Reduced treatment delays for drug-resistant TB/HIV co-infected patients

with decentralised care and rapid Xpert MTB/Rif test in Khayelitsha, South

Africa

Helen Cox, Jennifer Hughes, Sizulu Moyo, Johnny Daniels, Vivian Cox, Mark Nicol, Gilles

van Cutsem and Virginia Azevedo

The global TB situationThe global TB situationEstimated

incidence, 2012Estimated number

of deaths, 2012

940,000*(0.8–1.1 million)

8.6 million(8.3–9.0 million)

450,000(300,000–600,000)

All forms of TB

Multidrug-resistant TB

HIV-associated TB 1.1 million (1.0–1.2 million)

320,000(300,000–340,000)

Source: WHO Global Tuberculosis Report 2013 * Excluding deaths attributed to HIV/TB

170,000(100,000–240,000)

Access to DR-TB treatment

WHO TB Report 2013

~450,000 incident cases annually<20% of estimated cases receive treatment

Target

Projected

Actual

South Africa• 18% Adult HIV prevalence• 1,000/100,000/year TB

incidence• 63% of TB patients are HIV

infected• TB is the leading cause of

death (54,000 deaths in 2011)

• ~14,000 cases of rifampicin-resistant TB (DR-TB) diagnosed in 2012

The DR-TB treatment gap in South Africa

2007 2008 2009 2010 2011 20120

2000

4000

6000

8000

10000

12000

14000

16000

0

10

20

30

40

50

60

70

80

74298198

9070

7386

10085

14161

4349

46

72

56

46

Diagnosed % Treated

MD

R-TB

dia

gnos

ed

Star

ted

on tr

eatm

ent (

%)

Source: NDOH data 2013

Delays to DR-TB treatment initiation (South Africa)

Setting Diagnostic method Median time to DR-TB treatment initiation

Northern Cape, 2009-10 Line probe assay 62 days

Western Cape, 2007-11 Line probe assay 55 days

Western Cape Hospital, 2007

Culture and DST 44 days (hospitalised patients)

KwaZulu Natal, 2001-2008 Culture and DST 84 days

KwaZulu Natal, 2008-09 Culture and DST 72 days

References: Hanrahan et al PLoS ONE 2012; Jacobson et al Clin Inf Dis 2011; Bamford et al SAMJ 2010; Heller et al IJTBLD 2010; Loveday et al IJTBLD 2012

Xpert MTB/Rif for TB and DR-TB diagnosis rolled out across South Africa

GeneXpert instrument placement across South Africa, May 2014

Time to result = 1.5 hours

• Xpert for all presumptive TB• To date, 3.5 million specimens have been tested

Benefits of rapid treatment initiation

Gandhi et al, Am J Resp Crit Care Med 2010

Rapid mortality (<1 month) among diagnosed patients in Tugela Ferry, KZN

90% HIV infected

98% HIV infected

Aim

• To assess the impact of Xpert for RR-TB diagnosis on treatment initiation in the context of decentralised DR-TB treatment in Khayelitsha, Cape Town

Khayelitsha

• Population ~ 400,000• Antenatal HIV prevalence 37%

(26,000 pts on ART)• ~ 5,100 TB cases registered each

year (75% HIV infected)• DOTS treatment success ~ 80%• Approximately 200 rifampicin-

resistant cases/year (75% HIV-infected)

• 10 health facilities providing HIV/TB/DR-TB diagnosis and treatment

Khayelitsha decentralised model• Hospital admission only if clinically indicated• PHC doctors initiate treatment and review monthly in

local clinics• Daily DOT and nurse management in clinics• Integration with ART provision • Specialist paediatric outreach support• Local audiometry screening service• Individual counselling, home visits, support groups,

social worker support• Recording and reporting at sub-district level

Implementation of interventions by year

Year 2003-2006 

2007-2008 2009 2010 2011 2012 2013

DST Method

Culture (phenotypic)

Mix of Culture (phenotypic) and

LPA

LPA Xpert

RR-TB model of care

Centralized Limited decentralization

Improved programme

implementation

Full decentralization

Cases diagnosed and treated by year

2003-06 2007-08 2009 2010 2011 2012 2013

No. diagnosed

Not available 218 212 195 219 210

No. treated 158 257 182 182 173 191 197

% treated Not available 83.5 85.8 88.7 87.2 93.8

Consistently high % of patients initiate treatment

Time to DR-TB treatment

2003-06 Cult

2007-08 Cult

2007-08 LPA

2009 LPA

2010 LPA

2011 LPA

2012 Xpert

2013 Xpert

0

20

40

60

80

100

120

140

71 76 50 40 34 28

137M

edia

n da

ys to

trea

tmen

t ini

tiatio

n

Xpert impact

p<0.0001

p<0.0001

p<0.001

Decentralisation impact

LPA impact

HIV and time to RR-TB treatment

2009 2010 2011 2012 20130

5

10

15

20

25

30

35

40

45

34 34

25

12

8

42

34

28

14

7

HIV negative HIV positive

Med

ian

time

to D

R-TB

trea

tmen

t (da

ys)

HIV and % initiating RR-TB treatment

2010 2011 2012 20130

20

40

60

80

100 9891.2

100 100

85.189.6 86.3

94.4

HIV negative HIV-infected

Trea

tmen

t ini

tiatio

n am

ong

diag

nose

d (%

)

Xpert

Across 2012-13, significant difference in treatment initiation between HIV negative and HIV infected, p<0.0001

Time to treatment (2011-12, Xpert)

● deaths, censoredP=0.134, Not significant

Conclusions• Decentralisation of DR-TB treatment reduced

time to treatment from 2.5 months to <1 month• Xpert reduced time to treatment to a median of

7 days, with more than 90% of HIV infected RR-TB cases starting treatment

• Rapid diagnosis is likely to reduce early mortality among HIV infected, although earlier presentation still required

• With well functioning systems, new diagnostic tests can translate into reduced time to treatment

Acknowledgments

• City of Cape Town Health Department

• Western Cape Province• National Health

Laboratory Service• Staff in Khayelitsha clinics• Médecins sans Frontières

staff• DR-TB patients in

Khayelitsha