Page 306 - SAHCS HIVMed Journal Vol 20 No 1 2019

Page 2 of 7 Original Research

However, most of the literature related to respiratory TABLE 1: Questions to assess respiratory symptoms.
symptoms in HIV infection emanate from North America. Symptom Question
There are insufficient data on the extent of this problem in low- Cough: Do you cough several times on most days? Yes or no
and middle-income countries (LMIC) where the majority of Productive cough: Do you bring up phlegm or mucus on most days? Yes or no
PLHIV reside and where the burden of tuberculosis and Breathlessness: Which of the following statements best describes your
(≥ 2, MRC dyspnoea situation?
bacterial pneumonia is higher than in high-income countries scale) 1. Not troubled by breathlessness except on strenuous exercise
2. Short of breath when hurrying on the level or walking up a
(HIC) like North America. slight hill
3. Walks slower than most people on the level, stops after
1.5 km or so, or stops after minutes of walking at own pace
4. Stops for breath after walking about 100 m or after walking
We studied the frequency of respiratory symptoms in PLHIV a few minutes in level ground
whether or not on ART in an urban area in South Africa in 5. Too breathless to leave house, or breathless when undressing
comparison to an HIV-negative control group, as well as the Wheezing or Have you had attacks of wheezing or whistling in your chest at
any time in the last 12 months? Yes or no
whistling:
determinants of respiratory symptoms. Source: Adapted from The British Medical Research Council (MRC) Respiratory
Questionnaire, 10,11 the MRC dyspnoea scale, the World Health Survey, the ATS-DLD-
13
12
15,16
14
Methods 78-A and questions used in other publications
MRC, Medical Research Council.
We conducted a cross-sectional study in Johannesburg, were recruited from one of the RCTs, laboratory data were
South Africa, from July 2016 to November 2017. We recruited retrieved from the RCT visit closest to our study visit.
four groups of participants from the Johannesburg area:
HIV-positive participants not yet on ART, HIV-positive Data analysis
participants on first-line ART, HIV-positive participants on
second-line treatment and HIV-negative control participants. Outcomes were described as median with interquartile range
The HIV-positive participants were recruited from past or for continuous variables (all non-normally distributed) and
ongoing randomised controlled trials (RCTs) comparing count with percentage for categorical variables. Differences
different ART regimens in a governmental HIV care facility in continuous variables across the four groups were tested
6,7
in central Johannesburg. The control group was recruited using a Mann–Whitney U test and categorical variables using
by HIV-positive participants who invited their family or a Fisher’s exact test. The frequency of respiratory symptoms
friends with a negative or unknown HIV status with the across the four groups was displayed in bar charts.
same age range (+/−5 years) and sex to participate in the
study. All control participants underwent HIV counselling We combined the four respiratory symptoms in a composite
and testing according to the South African Department of outcome ‘any respiratory symptom’ that was defined as the
8
Health guidelines. If a control participant tested occurrence of at least one of the respiratory complaints,
HIV-positive and there was no history of ART use, they were namely coughing, bringing up phlegm, shortness of breath
counted in the first group (HIV-positive, ART-naïve) and and/or wheezing or whistling. We first analysed if the
referred to a local clinic to initiate ART. If they were on ART frequency of respiratory symptoms differed according to
already, they were included in the group on first- or second- HIV or ART status in three logistic regression models using
line ART, depending on their current ART regimen. ‘any respiratory symptom’ as outcome. In the first model, we
assessed the unadjusted association between HIV and ART
Data collection status and the occurrence of any respiratory symptom using
the HIV-negative group as the reference group. The second
Data were collected during a single visit. Information on model was adjusted for sex and age, and the third model
demographics and smoking was assessed with a modified additionally adjusted for body mass index (BMI), ever
9
version of the WHO STEPs instrument . Information on smoking, passive smoking, respiratory infections in the past
medical history, respiratory symptoms, working and living (pneumonia and/or tuberculosis) and bronchial hyper-
circumstances and occupational exposure to potentially reactivity. To investigate the influence of HIV- and ART-
harmful agents was obtained using following questionnaires: related characteristics, we repeated the models described
The British Medical Research Council (MRC) Respiratory above including the HIV-positive participants only and using
Questionnaire, 10,11 the MRC dyspnoea scale, the World the ART-naïve group as the reference group. In the third
12
Health Survey, the ATS-DLD-78-A and questions used in model, HIV viral load and CD4+ cell counts were added.
13
14
other publications 15,16 . Respiratory symptoms evaluated in
this study were ‘cough’, ‘bringing up phlegm’, ‘breathlessness’ Finally, we analysed which determinants were associated
and ‘wheezing or whistling’ according to the MRC Respiratory with any respiratory symptom. The following factors were
Questionnaire and the ATS-DLD-78. A questionnaire also considered in univariable analysis: HIV status, age, sex, BMI,
included ‘cough’ and ‘bringing up phlegm’ as symptoms ever smoking and passive smoking, respiratory illnesses in
which were evaluated separately (Table 1). the past such as tuberculosis and pneumonia, history of
bronchial hyper-reactivity and environmental factors
A physical examination was performed, which included (worked in the mining industry or worked in a dusty job or
measurements of height and weight. Blood sample was exposure to gas, chemical fumes or pesticides in work). All
collected for measurement of HIV viral load and CD4-cell factors with a p-value of < 0.2 in univariate analysis as well as
count (HIV-positive participants only). For participants who age and sex were then included in multivariable analysis

http://www.sajhivmed.org.za 299 Open Access