospital HIV clinic Jos, Nigeria from 2006 and 2017, were analyzed. Prevalence of KS over three consecutive four-year periods of increased ART coverage: 2006-2009, 2010-2013 and 2014-2017 were estimated. Logistic regression models were used to compare age- and sex-adjusted trends in KS prevalence rates at HIV care initiation over the three-time periods, and determine independent predictor of KS.Results: Among 16,431 adults who initiated HIV care from January 2006 to Decemeber 2017, mean age 35.1 (9.5) years, 65.7% female, and 51.1% with CD4 T-cell count < 200 cells/mm3; the overall prevalence of KS was 0.59 % (95% CI 0.48-0.72). KS prevalence was lowest at the time of HIV care initiation from 2006-2009 (0.39%; 95% CI 0.29-0.53), peaked at 1.12% (95% CI 0.82-1.52) from 2010-2013 and declined to 0.72% (95% CI 0.42-1.20) from 2014-2017. Compared to 2006-2009, age- and sex-adjusted odds for KS were 2.81 (95% CI 1.83-4.34, p<0.01) for 2010-2013 and 1.74 (95% CI 0.95-3.30, p=0.07) for 2014-2017. Independent predictors of KS were period of enrollment, male sex and low CD4 T-cell count. Conclusion: Expansion of ART access was not associated with a decline in KS prevalence among adults initiating ART in Nigeria.

diagnosis was uncertain or for patients with advanced KS who required chemotherapy or additional therapy for KS apart from ART. Tissue biopsy was usually done on-site, while histology was usually done at the histology department within the hospital. Diagnosis or treatment of KS was not routinely covered by the HIV treatment program, so patients needed to pay out-of-pocket.

Participants and Data Sources
Patients were included in the study if they had a confirmed HIV diagnosis, were 18 years or older at the first clinic visit, and had no history of prior ART use. Data of patients with only one clinic visit record were excluded because at least 2 visits were required to complete evaluation for clinic enrollment. We obtained data from the electronic medical record (EMR) of the JUTH HIV clinic. During all patient encounters, data were collected using standardized forms and transferred into the EMR.
Pharmacy records and laboratory results were also uploaded in the EMR and merged with patients' clinical records. Information within the database included demographic data obtained at the initial visit, diagnosis of AIDS-defining conditions (including KS), ART history and results of all laboratory tests, including CD4-T cell count, which were carried out on site. All new patients had an assessment, which was captured and stored in the EMR in an 'Initial Assessment Form'. This captured visit date, demographic information, WHO stage of HIV disease, presence of any listed AIDS-defining conditions including KS (Yes/No), prior ART exposure (Yes/No) and results of CD4 T-cell count. Because results of CD4 -T cell count were not available immediately, patients were asked to return within two to our weeks to complete the Initial Assessment Form. This Initial Assessment Form was the primary source of data for this study.

Study Variables
The outcome of this study, prevalent KS, was defined as KS diagnosed at the first clinic visit or within 30 days of the first visit [22]. The exposure variable, the period of initiation of HIV care, was categorized into three consecutive periods of four years representing different eras of ART eligibility expansion in Nigeria;2006-2009, 2010-2013. Based on previous research, the following factors known to influence risk for KS and were available in the EMR were included as covariates: age at initiation of HIV care [23,24], sex [25], HIV-risk factor [26], and CD4 T-cell count at the time of HIV diagnosis(baseline CD4 T-cell count) [6,27]. Age and baseline CD4 -Tcell count was analysed as a continuous and categorical variable. Age was categorized into three groups (< 30, 30-39, ≥ 40 years); while CD4 T-cell count had four categories representing the severity of   immunosuppression (<200, 200-349, 350-400 and ≥ 500 cells/mm 3 ). Sex had two categories (male or female). The main risk for HIV transmission was captured as: heterosexual relationship, blood transfusion, intravenous drug use, men who have sex with men and unknown.

Statistical Analyses
Descriptive statistics of the study sample were generated. Continuous variables with normal distribution were summarized as means (Standard Deviation [SD]), while continuous variables with skewed distribution were summarized as median (Interquartile Range (IQR)). Categorical variables were presented as counts (percentages; %). We performed comparisons of characteristics of patients with and without KS using the Student's t-test for normally distributed continuous variables, Mann-Whitney U test for skewed continuous variables and the chi-squared test for categorical variables.
Crude prevalence of KS was calculated by dividing the number of patients with KS by the number of patients who initiated HIV care during that period and presented as percentages (95% confidence interval). Crude prevalence of KS during the three periods (2006-2009, 2010-2013, and 2014-2017) and prevalence by sex was computed. Trends in KS prevalence, over the three periods, were examined using the Chi-squared trend test.
Independent predictors of prevalent KS were evaluated using logistic regression models. To minimize bias due to missing variables, the multiple imputation method was used to impute missing baseline CD4 T-cell count, the only covariate with missing values. The multiple imputation method used the distribution of observed data to generate multiple plausible values of the missing baseline CD4 T-cell values to create multiple complete datasets. Each dataset was then analysed to get a set of estimates which were then combined using the method proposed by Rubin [28] to obtain an overall estimate, which was used in subsequent logistic regression analyses. After the bivariate logistic regression,     Results from the multivariate logistic regression models showed that prevalent KS was independently associated with the period of enrollment, sex and CD4 T-cell count (Table 4). Overall, men had 53% higher odds of presenting with KS compared to women, and the odds for having KS at initial presentation declined by 13% for every 100 cells/mm 3  may not always be the case [2,29]. In a Ugandan cohort, an increase in KS prevalence during a period of ART expansion was attributed to a clinical trial for which patients with KS were actively recruited [2]. That was however not the case in the present cohort. Paradoxically, increased ART coverage may lead to a transient increase in KS prevalence at enrollment, as sick patients from previously unreached communities gain access to HIV care [29]. Community-based studies would be needed to evaluate this hypothesis. Irrespective of the cause, the trend is troubling and calls for continuous routine monitoring of KS burden in patients initiating HIV care. Several studies have reported a higher risk of HIV-associated KS in males compared to females [22][23][24]32]. In North America and Europe cohorts, it was attributed to the risk of co-transmission of HIV and HHV-8 through sex in men who had sex with men (MSM), who constituted the largest HIV risk group [32][33][34][35]. In Nigeria and most of SSA, heterosexual sexual activity is the major route of HIV transmission, yet many studies have reported a higher KS prevalence in males, suggesting that there are other unexplained mechanisms [22][23][24]. In addition to the higher odds for KS in males, a striking, but incidental finding in the current study is the CD4 T-cell dependent difference in KS risk between males and females. KS prevalence in females appeared to be more sensitive to changes in CD4 T-cell count changes, with a significantly lower risk for KS occurring between CD4 T-cell counts of 200-500 cell/mm 3

. In contrast, for advanced immunosuppression (CD4 T-cell count < 200 cells/mm 3 ) or normal
CD4 T-cell count (above 500 cell/mm 3 ); the difference was not significant. To the best of our knowledge, this is the first study to describe sex-based differences in the association between CD4 Tcell and KS prevalence. Sex-based differences in the association between CD4 T-cell and disease severity has however been reported in other diseases [25,36]. This finding needs to be verified in other populations and its implications for KS prevention examined. Despite these limitations, a key strength of this study is its large sample size and 12-year time span which provided sufficient power to precisely assess KS prevalence and evaluate changes over significant periods of ART expansion in Nigeria. While routinely collected clinical data were utilized, the JUTH HIV clinic has rigorous protocols to maintain a high-quality database, which is routinely used for research [13,20,37]. Recent funding support from the US National Institute for Health has also supported the validation of cancer cases within the database. Lastly, although the study was conducted among patients in just one clinic, the age, sex, and CD4 T-cell count distribution of patients in this cohort were similar to reports from other large HIV cohort across Nigeria [27,38].