Undernutrition and Associated Factors Among Seropositive Adults in ART Clinic Treatment Centre, Hiwot Fana Specialized University Hospital, Eastern Ethiopia

Background: Ethiopia is found with a high HIV epidemic. HIV/AIDS and malnourishment worsen one another from different perspectives. In Sub-Saharan Africa including Ethiopia despite remarkable efforts to reduce the problems, HIV/AIDS and malnutrition has remained the main challenges of health systems. Objective: To assess malnutrition and identify factors associated with undernutrition among adult patients attending HIV follow-up service at Hiwot Fana Specialized University Hospital (HFSUH), Harar town, Eastern Ethiopia. Method: Institutional based cross-sectional study was conducted from November 1 to December 30, 2016, among five hundred two respondents. The sample was selected by a simple random sampling method from the pre-ART and ART sample frame. The data were collected using a structured questionnaire through face to face interviews. We used descriptive statistics to describe the prevalence of malnutrition and mean, frequencies, and proportion of the variables. Factors associated with undernutrition were assessed by a logistic regression model using STATA 14. Result: The prevalence of undernutrition was 26.5% (95%CI: 24.5, 28.5) among HIV/AIDS patients. Being male (AOR 2.11, 95%CI: 1.34, 3.12), age group of 18-29 years (AOR 2.85, 95%CI: 1.48, 5.62), having an unemployed partner (AOR 2.31, 95%CI: 1.36, 3.75), opportunistic infection (AOR 2.94, 95%CI: 1.44, 6.03), and not attending HIV related dietary counseling session (AOR 3.22, 95%CI: 1.64, 6.31) were significantly associated with undernutrition. Conclusion: One-fourth of HIV/AIDS patients had undernutrition. Being male, younger age groups, having an unemployed partner, having opportunistic infection, and no dietary counseling were positively associated with undernutrition. All concerned bodies should have to give attention to these identified factors.


Introduction
Sub-Saharan Africa is one of the regions highly infected with Human Immunodeficiency Virus (HIV). Out of the 36.7 million projected individuals living with the virus globally in 2015, Sub-Saharan Africa accounts for around 69% [1]. Ethiopia is also found with a high HIV epidemic. As stated by the 2016 national HIV, estimates reported by WHO, Ethiopia accounts for an estimated HIV occurrence of 1.1% [0.8% -1.3%] among 15-49 years adults [2].
HIV/AIDS and malnourishment worsen one another from different perspectives. HIV patients might be malnourished due to various reasons. HIV reduces food intake resulting from appetite loss and difficulty eating, possibly as a result of infections, side effects of medication, or depression. It affects nutrients' absorption may be as the result of recurrent or chronic diarrhea and HIV-related intestinal cell damage. Moreover, HIV increases the patients' energy needs due to viral replication and opportunistic infections (OIs). Similarly, HIV changes the way the body uses the nutrients it receives or has stored. Then again, malnutrition compromises the immune system, making the body vulnerable to frequent illness and increasing its energy and nutrient demand, thereby accelerating disease progression [3].
In Sub-Saharan Africa, malnutrition among adult HIV/ AIDS patients is common and accounts for 10% to 33% [4][5][6]. The proportion of undernutrition among HIV adult patients was 19% in Senegal and 5.8% in Nigeria [5,7]. Ethiopia is a poor country and carries the highest occurrences of chronic food insecurity [8]. Consistently in Ethiopia, malnutrition among HIV adult patients was high ranging from 12.3% in Dilla university hospital to 25.2% in Butajira hospital [9,10].
Different socio-demographic and clinical related variables like wealth index, employment status, age, sex, WHO clinical Staging, CD4 count, ART interruption, and presence of opportunistic infection were some of the identified predictors of adult HIV patients' malnutrition [11][12][13].
In the Sub-Saharan Africa, despite significant efforts are made in HIV treatment, still HIV/AIDS and malnutrition has caused severe health problems [14]. Between 2000 and 2016, Ethiopia has reduced the new HIV infections by 90%, and AIDS-related mortality was also reduced by more than 50% [15]. Moreover, Ethiopian Ministry of Health has developed and started to implement National Guidelines for HIV/AIDS and Nutrition since 2007 [16]. However, irrespective of all the above efforts, the prevalence of HIV/AIDS still high (0.9%) in Ethiopia [15]. This could be due to low treatment effectiveness and non-adherence to the treatment and nutritional status of patients [17].
In general in Ethiopia, there is a limited national evidence, particularly in the study area related with malnutrition among HIV/AIDS patients. Therefore, this study was designed to add some value to the existing evidence by assessing malnutrition and factors associated with undernutrition among adult patients attending HIV follow up service at Hiwot Fana Specialized University Hospital in eastern Ethiopia.

Study Setting
Hiwot Fana Specialized University Hospital is one of the oldest hospitals in Ethiopia, which was established during the occupation of Italian soldiers (1928)(1929)(1930)(1931)(1932)(1933). Currently, the hospital expected to serve for more than twenty million peoples of the region and neighboring regions like Dire Dawa administrative council, Oromia, and Ethiopian Somali Regional State. The majority of the cases are referrals. We conducted this study among adult patients registered for HIV follow-up service at HFSUH located in Harari regional state, 526 kilometers far away from the capital city, Addis Ababa, and toward the eastern direction. The hospital serves as a teaching center of eastern Ethiopia and delivers different health services to the community like internal medicine, surgery, gynecology and obstetrics, TB/HIV services for community. It is the center for training undergraduate medical and other health science students and function as a research centers [18].

Study Design and Period
This hospital-based cross-sectional study was conducted in Haramaya University Hiwot Fana Specialized University Hospital in ART treatment center from November 1 to December 2016.

Study Population and Sample Size Determination
All seropositive individuals aged ≥18 years who had a regular follow up in the ART treatment center and enrolled for a minimum of six month at Hiwot Fana Specialized University Hospital in eastern Ethiopia were eligible for the study. In this study five hundred two individuals were randomly selected who age were 18 years and above and enrolled in HIV follow-up (Pre-ART and initiated ART) at Hiwot Fana Specialized University Hospital. However, patients who critically ill and HIV positive pregnant/lactating women were not included in the study. The sample size was determined using a single population proportion formula. The sample size was calculated using Open Epi 3.1 based on a single population proportion formula by considering the prevalence of undernutrtion 25.5% among HIV patients which was reported in Ethiopia [19], with assuming 4% marginal error, 95% Confidence level and 10% a nonresponse rate. The calculated sample size was 502.

Sampling Procedure
During our study period, there were 5,058 registered adult seropositive individuals enrolled in ART care clinic (Pre-ART and initiated ART) at HFSUH, Harar town, Eastern Ethiopia. Regarding their ART status, 3,823 patients were on ART, and 1,235 were on Pre-ART care. Then, the calculated sample size was proportional allocated to Pre-ART and ART status based on their total population. We recruited 122 patients from Pre-ART and 380 from those who initiated ART. Finally, eligible patients from respective ART status were selected using a simple random sampling method using their card number.

Data Collection Procedure
An English Version structured questionnaire was adopted from literature and translated into local languages in the study setting. Then, the local language versions were translated back English by fluent speaker to check for accuracy and consistency. Data collection was made through face-to-face interviews using interviewer-administered structured and reviewing patients' medical records. The tools was pre tested at another nearby public hospitals out of our study setting. The study subjects variables including sex, age, marital status, educational level, ethnicity, religion, patient and partner's employment status, house hold's average monthly income, functional status, food aid from community organizations, attended HIV related dietary counseling session, and eating frequency through interviewing patients. Besides, we obtained variables such as WHO clinical staging, Treatment Centre, Hiwot Fana Specialized University Hospital, Eastern Ethiopia current CD4 cell count (cell/mm 3 ), ART status, and opportunistic infection from the patients' medical records. However, the height and weight of the participants were measured and recorded at the start of the interview. Then we calculated the body mass index (BMI) of each patient as weight in kilograms divided by height squared (kg/m 2 ). Data were collected by three trained Bachelor of Science holder clinical nurses working in the ART clinic. The overall data collection was supervised by public health professional.

Dependent Variable
Status of malnutrition among HIV/AIDS adult patients

Independent Variables
Socio-demographic and clinical related characteristics of HIV/AIDS patients including sex, age, marital status, educational status, ethnicity, religion, patient's employment status, house hold's average monthly income, and partner's employment status, clinical staging, CD4 cell amount, ART position, opportunistic infection, functional status, accessed HIV related food aid from community organizations, attended HIV related dietary counseling session, and eating frequency.

Data Processing and Analysis
Data were cleaned, coded, and entered by using EPiData Version 3.1 software and analyzed by using STATA version 14 (Stata Corp, USA) statistical software. We have checked the data thoroughly for consistency and completeness before entry. We computed descriptive statistics to present the outcome and independent variables using mean, frequencies, and proportion.
Using the following nutritional status classifications based on BMI calculation: BMI (< 16 kg/m 2 )=Severe malnutrition; BMI (16 kg/m 2 -16.99 kg/m 2 )=Moderate malnutrition; BMI (17 kg/m 2 -18.49 kg/m 2 )=Mild malnutrition; BMI (18.5 kg/m 2 -24.99 kg/m 2 )=Normal nutritional status; BMI (25 kg/m 2 -29.99 kg/m 2 )=Overweight; and BMI (≥30 kg/m 2 )=Obesity, we merged severe, moderate, and mild malnutrition classifications as a single category (undernutrition) was created. Then, accordingly we categorized into undernutrition, normal nutritional status, and overweight and obesity. Finally, we used the category of patients with undernutrition and normal nutritional status (n=473) in logistic regression analysis to detect predictors of undernutrition. However, we reported only the magnitude of patients with overweight/obesity (n=29) in the result of this study. We further used variables with a P-value < 0.05 in bivariate analysis into the final multivariable model. Then the statistical level of significance was reported at P < 0.05. We used a p-value and odds ratio with a 95% confidence interval to ascertain the report. We checked the fitness of the model by using Hosmer and Lomshow goodness-of-fit.

Data Quality Control
In assuring quality of the data, we emphasis highly on designing and data collection instrument. We pretested the questionnaire on 5% of the sample and accordingly the questions were refined based on the feedback obtained during pretest to reach the final version. One day training was given to data collectors and supervisors regarding data collection procedures, study objectives, and other related issues. Then we checked the questionnaires for consistency and completeness before data entry. The principal investigator and supervisor have closely supervised data collector on daily bases.

Operational Definitions
Based on the body mass index (BMI) calculation, nutritional status was classified as follows [20].

Ethics Approval and Informed Consent
This study has been performed following the principles stated in the Declaration of Helsinki. We obtained ethical approval from the College of Health and Medical Science of Haramaya University Institutional Health Research Ethics Review Committee. We also obtained permission from the hospital managers and ART clinic unit coordinator before starting data collection as well as written informed consent was obtained from participants. We informed and respected the participants' right either to refuse or withdraw from the study if needed. We have highly assured the respondents' confidentiality and privacy by using a nameless questionnaire and performing the interview in a private room.

Socio-demographic and Clinical Characteristics of the Study Participants
We recruited a total of 502 respondents in this study. Their mean age was 36.73±(9.65) years with a minimum and maximum of 19 and 76 years. The majority of the participants (60.8%) were females. More than one-third of respondents (42.4%) found in the age group of 30-

Bivariate Logistic Regression Analysis of Factors Associated with Undernutrition
As presented in (Table 2), variables including sex, age group, partner's employment conditions, clinical staging, CD4 cell count, opportunistic infection, status of functions, dietary counseling session, eating frequency have a significant association with undernutrition. Then we entered all these variables into multivariate analysis.

Multivariate Logistic Regression Analysis of Factors Associated with Undernutrition
As displayed in (Table 3), in multivariate analysis after adjusting for all independent variables, sex, age group, partner's employment status, opportunistic infection, and attended HIV dietary counseling session were significantly associated with undernutrition. Male patients have higher odds of undernutrition compared to those female patients (AOR 2.11, 95%CI: 1.34, 3.12). Young age groups (18-29 yers) were nearly three times more likely undernutrition was nearly about three as compared to elder patients in the age group of ≥50 years (AOR 2.85, 95%CI: 1.48, 5.62). Patients who have unemployed partners were more than two times more likely undernourished than patients who have employed partners (AOR 2.31, 95%CI: 1.36, 3.75). Patients who have opportunistic infection were about three times more likely to have undernutrition compared to patients who did not develop opportunistic infection (AOR 2.94, 95%CI: 1.44, 6.03). Moreover, patients who did not attend HIV related dietary counseling session have higher odds of undernutrition as compared to patients who attended HIV related dietary counseling session (AOR 3.22, 95%CI: 1.64, 6.31). However, WHO clinical staging, current CD4 cell count, functional status, and eating frequency have a significant association with undernutrition in bivariate logistic regression but they have no association in the final multivariate analysis.

Discussion
The primary aim of our study was to determine the prevalence of undernutrition and its predictors among adult Sero-positive individuals enrolled in HIV follow-up service at HFSUH, Harar town, Eastern Ethiopia. Moreover, this study presented the magnitude of overweight or obesity among the study participants.
The prevalence of undernutrition and overweight or obesity was 26.5% and 5.8%, respectively. Out of all undernutrition, 4.6% have severe malnutrition (body mass index < 16 kg/m 2 ), 5.6% have moderate undernutrition (body mass index: 16 kg/m 2 -16.99 kg/m 2 ), and 16.3% have mild undernutrition (BMI: 17 kg/m 2 -18.49 kg/m 2 ). In the final model analysis, we identified sex, age group, partner's employment status, opportunistic infection, and attended dietary counseling session were independently associated with undernutrition.
The discrepancy in the magnitude of undernutrition might be due to difference in the study design, socio-economic status, residence, culture and feeding styles of different ethnic groups, clinical stages of HIV, current CD4 cell count, health care awareness of the respondents, and the difference in ART care services like routine nutritional screening and variations in measurements of nutritional status. For instance, in our study, the majority of participants (63.9%) were found at the clinical stage one. However, undernutrition is high among patients with advanced clinical stage of HIV as compared to stage one [24,35]. Similarly, in our study, 39.6% of respondents have a current CD4 cell count of ≥ 500 cell/mm 3 , and only 11.8% have < 200 cell/mm 3 . This could also create a variation as having CD4 < 200 cell/mm 3 was significantly associated with undernutrition [24]. Moreover, a study in Nigeria evaluated the nutritional status using objective method such as body mass index, and subjective methods with malnutrition universal screening tool (MUST) and subjective global assessment (SGA) tool. But in our study, we used only the objective body mass index method [31].
The multivariate analysis result of our study revealed that male patients have higher odds of having undernutrition as compared to those female patients (AOR 2.11, 95%CI: 1.34, 3.12) and the finding was consistent with the previous study findings in Ethiopia and other African countries [21,23,30]. This result could be due to different possible justifications. Firstly, in Ethiopia, females are more concerned about their family particularly to care for their children, and possibly for the sake of their children female patients could seek help early and will be more adhere to HIV-related treatments like ART, and food counseling given to them. Secondly, in Ethiopia, some of the previously reported predictors of undernutrition among HIV patients such as substance abuse including khat, alcohol, and cigarettes are more common among males than females [25,33].
Patients who were found in the age category of 18-29 years have higher odds of undernutrition compared to elder patients of ≥50 years old (AOR 2.85, 95%CI: 1.48, 5.62) and this finding was similar with studies in Zimbabwe and Ethiopia [30,34]. This can be attributed to 1. Young adults are more vulnerable to substance misuse which could affect their nutritional habits. 2. Younger adults have a poor coping mechanism to stressful life events like acquiring HIV/AIDS and its consequences and also, having emotional instability like persistent anxiety, and depression could negatively affect their dietary intake and treatment outcome [36].
In this study partners' employment status was one of the important associated factors of undernutrition. Patients who have unemployed partners were more likely undernourished than patients who have employed partners (AOR 2.31, 95%CI: 1.36, 3.75). This result was consistent with a previous study in Ethiopia [32]. This result might be due to the reality that unemployment indorses poverty which in turn could be the base for household's food insecurity. This finding implies that generating employment chances for adult individuals living with HIV/AIDS might have a contribution to reducing undernutrition.
Also, patients who have developed opportunistic infection have higher odds of undernutrition compared to those who did not develop opportunistic infection (AOR 2.94, 95%CI: 1.44, 6.03). This finding matched the results of previous studies in Ethiopia [25,32]. This result could be due to the complications of opportunistic infections like diarrhea could worsen patients' nutritional status. This finding indicates that appropriately managing opportunistic infections is vital. Our study also revealed that patients who hadn't attend HIV related dietary counseling session have higher odds of undernutrition compared to those who attended HIV related dietary counseling session (AOR 3.22, 95%CI: 1.64, 6.31). This result was consistent with a previous study conducted in Ghana [37].
Limitations of the Study First, we assessed the participants' nutritional status objectively with anthropometric measurement like body mass index. However, we did not use subjective methods such as subjective global assessment (SGA) tools and laboratory methods to measure some micronutrient deficiencies. So, using only anthropometric measurement (BMI) might not reflect the exact respondents' nutritional status measurement. Second, due to its cross-sectional nature, our study cannot reveal the cause-effect relationships between the outcome variable (undernutrition) and independent variables.

Conclusions
Our study indicated that one-fourth of the study participants had undernutrition, which was very high. Being male, young age group, partner's employment status, having opportunistic infections, and attended HIV related dietary counseling session were significantly associated with undernutrition. Therefore, all concerned bodies, hospitals, other governmental and non-governmental stakeholders have to give attention to the identified associated factors to improve the nutritional status of HIV patients. Patients who have developed opportunistic infections need getting attention. Moreover, the emphasis on generating employment chances and strengthening HIV related dietary counseling for individuals living with HIV/AIDS is crucial.

Data Availability
The datasets used in this study are obtainable from the corresponding author on rational request.