WHO/HIV/2013.24
© World Health Organization 2013
GRADE tables: What recommendation should be made on the duration of breastfeeding in breastfeeding populations, for women living with HIV? DIRECT 1-A: Maternal ARV/ART plus breastfeeding versus exclusive breastfeeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-21 Question: Should maternal 3-ARV without breastfeeding restrictions or exclusive breastfeeding (2010 WHO guidelines) be used for preventing mother-to-child HIV transmission? Settings: Malawi, South Africa, Uganda, United Republic of Tanzania, Zimbabwe Bibliography: Chasela 2010, Jamieson 2012 (BAN); de Vincenzi 2011 (Kesho Bora)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Maternal 3-ARV (6 months of exclusive breastfeeding)
6 months of exclusive breastfeeding
Relative (95% CI)
Absolute
Vertical transmission (2 weeks) 1 randomized no trials serious risk of bias no serious inconsistency serious1 very serious2 none 46/849 (5.4%) 36/662 (5.4%) RR 1 (0.65 0 fewer per CRITICAL ⊕OOO to 1.52) 1000 (from 19 VERY LOW fewer to 28 more)
Vertical transmission (6 weeks) 1 randomized no trials serious risk of bias no serious inconsistency serious1 very serious2 see below3 8/284 (2.8%) 16/279 (5.7%) RR 0.49 29 fewer per CRITICAL ⊕OOO (0.21 to 1000 (from 45 VERY LOW 1.13) fewer to 7 more)
Vertical transmission (26–28 weeks) 2 randomized no trials serious risk of bias no serious inconsistency serious1 serious4 see below3 81/1110 (7.3%) 96/913 (10.5%) RR 0.69 (0.52 to 0.92) 33 fewer per 1000 (from 8 fewer to 50 fewer) ⊕⊕OO LOW CRITICAL
1 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Vertical transmission (48–52 weeks) 2 randomized no trials serious risk of bias no serious inconsistency serious1 serious4 see below3 92/1098 (8.4%) 105/889 (11.8%) RR 0.71 (0.54 to 0.92) 34 fewer per 1000 (from 9 fewer to 54 fewer) ⊕⊕OO LOW CRITICAL
Mortality (48–52 weeks) 2 randomized no trials serious risk of bias no serious inconsistency serious1 very serious2 see below3 27/1261 (2.1%) 32/1074 (3%) RR 0.68 10 fewer per CRITICAL ⊕OOO (0.41 to 1000 (from 18 VERY LOW 1.13) fewer to 4 more)
Maternal morbidity (48 weeks) 1 randomized no trials serious risk of bias no serious inconsistency serious1 very serious2 none 48/849 (5.7%) 22/662 (3.3%) RR 1.7 (1.04 to 2.79) 23 more per 1000 (from 1 more to 59 more) IMPORTANT ⊕OOO VERY LOW
Infant morbidity (48 weeks) 1 randomized no trials serious risk of bias no serious inconsistency serious1 no serious imprecision none 230/849 (27.1%) 188/662 (28.4%) RR 0.95 (0.81 to 1.12) 14 fewer per ⊕⊕⊕O IMPORTANT 1000 (from 54 MODERATE fewer to 34 more)
Maternal severe adverse events (12 months) 1 randomized no trials serious risk of bias no serious inconsistency serious1 serious4 see below3 57/412 (13.8%) 48/412 (11.7%) RR 1.19 (0.83 to 1.7) 22 more per 1000 (from 20 fewer to 82 more) ⊕⊕OO LOW IMPORTANT
Infant severe adverse events (12 months) 1 randomized no trials serious risk of bias no serious inconsistency serious1 serious4 see below3 107/412 (26%) 110/412 (26.7%) RR 0.97 (0.77 to 1.22) 8 fewer per 1000 (from 61 fewer to 59 more) ⊕⊕OO LOW IMPORTANT
2 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1 2
Studies do not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Very few events. 3 In Kesho Bora, “mothers who intended to breastfeed”. 4 Few events.
DIRECT 1-B: Infant ARV plus breastfeeding versus exclusive breastfeeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: Infant ARV plus breastfeeding versus exclusive breastfeeding for preventing mother-to-child HIV transmission Settings: Ethiopia, India, Malawi, South Africa, Uganda, United Republic of Tanzania, Zimbabwe Bibliography: Chasela 2010, Jamieson 2012 (BAN); Coovadia 2012 (HPTN 046); Kumwenda 2008, Taha 2011 (PEPI); Bedri 2008, Omer 2011 (SWEN)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Infant ARV + breastfeeding
Exclusive breastfeeding
Relative (95% CI)
Absolute
Vertical transmission (6 months) 3 randomized no serious no serious trials risk of bias inconsistency serious2 serious3 see below1 81/2319 (3.5%) 140/2201 (6.4%) RR 0.52 31 fewer per 1000 (0.4 to 0.67) (from 21 fewer to 38 fewer) ⊕⊕OO LOW CRITICAL
Vertical transmission (9 months) 2 randomized no serious no serious trials5 risk of bias inconsistency serious2 serious3 see below1,4 122/2496 (4.9%) 118/2486 (4.7%) RR 1.03 1 more per 1000 (0.8 to 1.32) (from 9 fewer to 15 more) ⊕⊕OO LOW CRITICAL
Vertical transmission (12 months) 3 randomized no serious no serious trials risk of bias inconsistency serious2 no serious imprecision none 155/2230 (7%) 195/2107 (9.3%) RR 0.75 (0.61 to 0.91) 23 fewer per 1000 CRITICAL ⊕⊕⊕O (from 8 fewer to 36 MODERATE fewer)
3 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Vertical transmission (24 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 no serious imprecision see below5 193/2013 (9.6%) 272/2006 (13.6%) RR 0.71 (0.59 to 0.84) 39 fewer per 1000 CRITICAL ⊕⊕⊕O (from 22 fewer to 56 MODERATE fewer)
Mortality (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious6 none 11/977 (1.1%) 17/1047 (1.6%) RR 0.69 (0.33 to 1.47) 5 fewer per 1000 CRITICAL ⊕OOO (from 11 fewer to 8 VERY LOW more)
Mortality (9 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 serious3 see below5 55/1016 (5.4%) 71/1003 (7.1%) RR 0.76 (0.54 to 1.08) 17 fewer per 1000 (from 33 fewer to 6 months more) ⊕⊕OO LOW CRITICAL
Mortality (12 months) 2 randomized no serious no serious trials risk of bias inconsistency serious2 serious3 none 49/1829 (2.7%) 80/1709 (4.7%) RR 0.57 20 fewer per 1000 (0.4 to 0.82) (from 8 fewer to 28 fewer) ⊕⊕OO LOW CRITICAL
Mortality (18 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 no serious imprecision see below5 187/2013 (9.3%) 212/2006 (10.6%) RR 0.88 (0.73 to 1.06) 13 fewer per 1000 CRITICAL ⊕⊕⊕O (from 29 fewer to 6 MODERATE months more)
Maternal morbidity (48 weeks)7 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious6 None 32/852 (3.8%) 22/662 (3.3%) RR 1.13 (0.66 to 1.93) 4 more per 1000 IMPORTANT ⊕OOO (from 11 fewer to 31 VERY LOW more)
Infant morbidity (48 weeks)7
4 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
randomized no serious no serious trials risk of bias inconsistency
serious2
no serious imprecision
none
269/852 (31.6%)
188/662 (28.4%)
RR 1.11 31 more per 1000 ⊕⊕⊕O IMPORTANT (0.95 to 1.3) (from 14 fewer to 85 MODERATE more)
TB incidence (cases) (48 weeks) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious6 none 4/852 (0.47%) 4/662 (0.6%) RR 0.78 1 fewer per 1000 IMPORTANT ⊕OOO (0.2 to 3.1) (from 5 fewer to 13 VERY LOW more)
Severe adverse event possibly/probably related to drug (9 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 serious3 see below5 123/2013 (6.1%) 80/2006 (4%) RR 1.53 (1.16 to 2.02) 21 more per 1000 (from 6 months more to 41 more) ⊕⊕OO LOW IMPORTANT
1 2
In HPTN 046, these were mothers not receiving ART at randomization. Six months after randomization, 31% of intervention mothers and 32% of placebo mothers were receiving ART. Studies do not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. 3 Few events. 4 In HPTN 046, more than 95% of infants in both groups no longer breastfed by the 9-month study visit. 5 In PEPI, data from both treatment arms are combined. 6 Very few events. 7 Combined count of all clinical severe adverse events.
INDIRECT-1: Both arms on ART or ARV prophylaxis with different types or lengths of breastfeeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: ART with exclusive breastfeeding versus ART with other types or length of feeding for preventing mother-to-child HIV transmission Settings: Uganda Bibliography: Homsy 2010
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
ART with other Other ART with exclusive Relative Inconsistency Indirectness Imprecision types or length of considerations breastfeeding (95% CI) feeding
Absolute
5 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Vertical transmission (median 18.1 months) 1 observational studies1 no serious no serious risk of bias inconsistency serious2 very serious3 none5 0/58 (0%) 0/89 (0%) not pooled not pooled ⊕OOO CRITICAL VERY LOW
Mortality (median 18.1 months) 1 observational studies1 no serious no serious risk of bias inconsistency serious2 very serious4 none5 2/58 (3.4%) 22/89 (24.7%) RR 0.14 (0.03 to 0.57) 213 fewer per 1000 (from 106 fewer to 240 fewer) ⊕OOO CRITICAL VERY LOW
1 2
In Homsy 2010, data from mixed feeding and weaned/never breastfeeding arms are combined. Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. 3 No events. 4 Very few events. 5 Feeding status at 6 months.
INDIRECT-2: Neither arm on ART or ARV prophylaxis with different types or lengths of breastfeeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should exclusive breastfeeding or another type of breastfeeding be used for preventing mother-to-child HIV transmission? Settings: Côte d'Ivoire Bibliography: Leroy 2008 (DITRAME-PLUS versus DITRAME’s AZT arm)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency
Indirectness Imprecision
Other considerations
Maternal ARVs, breastfeeding 4 months
Maternal AZT, unrestricted breastfeeding5
Relative (95% CI)
Absolute
Vertical transmission (18 months) 11 observational no serious no serious studies risk of inconsistency bias no serious indirectness very serious2 none3 35/367 (9.5%) 47/238 (19.7%) RR 0.48 (0.32 to 0.72) 103 fewer per 1000 ⊕OOO CRITICAL (from 55 fewer to VERY 134 fewer) LOW
6 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1 2
DITRAME PLUS (AZT + sdNVP; AZT + 3TC + sdNVP), compared to DITRAME AZT (unrestricted breastfeeding, maternal AZT). Very few events. 3 Not graded up for large effect due to very serious imprecision. 5 Predominantly breastfed for a median duration of 7.3 months (interquartile range 6–10 months). Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should exclusive breastfeeding for 0–5 months versus exclusive breastfeeding for 6–11 months be used for preventing mother-to-child HIV transmission? Settings: United Republic of Tanzania Bibliography: Fawzi 2002, Natchu 2012 (Trial of Vitamins)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Exclusive breastfeeding 0–5 months
Exclusive breastfeeding 6–11 months
Relative (95% CI)
Absolute
Vertical transmission (60 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious3 see below1 47/585 (8%) 19/585 (3.2%) RR 2.47 (1.47 to 4.16) 48 more per 1000 (from 15 more to 103 more) ⊕OOO CRITICAL VERY LOW
Mortality (60 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious3 see below1 16/690 (2.3%) 43/690 (6.2%) RR 0.37 (0.21 to 0.65) 39 fewer per 1000 ⊕OOO CRITICAL (from 22 fewer to 49 VERY fewer) LOW
1 2
Only 30% were exclusive breastfeeding for ≥3 months. No ARVs or ART available. 3 Very few events.
7 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should exclusive breastfeeding for 6–11 months versus exclusive breastfeeding for 12–23 months be used for preventing mother-to-child HIV transmission? Settings: United Republic of Tanzania Bibliography: Fawzi 2002, Natchu 2012 (Trial of Vitamins)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
exclusive exclusive breastfeeding 6-11 breastfeeding 12-23 months months
Relative (95% CI)
Absolute
Vertical transmission (60 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious3 see below1 19/585 (3.2%) 39/585 (6.7%) RR 0.49 (0.28 to 0.83) 34 fewer per 1000 ⊕OOO CRITICAL (from 11 fewer to 48 VERY fewer) LOW
Mortality (60 months) 1 randomized no serious no serious trials risk of bias inconsistency serious2 very serious3 see below1 43/690 (6.2%) 54/690 (7.8%) RR 0.8 (0.54 16 fewer per 1000 ⊕OOO CRITICAL to 1.17) (from 36 fewer to 13 VERY more) LOW
1 2
Only 30% were exclusive breastfeeding for ≥3 months. No ART or ARVs available. 3 Very few events.
8 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should exclusive breastfeeding with abrupt cessation at 4 months versus exclusive breastfeeding ≥6 months be used for preventing mother-to-child HIV transmission? Settings: Zambia Bibliography: Kuhn 2005, Kuhn 2008, Arpadi 2009, Kuhn 2009, Fawzy 2011 (ZEBS)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Abrupt cessation (4 months)
Exclusive breastfeeding ≥6 months
Relative (95% CI)
Absolute
HIV-free survival (24 months) 1 randomized no serious no serious trials risk of inconsistency bias serious1 no serious imprecision none 247/328 (75.3%) 239/333 (71.8%) RR 1.05 (0.96 to 1.15) 36 months more per CRITICAL ⊕⊕⊕O 1000 (from 29 fewer to MODERATE 108 more)
Any diarrhoea (4.5 months) (denominator back-calculated from percentages) 12 randomized no serious no serious trials risk of inconsistency bias serious1 very serious3 none 34/158 (21.5%) 17/340 (5%) RR 4.3 (2.48 to 7.47) 165 more per 1000 IMPORTANT ⊕OOO (from 74 more to 324 VERY LOW more)
Any diarrhoea (6 months) (denominator back-calculated from percentages) 12 randomized no serious no serious trials risk of inconsistency bias serious1 very serious3 none 61/203 (30%) 52/369 (14.1%) RR 2.13 (1.54 to 2.96) 159 more per 1000 IMPORTANT ⊕OOO (from 76 months more VERY LOW to 276 months more)
Any diarrhoea (12 months) (denominator back-calculated from percentages) 12 randomized no serious no serious trials risk of inconsistency bias serious1 serious4 none 99/226 (43.8%) 109/280 (38.9%) RR 1.13 (0.91 to 1.39) 51 more per 1000 (from 35 fewer to 152 more) ⊕⊕OO LOW IMPORTANT
Any diarrhoea (18 months) (denominator back-calculated from percentages)
9 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
12
randomized no serious no serious trials risk of inconsistency bias
serious1
serious4
none
103/304 (33.9%)
45/157 (28.7%)
RR 1.18 (0.88 to 1.58)
52 more per 1000 (from IMPORTANT ⊕OOO 34 fewer to 166 months VERY LOW more)
Any diarrhoea (24 months) (denominator back-calculated from percentages) 12 randomized no serious no serious trials risk of inconsistency bias serious1 very serious3 none 98/434 (22.6%) 7/18 (38.9%) RR 0.58 (0.32 to 1.06) 163 fewer per 1000 IMPORTANT ⊕OOO (from 264 fewer to 23 VERY LOW more)
Weight-for-age Z scores (4.5 months) (better indicated by higher values) 12 randomized no serious no serious trials risk of inconsistency bias serious1 no serious imprecision none 167 330 – Mean difference 0.03 ⊕⊕⊕O IMPORTANT lower (0.05 to 0.01 MODERATE lower)
Weight-for-age Z scores (6 months) (better indicated by higher values) 12 randomized no serious no serious trials risk of inconsistency bias serious1 no serious imprecision none 213 353 – Mean difference 0.17 ⊕⊕⊕O IMPORTANT lower (0.18 to 0.16 MODERATE lower)
Weight-for-age Z scores (12 months) (better indicated by higher values) 12 randomized no serious no serious trials risk of inconsistency bias serious1 no serious imprecision none 240 266 – Mean difference 0.28 ⊕⊕⊕O IMPORTANT lower (0.29 to 0.27 MODERATE lower)
Weight-for-age Z scores (18 months) (better indicated by higher values) 12 randomized no serious no serious trials risk of inconsistency bias serious1 no serious imprecision none 347 114 – Mean difference 0.18 ⊕⊕⊕O IMPORTANT higher (0.16 to 0.2 MODERATE higher)
Weight-for-age Z scores (24 months) (better indicated by higher values) 12 randomized no serious no serious trials risk of inconsistency bias serious1 very serious5 none 442 9 – Mean difference 0.2 IMPORTANT ⊕OOO higher (0.09 lower to VERY LOW 0.49 higher)
1 2
ARVs not available until near end of trial. “As practiced” analysis.
10 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
3 4
Very few events. Few events. 5 Very few participants in exclusive breastfeeding group. Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should early (3 months) exclusive breastfeeding versus early (3 months) predominant breastfeeding be used for preventing mother-to-child HIV transmission?1 Settings: Zimbabwe Bibliography: Iliff 2005 (ZVITAMBO)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Early (3 months) exclusive breastfeeding
Early (3 months) predominant breastfeeding
Relative (95% CI)
Absolute
Vertical transmission (6 months) (numerator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious2 very serious3 none 2/156 (1.3%) 15/490 (3.1%) RR 0.42 18 fewer per 1000 ⊕OOO CRITICAL (0.1 to 1.81) (from 28 fewer to VERY 25 more) LOW
Vertical transmission (12 months) (numerator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious2 very serious3 none 5/156 (3.2%) 36/490 (7.3%) RR 0.44 (0.17 to 1.09) 41 fewer per 1000 ⊕OOO CRITICAL (from 61 fewer to 7 VERY more) LOW
Vertical transmission (18 months) (numerator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious2 very serious3 none 11/156 (7.1%) 42/490 (8.6%) RR 0.82 (0.43 to 1.56) 15 fewer per 1000 ⊕OOO CRITICAL (from 49 fewer to VERY 48 more) LOW
1 2
Predominant breastfeeding. Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Also, no ARVs or ART available. 3 Very few events.
11 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should early (3 months) exclusive breastfeeding versus early (3 months) mixed breastfeeding be used for preventing mother-to-child HIV transmission? Settings: Zimbabwe Bibliography: Iliff 2005 (ZVITAMBO)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Early (3 months) exclusive breastfeeding
Early (3 months) Relative mixed breastfeeding (95% CI)
Absolute
Vertical transmission (6 months) (denominator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious1 very serious2 none 2/156 (1.3%) 62/1414 (4.4%) RR 0.29 (0.07 to 1.18) 31 fewer per 1000 ⊕OOO CRITICAL (from 41 fewer to 8 VERY more) LOW
Vertical transmission (12 months) (denominator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious1 very serious2 none3 5/156 (3.2%) 119/1414 (8.4%) RR 0.38 (0.16 to 0.92) 52 fewer per 1000 ⊕OOO CRITICAL (from 7 fewer to 71 VERY fewer) LOW
Vertical transmission (18 months) (denominator back-calculated from percentages) 1 randomized no serious no serious trials risk of bias inconsistency very serious1 very serious2 none3 11/156 (7.1%) 197/1414 (13.9%) RR 0.51 (0.28 to 0.91) 68 fewer per 1000 (from 13 fewer to 100 fewer) ⊕OOO CRITICAL VERY LOW
1
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Also, no ARVs or ART available. 2 Very few events. 3 Not graded up for large effect due to multiple downgradings.
12 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-24 Question: Should exclusive breastfeeding versus mixed breastfeeding be used for preventing mother-to-child HIV transmission? Settings: KwaZulu Natal, South Africa Bibliography: Bobat 1997, Coutsoudis 1999
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias Inconsistency Indirectness Imprecision
Other considerations
Exclusive breastfeeding
Mixed breastfeeding
Relative (95% CI)
Absolute
Vertical transmission (3 months) (note: numerator back-calculated from percentages) 1 observational studies no serious no serious risk of bias inconsistency very serious1 very serious2 none 15/103 (14.6%) 69/288 (24%) 93 fewer per 1000 RR 0.61 (0.36 to 1.01) (from 153 fewer to 2 more) ⊕OOO CRITICAL VERY LOW
Vertical transmission (mean 22–23 months) 13 observational studies no serious no serious risk of bias inconsistency very serious1 very serious2 none 14/36 (38.9%) 24/76 (31.6%) RR 1.23 73 more per 1000 (0.73 to 2.08) (from 85 fewer to 341 more) ⊕OOO CRITICAL VERY LOW
1 2
Study does not directly address the relative effect of breastfeeding duration. Also, no ARVs or ART available. Very few events. 3 Median exclusive breastfeeding duration was 12 months (range 1–12). Median exclusive breastfeeding in mixed feeders was 2 months (range 1–12).
INDIRECT-3: ART or ARV prophylaxis with breastfeeding versus formula feeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: Maternal ART or ARV prophylaxis with breastfeeding versus formula feeding for preventing mother-to-child HIV transmission Settings: India, Malawi, Mozambique, Rwanda, United Republic of Tanzania Bibliography: Palombi 2007, Marazzi 2009 (DREAM); Peltier 2009; Alvarez-Uria 2012 Quality assessment No. of patients Effect Quality Importance
13 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Maternal ART or ARV prophylaxis with breastfeeding
Formula feeding
Relative (95% CI)
Absolute
Mortality (12 months) 1 observational studies no serious no serious risk of bias inconsistency serious very serious1 none 8/287 (2.8%) 115/287 (40.1%) RR 0.07 (0.03 to 0.14) 373 fewer per 1000 (from 345 fewer to 389 fewer) ⊕OOO VERY LOW CRITICAL
Vertical transmission (6 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 very serious1 none 2/251 (0.8%) 15/809 (1.9%) RR 0.43 (0.1 11 fewer per 1000 ⊕OOO to 1.87) (from 17 fewer to 16 VERY months more) LOW CRITICAL
Vertical transmission (9 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 very serious1 none 1/227 (0.44%) 0/305 (0%) RR 4.03 (0.16 to 98.38) ⊕OOO VERY LOW CRITICAL
HIV-free survival (12 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious3 none 126/131 (96.2%) 130/151 (86.1%) RR 1.12 103 more per 1000 ⊕OOO (1.04 to 1.2) (from 34 more to 172 VERY more) LOW CRITICAL
Mortality (9 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 very serious1 none 7/227 (3.1%) 17/305 (5.6%) RR 0.55 (0.23 to 1.31) 25 fewer per 1000 ⊕OOO (from 43 fewer to 17 VERY more) LOW CRITICAL
Weight-for-age Z scores ≤2.0 (6 months)
14 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
observational studies
no serious no serious risk of bias inconsistency
serious2
very serious1
none
28/251 (11.2%)
92/809 (11.4%)
RR 0.98 (0.66 to 1.46)
2 fewer per 1000 ⊕OOO IMPORTANT (from 39 fewer to 52 VERY more) LOW
1 2
Very few events. Studies do not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. 3 Few events. Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: Should extended infant ARV prophylaxis with exclusive breastfeeding or no extended prophylaxis and formula feeding be used for preventing mother-to-child HIV transmission? Settings: Botswana Bibliography: Thior 2006, Shapiro 2009 (Mashi)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Infant 6 months of AZT, Infant 1 month Other Inconsistency Indirectness Imprecision 6 months of exclusive AZT, 6 months of considerations breastfeeding formula feeding
Relative (95% CI)
Absolute
Vertical transmission between months 1–3 1 randomized no serious no serious trials risk of inconsistency bias serious1 very serious2 none 15/547 (2.7%) 1/537 (0.19%) RR 14.73 (1.95 to 111.09) 26 months more per ⊕OOO CRITICAL 1000 (from 2 more VERY to 205 more) LOW
Vertical transmission between months 4–6 1 randomized no serious no serious trials risk of inconsistency bias serious1 very serious2 none 6/547 (1.1%) 1/537 (0.19%) RR 5.89 (0.71 to 48.76) 9 more per 1000 ⊕OOO CRITICAL (from 1 fewer to 89 VERY more) LOW
Vertical transmission between months 7–24 1 randomized no serious no serious trials risk of inconsistency bias serious1 very serious2 none 3/547 (0.55%) 0/537 (0%) RR 6.87 (0.36 to 132.73) ⊕OOO CRITICAL VERY LOW
AZT: zidovudine.
15 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1 2
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Very few events.
INDIRECT-4: breastfeeding versus formula feeding without extended maternal or infant ART or ARV prophylaxis Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: In the absence of without extended maternal or infant ART or ARV prophylaxis , should exclusive breastfeeding or formula feeding be used for preventing mother-to-child HIV transmission? Settings: Côte d'Ivoire, KwaZulu Natal, South Africa Bibliography: Bobat 1997; Coutsoudis 1999; Leroy 2008 (DITRAME PLUS data); Kindra 2012
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Exclusive breastfeeding
Formula feeding without prophylaxis
Relative (95% CI)
Absolute
Vertical transmission (3 months) (numerator back-calculated from percentages) 1 observational studies1 no serious no serious risk of bias inconsistency serious2 serious5 none 30/206 (14.6%) 98/444 (22.1%) RR 0.66 (0.45 to 0.96) 75 fewer per 1000 (from 9 fewer to 121 fewer) ⊕OOO VERY LOW CRITICAL
Vertical transmission (18 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 very serious3 none 25/321 (7.8%) 47/238 (19.7%) RR 0.39 (0.25 to 0.62) 120 fewer per 1000 ⊕OOO (from 75 fewer to 148 VERY fewer) LOW CRITICAL
Vertical transmission (mean 21–23 months) 1 observational studies no serious no serious risk of bias inconsistency serious2 very serious3 none 28/72 (38.9%) 29/97 (29.9%) RR 1.3 (0.85 to 1.98) 90 more per 1000 ⊕OOO (from 45 fewer to 293 VERY more) LOW CRITICAL
Mid-upper arm circumference for age Z scores (14 weeks) (better indicated by higher values)
16 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
observational studies
no serious no serious risk of bias inconsistency
serious2
serious4
none
60
67
-
Mean difference 0.43 lower (0.79 to 0.07 lower)
⊕OOO IMPORTANT VERY LOW
Mid-upper arm circumference for age Z scores (6 months) (better indicated by higher values) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious4 none 60 67 Mean difference 0.38 higher (0.01 lower to 0.77 higher) ⊕OOO IMPORTANT VERY LOW
Mid-upper arm circumference for age Z scores (9 months) (better indicated by higher values) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious4 none 60 67 Mean difference 0.3 higher (0.09 lower to 0.69 higher) ⊕OOO IMPORTANT VERY LOW
BMI for age Z scores (14 weeks) (better indicated by higher values) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious4 none 60 67 Mean difference 0.56 higher (0.22 to 0.9 higher) ⊕OOO IMPORTANT VERY LOW
BMI for age Z scores (6 months) (better indicated by higher values) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious4 none 60 67 Mean difference 0.41 higher (0.01 to 0.81 higher) ⊕OOO IMPORTANT VERY LOW
BMI for age Z scores (9 months) (better indicated by higher values) 1 observational studies no serious no serious risk of bias inconsistency serious2 serious4 none 60 67 Mean difference 0.41 higher (0 to 0.82 higher) ⊕OOO IMPORTANT VERY LOW
1 2
In Bobat (1997) and Coutsoudis (1999), data from mixed feeding and formula feeding arms are combined. Studies do not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. 3 Very few events. 4 Few participants. 5 Few events.
17 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-23 Question: In the absence of without extended maternal or infant ART/ARV prophylaxis, should unrestricted breastfeeding or formula feeding be used for preventing mother-to-child HIV transmission? Settings: Kenya Bibliography: Nduati 2000
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
Unrestricted breastfeeding
formula feeding
Relative (95% CI)
Absolute
Vertical transmission (14 weeks) 1 randomized trials no serious no serious risk of bias inconsistency very serious1 very serious2 none3 47/197 (23.9%) 28/204 RR 1.74 (1.14 102 more per 1000 (from (13.7%) to 2.66) 19 more to 228 more) ⊕OOO CRITICAL VERY LOW
Vertical transmission (6 months) 1 randomized trials no serious no serious risk of bias inconsistency very serious1 very serious2 none3 53/197 (26.9%) 32/204 RR 1.72 (1.16 113 more per 1000 (from (15.7%) to 2.54) 25 more to 242 more) ⊕OOO CRITICAL VERY LOW
Vertical transmission (12 months) 1 randomized trials no serious no serious risk of bias inconsistency very serious1 serious4 none3 63/197 (32%) 36/204 RR 1.81 (1.26 143 more per 1000 (from (17.6%) to 2.6) 46 months more to 282 more) ⊕OOO CRITICAL VERY LOW
Vertical transmission (24 months) 1 randomized trials no serious no serious risk of bias inconsistency very serious3 serious4 none3 71/197 (36%) 41/204 RR 1.79 (1.29 159 more per 1000 (from (20.1%) to 2.5) 58 more to 301 more) ⊕OOO CRITICAL VERY LOW
HIV-free survival (24 months)
18 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
randomized trials
no serious no serious risk of bias inconsistency
very serious1 serious4
none
117/197 (59.4%)
144/204 RR 0.84 (0.73 113 fewer per 1000 (from (70.6%) to 0.97) 21 fewer to 191 fewer)
⊕OOO CRITICAL VERY LOW
1
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Also, no ARVs or ART available. 2 Very few events. 3 Not graded up for large effect due to multiple downgradings. 4 Few events.
INDIRECT-5: Both arms on ART or ARV with different regimens with same length of exclusive breastfeeding Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: Should maternal triple therapy with ABC or LPV/r with exclusive breastfeeding for 6 months be used for preventing mother-to-child HIV transmission? Settings: Botwana Bibliography: Shapiro 2010 (Mma'Bana)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
ABC, exclusive breastfeeding 6 months
LPV/r, exclusive breastfeeding 6 months
Relative (95% CI)
Absolute
Vertical transmission (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 6/283 (2.1%) 1/270 (0.37%) RR 5.72 (0.69 to 47.24) 17 more per 1000 ⊕OOO (from 1 fewer to 171 VERY more) LOW CRITICAL
Mortality (mother) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 1/285 (0.35%) 0/275 (0%) RR 2.9 (0.12 to 70.76) ⊕OOO VERY LOW CRITICAL
19 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Mortality (infant) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 7/283 (2.5%) 7/270 (2.6%) RR 0.95 (0.34 to 2.68) 1 fewer per 1000 ⊕OOO (from 17 fewer to 44 VERY more) LOW CRITICAL
≥1 grade 3 or 4 severe adverse event (mothers) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 17/285 (6%) 16/275 (5.8%) RR 1.03 (0.53 to 1.99) 2 more per 1000 ⊕OOO IMPORTANT (from 27 fewer to 58 VERY more) LOW
≥1 grade 3 or 4 severe adverse event (infants) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 28/283 (9.9%) 17/270 (6.3%) RR 1.57 36 months more per ⊕OOO IMPORTANT (0.88 to 2.8) 1000 (from 8 fewer VERY to 113 more) LOW
ABC, abacavir; LPV/r, ritonavir-boosed lopinavir 1 2
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Very few events.
20 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: Should maternal triple therapy with ABC or NVP with exclusive breastfeeding for 6 months be used for preventing mother-to-child HIV transmission Settings: Botswana Bibliography: Shapiro 2010 (Mma'Bana)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
ABC, exclusive breastfeeding 6 months
NVP, exclusive breastfeeding 6 months
Relative (95% CI)
Absolute
Vertical transmission (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 6/283 (2.1%) 1/156 (0.64%) RR 3.31 (0.4 to 27.22) 15 more per 1000 ⊕OOO (from 4 fewer to 168 VERY more) LOW CRITICAL
Mortality (mother) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 1/285 (0.35%) 3/170 (1.8%) RR 0.2 14 fewer per 1000 ⊕OOO (0.02 to 1.9) (from 17 fewer to 16 VERY months more) LOW CRITICAL
Mortality (infant) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 7/283 (2.5%) 7/156 (4.5%) RR 0.55 20 fewer per 1000 OOO (0.2 to 1.54) (from 36 fewer to 24 VERY more) LOW CRITICAL
≥1 grade 3 or 4 severe adverse event (mothers) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 17/285 (6%) 25/170 (14.7%) RR 0.41 (0.23 to 0.73) 87 fewer per 1000 (from 40 fewer to 113 fewer) OOO IMPORTANT VERY LOW
≥1 grade 3 or 4 severe adverse event (infants) (6 months)
21 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
randomized no serious no serious trials risk of bias inconsistency
serious1
very serious2
none
28/283 (9.9%)
13/156 (8.3%)
RR 1.19 (0.63 to 2.22)
16 months more per OOO IMPORTANT 1000 (from 31 fewer VERY to 102 more) LOW
ABC, abacavir; NVP, nevirapine 1 2
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Very few events.
Author(s): George W. Rutherford, Tara Horvath Date: 2012-11-29 Question: Maternal LPV/r versus NVP, breastfeeding 6 months for preventing mother-to-child HIV transmission Settings: Botswana Bibliography: Shapiro 2010 (Mma'Bana)
Quality assessment
No. of patients
Effect Quality Importance
No. of studies
Design
Risk of bias
Inconsistency Indirectness Imprecision
Other considerations
LPV/r, exclusive breastfeeding 6 months
NVP, exclusive breastfeeding 6 months
Relative (95% CI)
Absolute
Vertical transmission (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 1/270 (0.37%) 1/156 (0.64%) RR 0.58 (0.04 to 9.17) 3 fewer per 1000 OOO (from 6 fewer to 52 VERY more) LOW CRITICAL
Mortality (mother) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 0/275 (0%) 3/170 (1.8%) RR 0.09 (0 16 fewer per 1000 to 1.7) (from 18 fewer to 12 more) OOO VERY LOW CRITICAL
Mortality (infant) (6 months)
22 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS
WHO/HIV/2013.24
© World Health Organization 2013
1
randomized no serious no serious trials risk of bias inconsistency
serious1
very serious2
none
7/270 (2.6%)
7/156 (4.5%)
RR 0.58 (0.21 to 1.62)
19 fewer per 1000 (from 35 fewer to 28 more)
OOO VERY LOW
CRITICAL
≥1 grade 3 or 4 severe adverse event (mothers) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 16/275 (5.8%) 25/170 (14.7%) RR 0.4 (0.22 to 0.72) 88 fewer per 1000 (from 41 fewer to 115 fewer) OOO IMPORTANT VERY LOW
≥1 grade 3 or 4 severe adverse event (infants) (6 months) 1 randomized no serious no serious trials risk of bias inconsistency serious1 very serious2 none 17/270 (6.3%) 13/156 (8.3%) RR 0.76 (0.38 to 1.51) 20 fewer per 1000 (from 52 fewer to 42 more) OOO IMPORTANT VERY LOW
LPV/r, ritonavir-boosted lopinavir; NVP, nevirapine. 1 2
Study does not directly compare prolonged postpartum maternal ARV prophylaxis and standard breastfeeding to exclusive breastfeeding (2010 WHO guidelines) without ARV prophylaxis. Very few events.
23 This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS