Page 60 - SAHCS HIVMed Journal Vol 20 No 1 2019

Page 5 of 10 Review Article

# LTFU (5 years) - - - - - - 22 604 1081 - - - - - - Two South African studies, both of which reviewed data from
± 2002 to 2007, estimated 13% LTFU at 1 year, and 18.7% at a
4
median of 2.4 years, respectively. It is also important to note
6
# LTFU (12 months) 1375 200 17 4713 10 195 174 - 483 57 323 7 94 32 23 that the studies included in this review were conducted
during years when the earlier South African ART guidelines
were in place; the 2010 and 2013 guidelines utilised CD4
threshold for ART initiation of 200 and 350, respectively. The
2015 guidelines expanded ART to a CD4 threshold of 500, and
Baseline CD4, cells/mm 3 unless otherwise reported IQR Median 36–169 98 103–266 195 76–175 133 NR NR 56–192 125 NR NR NR NR 60–184 121 NR NR s.d. 9.3 mean 14.5% 8.0–19.1 12.6% 112–218 176 171–299 244 297–515 434 more recently, universal test and treat has been adopted
nationally. In more recent studies, Bock et al. reported 26%
7
LTFU at 2 years in three South African public sector clinics
through 2016; Grimsrud et al. reported 17% LTFU at 2 years in
8
a similar patient population. While overall studies are
reporting higher rates of LTFU as the ART thresholds have
changed, there is still not a widely adopted definition and
30.1–43.5
28.9–40.2
29.6–90.1
29.3–41.2
31.6–43.5
standardisation of measurement of LTFU. It is therefore hard
1.4–7.4
24–31
Age, in years unless otherwise reported to draw definitive conclusions about trends of LTFU rates
25–32
29–43
IQR
NR
NR
NR
NR
NR
over time.
61.3 months
Median
Previous studies have examined the need for standardisation
36.9
34.4
33.8
35.9
4.2
NR
NR
NR
NR
NR
27
35
28
of a LTFU definition in ART programmes in sub-Saharan
Africa, citing methodological concerns and a range of
82,83,84,85
outcomes depending on the definition used.
Indeed,
there have been a variety of study definitions with a vast
diversity of LTFU estimates. A review from 2007 of ART
programmes in sub-Saharan Africa estimated 20% attrition at
6 months and between 25% and 75% at 2 years depending on
the estimation method used. A subsequent review of studies
86
at 3 years; 59% of these patients were LTFU. A larger review
87
by the same authors subsequently estimated 35% attrition in
Africa by 36 months. In South Africa, a systematic review
88
from 2014 estimated approximately 33% attrition by 4 years
on ART using studies published between 2008 and 2013.
2
Province Gauteng Gauteng KwaZulu-Natal Gauteng, Mpumalanga Western Cape, Gauteng, KwaZulu-Natal Western Cape, Gauteng Western Cape, Eastern Cape, KwaZulu-Natal, Mpumalanga Western Cape Gauteng Gauteng KwaZulu-Natal Western Cape Gauteng Gauteng from 2007 to 2009 estimated 29.5% attrition (death or LTFU)
Chi et al. performed a meta-analysis that included patients
10
from 19 countries and 111 health facilities, concluding that a
Rural versus urban Urban Urban Rural Mixed Mixed Urban Mixed Urban Urban Urban Urban Urban Urban Urban standard definition of 180 days since last clinic visit was most
accurate in determining actual loss from care. Grimsrud
Dates of cohort 2004–2013 2010–2012 2009–2013 2004–2011 2004–2013 2007–2010 2004–2011 2002–2012 2004–2012 2004–2012 2006–2011 2008–2010 2010–2011 2013–2014 et al. examined the impact of using different definitions
11
of loss to follow-up on programme outcomes using data
from the International Epidemiological Databases to Evaluate
Study type Obs Obs Obs Obs Obs Obs Obs Obs Obs Obs Obs Obs Obs Int LTFU, lost to follow-up; Obs, observational study; Int, interventional study; NR, not reported; s.d., standard deviation; IQR, interquartile range. AIDS-Southern Africa, finding that utilising different
definitions led to significantly different estimates of those
Cohort size 12 840 1430 380 42 427 83 566 1556 90 071 5154 595 4266 151 250 93 100 LTFU, making it impossible to effectively compare rates from
different programmes if the same definition is not used; these
authors also recommended a standard definition of 180 days
since last clinic visit. We support this standardised definition.
LTFU TABLE 1: Characteristics of studies included in the meta analysis. definition 3 months 3 months 3 months 3 months 6 months 6 months 6 months 6 months 6 months 6 months Other 3 months Other Other A key limitation of this study was that despite including
67 studies that met our inclusion criteria, there was not large
Population Adults Adults Adults Adults Adults Adults Adults Adults Adults Children Children Pregnancy Pregnancy Pregnancy heterogeneity in terms of study locations. Many of them
were conducted in the same study site and often had
overlapping dates. There were a disproportionate number of
studies (two-thirds) in urban areas, primarily in Gauteng and
First author, ref Shearer 15 Clouse 16 Evangeli 17 Evans 18 Cornell 19 Wandeler 20 Fatti 22 Grimsrud 81 Mberi 21 Sengayi 66 Chhagan 70 Van Schalkwyk 77 Clouse 34 Schwartz 79 Western Cape provinces. Fifteen (22%) of them came from a
single clinic in Johannesburg (Themba Lethu). This overlap
led to a significant decrease in the number of studies we
http://www.sajhivmed.org.za 53 Open Access