Magnitude of virologic failure and associated factors among adult patients on antiretroviral therapy at Debre Markos Referral Hospital, Northwest Ethiopia, 2018


 Background : Viral load monitoring is used as an important marker for diagnosing early treatment failure in patients with HIV infection/AIDS. Ethiopia has started targeted viral load monitoring recently. However, factors leading to virological failure are not well understood. The aim of this study is to assess magnitude of virologic failure and associated factors among adult patient on ART at Debre Markos Referral Hospital, North West Ethiopia. Methods : Cross sectional study was conducted on 304 participants who had started first line HAAR. Data were collected from patients’ chart starting from ART commencement and face to face interview using semi structured questionnaire. Viral load from separated plasma were analyzed according to protocols. The collected data were analyzed using SPSS version 20. Binary and multiple logistic regression models were fitted to identify factors associated with virological failure among adult patient on ART and to control confounding effect. The results were presented as odds ratio (OR) with 95% confidence intervals. Independent associations were considered with p<0.05. Results: Magnitude of virological failure was 10.5%. Lower income, (AOR = 3.5, 95% CI = 1.2 - 10.5, P = 0.024), lack of social support (AOR = 2.9, 95% CI = 1.01 – 8.2, P = 0.024), Interruption of ART, (AOR = 3.5, 95% CI = 1.01 – 12.1, P = 0.046), drug adherence (AOR = 3.6, 95% CI = 1.1 – 11.3, P = 0.028), non-working functional status (AOR = 3.5, 95% CI = 1.2 – 9.7, P = 0.018), WHO stage III or IV (AOR = 2.9, 95% CI = 1.01 – 8.0, P = 0.040), CD4 count < 200 cells/ml (AOR = 3.0, 95% CI = 1.1 – 8.0, P = 0.031) and TB co-infection (AOR = 3.7, 95% CI = 1.2 – 11.3, P = 0.018) were significantly associated with virological failure. Conclusions and recommendations : The prevalence of virologic failure was high. Lower income, lack of social support, interruption of ART, drug non-adherence, baseline non-working functional status, WHO stage, CD4 count < 200 cells/ml and TB co-infection Conducting faith healing, TB co infection, WHO stage III and IV, adherence, and income were determinants of virologic failure. Therefore early identification of associated factors and monitoring of virologic failure has to be strengthened to benefit patients to prevent from further complication. Keywords : HIV, virological failure, ART, Ethiopia


Background
Human immunode ciency virus(HIV) is responsible for a worldwide pandemic, and it is the cause of acquired immune de ciency syndrome (AIDS) [1]. Antiretroviral (ARV) drugs refer to the medicines used to treat HIV whereas antiretroviral therapy (ART) refers to the use of a combination of three or more ARV drugs for treating HIV infection synonym with Higlhy active antiretroviral therapy (HAART) [2]. High viral load (virological failure) is de ned by a persistently detectable viral load more than 1000 copies/ml after at least 6 months of ART initiation [3].
Clinical failure is de ned as any WHO stage III or IV illness in the past 12 months. Immunological failure is de ned as current CD4 cell count less than CD4 cell count at time of ART initiation or Persistent CD4 levels below 100 cells/mm3 Treatment failure is when an antiretroviral regimen is unable to control HIV infection. which can be clinical failure, immunologic failure, virologic failure, or any combination of the three [4].
In 2013 the WHO recommended routine annual viral load monitoring (VLM) for all patients on ART, as the most accurate measure of treatment response. An elevated VL (> 1000 copies/ml) in a patient who has been on ART for at least six months can indicate either therapeutic failure due to antiretroviral resistance, and / or poor adherence to treatment [3,5].According to the viral load testing principle in Ethiopia, the three main reasons for viral load testing are: 1) routine testing; 2) suspected failure testing, and 3) repeat viral load testing after suspected treatment failure for patients who were virologically non-suppressed on rst time testing and underwent enhanced adherence support for 3 months and then repeat three to 6 months before switching to second line drug [6].
Since the introduction of combination drug regimens to treat human immunode ciency virus infection, known as highly active antiretroviral therapy (HAART), the rates of HIV-related morbidity and mortality have been markedly reduced [7].However, the presence of antiretroviral drug resistance mutations in the infecting viruses may decrease the effectiveness of antiretroviral treatment because the mutations reduce the chances of full There is indisputable success over the past twenty years in reducing HIV associated morbidity, mortality, transmission, stigma and improving the quality of life of people living with HIV [8]. But still investing and ensuring the sustainability of a viral load monitoring model must be a priority to detect early treatment failure [10]. Priority setting and health system reforms to manage HIV as a chronic disease must be upheld in government agenda [5]. The World Health Organization (WHO) cites poor access to services, complex drug regimens, pregnancy, mental health disorders, substance abuse, weak social support as major barriers to adherence [3].
The latest UNAIDS data, covering 160 countries, shows the last two years the number of people living with HIV on antiretroviral therapy has increased by about a third, reaching 17.0 million people, since the rst global treatment target was set in 2003, annual AIDS-related deaths have decreased by 43%. In the world's most affected region Africa, the number of people on treatment has more than doubled since 2010. AIDS related deaths in the region have decreased by 36% since 2010 [11]. In 2015 there were 2. In 2014, the Joint United Nations Program on HIV/AIDS (UNAIDS) and partners launched the3 90-90-90 targets; the aim was to diagnose 90% of all HIV-positive persons, provide antiretroviral therapy for 90% of those diagnosed, and achieve viral suppression for 90% of those treated by 2020[8]. In 2014/2015, of the 36.9 million of people with HIV globally, only 54% were diagnosed, 41% were on ART and 32% were virally suppressed i.e. virological failure 68 percent, demonstrating that we are still very far from achieving the three 90-90-90 targets [12]. The lowest achievement rates were in low income and middle income countries [5].This is estimated to result in 73% of people with HIV achieving viral suppression, a crucial step in ending the AIDS epidemic by 2030. However, 36.9 million people are living with HIV today and about 2.1 million new infections were recorded in 2015. Early placement of patients on combined ART and achievement of viral load suppression reduces mortality and HIV transmission and improves quality of life [13].
Adequately highlighted the gap in HIV diagnosis and provision of ART, which may be unattainable under the ambitious, UNAIDS 90-90-90 targets given the current trends [12]. Achieving the United Nations Program on HIV/AIDS (UNAIDS) 90-90-90 targets will require a viral load-informed care to ensure optimal HIV clinical follow-up and resistance systematic analysis of national HIV treatment cascades from 69 countries shows none of the countries had met the 3,90targets [14]. They found that diagnosis 90% of all HIV-positive people diagnosed ranged from 87% (the Netherlands) to 11% (Yemen). Treatment coverage 81% of all HIV-positive people on ART) ranged from 71% (Switzerland) to 3% (Afghanistan). Viral suppression 73% of all HIV-positive people virally suppressed was between 68% (Switzerland) and 7% (China) [13] .In 2014/2015, of the 36.9 million of people with HIV globally, only 54% were diagnosed, 41% were on ART and 32% were virally suppressed, demonstrating that we are still very far from achieving the 90-90-90 targets. The lowest achievement rates were in low income and middle income countries [13] Early detection of treatment failure, adherence counseling and appropriate switching to second-line therapy are key strengths of a viral load monitored model [15].
Identi ed gaps: Viral load monitoring has become the standard of care for monitoring the success of treatment and diagnosing early treatment failure of ART and has been explicitly recommended, when available, by the WHO since 2010.But in our settings in which there is no access to viral load testing, clinical monitoring alone or a combination of clinical and immunologic monitoring is used to assess response to ART and determine treatment failure which is late after many complication occur. The development of drug-resistant virus strains can be another threat if this virus starts to transmit in the population. Early detection of treatment failure is crucial to sustain the effectiveness of the rst-line therapy by doing as recommended by WHO. But still there is no study is conducted in the study area by testing clients VL to identify between the two factors. Because these technologies are expensive and introduced to our country recently so determining their VL after 6 month of ART initiation, proper initiation will show virologic failure and these study will identify factors associated with virologic failure.

Study area, setting and period
The study was conducted at Debre Markos Referral Hospital which is located 300 km and 265 km away from the capital city of Ethiopia: Addis Ababa and the Amhara regional city; Bihar Dar respectively. The major Health services provided in the hospital are outpatient, Inpatient and Emergency service. And in outpatient department services such as antiretroviral treatment, voluntary counseling and testing, mental health service, dental health service, radiology health service, laboratory service, emergency service, pharmacy service, cervical cancer screening, reproductive health training, There are ve Inpatient wards (Gynecological &Obstetric, Surgical, Medical, Pediatric and Eye unit) with 189 Hospital beds. There are 9 senior physicians, 24 general practitioners, 222 BSC and clinical nurses and other different professionals. The laboratory staffs are 31(technicians, technologist and MSC in microbiology and parasitology) are highly quali ed and the hospitals are well equipped with modern laboratory machine like viral load machine and FMOH select DMRH as one of the 20 viral load monitoring site which is expected to serve as referring site to determine viral load for all ART site in east Gojjam zone and partial west Gojjam .The study was conducted from February 8-May07/05/2018.

Study Design and population
Facility based cross sectional study design was conducted from February8-May07/2018. The source population was all adult PLHIV on Highly Active Antiretroviral Treatment registered and following their treatment in Debre Markos referral Hospital (DMRH) whilst the study population was adult PLHIV ≥ 15 years old who are taking HAART for ≥ 6 month in DMRH who have follow up during study period that can ful ll the inclusion criteria

Sample size determination and sampling technique
The sample size is determined p = prevalence from previous studies of virologic failure,10.7% in Feleg Hiwot referral hospital (30) and immunologic failure in DMRH is 21%,by taking virologic failure 10.7%(0.017) giving any particular out come to be with 5% marginal error and 95% con dence interval of certainty (alpha = 0.05). The sample size will be estimated based on single population proportion (p). The formula: n = (Zα/2) 2 P(1-P) (d) 2 Where: n = Sample size z = critical value 1.96, p = prevalence of virologic failure rate 10% d = precision (marginal error) = 0.05 Thus the sample size is n = (1.96)2 × 0.1(1-0.1) 0.0025 = 152 ,15.4 + 152 With 10% non response rate the total sample size will be 167.
All adult client age 15 years and over, HIV infected, continuing on ART for more than 6 month, and returning to clinic for a regular follow-up visit during study period. Simple random sampling method was used with lottery method from appointment registration book. Interview for most of the data and document review for clinical characteristics of patient and blood sample was drawn to determine VL in ART clinic before medication re ll by having consent from client (Table 1). Operational de nition High viral load: HIV RNA > 1000 copies/ml. Alcoholism: is de ned as drinking more than ve alcoholic beverages on at least one occasion within the past 30 days. Substance use: is de ned as persons reporting smoking, chat chewing and shisha in the last 30 day. Lack of social support: is lack of social support during illness to take to health facility. Lack of disclosure: is de ned as lack of telling to people due to fear of negative consequences. ART interruption:' is de ned as a treatment interruption for at least 1 week during the previous 6 months. Adult: Those patients with age 15 or more. Missed dose: missing one or more doses in the previous month.

Data Collection Tools
An adapted semi structured questionnaire was prepared in English and translated into Amharic, which is taken from federal MOH adherence assessment tool and viral load extraction tool from other published research [22].
Interviews were conducted from client coming to follow up in a private setting at the ART clinic before medication re ll. In the survey, an interviewer-administrated questionnaire was conducted to collect demographic data, substance use and psychosocial factors. socio economic and behavioral factors, drug related characteristics of clients but Clinical characteristics of HIV patient data were collected from patient chart adherences was assessed by patient self report based on pills count as good if take > 95%,poor if < 95% taken. After completing the questionnaire, patients underwent routine follow-up and blood work with the addition of a single viral load test 5 ml of whole blood was drawn and preserved in an EDTA tube, stored at appropriate temperatures for processing and tested for viral load by molecular machine(rochi).

Data Quality Assurance
The validity of the questionnaires was assured by proper designing and also pre-testing the questionnaire in 5% (15) of respondents other than those involved in the actual study in Debre markos health center. One days training was given to data collectors and supervisors at Debre Markos referal hospital. Questionnaire was checked for completeness on a daily basis by immediate supervisors. After checking all questionnaires for consistency and completeness the supervisors were submit the lled questionnaire to the principal investigator. Data collection and supervision was carried out by quali ed personnel, four nurses who were not working in the study site were recruited and trained for data collection and they had at least diploma in status and the supervisors was BSC holder in health .

Data processing, analysis and interpretation
All responses to the survey questionnaires or collected data was retranslated against the original English version and entered into Epi Data version 3.1 and analyzed by SPSS version 20. Data cleaning were made manually by removing missing/con icting ideas and responses to questions about relevant information.
Recoding and re-categorizing was made for relevant variables. Frequencies and summary statistics were used to describe the study population in relation to socio-demographic and other relevant variables. The strength of association between dependent and independent variables was assessed using adjusted odds ratio with 95% con dence interval. Bivariate and multivariate logistic regressions was performed to assess statistical association between dependent and independent variable.. Then, variables, which showed statistical association with p-value of less than 0.25 on bivariate analysis, were candidates for the multivariate logistic regression model. In the multiple logistic regression models signi cant association was declared at p-value < 0.05. Finally data was interpreted with the existed data by referring to literature reviews.

Socio-demographic characteristics
A total of 304 study participants were included in the study which give a response rate of 100.0%. At baseline, the mean age of the study subjects was 37.3 years ± SD, (SD = 10.6). The age was ranged from 16 to 72 years.
One hundred seventy three 179 (58.9%) were females. Regarding educational status 103 (33.9%) of the respondents completed college and above while 70 (23.0%) were not attended formal education. Majority, 243 (79.9%), of the respondents reside in rural residential area. Larger proportion of the respondents 100(32.9%) earn a monthly income of 1,500-3,199 ETB followed by 29.6% and 20.5% earning monthly income of ≥ 3200 ETB and < 1000 ETB respectively. These and the rest socio-economic and demographic characteristics of the respondents have been presented in Table 2. lack emotional support (Table 3).   Table 5.  Among the participants, 196 (64.5) were classi ed as working and the rest were ambulatory and bed reddened in accordance their magnitude at baseline which summed to give the proportion of non-working participants (35.5%) (Fig. 4).
Nearly two third, 196 (64.5%), of the participants started treatment at stage 1 and 2 of WHO disease classi cation; the rest 51 (16.8%) and 56 (18.4%) were suffered WHO stages 3 and 4 conditions respectively at the time of ART initiation (Fig. 5).
The median baseline CD4 cell count was 344.  (Table 6). Candidiasis, herpes zoster, recurrent pneumonia and diarrhea (> 1 month) were the four most frequent opportunistic infections developed by 54(32.7%), 49(29.7%), 48(29.1%) and 36(21.8%) of the respondents who had opportunistic infection respectively (Fig. 6) The median viral load was 57.5 (IQR: 127.8-43.0). Majority of the study participants, 272(89.5%), had viral load of < 1000 RNA copies/ml and the rest 32 had viral load of ≥ 1000 RNA copies/ml indicating the magnitude of virological failure to be 10.5% (Fig. 7). As summarized in the following table (Table 7), multivariate logistic regression model was tted to assess the independent effect of each of the factor found to be associated in the bivariate analysis. Average monthly income, lack of social support, discontinuation /interruption of ART to take other traditional herbal medicine, holy water etc…, drug adherence, baseline functional status, baseline WHO clinical stage, baseline CD4 count and baseline TB co-infection were remained to be independent determinant factors for virologic failure among adult patient on ART.   (Table 7).
However, the result of the current study, higher virological failure was reported studies done in Kenya 23.7% [22] and Cameron (23.2%) [23], 20.6% [37] and Swaziland 16% [28]. This difference may be due to differences in de nition of virological failure, where plasma viral load 400 copies/ml was used as cut-off point for virological failure [22,23,37]. Differences in the inclusion criteria used might also be another possible explanation, i.e., the study participants included in those studies were clinically or immunologically failed and had lower median age (9.5 years old) [37,38]. On the other hand, the nding of the current study was higher than studies in Bale (9.3%) [29], Gondar (4.1%) [24], Addis Ababa (1.3%) [41] and Jimma (5.3%) [36]. It was also lower than a metaanalysis done in Ethiopia that reported overall virological failure of 5.6 [42]. The possible explanation might be due to difference in interventions for optimization of patient's adherence, better treatment outcome and type of regimen given. For instance, nutritional support was given for the study participants [36] which were different from the current study setting. Furthermore, the differences might also be due to the study areas where the residents of these areas can also be with different socio-economic status.
In the current study virological failure was varied by monthly income of the respondents, patients who had had low monthly income were more likely to experience virological failure. In agreement with this nding there is good reason to expect that socio-economic variables should be associated with antiretroviral adherence and living conditions including availability of food which in turn can impact HIV disease activity (including viral replication) [38,39]. For instance, a study done in Jimma, Ethiopia [36] reported that patients with average family income of middle and highest were more likely to have an overall adherence and ART outcome than the lowest average family income. Likewise a study from India [39] reported that having the economic ability to ful l their basic needs and to receive their medication had a positive signi cant association with both ART drug adherence and treatment outcomes [39].
The nding of the current study showed that social support was another determinant factor for virological failure; Patients lacking psychosocial support from family and others exhibited poor virological suppression.
This implies that consistent psycho-social support have a positive impact on ART outcomes. In line with this studies done in Uganda [26], Swaziland [38] reported positive social support was associated with relatively good treatment adherence and in turn better viral load reduction. Moreover, WHO's ART guideline recommend behavioral intervention (social support, avoidance of discrimination, depression) to prevent treatment interruption and to control HIV activities/replication and bring better treatment success [3]. Social support, such as someone to help with the tasks of starting to rebuild a life, assistance with cooking and assistance to grow crops, all encouraged ART adherence, prevention of opportunistic infections and reduction of viral replication and nally better treatment outcome [33,34,39]. Similarly, it has been reported in other study [40] as social support was a constant predictor of could virological suppression because it help to avoid risky daily behaviors and activities of the patient.
Interruption medication, speci cally interruption of treatment to take other traditional herbal medicine, holy water etc…, was a strong determinant factor of virological failure. In agreement with previous studies conducted in Ethiopia [32,33,34], this study identi ed that those patients who had interrupted their ART medication and use traditional herbal medicine or faith heal (like Holy water) were at a higher risk to developing virological failure than their counterparts. This ning was supported and evidenced by results of the above mentioned studies conducted in Ethiopia. A study done in Gondar [32] showed that individuals who missed 3 doses of ART per month were associated with an increased risk of drug resistance and reduced immunity, this in turn resulted in the loss of the opportunity to suppress viral replication and leads to virological failure.
Poor ART adherence found to have a great impact on the occurrence of HIV treatment failure [3]. It is widely agreed that once treatment is initiated, it should not be interrupted [3,9,38,44]. Despite this fact, in the current study the magnitude of poor adherence was about 23.3% and this poor adherence to ART was found to be an important determinant of virological failure. Patients with poor medication adherence were 3.5 times as likely to develop virological failure as patients with good adherence. This is because low level of antiretroviral in the body owing to the non-adherence is not su cient to suppress viral replication, hence leads to detection of HIV RNA level in the blood [3,44]. Even though, there was differences in adherence measurement, this nding was comparable with ndings of studies conducted in Uganda [26], Kenya [22], Swaziland [38], Tigray [35], Gondar [32] and Jimma [36] where poor medication adherence was the main risk factor for virological failure. For example the study done in Gondar reported that patients with poor medication adherence were 16 times more likely to develop virological failure compared with patients with good adherence [32]. In order to explain the effect of adherence on ART viral suppression studies suggested that around 70% of patients on rst-line ART who had a rst high viral load will re-suppress following an adherence intervention [44], indicating nonadherence as the reason for the high viral load in the majority of cases.
In agreement with previous studies [32,[35][36][37][38][39][40][41][42], the current study found baseline non-working functional status, low CD4 cell count, advanced WHO clinical stage, and presence of opportunistic infections leads to ART virological failure. Despite this fact, these risk factors are highly interconnected, CD4 cell count is the backbone of immunity construction that helps the human body to protect from the disease and can prevent HIV replication [44]. As patients' immune status becomes compromised, the rate of viral replication increase and the chance of acquiring opportunistic infections is high which leads to advanced stage of the disease.
Consequently, the patient gives more emphasis to the current problem than the chronic HIV, stop taking drugs and interrupt follow-up (poor adherence) which cause HIV treatment failure/virological failure.
Baseline working status of the patents was one of the determinant factor of virologic failure; this study found that the patients who were not able to work (initial non-working functional status, i.e. ambulatory/bedridden) due to health problem had 3.5 times more risk of developing virological failure than patients who were able to work (initial working functional status). This nding was concordant with study done in Ethiopia [35,41,43].
This might be due to the fact that those worker patients may have better income that in turn creates opportunity to get better care and support. This nding could also be explained as patients who were unable to work might be in their advanced WHO stage of the disease aggravated from greater immunosuppression (low CD4 count) [5,13,44] or they might be depressed, feel hopelessness which might in turn lead to poor drug adherence then virological or in general treatment failure [37][38][39][40] [3,40,44]. Furthermore, clients with compromised immunity are more vulnerable to different opportunistic infections that sustain the vicious cycle of immunity and viral replication [44]. On the other hand, low CD4 count might re ect viral replication due to treatment interruption (poor adherence) or resistance [7,18].
In this study TB co-infection at start of ART was founded to be a strong determinant factor of virologic failure.
This nding was in line with study conducted in Haiti [19], Cameroon [23] and Uganda [26]. One of the possible explanation for this result might be the fact that the intimate linkage between HIV and Tb might compress the immunity of the patient (including decrease in count of CD4 cells) and enables the progression of HIV disease to advanced stage rapidly and disallowing patients from regular treatment intake and hence lead to virological failure [31,34].
Even though statistically signi cant association between virological failure and age, educational status, substance use, duration of ART (time on ART) and presence of other opportunistic infections was reported by a lot of previous studies [19,23,29,38], the result of the current study did not suggest an association between these factors and ART virological failure. The differences in study designs, where some of the previous studies were case-control and cohort study designs; study population (patients in more than one health facility, different sample size, different aged /some include children); and study areas might be possible explanation for the difference between the results of the current and previous studies.

Conclusions
The magnitude of virological failure among adult HIV patients on ART was high, it was higher than most the ndings of studies in the country and even in the region.
Having lower monthly income, lack of social support, interruption of ART to take traditional herbal medicine/Holy water, baseline non-working functional status, baseline advanced (III and IV) WHO stage,

Declarations
Ethics approval and consent to participate Ethical clearance was obtained from the ethical review committee of Debremarkos University, College of Health Sciences but the committee's reference number and human participant consent not applicable.

Consent for publication
The information requested does not applicable

Competing interests
The authors declare that they have no competing interests.

Availability of data and material
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Funding
Debre Markos University contributed all nancial need for the research.
Ethical consideration and consent Ethical approval and clearance was obtained from Debre Markos University health Science College. Formal letter was received from hospital to review patients' medical charts for data collection. Prior to the interview Verbal consent was obtained from the study participants and parents of age 15-17years old. No identi ers were collected from the clients to ensure their con dentiality.

Author's contributions
The authors' responsibilities were as follows: A.A participated in the design of the study, performed the data collection and the statistical analysis and served as the lead author of the manuscript. M.T and G.A supervised the study, and ensured quality of the data. TB and SB were involved in writing the manuscript. All authors also highly participated in preparing and revising this nal manuscript and approved the nal manuscript.
ACKNOWLEDGMENT I am very much indebted to my advisors Mr. Girma Alem and Mr. Mekuanint Taddele for their unreserved guidance and support.
I would like to thanks Debre markos University department of public health for giving me a permission to conduct this study at Debremarkos Referral Hospital.
I would like to thank my colleagues, for their constructive comments, suggestions, and supplementary ideas for this thesis development.
Last, but not least, I would like to thank Debremarkos Referal Hospital ART department workers for their help by giving necessary data to do this thesis.
Finally I would like to thank Debremarkos Referal Hospital card room workers for their help by identifying and giving the selected patient's follow up cards.