PHYSICAL ACTIVITY LEVELS AND ACTIVITY PREFERENCES OF A COHORT OF SOUTH AFRICAN INDIVIDUALS LIVING WITH HIV

Correspondence Author: R Roos Department of Physiotherapy, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa Email: ronel.roos@wits.ac.za ABSTRACT: Physical inactivity is a risk factor for many lifestyle diseases. It is still poorly understood in individuals living with HIV. The aims of this study were to evaluate the physical activity levels of individuals living with HIV on antiretroviral therapy (ARV) and to assess their physical activity preferences. An observation study was done from October 2010 to June 2012 at a large urban HIV clinic in Johannesburg where 205 individuals were consecutively sampled. Physical activity was assessed with the Yamax SW200 pedometer over a seven day period. The mean age of the sample was 38 (± 9.5) years, CD4 count 285.1 (± 157.9) and time on ARV 8.7 (± 2.3) months. Physical activity of the sample was reduced at 7673.2 (± 4017.7) steps per day with women walking less than men [6993.3 (± 3462.6) and 10076.3 (± 4885.6) respectively]. Eight individuals (3.9%) did formal sporting activities and a hundred and twenty three individuals (60%) did some form of exercise. Walking (45.5%, n=56) and running (26.8%, n=33) were most frequently reported. Physical inactivity was present in the majority of the sample and interventions including activity preferences of individuals could be strategies to address this health concern.


INTRODUCTION
Physical inactivity is a health concern worldwide due to the increased bur den related to this risk factor for many chronic noncommunicable diseases (Lee et al 2012). The World Health Survey reported the crude prevalence rates for physical inactivity in South Africa to be 43% for men and 47% for women in 2002-2003(Guthold et al 2008. If the burden of physical inacti vity is reviewed in relation to non communicable diseases and ischaemic heart disease (IHD) is used as an exam ple, 30% of IHD cases in South Africa are attributed to physical inactivity tionally high in females in comparison to males and it is estimated that 1/5 of childbearing aged women are HIV posi tive (Statistics South Africa 2010). The likelihood of South African individu als infected with HIV to live longer is also now more probable. This is due to antiretroviral initiation guidelines cur rently being aligned with international guidelines (Mayosi et al 2012). In this context, the current limited published literature related to physical activity creates an opportunity for investiga tive research into the different facets of physical activity and to provide infor mation on the South African situation. The aim of this study was therefore twofold: firstly to determine the physi cal activity levels of a cohort of South African individuals living with HIV initiated on highly active antiretrovi ral therapy (HAART); and secondly to evaluate said individuals' physical acti vity preferences to provide information on modes of activity. (Joubert et al 2007) compared to inter national statistics of 6% (Lee et al 2012). Screening of physical activity levels is said to be the first step in effectively managing this health concern as it pro vides information on the status quo and provides means for comparing different populations (Hallal et al 2012). Objective devices such as accelerometers and/or pedometers are the preferred suggested means to gather such data (Hallal et al 2012;TudorLocke and Lutes 2009).
Physical activity has been well researched in the general population but is still poorly understood in the HIV population. This is of concern particu larly in the South African context due to the number of individuals living with HIV. An estimated 17% of South Africans are living with HIV according to the 2011 World prevalence rates of HIV and only Swaziland, Botswana and Lesotho report higher statistics (USAID 2011). In the South African context HIV prevalence remains dispropor

METHOD STUDY DESIGN AND PARTICIPANT DESCRIPTION
An observational study was carried out at a large urban HIV outpatient clinic in Johannesburg South Africa from October 2010 to June 2012. Participants were sampled consecutively according to inclusion and exclusion criteria. The information presented in this paper forms part of a much larger study inves tigating the risk factors for IHD in individuals who are HIV positive. The inclusion, exclusion criteria and sam ple size calculation are therefore based on the objectives of the larger study. Individuals were included if they were between 20 65 years of age, on HAART treatment for six to twelve months and ambulatory without an assistive device. They were excluded if they had any past medical history of cardiovascular disease, difficulty walking, were cur rently pregnant, presented with any acute illness and/or current active oppor tunistic infection or were emotionally unstable. Sample size was calculated at 195 study participants using preva lence rates for hypertension in the South African context as a guide. No preva lence rates for IHD in the general and/ or HIV population in South Africa were available at the start of the project (Steyn, 2008). Alpha was set at 5% and power at 80%. The sample was increased with a factor of 100/95 to allow for any loss to followup of participants accounting for a final sample size of 205.

Procedure
The study was approved by the Univer sity of the Witwatersrand Human Ethics committee. Permission was received from the hospital, clinic management and all participants gave informed con sent prior to participation. Two research assistants gained informed consent from study participants. Figure 1 is a flow diagram of the procedure related to the study. Figure 2 is a flow diagram that sum marises retention of individuals from interest to participation until attendance of the second assessment visit. Table I consists of the demographics related to the study participants.  Each participant completed a demo graphic questionnaire that included questions pertaining to demographic background, HIV history and physical activity preferences. Study participants were asked two standardised ques tions to evaluate their mode of physical acti vity preferences: "Do you participate in any formal sport activity?" and "Do you do any exercises on your own?" Participants could answer yes or no to the questions posed and were then asked to elaborate on their answers. A "for mal sport activity" was considered any physical activity that was being super vised at the time such as by a coach and/ or physical trainer. Individuals' latest CD4 count value were collected from their clinic file and/or clinic laboratory database. Time spent on HAART and details of specific HAART medication were collected from each participant and his/her clinic file.
Physical activity level was assessed using the Yamax SW200 pedometer to provide information on walking beha viour (daily step count). The Yamax SW200 pedometer is considered a valid tool for assessing ambulation physical activity during freeliving physical activity (TudorLocke et al 2002). It is also often used as the criterion when evaluating other pedometers (Tudor Locke and Lutes 2009). The convergent validity of the Yamax SW200 pedometer with a dualmode CSA accelerometer was noted to be high at r = 0.74 -0.86 (TudorLocke et al 2002). Participants were asked to wear the pedometer for seven consecutive days from getting up in the morning until going to bed at night and to document their daily steps on a physical activity log sheet at night when they removed the pedometer. The physical activity log sheet was avail able in an English and isiZulu format and participants could choose which form they wanted to complete. They were encouraged not to alter their normal physical activity routine during the evaluation period. The hipmounted pedometer placement was standardised by instructing participants to place it on a belt or waistband, on the right side in the midline of the thigh, consistent with manufacturer's recommendations. Each morning prior to positioning the pedo meter they had to reset the pedometer monitor to zero. Reactivity related to the physical activity assessment method was calculated following a pilot study done on 24 individuals. No significant altera tion (p = 0.4) in physical activity level was observed between the first and last day of assessment in participants when wearing the hipmounted pedometer and documenting their findings on a physical activity log sheet during the pilot study. The accuracy of the pedometers was evaluated when participants were wear ing them prior to assessing their physi cal activity levels during the seven day period. This was assessed by means of a 10 metre walking trial in which pedo meter registered steps were compared to actual counted steps. The walking trial was done twice and the best agreement evaluated in the pilot study. The asso ciation between the pedometer counted steps and actual counted steps was moderate (r = 0.6 with p = 0.03). The participants attended two sessions with the first author and all consultations occurred in a private consultation room within the clinic. A research assistant acted as translator during the study if a participant was unable to understand English or Afrikaans.

STATISTICAL ANALYSIS
Data analysis was done with STATA 12.0 and IBM SPSS 20. Data were evaluated for normal distribution. Continuous data e.g. pedometer step count were summa rised as means and standard deviations. Categorical data e.g. gender, physical activity categories and exercise prefe rence responses were summarised as frequencies and percentages. A review of participants' responses provided the information regarding type of sporting activities and exercises performed by individuals. The following pedometer physical activity categories were used as a guide during interpretation of results: "sedentary" less than 5000 steps per day, "light active" between 5000 -7499 steps per day, "somewhat active" between 75009999 steps per day, "active" between 1000012499 steps per day and "very active" more than 12500 steps per day (TudorLocke and Bassett 2004). Findings were reported to one decimal value.
The significance of the study was set at p<0.05.

RESULTS
The sample consisted mostly of females [77.1% (n=158) Physical activity data for 195 partici pants were available for analysis due to the following reasons: three participants' data were excluded during analysis due to not completing seven days of pedo meter assessment, seven participants did not attend their second visit or return their pedometer and pedometer log sheet and three participants send a friend/ family member to return their pedometer and physical activity log sheet if they could not attend their second session. The mean pedometer step count finding of the sample was 7673.2 (±4017.7) with women (n=152) walking less [6993.3 (±3462.6)] than men (n=43) [10 076.3 (4885.6)]. The study participants' fre quencies according to pedometer phy sical activity categories were as follows: 25.4% (n=52) were "sedentary", 27.8% (n=57) were "light active", 20% (n=41) were "somewhat active", 11.2% (n=23) were considered "active" and 10.7% (n=22) were "very active". The majority of the sample therefore fell into the less than active category.
Eight participants (3.9%) participated in formal sporting activities and one hundred and twenty three participants (60%) incorporated some form of exer cise into their life. Table 2 is an outline of the physical activity preferences of those individuals who participated in formal or selfinitiated exercise.

DISCUSSION
The aim of this study was firstly to determine the physical activity level of a cohort of South African individuals living with HIV on HAART and secondly to evaluate said individuals' phy sical activity preferences to provide informa tion on modes of activity. Ambulation physical activity (walking behaviour) was of interest in this study as it forms part of a much larger project related to ischaemic heart disease risk and lower levels of physical activity such as activi ties of daily living were not collected. This study contributes to the body of knowledge specific to ambu lation phy sical activity and provides information of an urban South African group of individuals living with HIV when on HAART for six to twelve months. Male participants accumulated more steps per day compared to their female counter parts. This finding is similar to studies conducted in the general South African population concerning prevalence rates of physical inactivity and gender dif ferences (Guthold et al 2008). The pedo meter step count finding was less than that reported by Cook et al (2010) in South African individuals living in a rural area. The average step count in their sample was 12 471 steps/day. It is well known that epidemiological transi tion that consists partly of urbanisation result in behavioural change (Mensah 2008). The change is then often noted in individuals' diet and physical activity levels. The different settings could there fore have influenced the physical acti vity levels due to possible differences in transportation, geographical location and living environment of participants in this study. The HIV status of the rural population was also not reported and this added possible difference in immune function between the two study popu lations could explain the different step count findings. Increased viral load and decreased CD4 count is associated with HIV disease progression. The partici pants' mean CD4 count was more than 200 cells/μL but was still not optimal. It is known that an inverse relationship exists between physical activity levels and these two immune parameters in individuals (Bopp et al 2004). The sample was on HAART for 8.7 months and it may be possible that with longer HAART exposure their physical activity levels might improve over time due to stabilisation in immune parameters and improved wellness.
In some instances HAART might have negative effects on physical acti vity. Stavudine containing regimens are known to increase the risk of distal sensory polyneuropathy (DSP) in indi viduals living with HIV (Shurie and Deribew, 2010) and this could influ ence an individual's walking ability. The development of DSP in individuals on Stavudine might have been one of the factors that influenced the alteration of the National HAART initiation guide lines to now include Tenofovir contain ing regimen as the first line HAART regimen in South Africa (Department of Health, 2013). A percentage of the study sample was on Stavudine but at the time of this study were not reporting any symptoms of DSP however the possibi lity of DSP affecting physical activity is possible (Ites et al 2011). The physical activity levels are very similar to an inter national HIV group. RamirezMarrero et al (2008) investigated the physical acti vity levels in 58 Hispanic adults living with HIV in San Juan, Porto Rico. They reported a mean of 7 418 (± 2714) steps per day with men accumulating slightly less steps than the South African males at 7594 (±2817) and women slightly more at 7151 (±2589) steps/day. The beneficial effects of exercise in individuals infected with HIV have been well reported. These effects include improvement in quality of life, cardio metabolic risk factors, cardiorespiratory fitness and muscle strength (O'Brien et al 2010). A challenge that investigators are often faced with when conducting clini cal studies are study participants' adher ence to exercise and withdrawal from interventions. O'Brien et al (2010) report the following reasons for nonadherence and withdrawal compiled from a review of exercise studies in individuals liv ing with HIV. Reasons for withdrawal include: lack of interest, time and moti vation; economic issues; issues related to study participants' family; transpor tation difficulties and sessions clashing with work schedules. Reasons related to exercise non adherence included: illness; complaints that exercises were too difficult; time limitation; loss of interest and transportation difficulties (O'Brien et al 2010). Here loss to fol lowup reasons was very similar to the responses noted above. One interesting finding during the course of the study however was the difficulty contacting individuals telephonically. The majo rity of participants used mobile phones as communication devices. The clinic where the study was undertaken also found contacting individuals problem atic and explained that patients often buy new simcards where free airtime was included instead of buying airtime for an existing phone due to less cost involved and savings accrued when a new sim card is bought.
A possible strategy to lessen with drawal and nonadherence to exercise programmes is to implement activity that individuals prefer doing. Published information on physical activity pre ferences in a HIV population could not be found and information in a general population is also scarce. Burton, Khan and Brown (2012) evaluated physical activity preferences in adults at risk of being inactive in a general Australian population. The format (how), location (where) and social setting (with whom) of the physical activity preferences were assessed. The researchers noted that more than 75% of respondents preferred activities at no or limited cost, 80% preferred activities that could be done close to home and 75% of respondents preferred activities that could be done alone. Similarly in this study, the majo rity of participants did exercise that was not supervised and could also be done on their own. They most often included activities such as walking or running. Walking has often been reported to be the preferred mode of activity in indi viduals with type two diabetes (Forbes et al 2010) and sedentary individuals (Booth et al 1997). This might be due to walking not being dependent on learn ing a new skill, having no cost implica tions, not dependent on belonging to an exercise group and can be done close to home. Even though participants reported walking and running as their preferred activities, the majority of the sample still accumulated less than 10000 steps per day. Both these activities are depend ent on ambulation activity and are ideal for assessing and monitoring with a pedometer. If participants did these two activities regularly during the seven days of assessment it would have been reflected in their pedometer findings. It could therefore be argued that they reported some healthy activity behav iour but the actual amount of these two activities was not given.
The current study is of clinical sig nificance as it demonstrated that most study participants attempted increas ing their activity level by performing some type of exercise even though their immune parameters were not optimal. Knowledge that study participants gain during current clinic HAART initiation programmes on living a healthy life style as a means of improving immune function is therefore influencing indi Secondly, the focus of the study pertain ing to physical activity preferences was to find out the mode (method) of activ ity and not format, location and social setting as assessed by other authors. Future studies could possibly look into the wider scope of physical activity pre ferences to include these components.

CONCLUSION
In conclusion, this study provides information on the physical activity levels and mode of physical activity pre ferences of individuals living with HIV attending an urban HIV clinic. Physical inactivity was present in the majority of study participants as they took less than the recommended 10000 steps per day. Intervention programmes addressing this risk factor for ischaemic heart dis ease at a primary care level could poten tially lessen the future burden of disease related to inactivity. . "Any opinion, findings and conclusions or recom mendations expressed in this material are those of the author(s) and therefore the NRF does not accept any liability in regard thereto. The views and opinions expressed are not those of the MRC but of the author (s) of the material produced or publicized".