Articles

First-line antiretroviral therapy with a protease inhibitor

versus non-nucleoside reverse transcriptase inhibitor and
switch at higher versus low viral load in HIV-infected
children: an open-label, randomised phase 2/3 trial
The PENPACT-1 (PENTA 9/PACTG 390) Study Team*

Summary
Background Children with HIV will be on antiretroviral therapy (ART) longer than adults, and therefore the Lancet Infect Dis 2011;
durability of first-line ART and timing of switch to second-line are key questions. We assess the long-term outcome 11: 273–83

of protease inhibitor and non-nucleoside reverse transcriptase inhibitor (NNRTI) first-line ART and viral load Published Online
February 1, 2011
switch criteria in children.
DOI:10.1016/S1473-
3099(10)70313-3
Methods In a randomised open-label factorial trial, we compared effectiveness of two nucleoside reverse transcriptase See Comment page 254
inhibitors (NRTIs) plus a protease inhibitor versus two NRTIs plus an NNRTI and of switch to second-line ART at a *Details of the writing
viral load of 1000 copies per mL versus 30 000 copies per mL in previously untreated children infected with HIV from committee at the end of the
Europe and North and South America. Random assignment was by computer-generated sequentially numbered lists paper and full details in the
webappendix (pp 3–4)
stratified by age, region, and by exposure to perinatal ART. Primary outcome was change in viral load between baseline
and 4 years. Analysis was by intention to treat, which we defined as all patients that started treatment. This study is Correspondence to:
Linda Harrison, MRC Clinical
registered with ISRCTN, number ISRCTN73318385. Trials Unit, 222 Euston Road,
London NW1 2DA, UK
Findings Between Sept 25, 2002, and Sept 7, 2005, 266 children (median age 6·5 years; IQR 2·8–12·9) were randomly lijh@ctu.mrc.ac.uk
assigned treatment regimens: 66 to receive protease inhibitor and switch to second-line at 1000 copies per mL (PI-low),
65 protease inhibitor and switch at 30 000 copies per mL (PI-higher), 68 NNRTI and switch at 1000 copies per mL
(NNRTI-low), and 67 NNRTI and switch at 30 000 copies per mL (NNRTI-higher). Median follow-up was 5·0 years
(IQR 4·2–6·0) and 188 (71%) children were on first-line ART at trial end. At 4 years, mean reductions in viral load
were –3·16 log10 copies per mL for protease inhibitors versus –3·31 log10 copies per mL for NNRTIs (difference
–0·15 log10 copies per mL, 95% CI –0·41 to 0·11; p=0·26), and –3·26 log10 copies per mL for switching at the low
versus –3·20 log10 copies per mL for switching at the higher threshold (difference 0·06 log10 copies per mL, 95% CI
–0·20 to 0·32; p=0·56). Protease inhibitor resistance was uncommon and there was no increase in NRTI resistance
in the PI-higher compared with the PI-low group. NNRTI resistance was selected early, and about 10% more children
accumulated NRTI mutations in the NNRTI-higher than the NNRTI-low group. Nine children had new CDC stage-C
events and 60 had grade 3/4 adverse events; both were balanced across randomised groups.

Interpretation Good long-term outcomes were achieved with all treatments strategies. Delayed switching of protease-
inhibitor-based ART might be reasonable where future drug options are limited, because the risk of selecting for
NRTI and protease-inhibitor resistance is low.

Funding Paediatric European Network for Treatment of AIDS (PENTA) and Pediatric AIDS Clinical Trials Group
(PACTG/IMPAACT).

Introduction will need to receive ART into adulthood, and hence

In the early 2000s, paediatricians were divided in opinion
as to whether first-line antiretroviral therapy (ART)
should include either a protease inhibitor1 or a non-
nucleoside reverse transcriptase inhibitor (NNRTI).2
Furthermore, because of limited choice of antiretroviral
drugs for children and high failure rates of first-line
regimens,3–6 there was concern that if the switch to
second-line was early (soon after virological failure with
low viral load), treatment options would be quickly
exhausted. Although greater numbers of antiretroviral
drugs have become available, children starting ART early
in life (now recommended by all paediatric guidelines)7–9
potentially receive chronic treatment for many decades.
No studies have directly compared long-term clinical
outcome of first-line ART with protease inhibitors or
NNRTIs in children, and only one small randomised
trial, in adults, has investigated switching from first-line
to second-line therapy at different viral-load thresholds.10
Our trial, a collaboration between the Paediatric
European Network for Treatment of AIDS (PENTA) and
the Pediatric AIDS Clinical Trials Group (PACTG/
IMPAACT) in the USA, assesses the long-term outcome
of both these strategies with a randomised open-label
factorial design.

www.thelancet.com/infection Vol 11 April 2011 273

 Articles

Methods databases, which allowed access to the next number but

Participants
We did an international multicentre phase 2/3, random-
ised, open-label, two by two factorial trial enrolling
children infected with HIV-1 from clinical centres in
Europe and North and South America between Sept 25,
2002, and Sept 7, 2005. Children who either had not been
treated with antiretroviral drugs or had received
antiretroviral drugs for less than 56 days to reduce
mother-to-child transmission (excluding single-dose
nevirapine, protocol amendment) and who needed ART
were eligible. All parents or guardians and children, as
appropriate, gave written consent. The protocol was
approved by the relevant ethics committee or institutional
review board for each participating centre.
Randomisation and masking
Randomisation was stratified by age (<3 years or ≥3 years),
region (PENTA or PACTG centre), and by exposure to
perinatal ART to reduce mother-to-child transmission.
The computer-generated sequentially numbered ran-
domisation lists (with variable block sizes) were prepared
by the trial statistician (who had further involvement in
the trial, but was not involved in regimen allocation) and
securely incorporated within the PENTA and PACTG
not the whole list. Site personnel from participating
centres randomly assigned children by faxing or phoning
the PENTA trials unit or completing an online checklist
at the PACTG trials unit. Children were simultaneously
assigned (1:1) to start ART with two NRTIs plus either a
protease inhibitor or an NNRTI and to switch from first-
line to second-line ART at a viral-load threshold of
1000 copies per mL (low threshold) or 30 000 copies
per mL (higher threshold). This was an open label study
with no masking.
Procedures
First-line ART was defined as the initial randomly
assigned regimen, allowing drug substitutions (ideally
within the same class) for non-virological reasons
(eg, toxic effects). The initial regimen was chosen by the
treating clinician according to the randomly assigned
group. Children were switched to second-line ART if the
viral-load threshold (<1000 copies per mL or <30 000 copies
per mL) was not achieved by week 24, or if after an initial
decline in viral load by week 24 there was viral-load
rebound at or above the randomly assigned level,
confirmed within 2–5 weeks. Switch to second-line ART
was also required if the child experienced a new CDC

266 patients randomly assigned to study groups

66 assigned to PI-low 65 assigned to PI-higher 68 assigned to NNRTI-low 67 assigned to NNRTI-higher

3 excluded
1 major eligibility violation*
2 withdrew consent before
starting ART

66 started protease inhibitors
40 remained on first-line ART (34 did not
reach switch point, 6 reached switch point
but did not switch)
6 discontinued ART after first-line
20 switched to second-line (15 at switch
point, 5 after switch point)
65 started treatment (63 protease
inhibitors, 2 NNRTIs†)
56 remained on first-line ART (54 did not
reach switch point, 2 reached switch
point but did not switch)
1 discontinued ART after first-line
8 switched to second-line (1 before
switch point, 3 at switch point, 4 after
switch point)
68 started treatment (67 NNRTIs, 1 protease
inhibitors†)
48 remained on first-line ART (45 did not
reach switch point, 3 reached switch point
but did not switch)
3 discontinued ART after first-line
17 switched to second-line (6 before switch
point, 9 at switch point, 2 after
switch point)
64 started treatment (63 NNRTIs,
1 protease inhibitors†)
44 remained on first-line ART (44 did
not reach switch point)
5 discontinued ART after first-line
15 switched to second-line (5 before
switch point, 10 at switch point)

9 lost to follow-up 4 lost to follow-up 4 lost to follow-up, 7 lost to follow-up,

4 withdrew consent 1 missing viral load at
4 years

57 in follow-up at 4 years, included in 61 in follow-up at 4 years, included in 60 in follow-up at 4 years, included in 56 in follow-up at 4 years, included in
primary outcome primary outcome primary outcome primary outcome

4 lost to follow-up 5 lost to follow-up 1 lost to follow-up 4 lost to follow-up,

1 withdrew consent 1 withdrew consent
1 died 1 had viral loads after 4 years

53 in follow-up at end of study 56 in follow-up at end of study 57 in follow-up at end of study 52 in follow-up at end of study

Figure 1: Trial profile

PI=protease inhibitor (start ART with two NRTIs plus a PI). NNRTI=non-nucleoside reverse transcriptase inhibitors (start ART with two NRTIs plus an NNRTI). Low (threshold)=switch from first-line to
second-line ART at a viral-load threshold of ≥1000 copies per mL. Higher (threshold)=switch from first-line to second-line ART at a viral-load threshold of ≥30 000 copies per mL. ART=antiretroviral
therapy. NRTI=nucleoside reverse transcriptase inhibitor. *Had received antiretroviral drugs for 56 days or more for reduction of mother-to-child transmission. †Drug non-availability or refusal.

274 www.thelancet.com/infection Vol 11 April 2011

 Articles

stage-C event11 at or after 24 weeks of ART. Children
randomly assigned to receive first-line ART that contained
protease inhibitors were strongly encouraged to switch to
second-line ART that contained NNRTI and vice-versa—
ie, they were not mandated to use these second-line
regimens, but they were highly recommended.
Children were assessed at screening (week –2),
randomisation (week 0), weeks 2, 4, 8, 12, 16, 24, and then
every 12 weeks until the last child randomly assigned to
treatment reached 4 years of follow-up.
Our primary outcome was change in log10 HIV RNA
viral load between baseline (mean of viral loads at
screening and randomisation) and 4 years (mean of viral
loads at weeks 192 and 204). Measurements of viral loads
for the primary comparison were done in two centralised
laboratories with real-time PCR (lower cutoff 40 copies
per mL; Abbott RealTime, Abbott, IL, USA). Local viral
loads were used when samples were unavailable for
central testing.
Secondary outcomes were regimen switch, change in
the proportion of CD4 (CD4%) from baseline to 4 years,
viral load less than 400 copies per mL at week 24 on
first-line ART, viral load less than 400 copies per mL at
4 years, continued viral-load suppression (never
confirmed >400 copies per mL after 24 weeks) on first-
line ART, failure of second-line ART (defined as
confirmed viral load >30 000 copies per mL or
discontinuation of second-line ART), grade 3/4 adverse
events (non-HIV related), new CDC stage-C events, and
resistance (done on stored samples in two central
laboratories). Baseline resistance tests were done on
samples within 84 days before randomisation.
Resistance during follow-up was measured on the last
sample with a viral load greater than 1000 copies per mL
before switch, or at confirmed viral load greater than
1000 copies per mL before resuppression (to ensure a
fair comparison between the low-threshold and higher-
threshold groups). Additionally, resistance testing was
done on samples with a viral load greater than
1000 copies per mL at 4 years and trial end. Major
resistance mutations were defined according to the
December 2009 International AIDS Society–USA

Total (n=263) Protease-inhibitor NNRTI group Low-threshold group Higher-threshold group

group (n=131) (n=132) (n=134) (n=129)
Number of boys (%) 136 (52%) 69 (53%) 67 (51%) 71 (53%) 65 (50%)
Median age (years) 6·5 (IQR 2·8–12·9; 7·1 (2·8–13·7; 6·4 (2·7–11·0; 6·1 (2·5–13·1; 6·9 (3·1–12·5;
range 0·1–17·8) 0·2–17·8) 0·1–17·6) 0·3–17·8) 0·1–17·5)
Ethnic origin
White 69 (26%) 40 (31%) 29 (22%) 34 (25%) 35 (27%)
Black 129 (49%) 60 (46%) 69 (52%) 64 (48%) 65 (50%)
Hispanic, Latino, Asian, or mixed 65 (25%) 31 (24%) 34 (26%) 36 (27%) 29 (22%)
Route of infection
Vertical 209 (79%) 103 (79%) 106 (80%) 104 (78%) 105 (81%)
Parenteral 36 (14%) 20 (15%) 16 (12%) 19 (14%) 17 (13%)
Sexual contact 13 (5%) 6 (5%) 7 (5%) 9 (7%) 4 (3%)
Uncertain 5 (2%) 2 (2%) 3 (2%) 2 (1%) 3 (2%)
CDC Stage
N or A 128 (49%) 62 (47%) 66 (50%) 67 (50%) 61 (47%)
B or C 135 (51%) 69 (53%) 66 (50%) 67 (50%) 68 (53%)
Mean viral load (log10c/ml) 5·1 (SD 0·8) 5·1 (0·8) 5·1 (0·8) 5·1 (0·8) 5·1 (0·9)
Mean CD4% 18 (SD 11) 18 (11) 18 (11) 18 (11) 18 (11)
Mean weight-for-age Z score –0·8 (SD 1·5) –0·8 (1·5) –0·8 (1·5) –0·9 (1·5) –0·7 (1·5)
Mean height-for-age Z score –1·0 (SD 1·4) –1·0 (1·4) –1·0 (1·4) –1·0 (1·4) –1·0 (1·4)
ART exposure for PMTCT 39 (15%) 19 (15%) 20 (15%) 22 (16%) 17 (13%)
≥1 major resistance mutation 10/239 (4%) 5/116 (4%) 5/123 (4%) 7/121 (6%) 3/118 (3%)
HIV-1 subtype*
B 101 (41%) 52 (42%) 49 (39%) 52 (41%) 49 (41%)
C 25 (10%) 13 (11%) 12 (10%) 13 (10%) 12 (10%)
F 48 (19%) 25 (20%) 23 (18%) 25 (20%) 23 (19%)
A/CRF_AG/D/G 52 (21%) 21 (17%) 31 (25%) 25 (20%) 27 (22%)
Unclassified 23 (9%) 12 (10%) 11 (9%) 13 (10%) 10 (8%)

ART=antiretroviral therapy. PMTCT=prevention of mother-to-child transmission. CRF=circulating recombinant form. NNRTI=non-nucleoside reverse transcriptase inhibitor.
NRTI=nucleoside reverse transcriptase inhibitor. *Available from 239 baseline resistance tests, and ten resistance tests during follow-up (123 protease-inhibitor group,
126 NNRTI group; 128 low-threshold group, 121 higher-threshold group).

Table 1: Baseline characteristics

www.thelancet.com/infection Vol 11 April 2011 275

 Articles

of 0·3 log10 copies per mL (5% two-sided significance

Total Protease- NNRTI Low-threshold Higher-
(n=263) inhibitor group group threshold level) assuming 10% of viral loads were missing at
group (n=131) (n=132) (n=134) group (n=129) 4 years and all were detectable. If 40% of viral loads
Protease inhibitor and NNRTI were undetectable there was 90% power to detect a
Lopinavir plus ritonavir 65 (25%) 64 (49%) 1 (1%) 29 (22%) 36 (28%) difference of 0·5 log10 copies per mL.13 Our trial was not
Nelfinavir 64 (24%) 63 (48%) 1 (1%) 37 (28%) 27 (21%) formally powered to detect interactions between the
Fosamprenavir plus ritonavir 2 (1%) 2 (2%) ·· 1 (1%) 1 (1%)
protease inhibitor versus NNRTI and viral-load
or full-dose ritonavir threshold randomisations.
Efavirenz 82 (31%) 2 (2%) 80 (61%) 43 (32%) 39 (30%) All analyses used intention to treat, which we defined
Nevirapine 50 (19%) ·· 50 (38%) 24 (18%) 26 (20%) as all eligible patients that started treatment; statistical
NRTI backbone tests were two-sided and adjusted for stratification
Zidovudine and lamivudine 113 (43%) 53 (40%) 60 (45%) 54 (40%) 59 (46%) factors. Primary comparisons of change in viral load
Abacavir and lamivudine 62 (24%) 29 (22%) 33 (25%) 31 (23%) 31 (24%) from baseline to 4 years in NNRTI versus protease
Stavudine and lamivudine 53 (20%) 31 (24%) 22 (17%) 28 (21%) 25 (19%) inhibitors and low (≥1000 copies per mL) versus higher
Zidovudine and didanosine 25 (10%) 14 (11%) 11 (8%) 13 (10%) 12 (9%) (≥30 000 copies per mL) viral-load thresholds used
Other 10 (4%) 4 (3%) 6 (5%) 8 (6%) 2 (2%) ANCOVA, adjusted for baseline viral load; likelihood-
based interval regression accounted for undetectable
NNRTI=non-nucleoside reverse transcriptase inhibitor. NRTI=nucleoside reverse transcriptase inhibitor. viral-load measurements. A sensitivity analysis was
Table 2: Initial antiretroviral therapy done with multiple imputation to account for missing
viral loads at 4 years. Logistic regression was used to
assess binary outcomes (eg, viral load <400 copies
guidelines12 and high-level resistance to specific per mL at 4 years), continuous outcomes (eg, CD4%)
For the Stanford University HIV antiretroviral drugs by the Stanford scoring system. used normal linear regression, analysis of adverse
Drug Resistance Database see Interim data on safety, adherence to randomised events used Poisson regression, difference in viral load
http://hivdb.stanford.edu/
strategies, and efficacy of protease inhibitors versus at switch used median regression, and time to event
NNRTI and low versus higher threshold were reviewed (eg, switch) used Cox proportional-hazards regression.
regularly by an independent data and safety monitoring All major resistance mutations after baseline were
board that met roughly each year (five meetings in total). accumulated,14 and differences tested with Poisson
There were no formal statistical rules for recommending regression assuming children not fulfilling criteria for
stopping or for modifying the trial. testing did not develop mutations; a sensitivity analysis
was done, with multiple imputation to account for
Statistical analysis children with missing resistance tests. We tested for
The planned sample size of 256 children was based on interaction between the treatment and threshold
an SD of 0·7 log10 copies per mL for the mean change in randomisations (p<0·05 deemed significant). Stata
viral load from baseline to 4 years, and provided 90% statistical software (version 11.1) was used throughout.
power to detect a difference for each main comparison This study is registered with ISRCTN, number
(protease inhibitor vs NNRTI; higher vs low threshold) ISRCTN73318385.

1·00 PI Low
Proportion of children not yet switched

NNRTI Higher

0·75

0·50

0·25

0
0 24 48 72 96 120 144 168 192 216 240 264 288 0 24 48 72 96 120 144 168 192 216 240 264 288
Weeks from randomisation Weeks from randomisation
Number at risk Number at risk
PI 131 130 115 108 104 103 100 96 92 82 72 51 37 Low threshold 134 130 112 106 102 100 94 88 85 78 64 50 38
NNRTI 132 126 119 115 110 106 98 94 92 82 63 46 36 Higher threshold 129 126 122 117 112 109 104 102 99 86 71 47 35

Figure 2: Time to switch to second-line ART

Vertical lines indicate 4 years, primary endpoint. PI=protease inhibitor (start ART with two NRTIs plus a PI). NNRTI=non-nucleoside reverse transcriptase inhibitors
(start ART with two NRTIs plus an NNRTI). Low threshold=switch from first-line to second-line ART at a viral-load threshold of ≥1000 copies per mL. Higher
threshold=switch from first-line to second-line ART at a viral-load threshold of ≥30 000 copies per mL. ART=antiretroviral therapy. NRTI=nucleoside reverse
transcriptase inhibitor.

276 www.thelancet.com/infection Vol 11 April 2011

 Articles

Role of the funding source
Among the sponsors of the study, representatives of the
National Institutes of Health were part of the study team
and therefore this sponsor was involved in study design,
coordination, data collection, data analysis, data
interpretation, and writing of the report. The
corresponding author had full access to all the data in the
study and had final responsibility for the decision to
submit for publication.
Results
Our trial included 266 children (figure 1): 133 children
from Europe, 77 from North America, and 56 from South
America; participants were from 68 centres in
13 countries. Consent for two children was withdrawn
before being started on ART and one child had a major
eligibility violation (had received antiretroviral drugs for
56 days or more for reduction of mother-to-child
transmission), therefore 263 children were included in
the analysis. Baseline characteristics of participants—
were balanced across both randomisations (table 1).
Nosocomial transmission before age 2 years (Romania)
was the primary source of infection with HIV for
36 parenterally infected children. ART for reduction of
mother-to-child-transmission was used in 39 children
(15%), balanced across randomised groups. Only five
children (2%) received single-dose nevirapine (before
protocol amendment); most received zidovudine
prophylaxis alone. A small proportion of baseline samples
tested retrospectively for resistance had one major
mutation or more (table 1).
All 263 children were started on ART after random
assignment; 220 (84%) within 3 days with a maximum
delay of 63 days. Four children (2%) started with a
regimen different from their allocation (figure 1), either
due to drug non-availability or refusal, but they were

100
Percentage <400 copies per mL (95% CI)

90
80
70
60
50
40
30 PI–low
20 PI–higher
10 NNRTI–low
NNRTI–higher
0

100
Percentage <50 copies per mL (95% CI)

90
80
70
60
50
40
30
20
10
0

26
24
Mean change in CD4% (95% CI)

22
20
18
16
14
12
10
8
6
4
2
0
0 24 48 72 96 120 144 168 192 216 240 264 288
Weeks from randomisation

Figure 3: Virological suppression and CD4% changes during follow-up

Data shown to week 288 when 91 children (43 PI, 48 NNRTI; 47 low-threshold, 44 higher-threshold) were in follow-up. Vertical lines indicate 4 year, primary
endpoint. PI=protease inhibitor (start ART with two NRTIs plus a PI). NNRTI=non-nucleoside reverse transcriptase inhibitors (start ART with two NRTIs plus an
NNRTI). Low threshold=switch from first-line to second-line ART at a viral-load threshold of ≥1000 copies per mL. Higher threshold=switch from first-line to
second-line ART at a viral-load threshold of ≥30 000 copies per mL. ART=antiretroviral therapy. NRTI=nucleoside reverse transcriptase inhibitor.

www.thelancet.com/infection Vol 11 April 2011 277

 Articles

Total Protease-inhibitor NNRTI group Rate ratio Low-threshold Higher-threshold Rate ratio
(n=263) group (n=131) (n=132) (95% CI); p group (n=134) group (n=129) (95% CI); p
Children expected to have tests (number of tests) 108 (165) 56 (90) 52 (75) .. 60 (89) 48 (76) ..
Children with tests (number of tests) 91 (128*) 46 (69) 45 (59) .. 51 (67) 40 (61) ..
Number of children included in analysis 246 121 125 .. 125 121 ..
Number of major NNRTI mutations
None 204 (83%) 112 (93%) 92 (74%) 3·12 (1·84 to 104 (83%) 100 (83%) 1·15 (0·73 to
5·29); <0·001 1·80); 0·50
1–2 34 (14%) 7 (6%) 27 (22%) .. 18 (14%) 16 (13%) ..
3 or more 8 (3%) 2 (2%) 6 (5%) .. 3 (2%) 5 (4%) ..
High-level NNRTI resistance (% of all children)
Nevirapine 42 (17%) 9 (7%) 33 (26%) .. 21 (17%) 21 (17%) ..
Efavirenz 38 (15%) 9 (7%) 29 (23%) .. 20 (16%) 18 (15%) ..
Etravirine 3 (1%) ·· 3 (2%) .. 1 (1%) 2 (2%) ..
Number of major protease inhibitor mutations
None 230 (93%) 108 (89%) 122 (98%) 0·25 (0·10 to 114 (91%) 116 (96%) 0·62 (0·27 to
0·68); 0·01 1·42); 0·27
1–2 16 (7%) 13 (11%) 3 (2%) .. 11 (9%) 5 (4%) ..
High-level protease inhibitor resistance† (% of all children)
Nelfinavir 11 (4%) 9 (7%) 2 (2%) .. 6 (5%) 5 (4%) ..
Atazanavir 3 (1%) 2 (2%) 1 (1%) .. 1 (1%) 2 (2%) ..
Saquinavir 1 (<1%) 1 (1%) ·· .. ·· 1 (1%) ..

PI=protease inhibitor. NNRTI=non-nucleoside reverse transcriptase inhibitor. Low threshold=switch from first-line to second-line ART at a viral-load threshold of ≥1000 copies per mL. Higher threshold=switch
from first-line to second-line ART at a viral-load threshold of ≥30 000 copies per mL. *Nine samples could not be amplified, 28 had no stored sample available. †No children developed high-level resistance to
lopinavir, tipranavir, fosamprenavir, or darunavir.

Table 3: Accumulated NNRTI and protease inhibitor resistance to trial end

included in their allocated group for analysis. In the
protease inhibitor group, about half were started on
lopinavir plus ritonavir and the other half were started
on nelfinavir (table 2). In the NNRTI group, more than
half were started on efavirenz (80 children) with the
remaining participants started on nevirapine (50). As
NRTIs, most children received lamivudine, with
zidovudine, abacavir, or stavudine.
234 children (89%) were in follow-up at 4 years, the
primary endpoint. At the end of the study (Aug 31, 2009),
median follow-up was 5·0 years (IQR 4·2–6·0;
range 0·1–6·7). At trial end, 188 children (71%) were on
first-line ART. Of 75 children (29%) who stopped first-
line ART, 60 children (28 assigned to protease inhibitors,
32 assigned to NNRTI; 37 assigned to the 1000 copies
per mL threshold, 23 the 30 000 copies per mL threshold)
had switched to second-line ART (four subsequently
started third-line) and 15 had discontinued ART after
their first-line regimen (ten subsequently lost to follow-
up; figure 1). Only four children were on nelfinavir at
trial end; 36 had substituted lopinavir plus ritonavir (32 at
the 2007 nelfinavir recall),15 22 had switched to second-
line, and two had discontinued ART. 87 children (33%)
substituted drugs while on first-line ART (mainly because
of toxic effects or the nelfinavir recall).
37 (62%) of the 60 children switched to second-line
ART were switched at the protocol-defined point (36 met
virological criteria and one met clinical criteria),
12 (20%) switched before the strictly defined point, and
11 (18%) switched after. A further 11 children reached
their protocol threshold but did not switch (figure 1,
webappendix p 1).
Median viral load at switch to second-line ART was
6720 copies per mL (IQR 1380–26 100) for the low-
threshold group, compared with 35 712 copies per mL
(IQR 8060–72 800) for the higher-threshold group
(difference 0·73 log10 copies per mL, 95% CI 0·41–1·04;
p=0·002). Children in the higher-threshold group
switched later (HR 0·58, 95% CI 0·34–0·98; p=0·04);
the estimated time until 10% of children assigned to
the 1000 copies per mL strategy had switched was
54 weeks, whereas this was 95 weeks for children
assigned to the higher-threshold strategy (figure 2).
Mean CD4% at switch did not differ between the groups
(27% low threshold vs 23% higher threshold;
difference –3·6%, 95% CI –9·1 to 2·0; p=0·07) and
residual viraemia as measured by the time-averaged
area under the viral-load curve above 400 copies per mL
after 24 weeks was also similar (mean 0·28 log10 copies
per mL [SD 0·52] in the low-threshold group vs
0·27 log10 copies per mL [0·49] in the higher-threshold
group; p=0·90).
Mean changes in viral load from baseline to 4 years,
assessed in 234 children (89%), were –3·16 log10 copies
per mL for protease inhibitors versus –3·31 log10 for
NNRTIs (difference –0·15 log10 copies per mL, 95% CI

278 www.thelancet.com/infection Vol 11 April 2011

 Articles

Total PI-low PI-higher Rate ratio NNRTI-low NNRTI-higher Rate ratio

(n=263) (n=66) (n=65) (95% CI); p (n=68) (n=64) (95% CI); p
Children expected to have tests (number of tests) 108 (165) 33 (48) 23 (42) .. 27 (41) 25 (34) ..
Children with tests (number of tests) 91 (128*) 28 (38) 18 (31) .. 23 (29) 22 (30) ..
Number of children included in analysis 246 61 60 .. 64 61 ..
Number of major NRTI mutations†
None 192 (78%) 49 (80%) 51 (85%) 0·71 (0·37 to 50 (78%) 42 (69%) 2·53 (1·44 to
1·34); 0·31 4·45); 0·001
1–2 41 (17%) 9 (15%) 6 (10%) .. 14 (22%) 12 (20%) ..
3 or more 13 (5%) 3 (5%) 3 (5%) .. ·· 7 (11%) ..
Number of TAMs
None 227 (92%) 58 (95%) 56 (93%) .. 61 (95%) 52 (85%) ..
1–2 15 (6%) 3 (5%) 4 (7%) .. 3 (5%) 5 (8%) ..
3 or more 4 (2%) ·· ·· .. ·· 4 (7%) ..
High-level NRTI resistance (% of all children)
Zidovudine 2 (1%) ·· ·· .. ·· 2 (3%) ..
Didanosine 7 (3%) 3 (5%) 1 (2%) .. ·· 3 (5%) ..
Lamivudine/emtricitabine‡ 49 (20%) 12 (20%) 8 (13%) .. 12 (19%) 17 (28%) ..
Stavudine 3 (1%) ·· ·· .. ·· 3 (5%) ..
Abacavir 8 (3%) 3 (5%) 1 (2%) .. 1 (2%) 3 (5%) ..
Tenofovir 1 (<1%) 1 (2%) ·· .. ·· ·· ..

Six children developed the L74V mutation (three PI-low, one PI-higher, two NNRTI-higher) and two children developed the K65R mutation (one PI-low, one NNRTI-low). For our analysis we assumed children not
expected to have tests did not develop mutations and excluded children with missing tests. PI=protease inhibitor. NNRTI=non-nucleoside reverse transcriptase inhibitor. Low (threshold)=switch from first-line to
second-line ART at a viral-load threshold of ≥1000 copies per mL. Higher (threshold)=switch from first-line to second-line ART at a viral-load threshold of ≥30 000 copies per mL. NRTI=nucleoside reverse
transcriptase inhibitor. TAMs=thymidine-analogue mutations. *Nine samples could not be amplified, 28 had no stored sample available. †Pair-wise comparisons between PI-low and NNRTI-low: rate ratio 0·70,
95% CI 0·38 to 1·32; p=0·27; PI-higher and NNRTI-higher: rate ratio 2·52, 95% CI 1·41 to 4·49; p=0·002. ‡Since M184V/I mutation alone confers high-level resistance to lamivudine and emtricabine, this can be
interpreted as the number of children acquiring M184V/I.

Table 4: Accumulated NRTI resistance (test for interaction, p=0·003) to trial end

–0·41 to 0·11; p=0·26), and –3.26 log10 copies per mL for
the low threshold versus –3·20 log10 copies per mL
for the higher threshold (difference 0·06 log10 copies
per mL, 95% CI –0·20 to 0·32; p=0·56). Sensitivity
analysis, imputing missing data at 4 years, gave very
similar results (data not shown).
During follow-up, there were no differences between
study groups in the proportion of children with viral
loads less than 400 copies per mL (protease inhibitors vs
NNRTIs p=0·77; low-threshold vs higher-threshold
p=0·53) or less than 50 copies per mL (protease inhibitors
vs NNRTIs p=0·35; low-threshold vs higher-threshold
p=0·41; figure 3). At week 24, a higher proportion of
children had viral loads lower than 400 copies per mL
and were on first-line ART in the NNRTI group (80%)
compared with the protease-inhibitor group (73%;
OR 1·49, 95% CI 0·82–2·72; p=0·18); however, at 4 years
differences between study groups were negligible (82%
for the protease inhibitor group and 82% for the NNRTI
group, OR 0·97, 95% CI 0·49–1·91, p=0·91; 83% for the
low-threshold group and 80% for the high-threshold
group, OR 0·83, 95% CI 0·42–1·63, p=0·42).
Furthermore, the proportion of children with a viral load
lower than 400 copies per mL at 4 years was similar for all
initial protease inhibitors and NNRTIs (80% lopinavir
plus ritonavir, 84% other protease inhibitors [mainly
nelfinavir], 80% efavirenz, 84% nevirapine).
At trial end, 149 children (57%) had continued viral-load
suppression on first-line ART, with no difference by class
(74 assigned to protease inhibitors [56%] vs 75 assigned to
NNRTI [57%]; HR 0·97, 95% CI 0·67–1·40; p=0·84).
Second-line ART failed in 18 children (7%), with similar
failure rates across both randomisations: ten assigned to
protease inhibitors (8%) and eight assigned to NNRTIs
(6%; HR 0·78, 95% CI 0·31–1·97; p=0·57); 11 assigned to
switch at 1000 copies per mL (8%) and seven assigned to
switch at 30 000 copies per mL (5%; HR 0·62, 95% CI
0·24–1·59; p=0·34).
Mean increases in CD4% from baseline to 4 years were
13·7% for the protease inhibitors group versus 15·2% for
the NNRTI group (difference 1·5%, 95% CI –0·7 to 3·7;
p=0·19), and 15·1% for the switch at 1000 copies per mL
group versus 13·9% for the switch at 30 000 copies
per mL group (difference –1·1%, 95% CI –3·4 to 1·1;
p=0·27; figure 3). Mean weight-for-age Z score increased
from baseline to 4 years by 0·53 for the protease inhibitors
group versus 0·77 for the NNRTI group (p=0·05),
and 0·73 for the switch at 1000 copies per mL group
versus 0·58 for the switch at 30 000 copies per mL group
(p=0·21). Mean height-for-age Z score increases were 0·61
for the protease inhibitors group versus 0·74 for the
NNRTI group (p=0·27), and 0·65 for the switch at
1000 copies per mL group versus 0·70 for the switch at
30 000 copies per mL group (p=0·66).

www.thelancet.com/infection Vol 11 April 2011 279

 Articles
97 grade 3/4 adverse events were reported in
60 children, with no differences across randomisations
(28 in the protease inhibitor group, 32 in the NNRTI
group, rate ratio 1·12, 95% CI 0·68–1·87, p=0·72; 30 in
the switch at 1000 copies per mL group, 30 in the switch
at 30 000 copies per mL group, rate ratio 1·05, 95% CI
0·63–1·73; p=0·98; webappendix p 2). Only 17 grade 3/4
adverse events (in 17 children) led to modification of
ART. 69 serious adverse events (SAEs) were reported in
48 children, but only one was life threatening (acute
renal failure, non-ART related, in the protease
inhibitor-higher group). The number of children
experiencing an SAE did not differ significantly between
groups (23 in the protease inhibitor group, 25 in the
NNRTI group, p=0·84; 19 in the switch at 1000 copies
per mL group, 29 in the switch at 30 000 copies per mL
group, p=0·09).
One child died (NNRTI-low) at week 277 due to
presumptive malignant disease. There were 14 new CDC
stage-C events (two cases of cytomegalovirus, one of
Mycobacterium avium intracellulare, six of sepsis or
pneumonia, one of cryptosporidiosis, two of oesophageal
candidosis, one of Pneumocystis jirovecii pneumonia, and
one of lymphoma) in nine children (three in protease
inhibitor-low, three in protease inhibitor-higher, one
in NNRTI-low, two in NNRTI-higher).
Of the 108 children who met criteria for resistance
testing, 91 (84%) had tests on 128 samples. Children
assigned to switch at 1000 copies per mL compared with
30 000 copies per mL developed a similar number of
protease inhibitor and NNRTI resistance mutations
(table 3). However, there was a suggestion of an
interaction between ART strategy and viral-load threshold
for NNRTI mutations (p=0·02) since children in the
NNRTI-higher group developed more mutations than
the NNRTI-low group, but children in the protease
inhibitor-higher group developed fewer than the protease
inhibitor-low group (webappendix p 1); this might be
because children in the protease inhibitor-low group
switched faster to NNRTI second-line and failed viro-
logically. Protease inhibitor resistance was mainly in
children started on nelfinavir; only one child who was
started on lopinavir plus ritonavir developed protease
inhibitor resistance (low-level, V82A). Only three children
developed high-level resistance to etravirine (table 3).
For NRTI resistance, there was evidence of interaction
(p=0·003), with children assigned to NNRTIs and
switch at 30 000 copies per mL developing more
mutations (table 4). NRTI resistance was mainly
M184V/I (lamivudine/emtricitabine), with few major
resistance mutations to didanosine, abacavir, or
tenofovir. However, in the NNRTI-higher group, more
children developed three or more NRTI mutations,
conferring high-level resistance to zidovudine,
didanosine, stavudine, or abacavir (table 4). In children
started on ART with lopinavir plus ritonavir, no increase
in NRTI resistance was noted in the protease inhibitor-
280
higher group (one M184V, one M184V+thymidine-
analogue mutations [TAMs]) compared with protease
inhibitor-low (four M184V, one M184V+TAM+L74V/
Y115F). Sensitivity analysis, imputing data for children
with missing resistance tests, gave very similar results
(data not shown).
Discussion
To our knowledge, ours is the first trial to compare the
long-term virological, immunological, and clinical
outcomes of starting ART with protease inhibitors versus
NNRTI-containing regimens in children (panel). Previous
trials in adults that compared efavirenz with nelfinavir
(INITIO,16 ACTG 38418), and efavirenz with lopinavir plus
ritonavir (ACTG 5142)17 as initial ART regimens showed
superior viral-load efficacy of efavirenz compared with
both protease inhibitors. In our smaller pragmatic
paediatric trial, clinicians could choose which protease
inhibitor and NNRTI to use; most children were on either
nelfinavir or lopinavir plus ritonavir as the protease
inhibitor, and nevirapine or efavirenz as the NNRTI.
When we started the trial, nelfinavir was the main protease
inhibitor available for children, but because of recall of
specific batches as a result of chemical impurity15 in 2007,
by trial end most children who had not already switched to
second-line had changed nelfinavir to lopinavir plus
ritonavir. Our finding of almost identical rates of viral-load
suppression for nelfinavir and lopinavir plus ritonavir at
4 years, and rates very similar to starting with an NNRTI,
suggest that there is no difference between starting ART
with regimens based on either protease inhibitors or
NNRTIs in children. In our trial, NRTI backbones were
similarly distributed between protease inhibitor and
NNRTI arms, which is important since we previously
showed the superiority of abacavir-containing over
zidovudine plus lamivudine backbones in children
infected with HIV started on ART in the PENTA 5 trial.23,24
Considering that our trial started in 2002 with regimens
that contained drugs such as nelfinavir that would not be
thought optimum today,25 long-term viral, immunological,
and clinical responses were good; most children responded
well over a median of 5 years; 188 (71%) were still on first-
line at trial end, and second-line therapy failed in
only 18 (7%). The drug regimens of 17 children were
modified for grade 3/4 toxic effects, and by trial end only
one child, who was very immunosuppressed at baseline,
died. The small number of new CDC stage-C events were
equally divided between the randomised strategies.
We used a factorial design to also address the question
of when to switch ART. Although the trial was not
formally powered to detect interactions between initial
ART and viral-load threshold strategies, it provided a
unique opportunity to efficiently study when to switch,
which has never been addressed in a large trial in adults
or children. 47 adults previously treated with ART with
viral loads of 200–10 000 copies per mL were enrolled
into a small pilot trial (ACTG A5115)10 between 2002
www.thelancet.com/infection Vol 11 April 2011

and 2004, randomising to switch immediately or to defer
switching until viral load rose to 10 000 copies per mL or
CD4 decreased by more than 20%. Patients were on
various ART regimens and were highly ART experienced.
Those patients in the deferred group remained
immunologically stable over about 60 weeks but acquired
more resistance mutations. The authors concluded that
delaying switching was a possible strategy if future drug
options were limited.
Our rationale for choosing the viral-load switch criteria
was pragmatic. At the time of design, 1000 copies per mL
was increasingly adopted to define virological failure and
prompt switch to second-line therapy in adults. This was
not routine practice for paediatricians, who had concerns
that drug options would be quickly exhausted. UK and
Irish observational data showed median switch viral load
was much higher than 1000 copies per mL and no clear
viral-load threshold triggered the switch.26 We chose an
upper threshold of 30 000 copies per mL for the randomised
comparison because this was 1·5 log10 copies per mL
higher than 1000 copies per mL, thus above the range of
assay variation and deemed acceptable in contemporary
practice. During our trial, reports from adult cohort
studies27,28 suggested that individuals continuing the same
treatment with detectable viraemia accumulated increasing
resistance mutations, especially with NNRTI-based
therapy, which raised concern that practitioners might not
follow the 30 000 copies per mL switch strategy. However,
although there were more switches below 30 000 copies
per mL in the NNRTI-higher group compared with the
protease inhibitor-higher group, the viral load at switch
was maintained at the intended level.
We found no difference in 4-year viral load between the
1000 copies per mL and the 30 000 copies per mL
switching groups. Furthermore, we noted no significant
difference in accumulated major protease inhibitor or
NNRTI mutations between the two groups. Whereas
NNRTI resistance was common, protease resistance was
infrequent, and was selected mainly in those starting
nelfinavir in accordance with other data;25,29,30 only one
child who was started on lopinavir plus ritonavir developed
low-level protease inhibitor resistance (V82A). Most
children who developed NNRTI resistance had only one
mutation; only three had high-level resistance to etravirine
(one in the NNRTI-low group). Thus NNRTI mutations
were probably selected early during viral rebound, before
viral load reached 1000 copies per mL, with few additional
mutations over the year it took for viral load to reach
30 000 copies per mL. Our findings are consistent with
recent data from the UK in which major mutations were
detected at less than 1000 copies per mL.31
For NRTI resistance, there was an interaction between
initial ART and viral-load threshold strategies. Examining
the switching groups separately according to whether
ART was started with protease inhibitors or NNRTIs,
showed that more major mutations developed in the
NNRTI-higher group; three or more TAMs were only
www.thelancet.com/infection Vol 11 April 2011
Articles
Panel: Research in context
Systematic review
We searched PubMed with the terms “randomized” and “HIV” and “antiretroviral” and
“switch”. The only previous randomised trial comparing different viral-load thresholds for
switching from first to second-line antiretroviral therapy (ART) was in adults and failed to
complete recruitment (ACTG A5115).10 Several trials have compared starting ART with
protease inhibitor versus non-nucleoside reverse transcriptase inhibitor (NNRTI)-based
regimens. In adults, two PubMed searches (“randomized” and “HIV” and “regimens” and
“initial”; “randomized” and “HIV” and “protease inhibitor” and “non-nucleoside reverse
transcriptase inhibitor”) revealed the largest trials (ACTG 384,16 ACTG 5142,17 INITIO,18
FIRST19). To identify randomised trials in children we searched “randomized” and “HIV”
and “children” and “exposure”. Two short-term (24–52 weeks) trials in young children
with perinatal NNRTI exposure have compared nevirapine with lopinavir plus ritonavir
(IMPAACT 1060,20 NEVEREST21). The second cohort of IMPAACT 1060, in children without
perinatal NNRTI exposure, has also recently stopped.22
Interpretation
Our trial is the first long-term (median 5 years) trial comparing protease inhibitor-based
versus NNRTI-based ART across all ages of children, most of whom had not been exposed
to perinatal NNRTIs. It is also the only completed trial to assess viral-load thresholds for
switching to second-line ART in either adults or children. Trials comparing protease
inhibitor-based versus NNRTI-based regimens in adults were done with different
endpoints and were often drug specific; overall viral-load responses were slightly better
with efavirenz than a protease inhibitors. In other paediatric trials, IMPAACT 1060 found
lopinavir plus ritonavir was superior to nevirapine at 24 weeks in children younger than
3 years, with and without perinatal NNRTI exposure. NEVEREST, in children younger than
2 years, found that switching to nevirapine after first achieving viral suppression on
lopinavir plus ritonavir had higher rates of viral suppression (<50 copies per mL)
compared with staying on lopinavir plus ritonavir; however, children who did not achieve
viral suppression were more likely to experience viral rebound (>1000 copies per mL) if
they switched to nevirapine compared with remaining on lopinavir plus ritonavir.
seen in this group. Over a median of 5 years, continuing
NNRTI-based ART until viral load increased to
30 000 copies per mL selected the M184V/I mutation in
about 10% more of all randomly assigned children (25%
of those with a resistance test) and important mutations
to other NRTI drugs in an additional 5% (14% of those
tested). By contrast, development of resistance to NRTIs
seemed to be largely prevented by the presence of
lopinavir plus ritonavir in both the protease
inhibitor-higher and protease inhibitor-low groups, and
resistance was mainly to lamivudine (both children with
TAMs/L74V/Y115F had multiple regimen changes and
never suppressed virologically).
What do the PENPACT-1 trial results mean for long-
term ART strategies for HIV-infected children, particularly
in resource-limited settings, where most now live and
where drug options are limited and viral-load-monitoring
facilities mostly unavailable? First, long-term virological,
clinical, and immunological outcomes, even for
suboptimum ART, are excellent, even when starting at low
CD4 (108 [41%] had CD4 <15%). Second, development of
NNRTI resistance cannot be readily prevented even with
regular viral-load monitoring every 3 months because it
happens very soon after viral-load rebound; however,
281
 Articles
continuing on a failing NNRTI regimen is likely to
increase accumulation of NRTI mutations, compromising
their subsequent use. Third, results of our trial raise the
question, not directly addressed here, of the effect of
perinatal exposure to single-dose nevirapine to reduce
mother-to-child transmission: should infants started on
lopinavir plus ritonavir, as recommended at present, be
delayed from switching at virological failure rather than
switching to an NNRTI plus two NRTI second-line
regimen, where rapid development of resistance might be
a risk if NNRTI mutations have been archived? Further
research is needed. Finally, if NRTIs become superseded
by new drugs, such that they no longer have a place in
second or subsequent lines of ART, then the role for viral-
load or resistance monitoring on ART might be less
important since new regimens can be given with no
overlapping resistance.
In conclusion, in the absence of single-dose nevirapine
prophylaxis for reduction of mother-to-child transmission
and anticipated poor adherence (eg, during adolescence),
no difference exists between protease inhibitor and
NNRTI containing initial ART regimens in children;
both result in good long-term viral-load, immunological,
and clinical outcomes. Delaying switching until viral-load
levels are 30 000 copies per mL results in accumulation
of more NRTI resistance mutations with NNRTI-
combination therapy compared with switching at
1000 copies per mL; conversely, for children on ART
based on protease inhibitors, the absence of a difference
in the NRTI and protease inhibitor resistance suggests
that delayed switching might be reasonable in
circumstances and settings where future drug options
are limited.
Contributors
The PENPACT-1 trial was designed by AB, AC, CG, DMG, MH, RM,
AM, LM, YS, and GTW. The trial was co-ordinated in North America by
MH, RM, AM, LM, and MES, and in Europe and South America by AB,
HC, AC, CG, DMG, LHarp, YS, and GTW. HC and ASW did data and
safety monitoring board analyses, overseen by AB and MH. AB, HC,
LHarr, and MH wrote the trial analysis plan, which all authors reviewed;
HC and LHarr did the final analysis. All members of the writing
committee contributed to the interpretation of the data. DMG, LHarr,
AM, and GTW wrote the first draft of the paper. All members of the
writing committee commented extensively, revised the report critically,
and approved the final version.
PENPACT-1 writing committee (alphabetical)
Abdel Babiker, Hannah Castro (nee Green), Alexandra Compagnucci,
Susan Fiscus, Carlo Giaquinto, Diana M Gibb, Lynda Harper,
Linda Harrison, Michael Hughes, Ross McKinney, Ann Melvin,
Lynne Mofenson, Yacine Saidi, M Elizabeth Smith,
Gareth Tudor-Williams, A Sarah Walker (full details of the study team in
the webappendix pp 3–4).
Conflict of interest
SF received honoraria for serving on scientific advisory boards and
lectures from Abbott Molecular Inc and Roche Molecular Systems. MH
is a paid data and safety monitoring board member for Boehringer
Ingelheim, Medicines Development, Pfizer, and Tibotec. RM is a paid
member of the data and safety monitoring board for Gilead Sciences.
DMG and ASW are paid members of the data and safety monitoring
board for Tibotec. The other members of the writing committee declare
no conflicts of interest.
282
Acknowledgments
The PENPACT-1 trial was sponsored jointly by the Paediatric European
Network for Treatment of AIDS (PENTA) Foundation, Agènce Nationale
de Recherche sur le Sida et les hepatites virales (ANRS) and the Pediatric
AIDS Clinical Trials Group (PACTG), subsequently the International
Maternal Pediatric Adolescent AIDS Clinical Trials Group (IMPAACT).
Overall support for PACTG/IMPAACT was provided by the National
Institute of Allergy and Infectious Diseases (NIAID; U01 AI068632), the
Eunice Kennedy Shriver National Institute of Child Health and Human
Development (NICHD), and the National Institute of Mental Health
(NIMH; AI068632). The content is solely the responsibility of the authors
and does not necessarily represent the official views of the NIH. This
work was supported by the Statistical and Data Analysis Center at Harvard
School of Public Health, under the National Institute of Allergy and
Infectious Diseases cooperative agreement #5 U01 AI41110 with the
PACTG and #1 U01 AI068616 with the IMPAACT Group. Support of the
sites was provided by the National Institute of Allergy and Infectious
Diseases (NIAID) and the NICHD International and Domestic Pediatric
and Maternal HIV Clinical Trials Network funded by NICHD (contract
number N01-DK-9-001/HHSN267200800001C). PENTA is a coordinated
action of the European Commission/European Union, supported by the
seventh framework programme (FP7/2007-2013) under the Eurocoord
grant agreement number 260694, the sixth framework contract number
LSHP-CT-2006-018865, the fifth framework programme contract number
QLK2-CT-2000-00150, and by the PENTA Foundation. UK clinical sites
were supported by a grant from the MRC; those in Italy by a grant from
the Istituto Superiore di Sanita—Progetto Terapia Antivirale 2004, 2005.
GSK and BMS provided drugs in Romania. The trial was coordinated by
four trials centres: the Medical Research Council (MRC) Clinical Trials
Unit, London, UK (with support from the MRC); INSERM SC10, Paris,
France (supported by ANRS); Frontier Science, New York, USA; and
Westat, Maryland, USA (supported by NICHD). We thank all the children,
families, and staff from the centres participating in the PENPACT-1 trial.
References
1 Palella FJ Jr, Delaney KM, Moorman AC, et al. Declining morbidity
and mortality among patients with advanced human
immunodeficiency virus infection. N Engl J Med 1998; 338: 853–60.
2 Vigouroux C, Gharakhanian S, Salhi Y, et al. Adverse metabolic
disorders during highly active antiretroviral treatments (HAART) of
HIV disease. Diabetes Metab 1999; 25: 383–92.
3 Faye A, Bertone C, Teglas JP, et al. Early multitherapy including
a protease inhibitor for human immunodeficiency virus type
1-infected infants. Pediatr Infect Dis J 2002; 21: 518–25.
4 Luzuriaga K, McManus M, Mofenson L, Britto P, Graham B,
Sullivan JL. A trial of three antiretroviral regimens in HIV-1-infected
children. N Engl J Med 2004; 350: 2471–80.
5 Funk MB, Linde R, Wintergerst U, et al. Preliminary experiences
with triple therapy including nelfinavir and two reverse
transcriptase inhibitors in previously untreated HIV-infected
children. AIDS 1999; 13: 1653–58.
6 Watson DC, Farley JJ. Efficacy of and adherence to highly active
antiretroviral therapy in children infected with human
immunodeficiency virus type 1. Pediatr Infect Dis J 1999; 18: 682–89.
7 Panel on Antiretroviral Therapy and Medical Management of
HIV-Infected Children. Guidelines for the use of antiretroviral
agents in pediatric HIV infection. http://aidsinfo.nih.gov/
ContentFiles/PediatricGuidelines.pdf (accessed Dec 20, 2010).
8 WHO. Antiretroviral therapy for HIV infection in infants and
children: towards universal access (2010 version). http://www.who.
int/hiv/pub/paediatric/infants2010/en/index.html (accessed
Sept 14, 2010).
9 Paediatric European Network for Treatment of AIDS (PENTA).
PENTA 2009 guidelines for the use of antiretroviral therapy in
paediatric HIV-1 infection. HIV Med 2009; 10: 591–613.
10 Riddler SA, Jiang H, Tenorio A, et al. A randomized study of
antiviral medication switch at lower- versus higher-switch
thresholds: AIDS Clinical Trials Group Study A5115. Antivir Ther
2007; 12: 531–41.
11 Caldwell MB, Oxtoby MJ, Simonds RJ, Lindegren ML, Rogers MF.
1994 revised classification system for human immunodeficiency
virus infection in children less than 13 years of age.
MMWR Recomm Rep 1994; 43: 1–10.
www.thelancet.com/infection Vol 11 April 2011

12 Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug
resistance mutations in HIV-1: December 2009. Top HIV Med 2009;
17: 138–45.
13 Hughes MD. Analysis and design issues for studies using censored
biomarker measurements with an example of viral load
measurements in HIV clinical trials. Stat Med 2000; 19: 3171–91.
14 Pillay D, Green H, Matthias R, et al. Estimating HIV-1 drug
resistance in antiretroviral-treated individuals in the United
Kingdom. J Infect Dis 2005; 192: 967–73.
15 Roche. Media release: Roche recalls Viracept due to chemical impurity.
http://www.roche.com/med-cor-2007-06-06b (accessed Sept 30, 2010).
16 Shafer RW, Smeaton LM, Robbins GK, et al. Comparison of
four-drug regimens and pairs of sequential three-drug regimens as
initial therapy for HIV-1 infection. N Engl J Med 2003; 349: 2304–15.
17 Riddler SA, Haubrich R, DiRienzo AG, et al. Class-sparing
regimens for initial treatment of HIV-1 infection. N Engl J Med
2008; 358: 2095–106.
18 Yeni P, Cooper DA, Aboulker JP, et al. Virological and
immunological outcomes at 3 years after starting antiretroviral
therapy with regimens containing non-nucleoside reverse
transcriptase inhibitor, protease inhibitor, or both in INITIO:
open-label randomised trial. Lancet 2006; 368: 287–98.
19 MacArthur RD, Novak RM, Peng G, et al. A comparison of three
highly active antiretroviral treatment strategies consisting of
non-nucleoside reverse transcriptase inhibitors, protease inhibitors,
or both in the presence of nucleoside reverse transcriptase
inhibitors as initial therapy (CPCRA 058 FIRST Study): a long-term
randomised trial. Lancet 2006; 368: 2125–35.
20 Palumbo P, Lindsey JC, Hughes MD, et al. Antiretroviral treatment
for children with peripartum nevirapine exposure. N Engl J Med
2010; 363: 1510–20.
21 Coovadia A, Abrams EJ, Stehlau R, et al. Reuse of nevirapine in
exposed HIV-infected children after protease inhibitor-based viral
suppression: a randomized controlled trial. JAMA 2010;
304: 1082–90.
22 NIAID Web Bulletin. NIH Network identifies better treatment
regimen for HIV-infected infants: ritonavir-boosted lopinavir proves
superior to nevirapine. http://www.niaid.nih.gov/news/
newsreleases/2010/Pages/IMPAACTP1060.aspx (accessed
Dec 6, 2010).
www.thelancet.com/infection Vol 11 April 2011
Articles
23 Paediatric European Network for Treatment of AIDS (PENTA).
Comparison of dual nucleoside-analogue reverse-transcriptase
inhibitor regimens with and without nelfinavir in children with
HIV-1 who have not previously been treated: the PENTA 5
randomised trial. Lancet 2002; 359: 733–40.
24 Green H, Gibb DM, Walker AS, et al. Lamivudine/abacavir
maintains virological superiority over zidovudine/lamivudine and
zidovudine/abacavir beyond 5 years in children. AIDS 2007;
21: 947–55.
25 Walmsley S, Bernstein B, King M, et al. Lopinavir-ritonavir versus
nelfinavir for the initial treatment of HIV infection. N Engl J Med
2002; 346: 2039–46.
26 Lee KJ, Lyall EGH, Walker AS, et al. Wide disparity in switch to
second-line therapy in HIV-infected children in CHIPS. 8th
International Congress on Drug Therapy in HIV Infection;
Glasgow, UK; Nov 12–16, 2006. Abstract PL2.4.
27 Gupta R, Hill A, Sawyer AW, Pillay D. Emergence of drug resistance
in HIV type 1-infected patients after receipt of first-line highly active
antiretroviral therapy: a systematic review of clinical trials.
Clin Infect Dis 2008; 47: 712–22.
28 von Wyl V, Yerly S, Boni J, Burgisser P, et al. Emergence of HIV-1
drug resistance in previously untreated patients initiating
combination antiretroviral treatment: a comparison of different
regimen types. Arch Intern Med 2007; 167: 1782–90.
29 Hicks C, King MS, Gulick RM, et al. Long-term safety and durable
antiretroviral activity of lopinavir/ritonavir in treatment-naive
patients: 4 year follow-up study. AIDS 2004; 18: 775–79.
30 Chakraborty R, Smith CJ, Dunn D, et al. HIV-1 drug resistance in
HIV-1-infected children in the United Kingdom from 1998 to 2004.
Pediatr Infect Dis J 2008; 27: 457–59.
31 Mackie NE, Phillips AN, Kaye S, Booth C, Geretti AM.
Antiretroviral drug resistance in HIV-1-infected patients with
low-level viremia. J Infect Dis 2010; 201: 1303–07.
283