High Rates of Forward Transmission Events after Acute/Early HIV‐1 Infection

Abstract

A population-based phylogenetic approach was used to characterize human immunodeficiency virus (HIV)-transmission dynamics in Quebec.
HIV-1 pol sequences included primary HIV infections (PHIs; <6 months after seroconversion) from the Quebec PHI cohort (1998-2005; n=215) and the provincial genotyping program (2001-2005; n=481). Phylogenetic analysis determined sequence interrelationships among unique PHIs (n=593) and infections from untreated (n=135) and treated (n=660) chronically infected (CI) potential transmitter populations (2001-2005). Clinical features, risk factors, and drug resistance for clustered and nonclustered transmission events were ascertained.
Viruses from 49.4% (293/593) of PHIs cosegregated into 75 transmission chains with 2-17 transmissions/cluster. Half of the clusters included 2.7+/-0.8 (mean+/-SD) transmissions, whereas the remainder had 8.8+/-3.5 transmissions. Maximum periods for onward transmission in clusters were 15.2+/-9.5 months. Coclustering of untreated and treated CIs with PHIs were infrequent (6.2% and 4.8%, respectively). The ages, viremia, and risk factors were similar for clustered and nonclustered transmission events. Low prevalence of drug resistance in PHI supported amplified transmissions at early stages.
Early infection accounts for approximately half of onward transmissions in this urban North American study. Therapy at early stages of disease may prevent onward HIV transmission.

Full text

Amplified Transmission in Early HIV Infection •JID 2007:195 (1 April) •000
MAJOR ARTICLE
High Rates of Forward Transmission Events
after Acute/Early HIV-1 Infection
Bluma G. Brenner,
1
Michel Roger,
2
Jean-Pierre Routy,
3
Daniela Moisi,
1
Michel Ntemgwa,
1
Claudine Matte,
2
Jean-Guy Baril,
4
Re´jean Thomas,
5
Danielle Rouleau,
2
Julie Bruneau,
6
Roger Leblanc,
7
Mario Legault,
8
Cecile Tremblay,
9
Hugues Charest,
10
Mark A. Wainberg,
1
and the Quebec Primary HIV Infection Study Group
a
1
McGill AIDS Centre–Jewish General Hospital,
2
Centre Hospitalier de Universite´ de Montre´ al (CHUM)–Hoˆpital Notre-Dame,
3
McGill University
Health Centre,
4
Clinique Me´ dicale du Quartier Latin,
5
Clinique Me´ dicale l’Actuel,
6
CHUM–Hoˆ pital St. Luc,
7
Clinique Me´ dicale Goldberg, LeBlanc,
& Rosengren,
8
Fonds de la Recherche en Sante´ du Que´bec–SIDA Network,
9
CHUM–Hoˆ tel Dieu, and
10
Institut National de Sante´ Publique
du Que´bec, Montreal, Canada
(See the editorial commentary by Pillay and Fisher, on pages XXX–XX.)
Background. A population-based phylogenetic approach was used to characterize human immunodeficiency
virus (HIV)–transmission dynamics in Quebec.
Methods. HIV-1 pol sequences included primary HIV infections (PHIs; !6 months after seroconversion) from
the Quebec PHI cohort (1998–2005; ) and the provincial genotyping program (2001–2005; ).np215 np481
Phylogenetic analysis determined sequence interrelationships among unique PHIs ( ) and infections fromnp593
untreated ( ) and treated ( ) chronically infected (CI) potential transmitter populations (2001–np135 np660
2005). Clinical features, risk factors, and drug resistance for clustered and nonclustered transmission events were
ascertained.
Results. Viruses from 49.4% (293/593) of PHIs cosegregated into 75 transmission chains with 2–17 trans-
missions/cluster. Half of the clusters included ( ) transmissions, whereas the remainder had2.7 0.8 mean SD
transmissions. Maximum periods for onward transmission in clusters were months. Coclus-8.8 3.5 15.2 9.5
tering of untreated and treated CIs with PHIs were infrequent (6.2% and 4.8%, respectively). The ages, viremia,
and risk factors were similar for clustered and nonclustered transmission events. Low prevalence of drug resistance
in PHI supported amplified transmissions at early stages.
Conclusions. Early infection accounts for approximately half of onward transmissions in this urban North
American study. Therapy at early stages of disease may prevent onward HIV transmission.
An understanding of HIV-transmission dynamics is im-
portant in the design of effective prevention and treat-
ment interventions. A number of recent studies suggest
Received 6 September 2006; accepted 3 October 2006; electronically published
16 February 2007.
Potential conflicts of interest: none reported.
Presented in part: XIV International HIV Drug Resistance Workshop, Quebec
City, Canada, 7–11 June 2005 (abstract 112).
Financial support: Canadian Institutes for Health Research (grant MT-14738 for
resistance genotyping in the Quebec Primary HIV Infection [PHI] cohort study and
for research on PHI); Re´seau SIDA of the Fonds de la Recherche en Sante´du
Que´bec (funds to recruit patients into the Quebec PHI cohort study); Quebec
Ministry of Health (funds to the provincial genotypic resistance testing program).
a
Study group members are listed after the text.
Reprints or correspondence: Dr. Mark Wainberg, McGill AIDS Centre, 3755 Cote
Ste. Catherine Rd., Montreal, Quebec, Canada H3W 1G4 (mark.wainberg@
mcgill.ca).
The Journal of Infectious Diseases 2007;195:000–000
2007 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2007/19507-00XX$15.00
DOI: 10.1086/512088
that early stages of HIV infection may disproportion-
ately contribute to viral transmission and spread of the
epidemic [1–3]. Primary HIV infection (PHI) and early
stages of infection are associated with high viral burden
and viral set points in blood and semen, a major de-
terminant of HIV transmission [1, 2,4–6]. The Rekai-
Uganda surveillance study showed that 43.8% (10/23)
of new transmissions occurred in discordant partners
at 6–15 months subsequent to seroconversions of source
partners [6].
In contrast, other groups have used viral load/epi-
demiological/behavioral data to contend that the role
of PHIs in HIV transmission may be overestimated [7–
9]. Many cofactors influence transmission, including
access to antiretroviral therapy and medical care, high
risk behaviors, sexually transmitted diseases, and coin-
fections [7–9]. The findings of the North Carolina pro-
gram Screening and Tracing of Active Transmission

000 •JID 2007:195 (1 April) •Brenner et al.
(STAT) suggest that as many as half of identified sourcepartners
were chronically infected (CI) and that 37% of these were on
antiretroviral therapy [9].
Phylogenetic analysis of viral gene sequences has been used
as a molecular epidemiological approach to reconstruct trans-
mission events in early/acute infection [10–15]. Data from the
Swiss and UK PHI cohort studies have reported significant
clustering of viral sequences from 24% and 34% of recent in-
fections, respectively [13, 14]. Clustering of transmitted drug
resistance has also been reported in 10 PHI cases within the
San Francisco cohort [15].
These findings underscore the importance of tracing the eti-
ology of new HIV transmissions in different patient population
settings. Quebec is a unique venue for population-based mo-
lecular surveillance of HIV-1 transmission based on 2 major
initiatives of recent years. The provincial genotypic testing pro-
gram, recommended for drug-resistance testing in PHI, has
been in place since 2001. In addition, the Quebec PHI cohort,
established in 1997, is a large prospective longitudinal study of
viral evolution, drug resistance, transmission risk factors, and
disease progression after PHI [16, 17]. These initiatives offer a
unique opportunity to assess the role of the early phase of
infection in total HIV transmission and drug resistance in Que-
bec, a province with free universal access to health care, in-
cluding antiretroviral therapy.
The present study involved phylogenetic analysis on all ge-
notyped recent infections !6 months after seroconversions
( ), to estimate the relative importance of acute/earlynp593
infection in onward transmission chains. Comparative analysis
was performed on the treatment-naive and treated CI potential
transmitter populations ( and 660, respectively). In ad-np135
dition, data from the PHI cohort study were used to examine
the clinical and sociodemographic characteristics associated
with the features of transmission clusters in early infection.Our
findings underscore that acute/early infection accounts for ap-
proximately half of all HIV transmissions in Quebec.
PATIENTS AND METHODS
Study populations. The PHI study population is drawn from
2 PHI initiatives involving subjects having acute/early infection
(!6 months after seroconversion). Sequence data were available
from all participants in the Quebec cohort with confirmed PHI
(!6 months after serconversion; 1997–2005; ) [16–18].np215
Participants provided informed consent for blood collection
and resistance testing; they completed standardized nurse-ad-
ministered questionnaires describing risk factors, mode of
transmission, age, and estimated date of infection [16–18]. Viral
loads, drug-resistance profiles, and clinical epidemiological data
were included in the analysis.
The remainder of acute/early stage infections were identified
through the provincial genotyping program, at either of 2 Que-
bec reference laboratories (2001–2005; Hoˆpital Notre-Dame
[HND; ] and Jewish General Hospital [JGH;
np269 np
]). The clinical indication of PHI (!6 months after docu-
233
mented seroconversions) was noted on the laboratory requi-
sitions by prescribing physicians and validated by laboratory
personnel. On the basis of published data from l’Actuel, the
largest HIV clinic in Montreal, this PHI study population
( ) had been infected for an average of 4.9 months [17].
np180
Test requisitions also provided information on age, sex, viral
load, and date of first genotypic sampling.
Sequence and epidemiologic data compiled from these 2
initiatives were made anonymous by the assignment of irre-
trievable patient identifiers. A total of 717 sequences from pri-
mary/early infections included 593 unique subtype B infections,
3 identified source partners, 65 non–subtype B infections, and
56 repeat samplings. The sample repeats of participants ge-
notyped through both initiatives were useful tools to validate
clustering. Non–subtype B infections, largely from the recent
immigrant populations, were included on trees for comparative
purposes but excluded from all analyses [19].
There are an estimated 58,000 persons HIV-infected persons
in Canada. HIV became a reportable disease in Quebec in 2004
and includes ∼13,000–15,000 infected persons, of whom ∼3000
persons have been genotyped to date (viral load 1400 copies/
mL). Sequence data were obtained from a representative CI
potential transmitter population that included all genotyped
CIs from the JGH site (2001–2005), performing 40% of all
provincial genotyping. The treatment-naive and treated CIs in-
cluded single sequence determinations for all persons geno-
typed with clinical indications of CI (16 months after sero-
conversions; ) and first or subsequent treatment failure
np135
on an antiretroviral regimen ( ), respectively. All study
np660
initiatives were approved by clinic and hospital ethics com-
mittees, as well as by the Quebec Ministry of Health committee
on confidentiality and access of information.
Phylogenetic analysis. Genotyping was performed at either
HND or JGH using the same procedures as described above
to generate sequences for the HIV pol region spanning the
protease gene and reverse transcriptase (RT) codons 1–400 or
38–250 using Virco primers (Virco Lab) or the Bayer TRU-
GENE HIV-1 assay (Bayer Diagnostics) protocols, respectively.
All sequences were aligned to consensus HXB2 sequences, re-
moving gaps and cutting to identical sequence lengths using
BioEdit software [20]. Genotypic data identified minor andmajor
resistance mutations, based on the March/April 2005 Interna-
tional AIDS Society–USA resistance panel guidelines [21].
Phylogenetic interrelationships among viral sequences were
estimated using neighbor joining trees and maximum likeli-
hood methods with BioEdit and MEGA2 integrated molecular
evolutionary genetics analysis software [20, 22]. The existence
of clusters was ascertained using the statistical robustness of

Amplified Transmission in Early HIV Infection •JID 2007:195 (1 April) •000
Figure 1. Region of the primary HIV infection (PHI) phylogenetic tree, showing unique PHIs (U1–U15), clustered transmissions (np22), and repeat
patient samplings (np5). Cluster 48 depicts a transmission chain harboring drug resistance. Bootstrap values higher than 98% are indicated on the
branches.
the maximum likelihood topologies assessed by high bootstrap
values (198%) with 1000 resamplings and short branch lengths
(genetic distances 10.015%) [10, 22]. Infections in clusters were
validated for congruent polymorphisms and mutational motifs.
Comparative phylogenetic analysis evaluated coclustering of
unique subtype B genotyped primary/early infections (np
) with the representative potential transmitter population593
of treatment-naive ( ) or treated ( ) CIs. To min-np135 np660
imize any potential bias of nonclustering between the PHI and
CI patient population infection due to drug resistance, phy-
logenetic analysis was repeated after modifying CIs to wild-type
ancestral forms, changing protease codons 30, 50, 54, 82, 84,
and 90 and RT codons 41, 65, 67, 69, 70, 74, 103, 106, 151,
181, 184, and 215 to the wild-type codons present in the con-
sensus B sequence.
Statistical analysis. The maximum window period for
transmissions within clusters was estimated as the maximum
difference in time between the earliest and latest infections
within clusters. Differences in the viral load (log
10
copies/mL),
age, risk behaviors, and drug motifs among clustered and non-
clustered PHI transmissions and CI groups were ascertained
using Fisher’s exact tests and analysis of variance (ANOVA),
with GraphPad Prism software (version 4.0; available at: http:
//www.graphpad.com).
Sequence data. All sequences included in figure 1 were
deposited into GenBank under the sequential accession num-
bers EF011572–EF011609.
RESULTS
Clustering of PHIs. Sequence data (HIV-1 pol region) were
compiled from all PHIs (!6 months after seroconversion) iden-
tified through the provincial genotyping program ( ;
np502
2001–2005) and the Quebec PHI cohort study ( ; 1998–
np215
2005). A phylogenetic approach was used to identify the se-
quence interrelationships of these early/acute stage infections.
A region within this tree is presented in figure 1. As illus-
trated, many PHIs segregated into clusters having sequence
similarity based on the established criteria of high bootstrap
values (198%) and short branch lengths (genetic distances
10.015%) [10, 22]. Manual assessment of similarities in resis-
tance and polymorphism mutational motifs of sequences was

000 •JID 2007:195 (1 April) •Brenner et al.
Figure 2. Phylogenetic tree, showing clustered B (np293) and non–subtype B (np12) primary infections, and the corresponding phylogenetic
analysis, showing nonclustered B (np300) and non–subtype B (np53) infections.
Figure 3. The distribution of patients with primary HIV infection (PHI)
and chronically infected (CI) persons in and gthe 75 transmission clusters.
Overall, half of the transmission chains have 2–4 persons/cluster, whereas
the remaining individuals are in clusters having 5–17 persons/cluster.
Single clustering of CIs and nonclustered PHIs is also depicted.
evaluated to validate clusters. As an example, viruses in cluster
48 harbored transmitted drug resistance (figure 1). Irretrievable
nonnominative cross-identifiers identified repeat samplings of
viral sequences ( ) within clusters sequenced throughnp56
both genotyping initiatives. Four repeat samplings of infections
in transmission clusters are illustrated in figure 1. Non–subtype
B infections ( ), composing 9.8% of all recent infections,np65
were included on trees but excluded from subsequent analysis.
All together, 593 unique subtype B infections were identified.
Tree topology revealed that half (293/593) of all PHIs grouped
into 75 different transmission clusters, whereas the remaining
(300/593) infections represented unique sequences. The entire
phylogenetic tree, stratified according to clustered and non-
clustered transmissions is shown in figure 2. As shown in figures
2 and 3, clustered transmission events included between 2 and
17 infections per transmission cluster. Of note, 49% of clustered
transmission chains had 2–4 infections per cluster (i.e., 2.7
0.8 [ ]). The remaining 51% of clustered eventsmean SD
segregated into large clusters including 8.83.5 transmissions
per cluster (figures 2 and 3).
Sequence interrelationships between CI and PHI populations.
The clustering profiles of PHI transmissions were compared
with corresponding patterns observed for a representative po-
tential transmitter population of CI persons genotyped at the
JGH site (2001–2005). The treatment-naive CI population (n
p135) included persons genotyped for clinical indication of
CI (16 months after seroconversion, baseline before treatment).
The treated CI population ( ) included patients geno-np660
typed for reasons of first or subsequent treatment failure.
Infections from untreated and treated CI patient populations
rarely coclustered with PHIs (1% and 2.7%, respectively). In-
sofar as 70% of sequences from CI patients harbored drug
resistance, phylogenetic analysis was reevaluated after modi-
fying CIs to wild-type ancestral sequences. As depicted in figure
4, clustering of CIs was infrequent (3.2% [21/660], of cases),
with cluster sizes of ( ).3.1 1.6 mean SD
Phylogenetic analyses were then performed to determine
whether treatment-naive infections ( ) and ancestral se-np135
quences from CI treated persons ( ) coclustered withnp660
PHIs ( ). As depicted in figure 3, clustering of infectionsnp593
from treatment-naive ( ) and treated CI ( ) pa-np12 np17
tients with PHIs was infrequent. However, 25 and 12 of treat-
ment-naive and CI treated persons constituted new CI-PHI

Amplified Transmission in Early HIV Infection •JID 2007:195 (1 April) •000
Figure 4. Phylogenetic tree showing the wild-type ancestral forms of treated chronic infections ( )np660
coclusters containing 32 of the 300 nonclustered PHI trans-
missions (table 1).
The cumulative results of our phylogenetic results are sum-
marized in table 1. Small and large transmission chains were
largely attributable to onward transmission after recent infec-
tions, accounting for half of all transmission events. CIs, how-
ever, represented the source of new CI-PHI transmissions as
well as the intermediate partners in forward PHI-PHI trans-
mission events. Based on estimates that the JGH site performs
40% of all genotyping, untreated and treated CI populations
may account for 15% and 12% of onward transmissions,
respectively.
The viral loads for PHI and CI patient populations are also
summarized in table 1. A mean viral load of 4.1 log copies/mL
for the treated genotyped patient population is significantly
lower than the corresponding viral loads of 4.6 and 4.7 observed
for the PHI and treatment-naive populations, respectively
( ; , ANOVA; and , post hoc TukeyFp43.2 P!.0001 P!.001
tests). These results are similar to that reported by l’Actuel, the
major clinical center in Montreal [18]. A mean viral load of
2.58 was observed for the nongenotyped CI patient population;
this may be too low to facilitate forward transmission [18].
Time intervals for onward transmission. Because PHIs
were genotyped over 9 years (1997–2005), it was important to
further investigate whether early stages of infections were the
source of the majority of onward transmissions. The maximum
window periods for onward transmission after PIs were esti-
mated by determining the time intervals between the first and
last infections within each cluster. Forward transmission inter-
vals ranged from 1 to 37 months with overall transmission
intervals of ( ) months. It is important15.2 9.5 mean SD
to note that maximal transmission intervals are overestimated
because there were only 27 of 293 infections in which the first
infection was 124 months apart from other infections.

000 •JID 2007:195 (1 April) •Brenner et al.
Table 1. Clinical and clustering features for the primary HIV-infected (PHI) and chronically infected (CI) patient populations.
Study population
Age,
years
Viral load,
copies/mL
No. of patients in the study population clustering
with PHI transmission clusters (PHI transmissions, %)
Estimated
transmissions,
c
%
PHI
small clusters
a
PHI
large clusters
b
CI-PHI
new cluster
Genotyped PHI
d
37 10 4.64 0.83 293 (24.2) 300 (25.1) … 49.4
Genotyped naive CI
e
41 11 4.71 0.70 5 (0.8) 7 (1.2) 25 (4.2) ∼15.5
Genotyped treated CI
f
43 8 4.14 0.76
g
9 (1.5) 8 (1.3) 17 (2.0) ∼12.0
Nongenotyped CI
h
43 8 2.58
i
…………
NOTE. Data are mean SD values, unless otherwise indicated.
a
2–4 persons/cluster.
b
15 persons/cluster.
c
Based on 40% of genotyped CIs.
d
.np593
e
.np135
f
.np660
g
, for treated CI compared with PHI and naive CI subjects.P!.001
h
.np2328
i
Based on published findings [17].
Table 2. Clinical characteristics of primary HIV infections (PHIs) in clustered trans-
mission chains, compared with nonclustered unique infections.
Characteristic
Clustered
transmissions
a
(np144)
Clustered
transmissions
b
(np149)
Nonclustered
unique PHIs
(np300)
Viral load, mean SD, log copies/mL 4.70 0.93 4.68 0.80 4.66 0.78
Age, mean SD, years 38.8 9.9 34.8 8.6 37.2 9.8
Mode of transmission (np63) (np54) (np89)
MMS 53.9 (34) 74.0 (40)
c
57.0 (57)
IDU 34.9 (22) 13.0 (7) 28.0 (25)
HS 11.1 (7) 13.0 (7) 7.9 (7)
Sexual risk behavior (np41) (np47) (np57)
0 partners 14.6 (6) 10.6 (5) 5.2 (3)
1–4 partners 68.2 (28) 68.0 (32) 66.6 (38)
5–9 partners 7.3 (3) 2.1 (1) 8.8 (5)
110 partners 9.8 (4) 19.1 (9) 19.3 (11)
NOTE. Data are the percentage (no.) of subjects with characteristic, unless otherwise indicated. The
viral load and age of the patients were determined on the date of genotypic testing. Information on mode
of transmission and risk behavior was available through questionnaires completed by 195 participants in
the PHI cohort study. The sexual risk behavior of the population of men who have sex with men is the
no. of sexual partners during the 3-month period before PHI diagnosis. HS, heterosexual sex; IDU, injection
drug use; MMS, male-male sex.
a
2–4 persons/cluster.
b
15 persons/cluster.
c
The distribution of risk behavior in the chronic patient population is 57.5% for MMS, 28% for IDU,
and 14% for HS [17].
Clinical characteristics of nonclustered and clustered
transmissions. There were no differences in the viral load or
age distributions in infections associated with clustered or non-
clustered events (table 2). Epidemiologic data from the PHI
cohort study show that clustering could not be attributed to
differences in behavioral risk factors. The proportions of modes
of transmission (male-male sex, intravenous drug use, and het-
erosexual sex) were similar in clustered and nonclusteredtrans-
mission events (table 2). The overall incidence of high risk
sexual behavior with multiple partners was similar in non-
clustered and clustered transmission chains (table 2).
Drug-resistance mutational profiles were compared in the
genotyped PHI and CI patient populations (table 3). The ge-
notyped CI potential transmitter population had single-class
and multidrug resistance (MDR) in 17.5% and 52.5% of pa-
tients, respectively. This contrasts with the infrequent trans-

Amplified Transmission in Early HIV Infection •JID 2007:195 (1 April) •000
Table 3. Drug-resistance profiles in clustered and nonclustered
transmission events in genotyped primary HIV infection (PHI)
patients, compared with the treated chronically infected (CI) po-
tential transmitter population.
Drug-resistance
mutations
Patient population
P
a
CI treated
(np660)
PHI cluster
(np293)
PHI unique
(np300)
Wild type 189 (30) 250 (85.3) 245 (81.7) NS
Any resistance 461 (69.9) 43 (14.7) 55 (18.3) NS
NRTI only 79 (12) 8 (2.7) 16 (5.3) NS
NNRTI only 23 (3.5) 25 (8.5) 18 (6.0) NS
PI only 13 (2) 5 (1.7) 10 (3.3) NS
NRTI/NNRTI 81 (12.2) 3 (1.0) 6 (2.0) NS
NRTI/PI 121 (18.3) 0 3 (1.0) NS
NNRTI/PI 1 (0.1) 2 (0.7) 0 NS
NRTI/NNRTI/PI 145 (21.9) 0 2 (0.7) NS
Any NRTI only 425 (64.4) 11 (3.8) 27 (9.0) .009
Any NNRTI only 249 (37.8) 30 (10.2) 27 (9.0) NS
Any PI 279 (42.2) 7 (2.4) 15 (5.0) .092
NOTE. Data are no. (%) of subjects. The frequency of major and minor
resistance mutations associated with nucleoside reverse transcriptase inhib-
itors (NRTIs) and nonnucleoside reverse transcriptase inhibitors (NNRTIs), as
well as with protease inhibitors (PIs), were determined in the genotyped CI
and clustered and nonclustered PHI patient populations. Polymorphisms as-
sociated with resistance to NNRTIs (codons 98 and 179) and PIs (codons 10,
20, 36, 63, 71, 73, and 77) were excluded from analysis because there is a
high prevalence of these substitutions in treatment naive persons. NS, not
significant.
a
PHI clustered vs. unique.
mission of drug resistance and MDR in PHI (∼10% and ∼3%,
respectively) [16, 18].
At least half of transmitted drug resistance in PHI may reflect
forward transmission events as shown by the presence of drug-
resistance mutations in 14.7% of clustered transmissions (table
3). Clustering may lead to an overrepresentation of select resis-
tance-related mutational motifs. For example, one cluster
had 14 patients with the G190A mutation in RT. No significant
differences emerged in regard to the overall incidence of re-
sistance mutations within clustered and nonclustered isolates
(table 3). It is, however, noteworthy that the transmission of
mutations associated with resistance to nucleoside analogues
and protease inhibitors appeared less prevalent in clustered
transmissions (table 3).
DISCUSSION
Our results show that 49% of all PHI strains in the Quebec
HIV population form phylogenetic clusters, indicating that
early infection may account for a major proportion of onward
transmissions. This contrasts with the relatively low frequency
(!2%) of clustering observed within the CI genotyped popu-
lation, mostly representing individuals receiving long-term
therapy with antiretroviral drugs (ARVs). Thus, primary/early
infection, representing !10% of the total sequenced samples
in the provincial genotyping program, disproportionately ac-
counted for approximately half of onward transmission events.
Events surrounding acute/early infection may play a key role
in the spread of HIV. Our findings represent broad population-
based surveillance and are consistent with a recent report from
the Rekai study in Uganda in regard to primary/early infection
[6]. The latter suggested that the majority of new transmis-
sions may be due to contacts with individuals who were them-
selves in early stages of infection. In comparison with data
previously reported in the Swiss and British cohort studies,
the present study shows a higher incidence of PHI clustering
(∼50%) and larger sizes of clustered transmission events with
∼50% of transmission chains having ( )
8.8 3.5 mean SD
infections per cluster [13, 14]. Indeed, the role of acute in-
fection may be underestimated because ∼30% of PHIs remain
undiagnosed [17, 23, 24].
Some groups have postulated that cases of CI may be re-
sponsible for most HIV transmission and that early treatment
will have only limited impact on the spread of HIV [7–9, 25].
In stark contrast, we have rarely observed PHIs that can be
linked to CI transmitters. Our findings are different from those
reported in the North Carolina STAT study, on the basis of
partner identification after PHI [9]. The fact that all patients in
Quebec benefit from universal free access to medical care and
ARVs may reduce forward transmission from the treated poten-
tial transmitter population. Indeed, transmission of single-class
and primary MDR remains stable and relatively rare in Quebec,
representing 10% and 3% of all PHIs, respectively [16–18].
The potential benefits of highly active antiretroviral therapy
(HAART) in early infection may therefore be 2-fold. HAART
may lower the risk of onward transmission, and, in addition,
patients may potentially benefit from better immune control
and lower set points of viremia [1, 26, 27]. Initiation of HAART
during acute infection may be associated with durable virologic
and immunologic benefits for 172 weeks, compared with no
treatment [26]. Although some people who have newly diag-
nosed HIV infection may have already transmitted the virus to
others by the time of initiation of HAART, early treatment
intervention may nonetheless prevent significant numbers of
additional transmissions.
The high incidence of clustering of PHI-related transmission
events might be due to high risk sexual behavior. According to
the Quebec PHI cohort data, 23.8% of newly diagnosed men
who have sex with men had engaged in high risk sexual behavior
with ⭓5 partners before becoming infected, and such behaviors
did not significantly change subsequent to infection. These ob-
servations are consistent with those recently reported in regard
to increased coincidence of HIV with sexually transmitted dis-
eases and high risk sexual behavior [1, 7, 13, 14, 28]. However,
the present study showed that patients with high risk behavior
were found among small clusters, large clusters, and nonclus-

000 •JID 2007:195 (1 April) •Brenner et al.
tered infections. This suggests that different infections may vary
in transmissibility, and further study is necessary. However,
concerns arise from potential bias associated with self-reporting
of sensitive and stigmatized behavior and other factors.
The recent case report from New York City of a triple-class
MDR transmission in a man who has sex with men has raised
concerns as to the threat of transmission of replication-com-
petent drug-resistant variants of HIV [29]. Drug resistance ac-
quired in PHI is clonal and persists over time in the absence
of drug pressure [16, 30]. It is noteworthy that approximately
half of transmitted resistance can be attributed to clustered
infections but that transmission of viruses containing mutations
associated with resistance to nucleoside and protease inhibitors
was diminished in clustered infections, possibly because of re-
duced viral fitness.
The cost effectiveness and relevance of genotypic resistance
testing programs are often debated, because transmitted drug
resistance in PHI is relatively rare and resistance algorithms for
MDR in CI may be difficult to interpret. Our findings under-
score the importance of genotypic testing in PHI, as well as
of phylogenetic analysis, to evaluate evolving trends in HIV
transmission.
Several limitations must be considered with respect to these
results. Although some groups raise concerns in sequencing the
conserved HIV pol domain [31], our results are consistent with
those of other groups showing that there is adequate sequence
diversity to identify clustered transmission events [10–15, 32,
33]. We have also modified viral sequences in CIs to their
ancestral wild-type forms to eliminate any bias caused by re-
sistance mutations.
Although the present study compiled data from 3 sur-
veillance sites, we doubtless missed some PHIs in the Quebec
population during the study period (1996–2004). The incidence
of clustered transmission events in any given population de-
pends on how effectively the local health care system diagnoses
and tracks people with HIV infection. It is estimated that 30%–
50% of newly infected persons in North America may be un-
diagnosed and unaware of their serological status [17, 23, 24].
The nongenotyped CI patient population may be a source of
HIV-1 infection, although published findings by our group
show that this population generally has low viremia [17].
Taken together, our findings indicate that PHI can account
for a high proportion of HIV transmissions. Acute/early in-
fection is characterized by high viremia and high viral setpoints
in the absence of treatment [1]. Acute/early infections are often
undiagnosed, leading to high risk behavior, and unprotected
sex may facilitate transmission. Relatively homogeneous viral
quasispecies exist at early stages of infection, enhancing the
selective advantage of clonal transmissible species [16, 30].
Many multiresistant variants that arise in treated individuals
show reduced viral replicative fitness and transmissibility [30].
HIV-1–specific immunity may not arise during the first 2–4
months after PHI, and early immune responses may decrease
rapidly in the absence of treatment [34]. Treatment of CI pa-
tients reduces circulating viremia, a critical factor in HIV
transmission.
It is important to actively seek out recently infected persons
and to propose counselling to reduce high risk behavior during
this critical period [1, 3, 35–37]. Our findings further under-
score recommendations for genotyping in primary/early infec-
tion to document clustering of infection and to provide in-
formation on transmitted drug resistance, both as an issue in
public health and as a guide to future therapy [37].
QUEBEC PRIMARY HIV INFECTION STUDY
GROUP
The Quebec Primary HIV Infection Study Group includes R.
G. Lalonde, N. Gilmore, M. Klein, J. MacLeod, G. Smith, J.
Allan, C. Tsoukas, M. Potter, J. Falutz, and J. Cox (McGill
University Health Center); C. Fortin, A. de Pokomandy (Centre
Hospitalier de Universite´ de Montre´ al); B. Trottier, F. Asselin,
M. Boissonnault, L. Charest, H. Dion, S. Lavoie, D. Legault,
D. Longpre´, P. J. Maziade, M. E. Morin, D. Murphy, V. K.
Nguyen, R. O’Brien, and S. Ve´zina (Clinique Me´dicale l’Actuel);
P. C oˆte´ , S. Dufresne, P. Junod, F. Laplante, D. Poirier, Y. Parent,
M. A. Charron, B. Lessard, D. Tessier, E
´. Sasseville, A. Talbot,
and M. S. Joyal (Clinique Me´dicale du Quartier Latin); N.
Lapointe (Hoˆpital Ste-Justine); A. Dascal (Jewish General Hos-
pital); and M. Munoz (CLSC Coˆte des Neiges).
Acknowledgment
We thank all patients who participated in the study.
References
1. Pilcher CD, Tien HC, Eron JJ Jr, et al. Brief but efficient: acute infection
and the sexual transmission of HIV. Quest Study and Duke-UNC-
Emory Acute HIV Consortium. J Infect Dis 2004; 189:1785–92.
2. Quinn T, Wawer M, Sewankambo N, et al. Viral load and heterosexual
transmission of human immunodeficiency virus type 1. Rakai Project
Study Group. N Engl J Med 2000; 342:921–9.
3. Cohen MS, Pilcher CD. Amplified HIV transmission and new ap-
proaches to HIV prevention. J Infect Dis 2005; 191:1391–3.
4. Garcia PM, Kalish LA, Pitt J, et al. Maternal levels of plasma human
immunodeficiency virus type 1 RNA and the risk of perinatal trans-
mission. N Engl J Med 1999; 341:394–402.
5. Leynaert B, Downs AM, de Vincenzi I. Heterosexual transmission of
human immunodeficiency virus: variability of infectivity throughout
the course of infection. European Study Group on heterosexual trans-
mission of HIV. Am J Epidemiol 1998; 148:88–96.
6. Wawer MJ, Gray RH, Sewankambo NK, et al. Rates of HIV-1 trans-
mission per coital act, by stage of HIV-1 infection, in Rakai, Uganda.
J Infect Dis 2005; 191:1403–9.
7. Hayes RJ, White RG. Amplified HIV transmission during early-stage
infection [letter and reply]. J Infect Dis 2006; 193:604–6.
8. Hollingsworth D, Anderson R, Fraser C. Has the role of primary in-

Amplified Transmission in Early HIV Infection •JID 2007:195 (1 April) •000
fection in the epidemiology of HIV been overstated? [abstract 913] In:
Program and abstracts of the 13th Conference on Retroviruses and
Opportunistic Infections (Denver). 2006:386.
9. Pilcher C, Foust E, Ashby R, et al. Sexual transmission risk and rapid
public health intervention in acute infection [abstract 371]. In: Program
and abstracts of the 13th Conference on Retroviruses and Opportu-
nistic Infections (Denver). 2006:173.
10. Hue S, Clewley JP, Cane PA, Pillay D. HIV-1 pol gene variation is
sufficient for reconstruction of transmissions in the era of antiretroviral
therapy. AIDS 2004; 18:719–28.
11. Hue S, Clewley JP, Cane PA, Pillay D. Investigation of HIV-1 trans-
mission events by phylogenetic methods; requirement for scientific
rigour. AIDS 2005; 19:449–50.
12. Hue S, Pillay D, Clewley JP, Pybus OG. Genetic analysis reveals the
complex structure of HIV-1 transmission within defined groups. Proc
Natl Acad Sci 2005; 102:4425–9.
13. Pao D, Fisher M, Hue S, et al. Transmission of HIV-1 during primary
infection: relationship to sexual risk and sexually transmittedinfections.
AIDS 2005; 19:85–90.
14. Yerly S, Jost S, Telenti A, et al. Infrequent transmission of HIV-1 drug-
resistant variants. Antivir Ther 2004; 9:375–84.
15. Grant RM, Liegler TI, Hecht FM. Clusters and trends in primary re-
sistance in San Francisco 2001–2003. Antivir Ther 2004; 9:S104.
16. Brenner B, Routy JP, Quan Y, et al. Persistence of multidrug-resistant
HIV-1 in primary infection leading to superinfection. AIDS 2004; 18:
1653–60.
17. Routy JP, Machouf N, Edwardes MD, et al. Factors associated with a
decrease in the prevalence of drug resistance in newly HIV-1 infected
individuals in Montreal. AIDS 2004; 18:2305–12.
18. Salomon H, Quan Y, Brenner B, et al. Prevalence of HIV-1 resistant
to antiretroviral drugs in 81 individuals newly infected by sexual contact
or injecting drug use. AIDS 2000; 14:F17–23.
19. Akouamba BS, Viel J, Charest H, et al. HIV-1 genetic diversity and
clade introduction in an antenatal cohort in Montreal, Canada.Emerg-
ing Infect Dis 2005; 11:1230–4.
20. Hall TA. BioEdit: a user-friendly biological sequence alignment editor
and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser
1999; 41:95–8.
21. Johnson VA, Brun-Vezinet F, Clotet B, et al. Update of the drug re-
sistance mutations in HIV-1: 2005. Top HIV Med 2005; 13:51–7.
22. Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolu-
tionary genetics analysis software. Bioinformatics 2001; 17:1244–5.
23. Daar ES, Little S, Pitt J, et al. Diagnosis of primary HIV-1 infection.
Ann Intern Med 2001; 134:25–9.
24. Geduld J, Gatali M, Remis RS, Archibald CP. Estimates of HIV prev-
alence and incidence in Canada, 2002. Can Commun Dis Rep 2003;
29:197–207.
25. Xiridou M, Geskus R, de Wit J, Coutinho R, Kretzschmar M. Primary
HIV infection as source of HIV transmission within steady and casual
partnerships among homosexual men. AIDS 2004; 18:1311–20.
26. Hecht F, Wang L, Collier A, et al. Outcomes of HAART for acute/early
infection after treatment discontinuation [abstract 568]. In: Program
and abstracts of 12th Conference on Retroviruses and Opportunistic
Infections (Boston). 2005.
27. Smith DE, Walker BD, Cooper DA, Rosenberg ES, Kaldor JM. Is an-
tiretroviral treatment of primary HIV infection clinically justified on
the basis of current evidence? AIDS 2004; 18:709–18.
28. Dukers NH, Spaargaren J, Geskus RB, et al. HIV incidence on the
increase among homosexual men attending an Amsterdam sexually
transmitted disease clinic: using a novel approach for detecting recent
infections. AIDS 2002; 16:F19–24.
29. Markowitz M, Mohri H, Mehandru S, et al. Infection with multidrug
resistant, dual-tropic HIV-1 and rapid progression to AIDS: a case
report. Lancet 2005; 365:1031–8.
30. Brenner B, Routy JP, Petrella M, et al. Persistence and fitness of mul-
tidrug resistant human immunodeficiency virus type 1 acquired in
primary infection. J Virol 2002; 76:1753–61.
31. Sturmer M, Preiser W, Gute P, Nisius G, Doerr HW. Phylogenetic
analysis of HIV-1 transmission: pol gene sequences are insufficient to
clarify true relationships between patient isolates. AIDS 2004; 18:2109–13.
32. Hue S, Clewley JP, Cane PA, Pillay D. Investigation of HIV-1 trans-
mission events by phylogenetic analysis: requirement for scientific rig-
our. AIDS 2005; 19:449–50.
33. Yerly S, Vora S, Rizzardi P, et al. Acute HIV infection: impact on the
spread of HIV and transmission of drug resistance. AIDS 2001; 15:
2287–92.
34. Alter G, Tsoukas CM, Rouleau D, et al. Assessment of longitudinal
changes in HIV-specific effector activity in subjects undergoing un-
treated primary HIV infection. AIDS 2004; 18:1979–89.
35. Grassly NC, Garnett GP. The future of the HIV pandemic. Bull World
Health Organ 2005; 83:378–82.
36. Pilcher CD, McPherson JT, Leone PA, et al. Real-time, universalscreen-
ing for acute infection in a routine HIV counselling and testing pop-
ulation. JAMA 2002; 288:216–21.
37. Vandamme AM, Sonnerborg A, Ait-Khaled M, et al. Updated European
recommendations for the clinical use of HIV drug resistance testing.
Antivir Ther 2004; 9:829–48.