PRESENTED AT IAS 2013 – KUALA LUMPUR, MALAYSIA

Introduction•The changing epidemiology of opportunistic infections (OIs) since the introduction of

antiretroviral therapy (ART) has been well described in developed countries, where a greater than 50% reduction in AIDS-defining events has been reported.

•In low and middle income countries (LMIC), despite the global rollout of ART since the mid-2000s with more than 8 million patients on ART, OIs continue to be the major driver of HIV-associated morbidity and mortality, accounting for most of the estimated 1.7 million deaths globally in 2011.

•We conducted a systematic review and meta-analysis to evaluate the impact of ART on the incidence of 15 key OIs in adults in low and middle-income countries (LMIC).

Materials and methods•Eligible studies described the incidence of OIs and proportion on ART in Africa, Asia and

Latin America from 1990 to January 2011, and were identified using PubMed and Global Health databases and conference abstracts.

•ART status was categorised as naïve (NA), early ART (eART) (<12 months), late ART (lART) (≥ 12 months), and undefined time on ART (uART). Studies where the proportion on ART was ≥ 10% or <80%, or which did not provide data on ART use were excluded. For studies that presented data from multiple time points, the earliest estimate within each category was used.

•Incident risk (IR) was calculated using random effects meta-analyses for OIs with reports from 10 or more studies, and estimates from ART subgroups compared using meta-regression to calculate an adjusted odds ratio (aOR).

•The number of OI cases averted annually was estimated using AIM modelling for those OIs with a significant impact of ART and where the cost per case was known, applying IRs from NA and eART to the estimated HIV population by region with CD4 counts<200.

Results•We identified 4025 titles and abstracts, and 77 reports giving results from 57 cohorts or

trials and comprising 157 684 people were included (fig. 1).•Study size ranged from 54 to 61105 participants (median 602, IQR: 215-1191). A CD4

count less than 200 cells/mm3, or a World Health Organization (WHO) stage III or IV were the main criteria for ART initiation in most studies.

Andrea Low1, Georgios Gavriilidis2, Natasha Larke1, Marie-Renee Lajoie3, Olivier Drouin3, John Stover4, Lulu Muhe2, Philippa Easterbrook2

1 London School of Hygiene and Tropical Medicine, London, UK, 2 World Health Organization, Geneva, Switzerland, 3 McGill University, Montreal, Canada, 4 Futures Institute, Glastonbury

IMPACT OF ANTIRETROVIRALS ON THE INCIDENCE OF OPPORTUNISTIC INFECTIONS IN RESOURCE-LIMITED SETTINGS: A SYSTEMATIC REVIEW AND META-ANALYSIS

Conclusions•There was a substantial decrease in the incidence of most OIs with ART use in LMIC (from

60% to 95%), comparable to rates seen in Europe and North America. •This was greatest for oral candidiasis, toxoplasmosis, herpes zoster and Kaposis sarcoma,

with a lesser reduction for Cryptococcal meningitis and Pneumocystis jirovecci infection. Most of the impact is seen in the first year of treatment.

•There was no impact on the incidence of genital ulcer disease/HSV-2 or bacteremia.• The expansion of ART in LMIC has averted roughly 900,000 OIs annually, with yearly

savings of at least 32 million dollars, excluding savings for pulmonary tuberculosis and toxoplasmosis which are likely considerable.

•There were significant gaps in data from Latin America, and the predominance of data from large cohorts in Uganda, Cote D’Ivoire, and South Africa.

Acknowledgments• We thank Philippe Mayaud and Peter Vickerman for their comments and Gillian Bartlett for her statistical advice.

Figure 1. Flowchart of study selection for incidence and prevalence of OIs in low and middle-income countriesNOTE- HIC, high-income countries; ART, antiretroviral therapy; CMV, cytomegalovirus; OIs, opportunistic infections; TB, tuberculosis.

Figure 3. Summary incident risk of OIs in A. ART naïve and B. during early and unspecified ART, by region

NOTE- HIC, high-income countries; ART, antiretroviral therapy; CMV, cytomegalovirus; OIs, opportunistic infections; TB, tuberculosis.

Table 2. Estimated number of OI cases averted during the first year of ART by region

4025 titles identified (400 duplicates removed)3288 from PubMed

737 from Global Health124 Cochrane Systematic Reviews

1025 Conference Abstracts

481 reports excluded187 inappropriate population

73 no outcome of interest57 Middle Eastern or HIC48 no denominator cited

46 reviews30 data on mortality only

16 insufficient information on ART use

14 data on CMV only10 duplicated data

4566 abstracts excluded due to non-relevance

80 studies of prevalence59 for 12 OIs

42 for TB

127 reports included in quantitative synthesis 124 full-text reports

3 conference abstracts30 included in both prevalence and incidence

608 full-text articles or posters reviewed568 from PubMed and Global Health

7 Cochrane Reviews29 Conference Abstracts

77 studies of incidence40 for 12 OIs

58 for TB

A B

Figure 2. Low and middle-income countries reporting the incidence of OIs and ART status.NOTE- Beige-HIC; Grey- LMIC, no data; Blue- Latin America and Caribbean; Green- Sub-Saharan Africa; Purple- Asia; colour darkness reflects number of studies, with more studies darker.

•The most cases of OIs averted were oral candidiasis, herpes zoster, and pulmonary tuberculosis

•The majority of the cases averted were in Sub-Saharan Africa.

Opportunistic Infection

Number of cases averted CostAsia Latin

AmericaSub-Saha-

ran Africa

Total Cost per case (USD)

Total Savings

(USD)Pneumocystis jiroveci pneumonia (PCP) 11 421 3 665 51 757 66 844 53.97 3 607 546 Oral candidiasis 74 677 23 964 338 413 437 054 3.65 1 595 247 Herpes Zoster/Singles 35 581 11 418 161 244 208 243 11.14 2 319 831

Kaposi’s Sarcoma 5,711 1 833 25 879 33 422 23.50 785 412

Cerebral toxoplasmosis 3,514 1 128 15 925 20 567

Mycobacterium tuberculosis

UnspecifiedPulmonaryExtra-Pulmonary

22 84230 31012 739

7 3309 7264 088

103 515137 35657 729

133 687177 39274 556

182.76

234.99

24 432 654

17 519 977TOTAL 153 747 49 337 696 733 899 817 32 740 690

Table 1. Estimated summary incident risks for fifteen OIs by ART statusOpportunistic Infection

Summary of Incident Risk (%) (95%CI)(number of studies)

Adjusted odds ratioa (95% CI)

Early ART Late or Unspecified ART

p-valueb

Naïve Early ART(< 1 year)

Late ART (≥1 year)

Unspecified time on ART

Cryptococcus neoformans meningitis

1.7 (0.9-2.9) (13)

0.6 (0-2.0) (3)

0.9 (0.7-1.1) (6) 0.39 (0.09-1.63)c

0.56 (0.18-1.68)e

0.30

Pneumocystis jiroveci pneumonia

3.7 (0.6-9.4) (8)

1.1 (0.7-1.6) (3)

0.7 (0.3-1.2) (4) 0.12 (0.02-0.65)c

0.11 (0.02-0.77)e

0.04

Oral candidiasis 18.5 (10.7-27.8)(8)

1.5 (0.9-2.5)(1)

2.6 (1.1-4.6)(2)

0.02 (0.00-0.26)c

0.08 (0.01-0.50)e

0.02

Oesophageal candidiasis 2.9 (0.7-6.4)(6)

0.2 (0.1-0.7)(1)

3.8 (1.4-8.0)(1)

- NA

Herpes Zoster/ Shingles 10.5 (6.7-15.0)(6)

2.4 (1.5-3.5)(1)

4.3 (0.9-9.9)(3)

0.01 (0.00-0.83)c

0.11 (0.01-0.80)e

0.08

Herpes Simplex/ genital ulcer disease

7.2 (2.9-13.3)(8)

2.4 (0-10.6)(2)

8.1 (0.5-23.7)(3)

0.90 (0.05-17.28)c

1.16 (0.24-5.56)e

0.96

Kaposi’s sarcoma 1.4 (0.5-2.6) (12)

0.1 (0.1-1.1) (2)

1.2 (0.1-3.3) (3)

0.02 (0.00-0.90)c

2.62 (0.12-56.30)e

0.04

Cerebral toxoplasmosis 1.2 (0.3-2.5)(9)

0.4 (0.0-1.5)(3)

0.3 (0.1-0.7)(3)

0.02 (0.00-0.24)c

0.10 (0.01-1.33)e

0.01

Cryptosporidium diarrhoea

2.9 (1.2-5.4)(5)

0.1 (0-0.5)(1)

0.5 (0.1-1.1)(3)

- NA

Mycobacterium tuberculosis (all types)

8.6 (7.4-10.0)(31)

3.4 (2.6-4.4) (14)

1.6 (0.9-2.5)

(9)

5.3 (3.6-7.3) (12)

0.36 (0.22-0.57)c

0.16 (0.09-0.29)d

0.58 (0.37-0.93)e

<0.001

Pulmonary tuberculosis

9.0 (6.0-12.4) (12)

2.1 (0.9-3.7) (3)

1.1 (0-4.5)(2)

6.6 (5.0-8.5) (7)

0.17 (0.07-0.40)c

0.07 (0.02-0.20)d

0.80 (0.36-1.77)e

<0.001

Extra-pulmonary tuberculosis

3.7 (2.2-5.6) (10)

0.8 (0.1-1.9) (3)

1.0 (0-4.4) (2)

2.5 (1.0-4.5)(7)

0.16 (0.05-0.55)c

0.12 (0.03-0.55)d

0.88 (0.30-2.55)e

0.01

Bacterial pneumonia . 1.9 (1.1-3.1) (1)

. NA

Bacteraemia 4.7 (3.0-6.8) (4)

- 4.6 (0.3-21.8) (2)

. NA

Bacterial enteritis 5.3 (2.6-8.8) (6)

- 2.4 (0.1-11.1) (2)

- NA

Note-ART, antiretroviral therapy. NA, numbers insufficient to perform analysis. aReference is naïve population bDetermined by random effects logistic meta-regression; cImpact of Early ART; dImpact of Late ART; eImpact of unspecified time on ART.

The biggest impact was seen during the first year of ART for oral candida, zoster, KS, and toxoplasmosis

There was no impact of ART on genital ulcer disease and no evidence for bacteremia

The biggest impact of ART on TB was seen after the 1st year