Hepatic Function and its Association with Clinical Outcomes in Non-Adherent HIV-1 Adults ‎‎

INTRODUCTION

The Human immunodeficiency virus (HIV) remains a threat to global public health. Approximately 84.2 million HIV infections and 40.1 million HIV-related mortalities have been reported globally since the onset of HIV [1]. Currently, the global HIV incidence stands at 1.5 million HIV infections totaling to 38.4 million people living with HIV(PLHIV). Despite tremendous decline of HIV incidence in Eastern and Southern Africa region, the region remains disproportionately affected and threatened by a slow HIV response [2].  HIV is also a public health concern in Kenya which bears a heterogenous HIV burden largely clustered around the Lake Victoria basin [3]. Siaya county, located on the Lake Victoria basin bears a HIV prevalence of .12.3% , ranking it among the top five counties with highest HIV burden in Kenya [4].

HAART remains the global standard for HIV therapy and with optimal adherence it effectively suppresses viral replication to undetectable levels [5]. By 2021, approximately 28.7 million PLHIV worldwide received HAART of which 16.2 million were in the Eastern and Southern Africa region [6]. Approximately 83% of all PLHIV in Kenya have access to HAART [7]. Scale-up of HAART globally has led to a tremendous reduction in new HIV infections and decline in HIV related mortality over the past decade. Consequently, HIV has gradually evolved to a chronic manageable disease and great improvement in the quality of life among PLHIV realized.

Contrary to earlier guidelines requiring initiation of HAART at CD4+ cell count <200cells/mm³, HAART is currently indicated upon all patients who test positive for HIV regardless of their CD4+ count [8]. Consequently, HIV related mortalities have decreased dramatically and lifespans of all PLHIV on HAART prolonged. However, the burden of Non-HIV related morbidity and mortality has drastically increased in the same population [9]. This is largely attributed to the development of systemic derangements such as Immune reconstitution inflammatory syndrome, metabolic dysregulation, nephrotoxicity and hepatotoxicity subsequent to cumulative exposure to HAART [10].

Emerging evidence shows that liver related derangements, common among PLHIV are among the leading determinants of non-HIV mortality [11]. Such derangements arise from liver damage caused by multiple stimuli, eventually translating into acute or chronic liver disease [12]. Among the HAART naive patients, high HIV viraemia and low CD4+ counts mediate liver damage by promoting fibrogenesis and hepatocyte apoptosis [13]. HAART also advances structural liver damage via mitochondrial dysfunction, insulin resistance, hepatic steatosis and fibrosis [14]. Although current antiretroviral agents used in HAART  are effective in protecting against hepatic fibrosis, their efficiency is largely dependent on optimal adherence [15]. As such, non-adherence to HAART may predispose PLHIV to higher rates of liver derangements.

Liver injury among HIV infected patients can be detected using liver enzyme levels. For instance, among the HAART naive  HIV patients, elevation of gamma-glutamyl transferase (GGT) and alkaline phosphatase (ALP) without underlying liver disease indicates HIV mediated cholestatic liver damage (16). Likewise, elevated levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) serve as indicators of hepatocellular injury among the same population. Hypoalbuminemia, consistent with HIV progression also occurs among HAART naive patients [17]. Among the HAART experienced, mild and self-limiting to severe transaminitis as well as hyperbilirubinemia is common [18]. Therefore, HAART non-adherent patients are at risk of hepatic damage from both HIV and HAART. Despite this, monitoring of liver damage among PLHIV at baseline and during follow-up remains inadequate especially in resource limited settings. WHO guidelines and Kenyan national guidelines on antiretroviral therapy only indicate non-mandatory determination of AST and ALT levels [19]. This hampers early detection, diagnosis and management liver injury among PLHIV.

Previously, associations between higher HIV viral loads with elevated AST and ALT have been reported among HIV patients with poorly controlled viremia, reflecting HIV mediated liver injury [20]. In addition, low CD4+ counts are associated with increased incidence of transaminitis and rapid progression of fibrosis among the same population. However, effective HAART counters hepatocyte fibrosis and inflammation, restoring  levels of AST and ALT within normal ranges [21]. Furthermore, higher body mass index (BMI) is associated with decreased risk of hypoalbuminemia among HIV infected patients is [22]. While these associations have been reported among HIV patients naïve and adherent to HAART, it remains largely unknown whether they hold true among HIV patient’s non-adherent to HAART. As such, the current study sought to evaluate liver function of HAART non-adherent participants and associate them viral load, BMI and CD4+ counts.

MATERIALS AND METHODS

Study design: This was a cross-sectional analytical study.

Study site and population

This study was carried out among consenting HIV-1 adults attending Siaya county referral hospital, a public facility in Siaya county, Nyanza region of Kenya. Siaya has the third highest HIV burden of 15.3% after Kisumu (19.6%) and Homabay (17.5%) counties. It also bears the lowest levels of viral suppression amongst the three counties [4]. Study participants were enrolled through systematic random sampling.

Inclusion and exclusion criteria

Only individuals aged 18-60 years on first line HAART regime comprising of Tenofovir, Lamivudine and Efavirenz (TDF+3TC+EFV) for a period not less than 6 months were enrolled. The study excluded individuals with a history of hepatic, renal or cardiovascular disease, tuberculosis, hepatitis B virus, hepatitis C virus, diabetes, hypertension, cancer, substance abuse and pregnant women. Individuals on any other long-term medication other than HAART were also excluded. A total of 163 adults comprising of HIV-1 negative (n=51), HAART naïve (n=23), HAART adherent (n=47) and HAART non-adherent (n=42) were recruited into the study. Sample size was determined using the Daniel (1999) formula and previously reported HIV prevalence [4].

Demographic and anthropometric measures

An open-ended questionnaire was used to collect demographic data of study participants. A Seca 786 (Hamburg, Germany) stadiometer was used to measure body weight and height with participants wearing light clothes and standing in an upright posture. Weight measurement was made to the nearest gram and height to the nearest centimeter. Body mass index (BMI, kg/m²) was calculated by dividing weight (kg) with height (m) and categorized into underweight (2) normal (18.5 – 24.9 kg/m²) and overweight (>25.0 kg/m²) BMI groups based on World Health Organization (WHO) adult nutrition status.

Sample collection and processing.

An experienced phlebotomist collected 5ml of venous blood from participants’ antecubital vein into a syringe. 3ml of the blood was transferred to into an Ethylenediamine tetraacetic acid vacutainer (EDTA) and 2ml into plain BD vacutainer^® tubes (Becton Dickinson, Franklin Lakes, USA). The tubes were then labelled with the participants code, date and sample collection time. Blood samples in EDTA tubes were used for CD4+ T cell enumeration immediately after collection. Blood in plain vacutainer tubes was allowed to clot for five minutes prior to serum extraction by centrifugation at 3000 r.p.m for 5 minutes. Obtained serum was used for hepatic function tests.

HIV-1 diagnosis.

Confirmation of HIV-1 was done using rapid immunochromatographic test kit, Determine^TM (Abbott Laboratories, Tokyo, Japan) and first response^TM (Trinity Biotech Plc, Bray, Ireland).  In accordance with the Kenyan national HIV testing algorithm, Participants were considered HIV-1 infected if they had HIV positive results for Determine and HIV-1 positive results using first response kits.

Clinical chemistry measurements

Liver function tests were ran using an automated Mindray^TM BS-200 Clinical Chemistry analyzer (Mindray Medical Intl, Shenzhen, China) as per manufacturer’s instructions and standard operation procedures. Quality controls for selected liver function tests were done and parameters which failed were recalibrated.  Briefly, 1000ul of Serum was pitted into new, clean, unused Mindray^TM BS-200 cuvettes and loaded inside the sample chamber of the machine. Subsequently, measurement of serum Aspartate Aminotransferase, Alanine aminotransferase, Albumin, Total Protein, Gamma Glutamyl transpeptidase and C-reactive protein levels was done using the same machine.

Calculation of adherence to HAART

Adherence to HAART was assessed using comprehensive care clinic pharmacy refill records and percentage adherence calculated using the following formula.

Percentage adherence = (No. of doses patient should have taken – No. of doses missed) × 100

No. of doses the patient should have taken.

CD4+ T cell counts

The levels of CD4+ T cells were determined using a Pima™ point of care CD4 cytometer (Inverness medical™, Morges, Switzerland). 25.0 μl of well mixed whole blood sample collected in EDTA BD vacutainer^® tube was placed in the Pima CD4 test cartridge sample collector, capped and inserted into the Pima™ CD4 cytometer. Enumeration of CD4+ T cells was based on quantification of fluorescent signals emitted by fluorescent-labelled anti CD3 and CD4 specific monoclonal antibodies interacting with the sample in the cartridge. The total signals collected were converted into a CD4+ count displayed as cells/μl within 20 minutes. CD4+ counts were then classified into various HIV immunological stages of HIV as follows; ≥ 500 CD4+ cells/ μl as Stage I, 350 – 499 cells/ μl as Stage II, 200 – 349 cells/ μl as Stage III and < 200 cells/ μl as stage IV in accordance with WHO guidelines.

HIV viral load determination

Viral load was determined using an automated Abbott m2000 System (Abbott Molecular Inc., Illinois, USA). RNA was extracted from 200μl of serum sample. Reverse transcription of the RNA to cDNA was done followed by amplification of the cDNA using HIV-specific primers. The extract was fed into the analyzer which converted intensity of HIV probes to viral loads. WHO guidelines were used to categorize viral loads with viral loads

Statistical analysis

The data was analyzed using IBM SPSS version 25.0 (SPSS Inc. Chicago, USA). Categorical variables such as gender and duration of HAART were summarized into proportions and tabulated. Continuous variables (laboratory measures, age, height, weight, BMI) were described using medians (IQR). Chi-square test was used to determine differences in proportions. Kruskal wallis test was used to compare continuous data across the study groups followed by Bonferroni post hoc correction for multiple comparisons. Spearman correlation test was used to determine the association of each liver function marker with the viral load, CD4+ count and BMI.

RESULTS

Demographic and laboratory characteristics of study participants

A total of 163 adult participants comprising of males, (n = 73) and females (n = 90) were enrolled to the study. Of these, there were HIV-1 negative healthy controls (n= 51), HIV-1 positive HAART naive, (n=23), HIV-1 positive HAART adherent (n = 47) and HIV-1 positive HAART non-adherent (n = 42). A summary of demographic characteristics of study participants is presented in Table I. Age (P = 0.838) and height (P = 0.465) of study participants were similar across study groups but body weight differed across the study groups (P = <0.0001).  HAART non-adherent participants (median, 59.1 kg; IQR, 11.8; P<0.0001) weighed less than the HAART adherent group (median, 66.6 kg; IQR, 14.7). However, the HAART adherent group (median, 66.6 kg; IQR, 14.7; P<0.0001) weighed more than the HAART naive group (median,56.9 kg; IQR, 11.0). Likewise, body mass index (BMI) varied across groups (P = <0.0001) with the HAART non-adherent group (median, 21.1 kgm²; IQR, 3.72; P = 0.002) having lower BMI than the HIV-1 HAART adherent group (median, 23.5 kgm²; IQR, 5.49). In addition, HAART non-adherent group (median, 21.1 kgm²; IQR, 3.72; P<0.0001) and HIV-1 HAART naive (median, 20.4 kgm²; IQR, 5.11; P= 0.003) had lower BMI compared to HIV-1 negative healthy control group (median, 24.5 kgm²; IQR, 5.27).

Distribution of participants into underweight, normal and overweight BMI varied significantly across the groups (P = <0.0001). Majority of underweight participants were found in the HAART non-adherent while overweight participants were most frequent in the healthy control group. Varied distribution of participants amongst various CD4+ T cell categories was observed (P = <0.0001). The HAART non-adherent group accounted for majority of participants in WHO stage III. Similarly, distribution of participants between viral load categories differed significantly across the groups (P <0.0001). Proportion participants with low viral loads was highest in the HAART adherent group while the HAART non-adherent group had the highest proportion of participants with high viral loads.

Liver function parameters of study participants

Albumin levels varied significantly across groups (P < 0.0001) and were higher in the HAART non- adherent group (median, 3.6 g/dl; IQR, 1.7; P = 0.008) and HAART adherent group (median, 4.0 g/dl; IQR, 1.1; P < 0.0001) compared to the naive group (median, 3.2 g/dl; IQR, 1.2).  Table II summarizes the liver function parameters of study participants. Likewise, globulin levels differed across the groups (P< 0.0001) with the HAART non-adherent group having higher globulin levels (median, 3.4 g/dl; IQR, 1.9; P < 0.0001) relative to the HAART adherent group (median, 2.1 g/dl; IQR, 1.2). In addition, both the HAART adherent and HAART non-adherent groups had lower globulin compared to the HAART naive group (median, 3.4 g/dl; IQR, 1.9; P = 0.006 and median, 2.1 g/dl; IQR, 2.1; P < 0.0001 versus median, 4.1 g/dl; IQR, 1.0) respectively.

Total protein levels differed significantly among the groups (P< 0.0001) with between group analysis showing elevated levels of total proteins in the HAART non-adherent group (median, 7.3 g/dl; IQR, 0.8; P < 0.0001) compared to the HAART adherent group (median, 6.5 g/dl; IQR, 0.6). In contrast, the HAART adherent group (median, 6.5 g/dl; IQR, 0.6; P <0.0001) had lower total protein levels in relative to HAART naive (median, 7.5 g/dl; IQR, 1.0;) and healthy control group (median, 7.1 g/dl; IQR, 1.1). Levels of AST differed significantly across groups (P< 0.0001) and were higher in the HAART non-adherent (median, 31.6 IU/L; IQR, 8.0; P <0.0001), HAART adherent (median, 32.6 IU/L; IQR, 12.4; P <0.0001) and HAART naive (median, 33. 3 IU/L; IQR, 9.8; P <0.0001) relative to the healthy control group (median, 23.9 IU/L; IQR, 13.4). Likewise, levels of alanine aminotransferase (ALT) differed significantly across the groups (P = 0.003) with the HAART non-adherent group (median, 29.0 IU/L; IQR, 14.6; P<0.0001) having higher ALT levels relative to the healthy control group (median, 25.0 IU/L; IQR, 9.9).

The activity of GGT and ALP varied significantly in a similar fashion across groups at the same level of statistical significance (P < 0.0001). GGT activity was significantly higher in HAART non-adherent group (median 42.4 IU/L; IQR, 34.3; P<0.0001), HAART adherent group (median 51.7 IU/L; IQR, 32.0; P<0.0001) and in HAART naive group (median 46.3 IU/L; IQR, 31.0; P<0.0001) in comparison to healthy control group (median 28.0 IU/L; IQR, 25.3). Similarly, ALP activity was higher in the HAART non-adherent group (median 100.8 IU/L; IQR, 27.0; P<0.0001), HAART adherent group (median 108.7 IU/L; IQR, 27.6; P<0.0001) and in HAART naive group (median 102.5 IU/L; IQR, 31.0; P<0.0001) in comparison to HIV negative healthy controls (median 28.0 IU/L; IQR, 25.3).

Albumin, total protein, globulin, ALT and AST ratios of study participants.

Across group analysis revealed significant variations in albumin to total protein ratio (P <0.0001) as shown in Table II. Subsequent between group analysis revealed lower albumin to total protein values in HAART non-adherent group (median, 0.5; IQR, 0.2; P <0.0001) compared to the HAART adherent group (median, 0.6; IQR, 0.1). In contrast, albumin to total protein ratio was significantly higher in the HAART non-adherent (median, 0.5; IQR, 0.2; P = 0.009) and HAART adherent group (median, 0.6; IQR, 0.1; P <0.0001) in comparison to the HAART naive group (median, 0.4; IQR, 0.1). Similarly, values of albumin to globulin ratio were significantly higher in the HAART adherent group (median, 0.6; IQR, 0.1; P = 0.003) relative to the healthy control group (median, 0.5; IQR, 0.1). 

Correlation of liver function with Clinical outcomes HIV-1 HAART non-adherent

There was a weak significant positive correlation between globulin levels with CD4+ counts (𝜌 = 0.308, P = 0.048) among HIV-1 HAART non-adherent participants as shown in Table III.

Data shown are numbers (n) and proportion (%) of study subjects for categorical variables and median, interquartile range (IQR) for continuous variables. HIV-1 [-]; Human immunodeficiency virus type 1 negative. HIV-1 [+]; Human immunodeficiency virus type 1 positive. HAART; Highly active antiretroviral therapy, BMI; Body mass index. Data analysis was conducted using chi-square tests for proportions and Kruskal wallis test for continuous data. Thereafter, Bonferroni post-hoc test for was used for comparisons between study groups. Bonferroni correction was set at P a P< 0.0125 vs HIV-1[-] control, ^b P<0.0125 vs HIV-1 [-] naive, ^c PP values in bold are significant.

Table II: Liver function parameters of study participants

Data shown are median, and interquartile range (IQR) for continuous variables of study subjects. HIV-1 [-]; Human immunodeficiency virus type 1 negative. HIV-1 [+]; HIV-1 positive. HAART; Highly active antiretroviral therapy. AST; Aspartate aminotransferase. ALT; Alanine aminotransferase. GGT; Gamma glutamyl transpeptidase. ALP; Alkaline phosphatase. IU/L; international units per liter. Data analysis was conducted using Kruskal wallis test. Thereafter, Bonferroni post-hoc test was used for comparisons between study groups. Bonferroni correction was set at P ^aP< 0.0125 vs HIV-1[-] control, ^b P<0.0125 vs HIV-1 [-] naive, ^c PP values in bold are significant.

Table III. Correlation of liver function with Clinical outcome measures HIV-1 HAART non-adherent

Data presented are correlation coefficient (rho, 𝜌) with associated p-values. Statistical analysis was performed using Spearman’s rank correlation test. BMI; body mass index, ALB/GLB ratio; albumin-to-globulin ratio, ALB/T.PRT ratio; albumin-to-total protein ratio, AST; Aspartate aminotransferase, ALT; Alanine aminotransferase, ALP; alkaline phosphatase, GGT; gamma glutamyl transferase.

DISCUSSION