UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Anal HPV infection & diseaseCommon and preventable, but hard to treatMarra, E.

Link to publication

Creative Commons License (see https://creativecommons.org/use-remix/cc-licenses):Other

Citation for published version (APA):Marra, E. (2018). Anal HPV infection & disease: Common and preventable, but hard to treat.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 30 Jun 2020

CHAPTER 6

Virological and serological predictors of anal high-grade squamous intraepithelial

lesions among HIV-positive men who have sex with men

Elske Marra, Matthijs H. Siegenbeek van Heukelom, Annemiek Leeman,

Tim Waterboer, Chris J.L.M. Meijer, Peter J.F. Snijders, Audrey J. King,

Irina Cairo, Arne van Eeden, Wilma Brokking, Wim Quint, Pascal van der Weele,

Jan M. Prins, Henry J.C. de Vries, Maarten F. Schim van der Loeff

Submitted

120

Predictors of anal HSIL

Abstract Background Our objective was to identify virological and serological predictors of anal high-grade squamous intraepithelial lesions (HSIL) in HIV-positive men-who-have-sex-with-men (MSM). Methods HIV-positive MSM were recruited from a longitudinal study (2010-2013), during which anal self-swabs and serum were collected at up to five bi-annual visits. Swabs were HPV genotyped, and type-specific HPV viral load in anal swabs was determined. Serum antibodies to E6, E7, E1, E2 and L1 proteins of 7 hrHPV-types and HPV6 and 11 were analyzed. 193 participants had a high-resolution anoscopy (HRA) after the last study visit and were included in the current analysis. Anal HSIL was diagnosed by histopathological examination of anal biopsies. Causative HPV-type of anal HSIL was determined in whole tissue sections (WTS) and by laser capture micro-dissection if more than one HPV-type was found in WTS. Multivariable logistic regression was used to study whether persistent anal HPV infection, HPV viral load and seropositivity for HPV were predictors of anal HSIL, in general and for concordant causative HPV-type. Results Of 193 HIV-positive MSM, 50 (26%) were diagnosed with anal HSIL. HrHPV persistence in anal swabs was common, varying by hrHPV-type between 3-21%. Neither anal hrHPV viral load, nor seropositivity for L1, E6, E7, E1, or E2 was associated with anal HSIL. Only anal HPV persistence was independent associated with anal HSIL, in general and by concordant causative HPV-type. Conclusion Persistent HPV infection was strongly associated with anal HSIL, in general as well as for concordant HPV-type.

Introduction Infection with high risk human papillomavirus (hrHPV) can cause cervical cancer 1-4,

as well as anal 5,6, penile 7 and head-and-neck cancer 8,9. Men who have sex with men (MSM) are at increased risk of HPV associated anal cancer and its precursors 10. Incidence of anal cancer ranges from 5 per 100,000 per year among HIV-negative MSM, rising to 78 per 100,000 per year among HIV-positive MSM in the combination antiretroviral therapy (cART)-era 10-12.

Precursor lesions of anal cancer 12,13 are histopathologically graded as anal intraepithelial neoplasia (AIN) 1, 2, and 3 and categorized as low-grade squamous intraepithelial lesions (LSIL; AIN1) or high-grade squamous intraepithelial lesions (HSIL; AIN2/3). The gold standard for anal HSIL screening is high-resolution anoscopy (HRA) with biopsy of suspicious lesions 14-16. HRA is time-consuming, technically challenging and cumbersome for the patient. Selecting patients at risk of anal HSIL for HRA could be beneficial from a patient and health-care costs perspective. Yet, current alternative screening methods either lack sensitivity and specificity 10,17-19. Evaluation of demographic-, behavioral- and HIV-related factors for triage of HIV-positive MSM at higher risk of anal HSIL showed no uniform risk factors 12,15,20-23. Biomarkers, such as naturally acquired HPV antibodies, anal HPV persistence and anal HPV viral load, might also be used as predictors of anal HSIL. Serum antibodies against proteins of HPV are of interest as potential predictors of anal HSIL, as they were shown to be strong predictors of anal 24 and oropharyngeal cancer 25,26. Therefore, the aim of this study was to assess the predictive power of anal HPV persistence, anal hrHPV viral load and seropositivity for different HPV proteins (L1,E6,E7,E1,E2) for the presence of anal HSIL among HIV-positive MSM. Methods

Study participants Study design and sample collection of the HIV & HPV in MSM (H2M) study have

been previously described 19. In brief, HIV-negative and HIV-positive MSM aged ≥18 years were recruited for a prospective cohort study in 2010–2011 at three sites in Amsterdam, the Netherlands 19. Participants were followed up approximately every 6 months, for a maximum period of 24 months per person.

121

Predictors of anal HSIL

6

Abstract Background Our objective was to identify virological and serological predictors of anal high-grade squamous intraepithelial lesions (HSIL) in HIV-positive men-who-have-sex-with-men (MSM). Methods HIV-positive MSM were recruited from a longitudinal study (2010-2013), during which anal self-swabs and serum were collected at up to five bi-annual visits. Swabs were HPV genotyped, and type-specific HPV viral load in anal swabs was determined. Serum antibodies to E6, E7, E1, E2 and L1 proteins of 7 hrHPV-types and HPV6 and 11 were analyzed. 193 participants had a high-resolution anoscopy (HRA) after the last study visit and were included in the current analysis. Anal HSIL was diagnosed by histopathological examination of anal biopsies. Causative HPV-type of anal HSIL was determined in whole tissue sections (WTS) and by laser capture micro-dissection if more than one HPV-type was found in WTS. Multivariable logistic regression was used to study whether persistent anal HPV infection, HPV viral load and seropositivity for HPV were predictors of anal HSIL, in general and for concordant causative HPV-type. Results Of 193 HIV-positive MSM, 50 (26%) were diagnosed with anal HSIL. HrHPV persistence in anal swabs was common, varying by hrHPV-type between 3-21%. Neither anal hrHPV viral load, nor seropositivity for L1, E6, E7, E1, or E2 was associated with anal HSIL. Only anal HPV persistence was independent associated with anal HSIL, in general and by concordant causative HPV-type. Conclusion Persistent HPV infection was strongly associated with anal HSIL, in general as well as for concordant HPV-type.

Introduction Infection with high risk human papillomavirus (hrHPV) can cause cervical cancer 1-4,

as well as anal 5,6, penile 7 and head-and-neck cancer 8,9. Men who have sex with men (MSM) are at increased risk of HPV associated anal cancer and its precursors 10. Incidence of anal cancer ranges from 5 per 100,000 per year among HIV-negative MSM, rising to 78 per 100,000 per year among HIV-positive MSM in the combination antiretroviral therapy (cART)-era 10-12.

Precursor lesions of anal cancer 12,13 are histopathologically graded as anal intraepithelial neoplasia (AIN) 1, 2, and 3 and categorized as low-grade squamous intraepithelial lesions (LSIL; AIN1) or high-grade squamous intraepithelial lesions (HSIL; AIN2/3). The gold standard for anal HSIL screening is high-resolution anoscopy (HRA) with biopsy of suspicious lesions 14-16. HRA is time-consuming, technically challenging and cumbersome for the patient. Selecting patients at risk of anal HSIL for HRA could be beneficial from a patient and health-care costs perspective. Yet, current alternative screening methods either lack sensitivity and specificity 10,17-19. Evaluation of demographic-, behavioral- and HIV-related factors for triage of HIV-positive MSM at higher risk of anal HSIL showed no uniform risk factors 12,15,20-23. Biomarkers, such as naturally acquired HPV antibodies, anal HPV persistence and anal HPV viral load, might also be used as predictors of anal HSIL. Serum antibodies against proteins of HPV are of interest as potential predictors of anal HSIL, as they were shown to be strong predictors of anal 24 and oropharyngeal cancer 25,26. Therefore, the aim of this study was to assess the predictive power of anal HPV persistence, anal hrHPV viral load and seropositivity for different HPV proteins (L1,E6,E7,E1,E2) for the presence of anal HSIL among HIV-positive MSM. Methods

Study participants Study design and sample collection of the HIV & HPV in MSM (H2M) study have

been previously described 19. In brief, HIV-negative and HIV-positive MSM aged ≥18 years were recruited for a prospective cohort study in 2010–2011 at three sites in Amsterdam, the Netherlands 19. Participants were followed up approximately every 6 months, for a maximum period of 24 months per person.

122

Predictors of anal HSIL

Figure 1. Flow-chart of the H2M2 study among HIV-positive MSM. Within the H2M2 study, analyses were done in two different ways: (1) Studying predictors of anal HSIL on patient-level, in which the anal HSIL diagnosis was based on histology done by the pathology department of the clinic where the HRA was done; (2) Studying predictors of anal HSIL with known causative HPV type, in which the HPV type-specific HSIL diagnosis was based on histology and on results of whole tissue section analysis and, if indicated, laser capture microdissection of the anal biopsies. * No left-over tissue or HSIL not confirmed in re-assessment of available left-over tissue of the biopsy. Abbreviations: HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, WTS – whole tissue section, LCM – laser capture microdissection, MSM – men who have sex with men, H2M study – HPV and HIV in MSM study, AIN – anal intraepithelial neoplasia, HRA – high resolution anoscopy.

HRA was offered to HIV-positive participants in three clinics in Amsterdam during the study period: the Academic Medical Center (AMC), Onze Lieve Vrouwe Gasthuis (OLVG), and DC Klinieken. Only HIV-positive H2M participants who underwent a first HRA after their last H2M visit in one of these three clinics before December 2015 were included in this study (Figure 1). The Medical Ethics Committee of the AMC approved this study [MEC 07/182] and all participants provided written informed consent prior to enrolment.

Data collection At each H2M visit, participants completed a self-administered questionnaire.

Venous blood, and anal self-swabs were collected [regular flocked swab with 1 ml Universal Transport Medium (UTM); Copan, Brescia, Italy] 19. HIV-related data were obtained from the Dutch HIV Monitoring Foundation 27. Clinical information related to the HRA was obtained from the infectious diseases physician or dermatologist.

Human papillomavirus DNA detection and genotyping DNA detection and HPV genotyping of the anal samples has been previously

described 19. Briefly, DNA extraction was performed using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche, Mannheim, Germany). Broad-spectrum HPV DNA amplification was performed using the highly sensitive SPF10-PCR DEIA/LiPA25 system (version 1) 28. Persistent HPV infection was defined as at least three positive anal samples of the same HPV-type, with a maximum of one negative anal sample in between.

HPV viral load determination Type-specific hrHPV anal viral load was determined if a participant was positive for

HPV16,18,31,33,45,52,58 at the last H2M visit. Viral load was determined using a previously described type specific L1-targeting quantitative qPCR protocol optimized to approach SPF10-LiPA25 sensitivity levels 29,30. HPV viral loads were standardized for the amount of human cells present in each sample via a β-actin qPCR 30. qPCRs were performed in 20µl final volume using LightCycler TaqMan Master on the Roche LightCycler 480 platform (Roche Diagnostics, Almere, the Netherlands). The type-specific median anal HPV viral load was determined, after which categorical variables by hrHPV-type were made, grouping participants in three groups: no HPV infection,

123

Predictors of anal HSIL

6

Figure 1. Flow-chart of the H2M2 study among HIV-positive MSM. Within the H2M2 study, analyses were done in two different ways: (1) Studying predictors of anal HSIL on patient-level, in which the anal HSIL diagnosis was based on histology done by the pathology department of the clinic where the HRA was done; (2) Studying predictors of anal HSIL with known causative HPV type, in which the HPV type-specific HSIL diagnosis was based on histology and on results of whole tissue section analysis and, if indicated, laser capture microdissection of the anal biopsies. * No left-over tissue or HSIL not confirmed in re-assessment of available left-over tissue of the biopsy. Abbreviations: HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, WTS – whole tissue section, LCM – laser capture microdissection, MSM – men who have sex with men, H2M study – HPV and HIV in MSM study, AIN – anal intraepithelial neoplasia, HRA – high resolution anoscopy.

HRA was offered to HIV-positive participants in three clinics in Amsterdam during the study period: the Academic Medical Center (AMC), Onze Lieve Vrouwe Gasthuis (OLVG), and DC Klinieken. Only HIV-positive H2M participants who underwent a first HRA after their last H2M visit in one of these three clinics before December 2015 were included in this study (Figure 1). The Medical Ethics Committee of the AMC approved this study [MEC 07/182] and all participants provided written informed consent prior to enrolment.

Data collection At each H2M visit, participants completed a self-administered questionnaire.

Venous blood, and anal self-swabs were collected [regular flocked swab with 1 ml Universal Transport Medium (UTM); Copan, Brescia, Italy] 19. HIV-related data were obtained from the Dutch HIV Monitoring Foundation 27. Clinical information related to the HRA was obtained from the infectious diseases physician or dermatologist.

Human papillomavirus DNA detection and genotyping DNA detection and HPV genotyping of the anal samples has been previously

described 19. Briefly, DNA extraction was performed using the MagNA Pure LC Total Nucleic Acid Isolation Kit (Roche, Mannheim, Germany). Broad-spectrum HPV DNA amplification was performed using the highly sensitive SPF10-PCR DEIA/LiPA25 system (version 1) 28. Persistent HPV infection was defined as at least three positive anal samples of the same HPV-type, with a maximum of one negative anal sample in between.

HPV viral load determination Type-specific hrHPV anal viral load was determined if a participant was positive for

HPV16,18,31,33,45,52,58 at the last H2M visit. Viral load was determined using a previously described type specific L1-targeting quantitative qPCR protocol optimized to approach SPF10-LiPA25 sensitivity levels 29,30. HPV viral loads were standardized for the amount of human cells present in each sample via a β-actin qPCR 30. qPCRs were performed in 20µl final volume using LightCycler TaqMan Master on the Roche LightCycler 480 platform (Roche Diagnostics, Almere, the Netherlands). The type-specific median anal HPV viral load was determined, after which categorical variables by hrHPV-type were made, grouping participants in three groups: no HPV infection,

124

Predictors of anal HSIL

Tabl

e 1.

Cha

ract

eris

tics o

f the

stud

y po

pula

tion

of th

e H2

M2

(HIV

& H

PV in

MSM

2) s

tudy

by

anal

his

tolo

gica

l hig

h-gr

ade

squa

mou

s in

trae

pith

elia

l les

ion

(HSI

L) st

atus

.

To

tal (

N=1

93)

No

HSIL

(N=1

43)

HSIL

(N=5

0)

P va

lue#

Dem

ogra

phic

and

beha

viou

ral v

aria

bles

Age

in y

ears

at m

omen

t of H

RA, m

ean

(SD)

50

(1

0)

50

(10)

50

(9

) 0.

7 Sm

okin

g st

atus

at m

omen

t of H

RA a

0.4

Nev

er sm

oked

68

39

%

52

38%

16

40

%

Pr

evio

usly

smok

ed

63

36%

46

34

%

17

43%

Curr

ently

smok

ing

45

26%

38

28

%

7 18

%

Hi

gher

edu

catio

n (h

ighe

r pro

fess

iona

l edu

catio

n or

uni

vers

ity)*

0.

05

No

68

35%

56

39

%

12

24%

Yes

125

65%

87

61

%

38

76%

Livi

ng si

tuat

ion*

b

0.5

Livi

ng a

lone

95

50

%

71

50%

24

49

%

Li

ving

toge

ther

with

a st

eady

par

tner

91

48

%

66

46%

25

51

%

O

ther

5

3%

5 4%

0

0%

Co

untr

y of

birt

h* b

0.1

Net

herla

nds

145

76%

10

4 73

%

41

84%

Oth

er

46

24%

38

27

%

8 16

%

Se

xual

beh

avio

ural

var

iabl

es

Li

fetim

e nu

mbe

r of s

ex p

artn

ers,

med

ian

(IQR)

* a

400

(100

-100

0)

400

(150

-100

0)

350

(100

-100

0)

0.5

Life

time

num

ber o

f sex

par

tner

s, m

edia

n (IQ

R) *

a

0.8

<25

9 7%

6

5%

3 7%

25-9

9 19

14

%

13

10%

6

14%

100-

499

63

36%

48

36

%

15

34%

≥500

85

48

%

65

49%

20

45

%

Co

ndom

use

dur

ing

anal

sex

in th

e pr

eced

ing

6 m

onth

s* c

0.3

Nev

er

17

11%

14

12

%

3 7%

Som

etim

es

90

58%

62

54

%

28

68%

Alw

ays

48

31%

38

33

%

10

24%

HIV-

rela

ted

varia

bles

(all

arou

nd ti

me

of H

RA)

Curr

ently

usin

g cA

RT

0.6

No

5 3%

3

2%

2 4%

Yes

188

97%

14

0 98

%

48

96%

Dura

tion

of c

ART

use

in y

ears

, med

ian

(IQR)

d 8.

6 (5

.0-1

5.9)

9.

7 (5

.0-1

6.1)

7.

9 (4

.7-1

2.0)

0.

3 Ye

ars l

ivin

g w

ith v

iral s

uppr

essio

n e

0.4

<3 y

ears

30

16

%

21

15%

9

18%

3-10

yea

rs

88

46%

63

44

%

25

51%

>10

year

s 73

38

%

58

41%

15

31

%

N

adir

CD4

T-ce

ll co

unt,

cells

/µl,

mea

n (S

D) f

245

(134

) 24

0 (1

40)

258

(116

) 0.

4 CD

4 T-

cell

coun

t cel

ls/µl

, mea

n (S

D) g

681

(250

) 68

1 (2

57)

681

(231

) 0.

9 HI

V-RN

A pl

asm

a, c

opie

s/m

l

0.

5 <5

0 18

2 94

%

136

95%

45

92

%

≥5

0 11

6%

7

5%

4 8%

AIN

diag

nosis

AIN

dia

gnos

is at

firs

t HRA

No

biop

sies t

aken

42

22

%

42

29%

0

0%

N

o dy

spla

sia

68

35%

68

48

%

0 0%

AIN

1 33

17

%

33

23%

0

0%

AI

N2

25

13%

0

0%

25

50%

AIN

3 25

13

%

0 0%

25

50

%

Tim

e be

twee

n la

st H

2M v

isit a

nd H

RA in

yea

rs, m

edia

n (IQ

R)

1.3

(1.0

-1.8

) 1.

3 (0

.9-1

.8)

1.4

(0.9

-2.0

) 0.

1

Data

are

repr

esen

ted

as n

(%) u

nles

s ot

herw

ise in

dica

ted.

*

Varia

ble

from

the

base

line

visit

of H

2M st

udy;

# S

igni

fican

ce o

f diff

eren

ces b

etw

een

part

icip

ants

with

and

with

out a

nal H

SIL

was

ass

esse

d by

Chi

-squ

are

test

s or

Fish

er’s

exa

ct te

st fo

r cat

egor

ical

var

iabl

es a

nd W

ilcox

on ra

nk-s

um te

sts f

or c

ontin

uous

var

iabl

es.

(a) 1

7 m

issin

gs; (

b) 2

miss

ings

; (c)

38

miss

ing;

(d) 5

miss

ings

; (e)

2 m

issin

gs; (

f) 1

miss

ing;

(g) 2

6 m

issin

gs; (

h) 4

miss

ings

Ab

brev

iatio

ns: H

IV –

hum

an im

mun

odef

icie

ncy

viru

s; H

PV –

hum

an p

apill

omav

irus,

HSI

L –h

igh-

grad

e sq

uam

ous i

ntra

epith

elia

l les

ion,

MSM

– m

en w

ho h

ave

sex

with

men

, H2M

stud

y –

HPV

and

HIV

in M

SM st

udy,

AIN

– a

nal i

ntra

epith

elia

l neo

plas

ia, H

RA –

hig

h re

solu

tion

anos

copy

, SD

– st

anda

rd d

evia

tion,

IQR

– in

terq

uart

ile ra

nge.

125

Predictors of anal HSIL

6

Tabl

e 1.

Cha

ract

eris

tics o

f the

stud

y po

pula

tion

of th

e H2

M2

(HIV

& H

PV in

MSM

2) s

tudy

by

anal

his

tolo

gica

l hig

h-gr

ade

squa

mou

s in

trae

pith

elia

l les

ion

(HSI

L) st

atus

.

To

tal (

N=1

93)

No

HSIL

(N=1

43)

HSIL

(N=5

0)

P va

lue#

Dem

ogra

phic

and

beha

viou

ral v

aria

bles

Age

in y

ears

at m

omen

t of H

RA, m

ean

(SD)

50

(1

0)

50

(10)

50

(9

) 0.

7 Sm

okin

g st

atus

at m

omen

t of H

RA a

0.4

Nev

er sm

oked

68

39

%

52

38%

16

40

%

Pr

evio

usly

smok

ed

63

36%

46

34

%

17

43%

Curr

ently

smok

ing

45

26%

38

28

%

7 18

%

Hi

gher

edu

catio

n (h

ighe

r pro

fess

iona

l edu

catio

n or

uni

vers

ity)*

0.

05

No

68

35%

56

39

%

12

24%

Yes

125

65%

87

61

%

38

76%

Livi

ng si

tuat

ion*

b

0.5

Livi

ng a

lone

95

50

%

71

50%

24

49

%

Li

ving

toge

ther

with

a st

eady

par

tner

91

48

%

66

46%

25

51

%

O

ther

5

3%

5 4%

0

0%

Co

untr

y of

birt

h* b

0.1

Net

herla

nds

145

76%

10

4 73

%

41

84%

Oth

er

46

24%

38

27

%

8 16

%

Se

xual

beh

avio

ural

var

iabl

es

Li

fetim

e nu

mbe

r of s

ex p

artn

ers,

med

ian

(IQR)

* a

400

(100

-100

0)

400

(150

-100

0)

350

(100

-100

0)

0.5

Life

time

num

ber o

f sex

par

tner

s, m

edia

n (IQ

R) *

a

0.8

<25

9 7%

6

5%

3 7%

25-9

9 19

14

%

13

10%

6

14%

100-

499

63

36%

48

36

%

15

34%

≥500

85

48

%

65

49%

20

45

%

Co

ndom

use

dur

ing

anal

sex

in th

e pr

eced

ing

6 m

onth

s* c

0.3

Nev

er

17

11%

14

12

%

3 7%

Som

etim

es

90

58%

62

54

%

28

68%

Alw

ays

48

31%

38

33

%

10

24%

HIV-

rela

ted

varia

bles

(all

arou

nd ti

me

of H

RA)

Curr

ently

usin

g cA

RT

0.6

No

5 3%

3

2%

2 4%

Yes

188

97%

14

0 98

%

48

96%

Dura

tion

of c

ART

use

in y

ears

, med

ian

(IQR)

d 8.

6 (5

.0-1

5.9)

9.

7 (5

.0-1

6.1)

7.

9 (4

.7-1

2.0)

0.

3 Ye

ars l

ivin

g w

ith v

iral s

uppr

essio

n e

0.4

<3 y

ears

30

16

%

21

15%

9

18%

3-10

yea

rs

88

46%

63

44

%

25

51%

>10

year

s 73

38

%

58

41%

15

31

%

N

adir

CD4

T-ce

ll co

unt,

cells

/µl,

mea

n (S

D) f

245

(134

) 24

0 (1

40)

258

(116

) 0.

4 CD

4 T-

cell

coun

t cel

ls/µl

, mea

n (S

D) g

681

(250

) 68

1 (2

57)

681

(231

) 0.

9 HI

V-RN

A pl

asm

a, c

opie

s/m

l

0.

5 <5

0 18

2 94

%

136

95%

45

92

%

≥5

0 11

6%

7

5%

4 8%

AIN

diag

nosis

AIN

dia

gnos

is at

firs

t HRA

No

biop

sies t

aken

42

22

%

42

29%

0

0%

N

o dy

spla

sia

68

35%

68

48

%

0 0%

AIN

1 33

17

%

33

23%

0

0%

AI

N2

25

13%

0

0%

25

50%

AIN

3 25

13

%

0 0%

25

50

%

Tim

e be

twee

n la

st H

2M v

isit a

nd H

RA in

yea

rs, m

edia

n (IQ

R)

1.3

(1.0

-1.8

) 1.

3 (0

.9-1

.8)

1.4

(0.9

-2.0

) 0.

1

Data

are

repr

esen

ted

as n

(%) u

nles

s ot

herw

ise in

dica

ted.

*

Varia

ble

from

the

base

line

visit

of H

2M st

udy;

# S

igni

fican

ce o

f diff

eren

ces b

etw

een

part

icip

ants

with

and

with

out a

nal H

SIL

was

ass

esse

d by

Chi

-squ

are

test

s or

Fish

er’s

exa

ct te

st fo

r cat

egor

ical

var

iabl

es a

nd W

ilcox

on ra

nk-s

um te

sts f

or c

ontin

uous

var

iabl

es.

(a) 1

7 m

issin

gs; (

b) 2

miss

ings

; (c)

38

miss

ing;

(d) 5

miss

ings

; (e)

2 m

issin

gs; (

f) 1

miss

ing;

(g) 2

6 m

issin

gs; (

h) 4

miss

ings

Ab

brev

iatio

ns: H

IV –

hum

an im

mun

odef

icie

ncy

viru

s; H

PV –

hum

an p

apill

omav

irus,

HSI

L –h

igh-

grad

e sq

uam

ous i

ntra

epith

elia

l les

ion,

MSM

– m

en w

ho h

ave

sex

with

men

, H2M

stud

y –

HPV

and

HIV

in M

SM st

udy,

AIN

– a

nal i

ntra

epith

elia

l neo

plas

ia, H

RA –

hig

h re

solu

tion

anos

copy

, SD

– st

anda

rd d

evia

tion,

IQR

– in

terq

uart

ile ra

nge.

126

Predictors of anal HSIL

positive with HPV viral load ≤median, and positive with HPV viral load >median HPV viral load.

Serology All serum samples were stored at -80°C until analysis. HPV antibody detection

against L1, E6, E7 for HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58, and against E1 and E2 for HPV-types 16 and 18 was conducted simultaneously using a glutathione S-transferase multiplex assay 31. Antibody specific seropositivity for all HPV antigens was calculated based on standardized serology cutoff values (Supplementary table 1).

HRA procedure The HRA procedure has been described previously 20,32,33. Briefly, HRA consisted of a

digital rectal examination followed by intra- and perianal inspection with a high-resolution colposcope after repeatedly applying acetic acid (3–5% solution) and staining with Lugol’s iodine when indicated. Areas suspicious for squamous intraepithelial lesions (SIL) were biopsied.

Histopathology Biopsied lesions were graded by specialized pathologists. In two out of three clinics,

pathologists used p16 staining for AIN2 grading of biopsies, as recommended by the College of American Pathologists 16. The highest grade biopsy defined the overall diagnosis. Available left-over tissue of biopsies of participants with anal HSIL was reevaluated at DDL Diagnostic Laboratory, Rijswijk, the Netherlands. Reassessment entailed morphological examination of a newly cut Haematoxylin and Eosin (H/E) stained slide with supportive use of p16, by two specialized pathologists. HPV-types in the lesion were determined by genotyping using the SPF10-PCR DEIA/LiPA25 system (version 1) on the whole tissue section (WTS) and, if WTS analysis yielded more than one HPV-type, on specific lesions after laser capture micro-dissection 28,33,34.The single HPV-type found in a lesion was called the causative HPV-type.

Statistical analyses Differences in characteristics between men without anal HSIL (no anal HSIL/AIN1)

and with anal HSIL (AIN2/AIN3) were assessed by using Chi-squared or Fisher’s exact tests for categorical variables and Wilcoxon rank-sum tests for continuous variables.

Prediction of anal HSIL Analyses on prediction of anal HSIL were conducted on two levels: patient-level and

causative HPV-type-level (Figure 1). For the patient-level analyses a dataset was created of one record per participant

with the outcome anal HSIL (regardless of causative type) vs. no anal HSIL. Univariable and multivariable logistic regression was conducted to study the predictive power of each biomarker for anal HSIL diagnosis. Variables associated at P<0.2 in the univariable model were assessed in multivariable analysis, using a backward selection method.

For the causative HPV-type-level analyses a dataset was created of one record per HPV-type per patient. Each record included a value for HSIL: 1 if the patient had HSIL caused by that particular type and 0 if the patient did not have HSIL caused by that particular HPV-type. Of some lesions the HSIL diagnosis could not be confirmed (Figure 1), because no left-over tissue was available, or no HSIL was detected in the left-over material of the lesion. As these HSIL could not be assigned to an HPV-type, these men were excluded from the analyses. Univariable and multivariable logistic regression analyses using generalized estimating equations (GEE) was done to determine the association of type-specific biomarkers with type-specific anal HSIL diagnosis. GEE accounted for multiple observations (different HPV-types) within the same participant. The main dataset consisted of data on HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58. As there were 193 men in this study, this dataset contained 1737 records. Biomarkers associated in univariable regression at P<0.2 were included in multivariable logistic regression using GEE. A backward selection procedure was done to obtain a parsimonious model.

In both types of analyses age, number of years living with viral suppression 20 and time between last H2M visit and HRA were included as potential determinants. If determinants were found in any of the multivariable analyses, accuracy of the determinant as a predictor was assessed using a receiver operator curve (ROC) to estimate the area under the ROC curve (AUC).

Sensitivity analyses on both patient-level and causative HPV-type-level were done by using AIN2 vs. no anal HSIL, and AIN3 vs. no anal HSIL as outcome. Further sensitivity analyses on HPV-type-level were done on datasets with different combinations of HPV-types: (1) HPV-types 16 and 18; (2) HPV-types 16, 18, 31, 33, 45,

127

Predictors of anal HSIL

6

positive with HPV viral load ≤median, and positive with HPV viral load >median HPV viral load.

Serology All serum samples were stored at -80°C until analysis. HPV antibody detection

against L1, E6, E7 for HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58, and against E1 and E2 for HPV-types 16 and 18 was conducted simultaneously using a glutathione S-transferase multiplex assay 31. Antibody specific seropositivity for all HPV antigens was calculated based on standardized serology cutoff values (Supplementary table 1).

HRA procedure The HRA procedure has been described previously 20,32,33. Briefly, HRA consisted of a

digital rectal examination followed by intra- and perianal inspection with a high-resolution colposcope after repeatedly applying acetic acid (3–5% solution) and staining with Lugol’s iodine when indicated. Areas suspicious for squamous intraepithelial lesions (SIL) were biopsied.

Histopathology Biopsied lesions were graded by specialized pathologists. In two out of three clinics,

pathologists used p16 staining for AIN2 grading of biopsies, as recommended by the College of American Pathologists 16. The highest grade biopsy defined the overall diagnosis. Available left-over tissue of biopsies of participants with anal HSIL was reevaluated at DDL Diagnostic Laboratory, Rijswijk, the Netherlands. Reassessment entailed morphological examination of a newly cut Haematoxylin and Eosin (H/E) stained slide with supportive use of p16, by two specialized pathologists. HPV-types in the lesion were determined by genotyping using the SPF10-PCR DEIA/LiPA25 system (version 1) on the whole tissue section (WTS) and, if WTS analysis yielded more than one HPV-type, on specific lesions after laser capture micro-dissection 28,33,34.The single HPV-type found in a lesion was called the causative HPV-type.

Statistical analyses Differences in characteristics between men without anal HSIL (no anal HSIL/AIN1)

and with anal HSIL (AIN2/AIN3) were assessed by using Chi-squared or Fisher’s exact tests for categorical variables and Wilcoxon rank-sum tests for continuous variables.

Prediction of anal HSIL Analyses on prediction of anal HSIL were conducted on two levels: patient-level and

causative HPV-type-level (Figure 1). For the patient-level analyses a dataset was created of one record per participant

with the outcome anal HSIL (regardless of causative type) vs. no anal HSIL. Univariable and multivariable logistic regression was conducted to study the predictive power of each biomarker for anal HSIL diagnosis. Variables associated at P<0.2 in the univariable model were assessed in multivariable analysis, using a backward selection method.

For the causative HPV-type-level analyses a dataset was created of one record per HPV-type per patient. Each record included a value for HSIL: 1 if the patient had HSIL caused by that particular type and 0 if the patient did not have HSIL caused by that particular HPV-type. Of some lesions the HSIL diagnosis could not be confirmed (Figure 1), because no left-over tissue was available, or no HSIL was detected in the left-over material of the lesion. As these HSIL could not be assigned to an HPV-type, these men were excluded from the analyses. Univariable and multivariable logistic regression analyses using generalized estimating equations (GEE) was done to determine the association of type-specific biomarkers with type-specific anal HSIL diagnosis. GEE accounted for multiple observations (different HPV-types) within the same participant. The main dataset consisted of data on HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58. As there were 193 men in this study, this dataset contained 1737 records. Biomarkers associated in univariable regression at P<0.2 were included in multivariable logistic regression using GEE. A backward selection procedure was done to obtain a parsimonious model.

In both types of analyses age, number of years living with viral suppression 20 and time between last H2M visit and HRA were included as potential determinants. If determinants were found in any of the multivariable analyses, accuracy of the determinant as a predictor was assessed using a receiver operator curve (ROC) to estimate the area under the ROC curve (AUC).

Sensitivity analyses on both patient-level and causative HPV-type-level were done by using AIN2 vs. no anal HSIL, and AIN3 vs. no anal HSIL as outcome. Further sensitivity analyses on HPV-type-level were done on datasets with different combinations of HPV-types: (1) HPV-types 16 and 18; (2) HPV-types 16, 18, 31, 33, 45,

128

Predictors of anal HSIL

52, 58; (3) HPV-types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59; (4) HPV-types 6 and 11. Also two other definitions of persistence were tested, having: (1) at least four positive anal HPV samples; (2) at least two consecutive positive anal samples. Also two definitions of single positive visits were assessed: (1) at least one positive visit; (2) positive at the last H2M visit. Finally, sensitivity analyses were done on hrHPV viral load to assess the association between anal HSIL and hrHPV viral load as a continuous variable. For this, standardized Z-scores of the log-transformed type-specific anal hrHPV viral load were used to account for natural differences in hrHPV viral load between different hrHPV-types. The association of anal hrHPV viral load with anal HSIL was assessed using restricted cubic splines, with knots at the 20th, 40th, 60th, and 80th percentiles.

Results

In total, 193 HIV-positive MSM were included in this analysis, of whom 50 (26%) had anal HSIL based on the histopathological diagnosis of the clinic’s own pathology department (Figure 1). The mean age was 50 years (standard deviation (SD) 10). Nearly all MSM were on cART (n=188; 97%), the mean CD4 T-cell count was 681 cells/µl (SD 250) and the mean nadir CD4 T-cell count was 245 cells/µl (SD 134). Educational level was the only characteristic in which a borderline significant difference was found between participants with and without anal HSIL (p=0.05) (Table 1).

Anal HPV persistence of the two low-risk (6, 11) and seven hrHPV-types (16, 18, 31,

33, 45, 52, 58) was common, with persistence for HPV16 in 19% of study participants. The highest proportion persistence was found for HPV6 (22%) (Table 2 & Supplementary table 2). The highest hrHPV viral load was found for HPV16, with a median 57.4 DNA copies/human cell (interquartile range (IQR): 8.7-586.2) (Table 2). L1 seropositivity at last H2M visit was common, with 37 participants (19%) being HPV16 seropositive. E1, E2, E6 and E7 seropositivity was rare. HPV16 seropositivity was 3% for E1, 1% for E2, 3% for E6, and 1% for E7 (Table 2).

Patient-level results In univariable logistic regression analyses on patient-level, age, years living with

viral suppression and time between last H2M visit and HRA were not significantly associated with anal HSIL (Table 3). Of the assessed biomarkers, only HPV16 and HPV35 persistence were significantly associated with anal HSIL (odds ratio (OR) 2.46,

Tabl

e 2.

Fre

quen

cies

of H

PV ty

pe-s

peci

fic b

iom

arke

rs a

nd c

ausa

tive

HPV

type

s of a

nal H

SIL

amon

g 19

3 HI

V-po

sitiv

e M

SM

HPV6

HP

V11

HPV1

6 HP

V18

HPV3

1 HP

V33

HPV4

5 HP

V52

HPV5

8

N

n %

n

%

n %

n

%

n %

n

%

n %

n

%

n %

An

al H

PV p

ersis

tenc

e

At le

ast 4

pos

itive

vi

sits

155

30

17%

13

7%

23

13

%

11

6%

19

11%

9

5%

7 4%

13

7%

3

2%

3 po

sitiv

e vi

sits *

16

2 41

22

%

18

10%

34

19

%

14

8%

29

16%

16

9%

18

10

%

28

15%

6

3%

2 co

nsec

utiv

e po

sitiv

e vi

sits

167

46

24%

24

13

%

48

25%

22

12

%

39

20%

21

11

%

21

11%

42

22

%

10

5%

At le

ast o

nce

posit

ive

169

65

34%

33

17

%

71

37%

46

24

%

76

40%

39

20

%

43

23%

77

40

%

22

12%

Posit

ive

in la

st v

isit

169

36

19%

20

10

%

38

20%

20

10

%

34

18%

18

9%

24

13

%

28

15%

9

5%

Anal

HPV

Vira

l loa

d in

copi

es/h

uman

cell

**

med

ian/

IQR

-

-

- -

57

(9-5

86)

15

(0.1

-233

) 1

5 (2

-283

) 7

(0

.1-2

3)

8

(1-1

19)

1

(0.6

-73)

24

(3

-467

) N

o in

fect

ion

16

0 84

%

176

92%

17

1 90

%

177

93%

16

9 88

%

174

91%

18

3 96

%

≤ th

e m

edia

n

15

8%

7 4%

10

5%

7

4%

11

6%

9 5%

4

2%

> th

e m

edia

n

16

8%

8 4%

10

5%

7

4%

11

6%

8 4%

4

2%

HPV

Sero

posit

ivity

**

L1

19

3 47

24

%

36

19%

37

19

%

17

9%

13

7%

6 3%

33

17

%

8 4%

15

8%

E1

19

3 -

- -

- 6

3%

7 4%

-

- -

- -

- -

- -

- E2

19

3 -

-

- -

2 1%

3

2%

-

- -

-

-

- -

-

-

- E6

19

3 0

0%

1 0.

5%

5 3%

0

0%

3 2%

1

0.5%

0

0%

0 0%

1

0.5%

E7

19

3 1

0.5%

1

0.5%

2

1%

0 0%

1

0.5%

6

3%

0 0%

1

0.5%

1

0.5%

HP

V ty

pe in

AIN

lesio

ns *

**

AIN

2

3 2%

0

0%

5 3%

1

0.5%

3

2%

1 0.

5%

1 0.

5%

1 0.

5%

2 1%

AIN

3

2 1%

0

0%

3 2%

1

0.5%

1

0.5%

1

0.5%

1

0.5%

0

0%

0 0%

Ab

brev

iatio

ns: A

bbre

viat

ions

: HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

–hig

h-gr

ade

squa

mou

s int

raep

ithel

ial l

esio

n, M

SM –

men

who

hav

e se

x w

ith m

en, H

2M

stud

y –

HPV

and

HIV

in M

SM st

udy,

AIN

– a

nal i

ntra

epith

elia

l neo

plas

ia -

IQR

– in

terq

uart

ile ra

nge.

- :

not

mea

sure

d; *

3 p

ositi

ve v

isits

with

max

. 1 n

egat

ive

visit

in b

etw

een;

**

Mea

sure

d at

last

H2M

visi

t: fo

r mos

t of t

he m

en th

is w

as v

isit #

5, b

ut fo

r som

e th

is w

as v

isit #

4, o

r #3,

or #

2; *

** a

s es

tabl

ished

by

who

le ti

ssue

sect

ion

geno

typi

ng a

nd la

ser c

aptu

re m

icro

diss

ectio

n if

mor

e th

an o

ne H

PV ty

pe w

as fo

und

in w

hole

tiss

ue se

ctio

n.

129

6

52, 58; (3) HPV-types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59; (4) HPV-types 6 and 11. Also two other definitions of persistence were tested, having: (1) at least four positive anal HPV samples; (2) at least two consecutive positive anal samples. Also two definitions of single positive visits were assessed: (1) at least one positive visit; (2) positive at the last H2M visit. Finally, sensitivity analyses were done on hrHPV viral load to assess the association between anal HSIL and hrHPV viral load as a continuous variable. For this, standardized Z-scores of the log-transformed type-specific anal hrHPV viral load were used to account for natural differences in hrHPV viral load between different hrHPV-types. The association of anal hrHPV viral load with anal HSIL was assessed using restricted cubic splines, with knots at the 20th, 40th, 60th, and 80th percentiles.

Results

In total, 193 HIV-positive MSM were included in this analysis, of whom 50 (26%) had anal HSIL based on the histopathological diagnosis of the clinic’s own pathology department (Figure 1). The mean age was 50 years (standard deviation (SD) 10). Nearly all MSM were on cART (n=188; 97%), the mean CD4 T-cell count was 681 cells/µl (SD 250) and the mean nadir CD4 T-cell count was 245 cells/µl (SD 134). Educational level was the only characteristic in which a borderline significant difference was found between participants with and without anal HSIL (p=0.05) (Table 1).

Anal HPV persistence of the two low-risk (6, 11) and seven hrHPV-types (16, 18, 31,

33, 45, 52, 58) was common, with persistence for HPV16 in 19% of study participants. The highest proportion persistence was found for HPV6 (22%) (Table 2 & Supplementary table 2). The highest hrHPV viral load was found for HPV16, with a median 57.4 DNA copies/human cell (interquartile range (IQR): 8.7-586.2) (Table 2). L1 seropositivity at last H2M visit was common, with 37 participants (19%) being HPV16 seropositive. E1, E2, E6 and E7 seropositivity was rare. HPV16 seropositivity was 3% for E1, 1% for E2, 3% for E6, and 1% for E7 (Table 2).

Patient-level results In univariable logistic regression analyses on patient-level, age, years living with

viral suppression and time between last H2M visit and HRA were not significantly associated with anal HSIL (Table 3). Of the assessed biomarkers, only HPV16 and HPV35 persistence were significantly associated with anal HSIL (odds ratio (OR) 2.46,

Tabl

e 2.

Fre

quen

cies

of H

PV ty

pe-s

peci

fic b

iom

arke

rs a

nd c

ausa

tive

HPV

type

s of a

nal H

SIL

amon

g 19

3 HI

V-po

sitiv

e M

SM

HPV6

HP

V11

HPV1

6 HP

V18

HPV3

1 HP

V33

HPV4

5 HP

V52

HPV5

8

N

n %

n

%

n %

n

%

n %

n

%

n %

n

%

n %

An

al H

PV p

ersis

tenc

e

At le

ast 4

pos

itive

vi

sits

155

30

17%

13

7%

23

13

%

11

6%

19

11%

9

5%

7 4%

13

7%

3

2%

3 po

sitiv

e vi

sits *

16

2 41

22

%

18

10%

34

19

%

14

8%

29

16%

16

9%

18

10

%

28

15%

6

3%

2 co

nsec

utiv

e po

sitiv

e vi

sits

167

46

24%

24

13

%

48

25%

22

12

%

39

20%

21

11

%

21

11%

42

22

%

10

5%

At le

ast o

nce

posit

ive

169

65

34%

33

17

%

71

37%

46

24

%

76

40%

39

20

%

43

23%

77

40

%

22

12%

Posit

ive

in la

st v

isit

169

36

19%

20

10

%

38

20%

20

10

%

34

18%

18

9%

24

13

%

28

15%

9

5%

Anal

HPV

Vira

l loa

d in

copi

es/h

uman

cell

**

med

ian/

IQR

-

-

- -

57

(9-5

86)

15

(0.1

-233

) 1

5 (2

-283

) 7

(0

.1-2

3)

8

(1-1

19)

1

(0.6

-73)

24

(3

-467

) N

o in

fect

ion

16

0 84

%

176

92%

17

1 90

%

177

93%

16

9 88

%

174

91%

18

3 96

%

≤ th

e m

edia

n

15

8%

7 4%

10

5%

7

4%

11

6%

9 5%

4

2%

> th

e m

edia

n

16

8%

8 4%

10

5%

7

4%

11

6%

8 4%

4

2%

HPV

Sero

posit

ivity

**

L1

19

3 47

24

%

36

19%

37

19

%

17

9%

13

7%

6 3%

33

17

%

8 4%

15

8%

E1

19

3 -

- -

- 6

3%

7 4%

-

- -

- -

- -

- -

- E2

19

3 -

-

- -

2 1%

3

2%

-

- -

-

-

- -

-

-

- E6

19

3 0

0%

1 0.

5%

5 3%

0

0%

3 2%

1

0.5%

0

0%

0 0%

1

0.5%

E7

19

3 1

0.5%

1

0.5%

2

1%

0 0%

1

0.5%

6

3%

0 0%

1

0.5%

1

0.5%

HP

V ty

pe in

AIN

lesio

ns *

**

AIN

2

3 2%

0

0%

5 3%

1

0.5%

3

2%

1 0.

5%

1 0.

5%

1 0.

5%

2 1%

AIN

3

2 1%

0

0%

3 2%

1

0.5%

1

0.5%

1

0.5%

1

0.5%

0

0%

0 0%

Ab

brev

iatio

ns: A

bbre

viat

ions

: HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

–hig

h-gr

ade

squa

mou

s int

raep

ithel

ial l

esio

n, M

SM –

men

who

hav

e se

x w

ith m

en, H

2M

stud

y –

HPV

and

HIV

in M

SM st

udy,

AIN

– a

nal i

ntra

epith

elia

l neo

plas

ia -

IQR

– in

terq

uart

ile ra

nge.

- :

not

mea

sure

d; *

3 p

ositi

ve v

isits

with

max

. 1 n

egat

ive

visit

in b

etw

een;

**

Mea

sure

d at

last

H2M

visi

t: fo

r mos

t of t

he m

en th

is w

as v

isit #

5, b

ut fo

r som

e th

is w

as v

isit #

4, o

r #3,

or #

2; *

** a

s es

tabl

ished

by

who

le ti

ssue

sect

ion

geno

typi

ng a

nd la

ser c

aptu

re m

icro

diss

ectio

n if

mor

e th

an o

ne H

PV ty

pe w

as fo

und

in w

hole

tiss

ue se

ctio

n.

130

Predictors of anal HSIL

95% confidence interval (CI) 1.12-5.39, and OR 3.28, 95%CI 1.16-9.32, respectively). Neither anal hrHPV viral load nor any of the serological markers were associated with anal HSIL. Univariable associations were largely the same using AIN2 or AIN3 as outcomes, instead of HSIL (Table 3). In multivariable logistic regression, anal HPV16 and HPV35 persistence remained significantly associated with anal HSIL (adjusted OR 2.41, 95%CI 1.09-5.35, and adjusted OR 3.19, 95%CI 1.10-9.23, respectively). In multivariable logistic regression with AIN2 as outcome, anal HPV16 persistence and HPV18 L1 seropositivity were significantly associated with the outcome. In multivariable logistic regression with AIN3 as outcome, HPV33 and HPV35 persistence were significantly associated with the outcome (Table 4). Anal HPV16 persistence was a very weak predictor of anal HSIL, with an area under the ROC-curve of 0.58. Sensitivity analyses with different definitions of persistence gave similar results (data not shown).

HPV-type-level results In the 50 patients with HSIL, HSILs could be attributed to a causative HPV-type in 26

patients, including 23 AIN2 lesions and 12 AIN3 lesions. In the other 24 patients HSIL was not diagnosed in the left-over material at the second histopathological examination. Characteristics of this study population are provided in Supplementary table 3.

Including HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58, in univariable logistic regression using GEE, age, years living with viral suppression and time between last H2M visit and HRA were not significantly associated with anal HSIL with known causative HPV-type (Supplementary table 4). Neither anal hrHPV viral load nor any of the serological markers were associated with anal HSIL by the concordant HPV-type. Type-specific anal HPV persistence was the only biomarker significantly associated with anal HSIL caused by the concordant HPV-type (OR 14.68, 95%CI 6.16-35.02, P<0.001) (Figure 2). The area under the ROC curve for the prediction of type-specific anal HSIL by type-specific persistence was 0.60 (data not shown).

In the datasets including different combinations of HPV-types (including HPV-types 16 and 18; HPV-types 16, 18, 31, 33, 45, 52, 58; HPV-types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59), anal HPV persistence remained the only significantly associated biomarker of anal HSIL caused by the concordant HPV-type (Figure 2). The associations between anal HPV persistence, anal hrHPV viral load and seropositivity for L1, E6, E7,

E1, E2 remained largely the same when AIN2 or AIN3 were used as outcomes (Supplementary figure 1). However, type-specific anal HPV persistence was no longer significantly associated when AIN3 was used as outcome.

As the persistence definition is arbitrary, we assessed the effect of different definitions of persistence on the association with anal HSIL: at least 4 positive visits, 3 positive visits with a maximum of one negative visit in between (main definition used), 2 consecutive positive visits, at least once positive, positive at last H2M visit. We assessed all definitions in the four datasets including different combinations of HPV-types, for all three outcome measures (HSIL vs no HSIL, AIN2 vs no HSIL, AIN3 vs no HSIL). The association between anal HPV persistence and anal HSIL caused by the concordant HPV-type remained strong for all definitions of persistence, irrespective of the included HPV-types (Figure 3). The strength of the association between anal HPV persistence and type-concordant anal HSIL diagnosis could not be assessed when only HPV6 and 11 were included, due to limited power. However, all participants with anal HSIL caused by HPV6 or 11 had a persistent anal HPV infection of that type. All associations with persistence remained largely the same when AIN2 or AIN3 were used as outcome (data not shown). Assessment of anal hrHPV viral load as a continuous variable showed no significant association (P=0.3) with anal HSIL caused by the concordant HPV-type (Supplementary figure 2).

Discussion In this study we assessed associations of HPV related biomarkers and anal HSIL

among HIV-positive MSM in Amsterdam, the Netherlands. Of the biomarkers anal HPV persistence, anal hrHPV viral load, and seropositivity for L1 and various E antigens, only anal HPV16 and HPV35 persistence were significantly associated on a patient level. On HPV-type level, anal HPV persistence was the only biomarker significantly associated with anal HSIL caused by the concordant HPV-type. The discriminatory power, expressed as the AUC of the ROC, of persistence as biomarker for anal HSIL was too weak to use persistence as screening tool for anal HSIL screening in HIV-positive MSM.

Persistence of anal HPV infection is considered a requirement for development of anal HSIL and eventually progression to anal cancer 15,35,36. To our knowledge, the use of anal HPV persistence as a biomarker for the presence of anal HSIL has not been studied before.

131

Predictors of anal HSIL

6

95% confidence interval (CI) 1.12-5.39, and OR 3.28, 95%CI 1.16-9.32, respectively). Neither anal hrHPV viral load nor any of the serological markers were associated with anal HSIL. Univariable associations were largely the same using AIN2 or AIN3 as outcomes, instead of HSIL (Table 3). In multivariable logistic regression, anal HPV16 and HPV35 persistence remained significantly associated with anal HSIL (adjusted OR 2.41, 95%CI 1.09-5.35, and adjusted OR 3.19, 95%CI 1.10-9.23, respectively). In multivariable logistic regression with AIN2 as outcome, anal HPV16 persistence and HPV18 L1 seropositivity were significantly associated with the outcome. In multivariable logistic regression with AIN3 as outcome, HPV33 and HPV35 persistence were significantly associated with the outcome (Table 4). Anal HPV16 persistence was a very weak predictor of anal HSIL, with an area under the ROC-curve of 0.58. Sensitivity analyses with different definitions of persistence gave similar results (data not shown).

HPV-type-level results In the 50 patients with HSIL, HSILs could be attributed to a causative HPV-type in 26

patients, including 23 AIN2 lesions and 12 AIN3 lesions. In the other 24 patients HSIL was not diagnosed in the left-over material at the second histopathological examination. Characteristics of this study population are provided in Supplementary table 3.

Including HPV-types 6, 11, 16, 18, 31, 33, 45, 52, 58, in univariable logistic regression using GEE, age, years living with viral suppression and time between last H2M visit and HRA were not significantly associated with anal HSIL with known causative HPV-type (Supplementary table 4). Neither anal hrHPV viral load nor any of the serological markers were associated with anal HSIL by the concordant HPV-type. Type-specific anal HPV persistence was the only biomarker significantly associated with anal HSIL caused by the concordant HPV-type (OR 14.68, 95%CI 6.16-35.02, P<0.001) (Figure 2). The area under the ROC curve for the prediction of type-specific anal HSIL by type-specific persistence was 0.60 (data not shown).

In the datasets including different combinations of HPV-types (including HPV-types 16 and 18; HPV-types 16, 18, 31, 33, 45, 52, 58; HPV-types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59), anal HPV persistence remained the only significantly associated biomarker of anal HSIL caused by the concordant HPV-type (Figure 2). The associations between anal HPV persistence, anal hrHPV viral load and seropositivity for L1, E6, E7,

E1, E2 remained largely the same when AIN2 or AIN3 were used as outcomes (Supplementary figure 1). However, type-specific anal HPV persistence was no longer significantly associated when AIN3 was used as outcome.

As the persistence definition is arbitrary, we assessed the effect of different definitions of persistence on the association with anal HSIL: at least 4 positive visits, 3 positive visits with a maximum of one negative visit in between (main definition used), 2 consecutive positive visits, at least once positive, positive at last H2M visit. We assessed all definitions in the four datasets including different combinations of HPV-types, for all three outcome measures (HSIL vs no HSIL, AIN2 vs no HSIL, AIN3 vs no HSIL). The association between anal HPV persistence and anal HSIL caused by the concordant HPV-type remained strong for all definitions of persistence, irrespective of the included HPV-types (Figure 3). The strength of the association between anal HPV persistence and type-concordant anal HSIL diagnosis could not be assessed when only HPV6 and 11 were included, due to limited power. However, all participants with anal HSIL caused by HPV6 or 11 had a persistent anal HPV infection of that type. All associations with persistence remained largely the same when AIN2 or AIN3 were used as outcome (data not shown). Assessment of anal hrHPV viral load as a continuous variable showed no significant association (P=0.3) with anal HSIL caused by the concordant HPV-type (Supplementary figure 2).

Discussion In this study we assessed associations of HPV related biomarkers and anal HSIL

among HIV-positive MSM in Amsterdam, the Netherlands. Of the biomarkers anal HPV persistence, anal hrHPV viral load, and seropositivity for L1 and various E antigens, only anal HPV16 and HPV35 persistence were significantly associated on a patient level. On HPV-type level, anal HPV persistence was the only biomarker significantly associated with anal HSIL caused by the concordant HPV-type. The discriminatory power, expressed as the AUC of the ROC, of persistence as biomarker for anal HSIL was too weak to use persistence as screening tool for anal HSIL screening in HIV-positive MSM.

Persistence of anal HPV infection is considered a requirement for development of anal HSIL and eventually progression to anal cancer 15,35,36. To our knowledge, the use of anal HPV persistence as a biomarker for the presence of anal HSIL has not been studied before.

132

Predictors of anal HSIL

Tabl

e 3

- Uni

varia

ble

logi

stic

regr

essi

on o

f det

erm

inan

ts o

f ana

l HSI

L on

pat

ient

-leve

l.

Uni

varia

ble

logi

stic

re

gres

sion

- HSI

L vs

. no

HSIL

(N=1

93)

Uni

varia

ble

logi

stic

re

gres

sion

- AIN

2 vs

. no

HSIL

(N=1

68)

Uni

varia

ble

logi

stic

re

gres

sion

- AIN

3 vs

. no

HSIL

(N=1

68)

OR

(95%

CI)

P O

R (9

5% C

I) P

OR

(95%

CI)

P De

mog

raph

ic va

riabl

es

Ag

e

0.

6

0.

2

0.

8 ≤4

4 ye

ars

ref

ref

ref

45-5

4 ye

ars

0.70

(0

.32-

1.54

)

0.40

(0

.14-

1.11

)

1.30

(0

.43-

3.97

)

≥55

year

s 0.

64

(0.2

7-1.

50)

0.

48

(0.1

6-1.

38)

0.

95

(0.2

8-3.

27)

HI

V-re

late

d va

riabl

es

Ye

ars l

ivin

g w

ith v

iral s

uppr

essio

n §

0.4

0.09

0.

2 <

3 ye

ar Ϯ

re

f

re

f

re

f

3-

10 y

ears

0.

93

(0.3

7-2.

30)

0.

46

(0.1

6-1.

29)

4.

67

(0.5

8-37

.65)

>10

year

s 0.

60

(0.2

3-1.

58)

0.

27

(0.0

8-0.

88)

3.

26

(0.3

9-27

.30)

Tim

e va

riabl

es

Ti

me

betw

een

last

H2M

visi

t and

HRA

in y

ears

0.

2

0.

3

0.

6 <

1 ye

ar

ref

ref

ref

1-2

year

s 1.

89

(0.8

2-4.

34)

2.

12

(0.6

6-6.

80)

1.

70

(0.5

8-4.

99)

≥2

yea

rs

2.19

(0

.80-

6.03

)

2.69

(0

.69-

10.4

7)

1.

79

(0.4

7-6.

82)

An

al H

PV P

ersis

tenc

e* #

HPV6

per

siste

nce

1.98

(0

.94-

4.19

) 0.

08

2.54

(1

.00-

6.44

) 0.

06

1.49

(0

.54-

4.16

) 0.

5 HP

V11

pers

isten

ce

1.51

(0

.53-

4.29

) 0.

4 1.

48

(0.3

8-5.

68)

0.6

1.55

(0

.40-

5.98

) 0.

5 HP

V16

pers

isten

ce

2.46

(1

.12-

5.39

) 0.

03

4.14

(1

.62-

10.6

1)

0.00

4 1.

22

(0.3

8-3.

97)

0.7

HPV1

8 pe

rsist

ence

2.

34

(0.7

7-7.

14)

0.1

3.20

(0

.88-

11.6

2)

0.1

1.52

(0

.30-

7.67

) 0.

6 HP

V31

pers

isten

ce

0.91

(0

.36-

2.28

) 0.

8 0.

74

(0.2

0-2.

70)

0.6

1.09

(0

.34-

3.52

) 0.

9 HP

V33

pers

isten

ce

2.47

(0

.86-

7.05

) 0.

1 1.

28

(0.2

6-6.

34)

0.8

3.92

(1

.18-

13.0

1)

0.04

HP

V35

pers

isten

ce

3.28

(1

.16-

9.32

) 0.

03

2.29

(0

.56-

9.31

) 0.

3 4.

44

(1.3

1-15

.08)

0.

03

HPV3

9 pe

rsist

ence

0.

40

(0.0

5-3.

34)

0.3

0.80

(0

.09-

6.82

) 0.

8 n.

a.

n.a.

n.

a.

HPV4

5 pe

rsist

ence

1.

51

(0.5

3-4.

29)

0.4

1.48

(0

.38-

5.68

) 0.

6 1.

55

(0.4

0-5.

98)

0.5

HPV5

1 pe

rsist

ence

1.

78

(0.8

2-3.

88)

0.2

2.80

(1

.10-

7.14

) 0.

04

0.98

(0

.31-

3.15

) 0.

9

HPV5

2 pe

rsist

ence

1.

46

(0.6

1-3.

49)

0.4

2.05

(0

.72-

5.83

) 0.

2 0.

92

(0.2

5-3.

41)

0.9

HPV5

6 pe

rsist

ence

0.

61

(0.2

0-1.

90)

0.4

0.60

(0

.13-

2.75

) 0.

5 0.

62

(0.1

3-2.

89)

0.5

HPV5

8 pe

rsist

ence

0.

56

(0.0

6-4.

87)

0.6

1.13

(0

.13-

10.1

3)

0.9

n.a.

n.

a.

n.a.

HP

V59

pers

isten

ce

n.a.

n.

a.

n.a.

n.

a.

n.a.

n.

a.

n.a.

n.

a.

n.a.

An

al H

PV v

iral l

oad

HPV1

6 vi

ral l

oad

at la

st H

2M st

udy

visit

0.9

0.4

0.6

No

HPV

16 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(57.

4 DN

A co

pies

/ hum

an c

ell)

1.07

(0

.32-

3.56

)

0.55

(0

.07-

4.50

)

1.57

(0

.40-

6.11

)

> th

e m

edia

n

(57.

4 DN

A co

pies

/ hum

an c

ell)

1.34

(0

.44-

4.09

)

2.20

(0

.64-

7.59

)

0.52

(0

.06-

4.27

)

HPV1

8 vi

ral l

oad

at la

st H

2M st

udy

visit

0.08

0.

002

0.4

No

HPV

18 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(14.

5 DN

A co

pies

/ hum

an c

ell)

1.26

(0

.24-

6.71

)

n.a.

2.

35

(0.4

3-12

.83)

> th

e m

edia

n

(14.

5 DN

A co

pies

/ hum

an c

ell)

5.23

(1

.20-

22.8

0)

11

.25

(2.4

9-50

.74)

n.a.

n.

a.

HP

V31

vira

l loa

d at

last

H2M

stud

y vi

sit

0.8

0.8

No

HPV

31 in

fect

ion

ref

0

.9

ref

ref

≤ th

e m

edia

n

(14.

6 DN

A co

pies

/ hum

an c

ell)

1.20

(0

.30-

4.84

)

0.75

(0

.09-

6.38

)

1.71

(0

.33-

8.82

)

> th

e m

edia

n

(14.

6 DN

A co

pies

/ hum

an c

ell)

0.70

(0

.14-

3.42

)

n.a.

n.

a.

1.

50

(0.3

0-7.

56)

HP

V33

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

1

0.

4

0.

2 N

o HP

V 33

infe

ctio

n re

f

re

f

re

f

≤

the

med

ian

(7

.40

DNA

copi

es/ h

uman

cel

l) 4.

16

(0.8

9-19

.30)

4.06

(0

.64-

25.7

0)

4.

25

(0.6

7-26

.98)

> th

e m

edia

n

(7.4

0 DN

A co

pies

/ hum

an c

ell)

2.34

(0

.50-

10.8

6)

1.

52

(0.1

6-14

.26)

3.19

(0

.55-

18.5

2)

HP

V45

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

6

0.

4

0.

8 N

o HP

V 45

infe

ctio

n re

f

re

f

re

f

133

Predictors of anal HSIL

6

Tabl

e 3

- Uni

varia

ble

logi

stic

regr

essi

on o

f det

erm

inan

ts o

f ana

l HSI

L on

pat

ient

-leve

l.

Uni

varia

ble

logi

stic

re

gres

sion

- HSI

L vs

. no

HSIL

(N=1

93)

Uni

varia

ble

logi

stic

re

gres

sion

- AIN

2 vs

. no

HSIL

(N=1

68)

Uni

varia

ble

logi

stic

re

gres

sion

- AIN

3 vs

. no

HSIL

(N=1

68)

OR

(95%

CI)

P O

R (9

5% C

I) P

OR

(95%

CI)

P De

mog

raph

ic va

riabl

es

Ag

e

0.

6

0.

2

0.

8 ≤4

4 ye

ars

ref

ref

ref

45-5

4 ye

ars

0.70

(0

.32-

1.54

)

0.40

(0

.14-

1.11

)

1.30

(0

.43-

3.97

)

≥55

year

s 0.

64

(0.2

7-1.

50)

0.

48

(0.1

6-1.

38)

0.

95

(0.2

8-3.

27)

HI

V-re

late

d va

riabl

es

Ye

ars l

ivin

g w

ith v

iral s

uppr

essio

n §

0.4

0.09

0.

2 <

3 ye

ar Ϯ

re

f

re

f

re

f

3-

10 y

ears

0.

93

(0.3

7-2.

30)

0.

46

(0.1

6-1.

29)

4.

67

(0.5

8-37

.65)

>10

year

s 0.

60

(0.2

3-1.

58)

0.

27

(0.0

8-0.

88)

3.

26

(0.3

9-27

.30)

Tim

e va

riabl

es

Ti

me

betw

een

last

H2M

visi

t and

HRA

in y

ears

0.

2

0.

3

0.

6 <

1 ye

ar

ref

ref

ref

1-2

year

s 1.

89

(0.8

2-4.

34)

2.

12

(0.6

6-6.

80)

1.

70

(0.5

8-4.

99)

≥2

yea

rs

2.19

(0

.80-

6.03

)

2.69

(0

.69-

10.4

7)

1.

79

(0.4

7-6.

82)

An

al H

PV P

ersis

tenc

e* #

HPV6

per

siste

nce

1.98

(0

.94-

4.19

) 0.

08

2.54

(1

.00-

6.44

) 0.

06

1.49

(0

.54-

4.16

) 0.

5 HP

V11

pers

isten

ce

1.51

(0

.53-

4.29

) 0.

4 1.

48

(0.3

8-5.

68)

0.6

1.55

(0

.40-

5.98

) 0.

5 HP

V16

pers

isten

ce

2.46

(1

.12-

5.39

) 0.

03

4.14

(1

.62-

10.6

1)

0.00

4 1.

22

(0.3

8-3.

97)

0.7

HPV1

8 pe

rsist

ence

2.

34

(0.7

7-7.

14)

0.1

3.20

(0

.88-

11.6

2)

0.1

1.52

(0

.30-

7.67

) 0.

6 HP

V31

pers

isten

ce

0.91

(0

.36-

2.28

) 0.

8 0.

74

(0.2

0-2.

70)

0.6

1.09

(0

.34-

3.52

) 0.

9 HP

V33

pers

isten

ce

2.47

(0

.86-

7.05

) 0.

1 1.

28

(0.2

6-6.

34)

0.8

3.92

(1

.18-

13.0

1)

0.04

HP

V35

pers

isten

ce

3.28

(1

.16-

9.32

) 0.

03

2.29

(0

.56-

9.31

) 0.

3 4.

44

(1.3

1-15

.08)

0.

03

HPV3

9 pe

rsist

ence

0.

40

(0.0

5-3.

34)

0.3

0.80

(0

.09-

6.82

) 0.

8 n.

a.

n.a.

n.

a.

HPV4

5 pe

rsist

ence

1.

51

(0.5

3-4.

29)

0.4

1.48

(0

.38-

5.68

) 0.

6 1.

55

(0.4

0-5.

98)

0.5

HPV5

1 pe

rsist

ence

1.

78

(0.8

2-3.

88)

0.2

2.80

(1

.10-

7.14

) 0.

04

0.98

(0

.31-

3.15

) 0.

9

HPV5

2 pe

rsist

ence

1.

46

(0.6

1-3.

49)

0.4

2.05

(0

.72-

5.83

) 0.

2 0.

92

(0.2

5-3.

41)

0.9

HPV5

6 pe

rsist

ence

0.

61

(0.2

0-1.

90)

0.4

0.60

(0

.13-

2.75

) 0.

5 0.

62

(0.1

3-2.

89)

0.5

HPV5

8 pe

rsist

ence

0.

56

(0.0

6-4.

87)

0.6

1.13

(0

.13-

10.1

3)

0.9

n.a.

n.

a.

n.a.

HP

V59

pers

isten

ce

n.a.

n.

a.

n.a.

n.

a.

n.a.

n.

a.

n.a.

n.

a.

n.a.

An

al H

PV v

iral l

oad

HPV1

6 vi

ral l

oad

at la

st H

2M st

udy

visit

0.9

0.4

0.6

No

HPV

16 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(57.

4 DN

A co

pies

/ hum

an c

ell)

1.07

(0

.32-

3.56

)

0.55

(0

.07-

4.50

)

1.57

(0

.40-

6.11

)

> th

e m

edia

n

(57.

4 DN

A co

pies

/ hum

an c

ell)

1.34

(0

.44-

4.09

)

2.20

(0

.64-

7.59

)

0.52

(0

.06-

4.27

)

HPV1

8 vi

ral l

oad

at la

st H

2M st

udy

visit

0.08

0.

002

0.4

No

HPV

18 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(14.

5 DN

A co

pies

/ hum

an c

ell)

1.26

(0

.24-

6.71

)

n.a.

2.

35

(0.4

3-12

.83)

> th

e m

edia

n

(14.

5 DN

A co

pies

/ hum

an c

ell)

5.23

(1

.20-

22.8

0)

11

.25

(2.4

9-50

.74)

n.a.

n.

a.

HP

V31

vira

l loa

d at

last

H2M

stud

y vi

sit

0.8

0.8

No

HPV

31 in

fect

ion

ref

0

.9

ref

ref

≤ th

e m

edia

n

(14.

6 DN

A co

pies

/ hum

an c

ell)

1.20

(0

.30-

4.84

)

0.75

(0

.09-

6.38

)

1.71

(0

.33-

8.82

)

> th

e m

edia

n

(14.

6 DN

A co

pies

/ hum

an c

ell)

0.70

(0

.14-

3.42

)

n.a.

n.

a.

1.

50

(0.3

0-7.

56)

HP

V33

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

1

0.

4

0.

2 N

o HP

V 33

infe

ctio

n re

f

re

f

re

f

≤

the

med

ian

(7

.40

DNA

copi

es/ h

uman

cel

l) 4.

16

(0.8

9-19

.30)

4.06

(0

.64-

25.7

0)

4.

25

(0.6

7-26

.98)

> th

e m

edia

n

(7.4

0 DN

A co

pies

/ hum

an c

ell)

2.34

(0

.50-

10.8

6)

1.

52

(0.1

6-14

.26)

3.19

(0

.55-

18.5

2)

HP

V45

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

6

0.

4

0.

8 N

o HP

V 45

infe

ctio

n re

f

re

f

re

f

134

Predictors of anal HSIL

≤ th

e m

edia

n

(8.0

DN

A co

pies

/ hum

an c

ell)

1.62

(0

.45-

5.81

)

2.68

(0

.64-

11.2

2)

0.

74

(0.0

9-6.

33)

>

the

med

ian

(8

.0 D

NA

copi

es/ h

uman

cel

l) 0.

63

(0.1

3-3.

04)

1.

39

(0.2

8-6.

90)

n.

a.

n.a.

HPV5

2 vi

ral l

oad

at la

st H

2M st

udy

visit

0.6

0.9

0.5

No

HPV

52 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(1.3

DN

A co

pies

/ hum

an c

ell)

1.39

(0

.33-

5.79

)

0.93

(0

.11-

8.07

)

1.86

(0

.35-

9.76

)

> th

e m

edia

n

(1.3

DN

A co

pies

/ hum

an c

ell)

0.40

(0

.05-

3.32

)

0.80

(0

.09-

6.77

)

n.a.

n.

a.

HP

V58

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

3

0.

1

No

HPV

58 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(24.

2 DN

A co

pies

/ hum

an c

ell)

2.81

(0

.39-

20.5

2)

5.

87

(0.7

9-43

.77)

n.a.

n.

a.

>

the

med

ian

(2

4.2

DNA

copi

es/ h

uman

cel

l) n.

a.

n.a.

n.a.

n.

a.

n.

a.

n.a.

HPV

antib

odie

s

L1 se

ropo

sitiv

ity a

t las

t H2M

stud

y vi

sit

HPV6

0.

73

(0.3

1-1.

67)

0.5

0.87

(0

.26-

2.96

) 0.

8 0.

64

(0.2

2-1.

87)

0.4

HPV1

1 1.

16

(0.4

9-2.

76)

0.7

1.39

(0

.40-

4.79

) 0.

6 1.

02

(0.3

5-3.

03)

0.9

HPV1

6 1.

34

(0.5

7-3.

12)

0.5

1.39

(0

.40-

4.79

) 0.

6 1.

31

(0.4

7-3.

66)

0.6

HPV1

8 2.

68

(0.9

4-7.

65)

0.07

4.

44

(1.1

8-16

.75)

0.

04

1.75

(0

.44-

6.96

) 0.

4 HP

V31

1.56

(0

.45-

5.38

) 0.

5 2.

07

(0.4

0-10

.73)

0.

4 1.

24

(0.2

5-6.

19)

0.8

HPV3

3 1.

72

(0.3

0-9.

80)

0.6

2.25

(0

.23-

21.6

7)

0.5

1.39

(0

.15-

13.0

8)

0.8

HPV4

5 1.

10

(0.4

5-2.

70)

0.8

1.05

(0

.27-

4.04

) 0.

9 1.

13

(0.3

8-3.

36)

0.8

HPV5

2 1.

13

(0.2

2-5.

84)

0.9

3.19

(0

.58-

17.6

1)

0.2

n.a.

n.

a.

n.a.

HP

V58

0.83

(0

.22-

3.10

) 0.

8 1.

51

(0.3

0-7.

58)

0.6

0.43

(0

.05-

3.51

) 0.

4 E6

sero

posit

ivity

at l

ast H

2M st

udy

visit

**

HPV1

6 0.

84

(0.0

9-7.

72)

0.9

n.a.

n.

a.

n.a.

1.

39

(0.1

5-13

.08)

0.

8 E7

sero

posit

ivity

at l

ast H

2M st

udy

visit

**

HPV3

3 0.

66

(0.0

7-5.

86)

0.7

n.a.

n.

a.

n.a.

1.

10

(0.1

2-9.

94)

0.9

E1 se

ropo

sitiv

ity a

t las

t H2M

stud

y vi

sit**

HP

V18

1.36

(0

.25-

7.35

) 0.

7 1.

78

0.19

-16.

47

0.9

1.10

(0

.12-

9.94

) 0.

9 Ab

brev

iatio

ns: C

I - c

onfid

ence

inte

rval

; HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

– hi

gh-g

rade

squa

mou

s int

raep

ithel

ial l

esio

n,

LCM

– la

ser c

aptu

re m

icro

diss

ectio

n, M

SM –

men

who

hav

e se

x w

ith m

en, A

IN –

ana

l int

raep

ithel

ial n

eopl

asia

. *

Pers

isten

ce is

def

ined

as:

at l

east

thre

e an

al sa

mpl

es p

ositi

ve fo

r the

sam

e HP

V ty

pe. W

e al

low

ed o

ne n

egat

ive

sam

ple

betw

een

two

posit

ive

sam

ple

per

part

icip

ant a

s lon

g as

he

had

at le

ast t

hree

pos

itive

ana

l sam

ples

. **

Bas

ed o

n sm

all n

umbe

rs (c

ells

with

zero

obs

erva

tions

) som

e as

soci

atio

ns w

ere

not a

sses

sabl

e, a

nd th

ese

biom

arke

rs a

re n

ot re

port

ed in

this

tabl

e. T

his

conc

erns

: E6

sero

posit

ivity

for a

ll HP

V ty

pes o

ther

than

HPV

16, E

7 se

ropo

sitiv

ity fo

r all

HPV

type

oth

er th

an H

PV33

, E1

sero

posit

ivity

for H

PV16

, E2

sero

posit

ivity

fo

r bot

h HP

V16

and

HPV1

8.

# 2

part

icip

ants

wer

e no

t inc

lude

d in

the

univ

aria

ble

mod

els o

n pe

rsist

ence

, bec

ause

they

had

no

anal

sam

ple.

Ϯ

Part

icip

ants

who

nev

er h

ad a

n un

dete

ctab

le v

iral l

oad

are

incl

uded

in th

e ca

tego

ry <

3 ye

ars u

ndet

ecta

ble

vira

l loa

d.

§ vi

ral s

uppr

essio

n w

as d

efin

ed a

s hav

ing

a vi

ral l

oad

of <

200

copi

es/m

l in

test

s fro

m 1

Aug

ust 1

999

onw

ards

allo

win

g fo

r a o

netim

e bl

ip in

vira

l loa

d be

twee

n 20

0 an

d 40

0 co

pies

/ml.

For s

ampl

es te

sted

prio

r to

1 Au

gust

199

9 th

e cu

t-of

f of d

etec

tabi

lity

of th

e la

bora

tory

ass

ay th

at w

as u

sed

for t

hat s

ampl

e is

the

cut-

off f

or

vira

l sup

pres

sion.

135

Predictors of anal HSIL

6

≤ th

e m

edia

n

(8.0

DN

A co

pies

/ hum

an c

ell)

1.62

(0

.45-

5.81

)

2.68

(0

.64-

11.2

2)

0.

74

(0.0

9-6.

33)

>

the

med

ian

(8

.0 D

NA

copi

es/ h

uman

cel

l) 0.

63

(0.1

3-3.

04)

1.

39

(0.2

8-6.

90)

n.

a.

n.a.

HPV5

2 vi

ral l

oad

at la

st H

2M st

udy

visit

0.6

0.9

0.5

No

HPV

52 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(1.3

DN

A co

pies

/ hum

an c

ell)

1.39

(0

.33-

5.79

)

0.93

(0

.11-

8.07

)

1.86

(0

.35-

9.76

)

> th

e m

edia

n

(1.3

DN

A co

pies

/ hum

an c

ell)

0.40

(0

.05-

3.32

)

0.80

(0

.09-

6.77

)

n.a.

n.

a.

HP

V58

vira

l loa

d at

last

H2M

stud

y vi

sit

0.

3

0.

1

No

HPV

58 in

fect

ion

ref

ref

ref

≤ th

e m

edia

n

(24.

2 DN

A co

pies

/ hum

an c

ell)

2.81

(0

.39-

20.5

2)

5.

87

(0.7

9-43

.77)

n.a.

n.

a.

>

the

med

ian

(2

4.2

DNA

copi

es/ h

uman

cel

l) n.

a.

n.a.

n.a.

n.

a.

n.

a.

n.a.

HPV

antib

odie

s

L1 se

ropo

sitiv

ity a

t las

t H2M

stud

y vi

sit

HPV6

0.

73

(0.3

1-1.

67)

0.5

0.87

(0

.26-

2.96

) 0.

8 0.

64

(0.2

2-1.

87)

0.4

HPV1

1 1.

16

(0.4

9-2.

76)

0.7

1.39

(0

.40-

4.79

) 0.

6 1.

02

(0.3

5-3.

03)

0.9

HPV1

6 1.

34

(0.5

7-3.

12)

0.5

1.39

(0

.40-

4.79

) 0.

6 1.

31

(0.4

7-3.

66)

0.6

HPV1

8 2.

68

(0.9

4-7.

65)

0.07

4.

44

(1.1

8-16

.75)

0.

04

1.75

(0

.44-

6.96

) 0.

4 HP

V31

1.56

(0

.45-

5.38

) 0.

5 2.

07

(0.4

0-10

.73)

0.

4 1.

24

(0.2

5-6.

19)

0.8

HPV3

3 1.

72

(0.3

0-9.

80)

0.6

2.25

(0

.23-

21.6

7)

0.5

1.39

(0

.15-

13.0

8)

0.8

HPV4

5 1.

10

(0.4

5-2.

70)

0.8

1.05

(0

.27-

4.04

) 0.

9 1.

13

(0.3

8-3.

36)

0.8

HPV5

2 1.

13

(0.2

2-5.

84)

0.9

3.19

(0

.58-

17.6

1)

0.2

n.a.

n.

a.

n.a.

HP

V58

0.83

(0

.22-

3.10

) 0.

8 1.

51

(0.3

0-7.

58)

0.6

0.43

(0

.05-

3.51

) 0.

4 E6

sero

posit

ivity

at l

ast H

2M st

udy

visit

**

HPV1

6 0.

84

(0.0

9-7.

72)

0.9

n.a.

n.

a.

n.a.

1.

39

(0.1

5-13

.08)

0.

8 E7

sero

posit

ivity

at l

ast H

2M st

udy

visit

**

HPV3

3 0.

66

(0.0

7-5.

86)

0.7

n.a.

n.

a.

n.a.

1.

10

(0.1

2-9.

94)

0.9

E1 se

ropo

sitiv

ity a

t las

t H2M

stud

y vi

sit**

HP

V18

1.36

(0

.25-

7.35

) 0.

7 1.

78

0.19

-16.

47

0.9

1.10

(0

.12-

9.94

) 0.

9 Ab

brev

iatio

ns: C

I - c

onfid

ence

inte

rval

; HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

– hi

gh-g

rade

squa

mou

s int

raep

ithel

ial l

esio

n,

LCM

– la

ser c

aptu

re m

icro

diss

ectio

n, M

SM –

men

who

hav

e se

x w

ith m

en, A

IN –

ana

l int

raep

ithel

ial n

eopl

asia

. *

Pers

isten

ce is

def

ined

as:

at l

east

thre

e an

al sa

mpl

es p

ositi

ve fo

r the

sam

e HP

V ty

pe. W

e al

low

ed o

ne n

egat

ive

sam

ple

betw

een

two

posit

ive

sam

ple

per

part

icip

ant a

s lon

g as

he

had

at le

ast t

hree

pos

itive

ana

l sam

ples

. **

Bas

ed o

n sm

all n

umbe

rs (c

ells

with

zero

obs

erva

tions

) som

e as

soci

atio

ns w

ere

not a

sses

sabl

e, a

nd th

ese

biom

arke

rs a

re n

ot re

port

ed in

this

tabl

e. T

his

conc

erns

: E6

sero

posit

ivity

for a

ll HP

V ty

pes o

ther

than

HPV

16, E

7 se

ropo

sitiv

ity fo

r all

HPV

type

oth

er th

an H

PV33

, E1

sero

posit

ivity

for H

PV16

, E2

sero

posit

ivity

fo

r bot

h HP

V16

and

HPV1

8.

# 2

part

icip

ants

wer

e no

t inc

lude

d in

the

univ

aria

ble

mod

els o

n pe

rsist

ence

, bec

ause

they

had

no

anal

sam

ple.

Ϯ

Part

icip

ants

who

nev

er h

ad a

n un

dete

ctab

le v

iral l

oad

are

incl

uded

in th

e ca

tego

ry <

3 ye

ars u

ndet

ecta

ble

vira

l loa

d.

§ vi

ral s

uppr

essio

n w

as d

efin

ed a

s hav

ing

a vi

ral l

oad

of <

200

copi

es/m

l in

test

s fro

m 1

Aug

ust 1

999

onw

ards

allo

win

g fo

r a o

netim

e bl

ip in

vira

l loa

d be

twee

n 20

0 an

d 40

0 co

pies

/ml.

For s

ampl

es te

sted

prio

r to

1 Au

gust

199

9 th

e cu

t-of

f of d

etec

tabi

lity

of th

e la

bora

tory

ass

ay th

at w

as u

sed

for t

hat s

ampl

e is

the

cut-

off f

or

vira

l sup

pres

sion.

136

Predictors of anal HSIL

Tabl

e 4

- Mul

tivar

iabl

e lo

gist

ic re

gres

sion

on

patie

nt-le

vel o

f det

erm

inan

ts o

f ana

l HSI

L.

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - H

SIL

vs. n

o HS

IL

(N=1

83)

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - A

IN2

vs. n

o HS

IL

(N=1

58)

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - A

IN3

vs. n

o HS

IL

(N=1

58)

aO

R (9

5% C

I) P

aOR

(95%

CI)

P aO

R (9

5% C

I) P

HPV1

6 pe

rsist

ence

*#

2.41

(1

.09-

5.35

) 0.

03

3.99

(1

.18-

13.4

6)

0.03

HPV3

3 pe

rsist

ence

*#

4.29

(1

.25-

14.7

5)

0.02

HPV3

5 pe

rsist

ence

*#

3.19

(1

.10-

9.23

) 0.

03

4.

85

(1.3

9-16

.99)

0.

01

HPV1

8 L1

sero

posit

ivity

at l

ast H

2M v

isit

5.

28

(1.2

7-21

.93)

0.

02

Ab

brev

iatio

ns: C

I - c

onfid

ence

inte

rval

; HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

–hig

h-gr

ade

squa

mou

s int

raep

ithel

ial

lesio

n, M

SM –

men

who

hav

e se

x w

ith m

en, A

IN –

ana

l int

raep

ithel

ial n

eopl

asia

. *

Pers

isten

ce is

def

ined

as:

at l

east

thre

e an

al sa

mpl

es p

ositi

ve fo

r the

sam

e HP

V ty

pe. W

e al

low

ed o

ne n

egat

ive

sam

ple

betw

een

two

posit

ive

sam

ple

per

part

icip

ant a

s lon

g as

he

had

at le

ast t

hree

pos

itive

ana

l sam

ples

. #

10 p

artic

ipan

ts w

ere

not i

nclu

ded

in th

e m

ultiv

aria

ble

mod

els,

bec

ause

they

had

<3

H2M

visi

ts re

sulti

ng in

no

estim

atio

n fo

r per

siste

nce

acco

rdin

g to

th

is de

finiti

on, o

r the

y ha

d no

ana

l sam

ple.

The hrHPV viral load in anal swabs has previously been found to be a determinant for histologically proven anal HSIL among HIV-positive and among HIV-negative MSM 37. In the analysis where we categorized anal hrHPV viral load, the observed association with anal HSIL merely reflected anal hrHPV infection; when we assessed hrHPV viral load as a continuous variable no significant association was found. Thus, we could not confirm the previously found association between anal hrHPV viral load and anal HSIL. This might be explained by the inclusion of anal samples obtained at the last H2M visit. In some cases this was 1.3 years prior to HRA, while Jin and colleagues 37 assessed anal hrHPV viral load of the swab taken at the same consultation as the HRA was done. The rate of spontaneous regression of anal HSIL was 19.2/100 person-years in an Australian study among 574 HIV-positive MSM 38. Therefore, some anal HSIL lesions in our study might have regressed by the time HRA was performed.

All participants with anal HSIL caused by HPV6 or 11 had a persistent anal HPV infection of that type. Furthermore, the strength of the association between anal HPV persistence and anal HSIL was not affected by including or excluding HPV6 and 11 from analyses. HPV6 and 11 are known as low-risk HPV-types, but they have been previously found to cause anal HSIL 39, and also anal cancer 6,40,41. The progression rate to anal cancer of anal HSIL caused by low risk HPV-types is not clear.

HPV E6 seropositivity has been found to be predictive for HPV-induced oropharyngeal- 25,26, oropharynx- 42, and anal cancer 24, but we could not confirm this for anal HSIL.

Strengths and limitations This study is unique because it is the first longitudinal study able to assess the

association of a range of potential virological biomarkers with histologically proven anal HSIL. Moreover, in a selection of lesions we were able to determine the causative HPV-type, thus allowing to analyze data both on patient-level and on causative HPV-type-level. A limitation of this study is the interval between sample collection for the biomarkers and establishing the histopathological HSIL endpoint via HRA. However, when correcting for time between last H2M visit and HRA, results remained largely the same. Additionally, positivity of some biomarkers, like E6 or E7 antibody positivity, was relatively rare, resulting in the inability to assess the value of these biomarkers in a

137

Predictors of anal HSIL

6

Tabl

e 4

- Mul

tivar

iabl

e lo

gist

ic re

gres

sion

on

patie

nt-le

vel o

f det

erm

inan

ts o

f ana

l HSI

L.

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - H

SIL

vs. n

o HS

IL

(N=1

83)

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - A

IN2

vs. n

o HS

IL

(N=1

58)

Mul

tivar

iabl

e lo

gist

ic

regr

essio

n - A

IN3

vs. n

o HS

IL

(N=1

58)

aO

R (9

5% C

I) P

aOR

(95%

CI)

P aO

R (9

5% C

I) P

HPV1

6 pe

rsist

ence

*#

2.41

(1

.09-

5.35

) 0.

03

3.99

(1

.18-

13.4

6)

0.03

HPV3

3 pe

rsist

ence

*#

4.29

(1

.25-

14.7

5)

0.02

HPV3

5 pe

rsist

ence

*#

3.19

(1

.10-

9.23

) 0.

03

4.

85

(1.3

9-16

.99)

0.

01

HPV1

8 L1

sero

posit

ivity

at l

ast H

2M v

isit

5.

28

(1.2

7-21

.93)

0.

02

Ab

brev

iatio

ns: C

I - c

onfid

ence

inte

rval

; HIV

– h

uman

imm

unod

efic

ienc

y vi

rus;

HPV

– h

uman

pap

illom

aviru

s, H

SIL

–hig

h-gr

ade

squa

mou

s int

raep

ithel

ial

lesio

n, M

SM –

men

who

hav

e se

x w

ith m

en, A

IN –

ana

l int

raep

ithel

ial n

eopl

asia

. *

Pers

isten

ce is

def

ined

as:

at l

east

thre

e an

al sa

mpl

es p

ositi

ve fo

r the

sam

e HP

V ty

pe. W

e al

low

ed o

ne n

egat

ive

sam

ple

betw

een

two

posit

ive

sam

ple

per

part

icip

ant a

s lon

g as

he

had

at le

ast t

hree

pos

itive

ana

l sam

ples

. #

10 p

artic

ipan

ts w

ere

not i

nclu

ded

in th

e m

ultiv

aria

ble

mod

els,

bec

ause

they

had

<3

H2M

visi

ts re

sulti

ng in

no

estim

atio

n fo

r per

siste

nce

acco

rdin

g to

th

is de

finiti

on, o

r the

y ha

d no

ana

l sam

ple.

The hrHPV viral load in anal swabs has previously been found to be a determinant for histologically proven anal HSIL among HIV-positive and among HIV-negative MSM 37. In the analysis where we categorized anal hrHPV viral load, the observed association with anal HSIL merely reflected anal hrHPV infection; when we assessed hrHPV viral load as a continuous variable no significant association was found. Thus, we could not confirm the previously found association between anal hrHPV viral load and anal HSIL. This might be explained by the inclusion of anal samples obtained at the last H2M visit. In some cases this was 1.3 years prior to HRA, while Jin and colleagues 37 assessed anal hrHPV viral load of the swab taken at the same consultation as the HRA was done. The rate of spontaneous regression of anal HSIL was 19.2/100 person-years in an Australian study among 574 HIV-positive MSM 38. Therefore, some anal HSIL lesions in our study might have regressed by the time HRA was performed.

All participants with anal HSIL caused by HPV6 or 11 had a persistent anal HPV infection of that type. Furthermore, the strength of the association between anal HPV persistence and anal HSIL was not affected by including or excluding HPV6 and 11 from analyses. HPV6 and 11 are known as low-risk HPV-types, but they have been previously found to cause anal HSIL 39, and also anal cancer 6,40,41. The progression rate to anal cancer of anal HSIL caused by low risk HPV-types is not clear.

HPV E6 seropositivity has been found to be predictive for HPV-induced oropharyngeal- 25,26, oropharynx- 42, and anal cancer 24, but we could not confirm this for anal HSIL.

Strengths and limitations This study is unique because it is the first longitudinal study able to assess the

association of a range of potential virological biomarkers with histologically proven anal HSIL. Moreover, in a selection of lesions we were able to determine the causative HPV-type, thus allowing to analyze data both on patient-level and on causative HPV-type-level. A limitation of this study is the interval between sample collection for the biomarkers and establishing the histopathological HSIL endpoint via HRA. However, when correcting for time between last H2M visit and HRA, results remained largely the same. Additionally, positivity of some biomarkers, like E6 or E7 antibody positivity, was relatively rare, resulting in the inability to assess the value of these biomarkers in a

138

Predictors of anal HSIL

Figure 2. Associations between biomarkers and anal HSIL with known causative HPV type among 169 HIV-positive MSM in Amsterdam, the Netherlands (H2M2 study); results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. For E2, E6, E7 seropositivity, no odds ratio and 95% confidence intervals are shown because these biomarkers were unassessable due to small numbers. Furthermore, not all biomarkers were measured for all HPV types. Therefore not all biomarkers could be assessed in all datasets. * Persistence is defined as: at least three anal samples positive for the same HPV type. We allowed one negative sample between two positive samples per participant as long as there were at least three positive anal samples. HPV viral load at last H2M visit was only measured for HPV16 and HPV18. Seropositivity at last H2M visit is defined as seropositive above the standard cut-off at the last H2M study visit. L1, E6 and E7 antibodies were measured for HPV types 6, 11, 16, 18, 31, 33, 45, 52, and 58; antibodies against E1 and E2 were only measured for HPV types 16 and 18. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

Figure 3. Number of visits with HPV positive anal sample and its association with anal HSIL with known causative HPV type; results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. In the analyses of “At least 4 positive visits” 155 men were included. In the analysis of “3 positive visits” 162 men were included; in the analysis with “2 consecutive visits” 167 men were included; and in the analyses “At least one positive visit” and “Positive at last H2M visit” 169 men were included. * At least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

limited sample size. Finally, we were not able to identify a causal HPV type for some patients with anal HSIL. Conclusion

Anal HPV persistence is strongly associated with anal HSIL among HIV-positive MSM, compatible with the current theory on the etiogenesis of anal HSIL. Although

139

Predictors of anal HSIL

6

Figure 2. Associations between biomarkers and anal HSIL with known causative HPV type among 169 HIV-positive MSM in Amsterdam, the Netherlands (H2M2 study); results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. For E2, E6, E7 seropositivity, no odds ratio and 95% confidence intervals are shown because these biomarkers were unassessable due to small numbers. Furthermore, not all biomarkers were measured for all HPV types. Therefore not all biomarkers could be assessed in all datasets. * Persistence is defined as: at least three anal samples positive for the same HPV type. We allowed one negative sample between two positive samples per participant as long as there were at least three positive anal samples. HPV viral load at last H2M visit was only measured for HPV16 and HPV18. Seropositivity at last H2M visit is defined as seropositive above the standard cut-off at the last H2M study visit. L1, E6 and E7 antibodies were measured for HPV types 6, 11, 16, 18, 31, 33, 45, 52, and 58; antibodies against E1 and E2 were only measured for HPV types 16 and 18. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

Figure 3. Number of visits with HPV positive anal sample and its association with anal HSIL with known causative HPV type; results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. In the analyses of “At least 4 positive visits” 155 men were included. In the analysis of “3 positive visits” 162 men were included; in the analysis with “2 consecutive visits” 167 men were included; and in the analyses “At least one positive visit” and “Positive at last H2M visit” 169 men were included. * At least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

limited sample size. Finally, we were not able to identify a causal HPV type for some patients with anal HSIL. Conclusion

Anal HPV persistence is strongly associated with anal HSIL among HIV-positive MSM, compatible with the current theory on the etiogenesis of anal HSIL. Although

140

Predictors of anal HSIL

strongly associated, anal HPV persistence lacks discriminatory power and thus is not suitable as tool to select patients at high risk for anal HSIL for HRA screening.

Acknowledgements

The authors would like to thank Sofie Mooij and Titia Heijman for assisting in the design of the study and datacollection, Manon van Dijk for assistence in data-analyses, Wilma Vermeulen, Elske van Logchem, and Naomi van Marm for laboratory assays, and Hans-Erik Nobel, Olivier Richel, Karien Goosens, Aafien Hendriks, Irene Holtslag, Jacqueline Engelen, Jan Gert Geerdink, Aafke Bosma, Angelique Toonen, Bart Nanninga, and Marjolein Broekhuizen for performing HRA.

References 1. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human

papillomavirus and cervical cancer. J Clin Pathol. Apr 2002;55(4):244-265. 2. Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of

invasive cervical cancer worldwide. J Pathol. Sep 1999;189(1):12-19. 3. Nobbenhuis MA, Walboomers JM, Helmerhorst TJ, et al. Relation of human papillomavirus

status to cervical lesions and consequences for cervical-cancer screening: a prospective study. Lancet. Jul 3 1999;354(9172):20-25.

4. Depuydt CE, Thys S, Beert J, Jonckheere J, Salembier G, Bogers JJ. Linear viral load increase of a single HPV-type in women with multiple HPV infections predicts progression to cervical cancer. Int J Cancer. Nov 01 2016;139(9):2021-2032.

5. De Vuyst H, Clifford GM, Nascimento MC, Madeleine MM, Franceschi S. Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: A meta‐analysis. International Journal of Cancer. 2009;124(7):1626-1636.

6. Alemany L, Saunier M, Alvarado-Cabrero I, et al. Human papillomavirus DNA prevalence and type distribution in anal carcinomas worldwide. Int J Cancer. Jan 1 2015;136(1):98-107.

7. Partridge JM, Koutsky LA. Genital human papillomavirus infection in men. Lancet Infect Dis. Jan 2006;6(1):21-31.

8. Morbini P, Benazzo M. Human papillomavirus and head and neck carcinomas: focus on evidence in the babel of published data. Acta Otorhinolaryngol Ital. Aug 2016;36(4):249-258.

9. Rietbergen MM, Leemans CR, Bloemena E, et al. Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm. Int J Cancer. Apr 1 2013;132(7):1565-1571.

10. Machalek DA, Poynten M, Jin F, et al. Anal human papillomavirus infection and associated neoplastic lesions in men who have sex with men: a systematic review and meta-analysis. Lancet Oncol. May 2012;13(5):487-500.

11. Silverberg MJ, Lau B, Justice AC, et al. Risk of anal cancer in HIV-infected and HIV-uninfected individuals in North America. Clin Infect Dis. Apr 2012;54(7):1026-1034.

12. de Pokomandy A, Rouleau D, Ghattas G, et al. HAART and progression to high-grade anal intraepithelial neoplasia in men who have sex with men and are infected with HIV. Clin Infect Dis. May 2011;52(9):1174-1181.

13. Berry JM, Jay N, Cranston RD, et al. Progression of anal high-grade squamous intraepithelial lesions to invasive anal cancer among HIV-infected men who have sex with men. Int J Cancer. Mar 1 2014;134(5):1147-1155.

14. Nathan M, Singh N, Garrett N, Hickey N, Prevost T, Sheaff M. Performance of anal cytology in a clinical setting when measured against histology and high-resolution anoscopy findings. AIDS. Jan 28 2010;24(3):373-379.

15. Palefsky JM, Holly EA, Efirdc JT, et al. Anal intraepithelial neoplasia in the highly active antiretroviral therapy era among HIV-positive men who have sex with men. AIDS. Sep 2 2005;19(13):1407-1414.

16. Darragh TM, Colgan TJ, Cox JT, et al. The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions: background and consensus

Figure 3. Number of visits with HPV positive anal sample and its association with anal HSIL with known causative HPV type; results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. In the analyses of “At least 4 positive visits” 155 men were included. In the analysis of “3 positive visits” 162 men were included; in the analysis with “2 consecutive visits” 167 men were included; and in the analyses “At least one positive visit” and “Positive at last H2M visit” 169 men were included. * At least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

limited sample size. Finally, we were not able to identify a causal HPV type for some patients with anal HSIL. Conclusion

Anal HPV persistence is strongly associated with anal HSIL among HIV-positive MSM, compatible with the current theory on the etiogenesis of anal HSIL. Although

141

Predictors of anal HSIL

6

strongly associated, anal HPV persistence lacks discriminatory power and thus is not suitable as tool to select patients at high risk for anal HSIL for HRA screening.

Acknowledgements

The authors would like to thank Sofie Mooij and Titia Heijman for assisting in the design of the study and datacollection, Manon van Dijk for assistence in data-analyses, Wilma Vermeulen, Elske van Logchem, and Naomi van Marm for laboratory assays, and Hans-Erik Nobel, Olivier Richel, Karien Goosens, Aafien Hendriks, Irene Holtslag, Jacqueline Engelen, Jan Gert Geerdink, Aafke Bosma, Angelique Toonen, Bart Nanninga, and Marjolein Broekhuizen for performing HRA.

References 1. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. The causal relation between human

papillomavirus and cervical cancer. J Clin Pathol. Apr 2002;55(4):244-265. 2. Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of

invasive cervical cancer worldwide. J Pathol. Sep 1999;189(1):12-19. 3. Nobbenhuis MA, Walboomers JM, Helmerhorst TJ, et al. Relation of human papillomavirus

status to cervical lesions and consequences for cervical-cancer screening: a prospective study. Lancet. Jul 3 1999;354(9172):20-25.

4. Depuydt CE, Thys S, Beert J, Jonckheere J, Salembier G, Bogers JJ. Linear viral load increase of a single HPV-type in women with multiple HPV infections predicts progression to cervical cancer. Int J Cancer. Nov 01 2016;139(9):2021-2032.

5. De Vuyst H, Clifford GM, Nascimento MC, Madeleine MM, Franceschi S. Prevalence and type distribution of human papillomavirus in carcinoma and intraepithelial neoplasia of the vulva, vagina and anus: A meta‐analysis. International Journal of Cancer. 2009;124(7):1626-1636.

6. Alemany L, Saunier M, Alvarado-Cabrero I, et al. Human papillomavirus DNA prevalence and type distribution in anal carcinomas worldwide. Int J Cancer. Jan 1 2015;136(1):98-107.

7. Partridge JM, Koutsky LA. Genital human papillomavirus infection in men. Lancet Infect Dis. Jan 2006;6(1):21-31.

8. Morbini P, Benazzo M. Human papillomavirus and head and neck carcinomas: focus on evidence in the babel of published data. Acta Otorhinolaryngol Ital. Aug 2016;36(4):249-258.

9. Rietbergen MM, Leemans CR, Bloemena E, et al. Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm. Int J Cancer. Apr 1 2013;132(7):1565-1571.

10. Machalek DA, Poynten M, Jin F, et al. Anal human papillomavirus infection and associated neoplastic lesions in men who have sex with men: a systematic review and meta-analysis. Lancet Oncol. May 2012;13(5):487-500.

11. Silverberg MJ, Lau B, Justice AC, et al. Risk of anal cancer in HIV-infected and HIV-uninfected individuals in North America. Clin Infect Dis. Apr 2012;54(7):1026-1034.

12. de Pokomandy A, Rouleau D, Ghattas G, et al. HAART and progression to high-grade anal intraepithelial neoplasia in men who have sex with men and are infected with HIV. Clin Infect Dis. May 2011;52(9):1174-1181.

13. Berry JM, Jay N, Cranston RD, et al. Progression of anal high-grade squamous intraepithelial lesions to invasive anal cancer among HIV-infected men who have sex with men. Int J Cancer. Mar 1 2014;134(5):1147-1155.

14. Nathan M, Singh N, Garrett N, Hickey N, Prevost T, Sheaff M. Performance of anal cytology in a clinical setting when measured against histology and high-resolution anoscopy findings. AIDS. Jan 28 2010;24(3):373-379.

15. Palefsky JM, Holly EA, Efirdc JT, et al. Anal intraepithelial neoplasia in the highly active antiretroviral therapy era among HIV-positive men who have sex with men. AIDS. Sep 2 2005;19(13):1407-1414.

16. Darragh TM, Colgan TJ, Cox JT, et al. The Lower Anogenital Squamous Terminology Standardization Project for HPV-Associated Lesions: background and consensus

Figure 3. Number of visits with HPV positive anal sample and its association with anal HSIL with known causative HPV type; results from univariable logistic regression using GEE with different HPV types included. Odds ratios are displayed on a logarithmic scale. In the analyses of “At least 4 positive visits” 155 men were included. In the analysis of “3 positive visits” 162 men were included; in the analysis with “2 consecutive visits” 167 men were included; and in the analyses “At least one positive visit” and “Positive at last H2M visit” 169 men were included. * At least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Abbreviations: OR – odds ratio; HIV – human immunodeficiency virus; HPV – human papillomavirus, HSIL –high-grade squamous intraepithelial lesion, MSM – men who have sex with men, GEE – generalized estimating equations.

limited sample size. Finally, we were not able to identify a causal HPV type for some patients with anal HSIL. Conclusion

Anal HPV persistence is strongly associated with anal HSIL among HIV-positive MSM, compatible with the current theory on the etiogenesis of anal HSIL. Although

142

Predictors of anal HSIL

recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. J Low Genit Tract Dis. Jul 2012;16(3):205-242.

17. Cachay ER, Agmas W, Mathews WC. Relative accuracy of cervical and anal cytology for detection of high grade lesions by colposcope guided biopsy: a cut-point meta-analytic comparison. PLoS One. 2012;7(7):e38956.

18. Botes LP, Pett S, Carr A, et al. Anal cytological abnormalities are poor predictors of high-grade intraepithelial neoplasia amongst HIV-positive men who have sex with men. Sex Health. Mar 2013;10(1):9-17.

19. van Aar F, Mooij SH, van der Sande MA, et al. Anal and penile high-risk human papillomavirus prevalence in HIV-negative and HIV-infected MSM. AIDS. Nov 28 2013;27(18):2921-2931.

20. Siegenbeek van Heukelom ML, Marra E, de Vries HJC, Schim van der Loeff MF, Prins JM. Risk factors for anal high-grade squamous intraepithelial lesions in HIV-positive MSM: is targeted screening possible? AIDS. Oct 23 2017;31(16):2295-2301.

21. Melo VH, Guimaraes MD, Rocha GM, et al. Prevalence and risk factors associated with anal intraepithelial neoplasia among HIV-positive men in Brazil. J Low Genit Tract Dis. Apr 2014;18(2):128-135.

22. Alvarez J, de Pokomandy A, Rouleau D, et al. Episomal and integrated human papillomavirus type 16 loads and anal intraepithelial neoplasia in HIV-seropositive men. AIDS. Sep 24 2010;24(15):2355-2363.

23. Libois A, Feoli F, Nkuize M, et al. Prolonged antiretroviral therapy is associated with fewer anal high-grade squamous intraepithelial lesions in HIV-positive MSM in a cross-sectional study. Sex Transm Infect. Feb 2017;93(1):15-17.

24. Kreimer AR, Brennan P, Lang Kuhs KA, et al. Human papillomavirus antibodies and future risk of anogenital cancer: a nested case-control study in the European prospective investigation into cancer and nutrition study. J Clin Oncol. Mar 10 2015;33(8):877-884.

25. Kreimer AR, Johansson M, Waterboer T, et al. Evaluation of human papillomavirus antibodies and risk of subsequent head and neck cancer. J Clin Oncol. Jul 20 2013;31(21):2708-2715.

26. Kreimer AR, Johansson M, Yanik EL, et al. Kinetics of the Human Papillomavirus Type 16 E6 Antibody Response Prior to Oropharyngeal Cancer. J Natl Cancer Inst. Aug 1 2017;109(8).

27. van Sighem AI, van de Wiel MA, Ghani AC, et al. Mortality and progression to AIDS after starting highly active antiretroviral therapy. AIDS. Oct 17 2003;17(15):2227-2236.

28. Molijn A, Kleter B, Quint W, van Doorn LJ. Molecular diagnosis of human papillomavirus (HPV) infections. J Clin Virol. Mar 2005;32 Suppl 1:S43-51.

29. Van der Weele P, Breeuwsma M, Van Logchem E, et al. Bivalent HPV vaccines lead to reduced (vaccine type) incident and persistent HPV infections and lower viral load in young Dutch females. Submitted. 2018.

30. van der Weele P, van Logchem E, Wolffs P, et al. Correlation between viral load, multiplicity of infection, and persistence of HPV16 and HPV18 infection in a Dutch cohort of young women. J Clin Virol. Oct 2016;83:6-11.

31. Waterboer T, Sehr P, Michael KM, et al. Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem. Oct 2005;51(10):1845-1853.

32. Richel O, de Vries HJ, van Noesel CJ, Dijkgraaf MG, Prins JM. Comparison of imiquimod, topical fluorouracil, and electrocautery for the treatment of anal intraepithelial neoplasia in HIV-positive men who have sex with men: an open-label, randomised controlled trial. Lancet Oncol. Apr 2013;14(4):346-353.

33. Richel O, Quint KD, Lindeman J, et al. One lesion, one virus: individual components of high-grade anal intraepithelial neoplasia in HIV-positive men contain a single HPV type. J Infect Dis. Jul 01 2014;210(1):111-120.

34. Quint W, Jenkins D, Molijn A, et al. One virus, one lesion--individual components of CIN lesions contain a specific HPV type. J Pathol. May 2012;227(1):62-71.

35. Schiffman M, Kjaer SK. Chapter 2: Natural history of anogenital human papillomavirus infection and neoplasia. JNCI Monographs. 2003;2003(31):14-19.

36. Moscicki AB, Schiffman M, Burchell A, et al. Updating the natural history of human papillomavirus and anogenital cancers. Vaccine. Nov 20 2012;30 Suppl 5:F24-33.

37. Jin F, Roberts JM, Grulich AE, et al. The performance of human papillomavirus biomarkers in predicting anal high-grade squamous intraepithelial lesions in gay and bisexual men. AIDS. Jun 1 2017;31(9):1303-1311.

38. Tong WW, Jin F, McHugh LC, et al. Progression to and spontaneous regression of high-grade anal squamous intraepithelial lesions in HIV-infected and uninfected men. AIDS. Sep 10 2013;27(14):2233-2243.

39. Siegenbeek van Heukelom ML, Richel O, de Vries HJ, et al. Low- and high-risk human papillomavirus genotype infections in intra-anal warts in HIV-positive men who have sex with men. Br J Dermatol. Mar 19 2016.

40. Guimera N, Lloveras B, Lindeman J, et al. The occasional role of low-risk human papillomaviruses 6, 11, 42, 44, and 70 in anogenital carcinoma defined by laser capture microdissection/PCR methodology: results from a global study. Am J Surg Pathol. Sep 2013;37(9):1299-1310.

41. Bruni L, Barrionuevo-Rosas L, Albero G, et al. Human Papillomavirus and Related Diseases in the World. Summary Report 27 July 2017. 2017; http://www.hpvcentre.net/statistics/reports/XWX.pdf. Accessed 24-8, 2017.

42. Schroeder L, Wichmann G, Willner M, et al. Antibodies against human papillomaviruses as diagnostic and prognostic biomarker in patients with neck squamous cell carcinoma from unknown primary tumor. Int J Cancer. Nov 21 2017.

143

Predictors of anal HSIL

6

recommendations from the College of American Pathologists and the American Society for Colposcopy and Cervical Pathology. J Low Genit Tract Dis. Jul 2012;16(3):205-242.

17. Cachay ER, Agmas W, Mathews WC. Relative accuracy of cervical and anal cytology for detection of high grade lesions by colposcope guided biopsy: a cut-point meta-analytic comparison. PLoS One. 2012;7(7):e38956.

18. Botes LP, Pett S, Carr A, et al. Anal cytological abnormalities are poor predictors of high-grade intraepithelial neoplasia amongst HIV-positive men who have sex with men. Sex Health. Mar 2013;10(1):9-17.

19. van Aar F, Mooij SH, van der Sande MA, et al. Anal and penile high-risk human papillomavirus prevalence in HIV-negative and HIV-infected MSM. AIDS. Nov 28 2013;27(18):2921-2931.

20. Siegenbeek van Heukelom ML, Marra E, de Vries HJC, Schim van der Loeff MF, Prins JM. Risk factors for anal high-grade squamous intraepithelial lesions in HIV-positive MSM: is targeted screening possible? AIDS. Oct 23 2017;31(16):2295-2301.

21. Melo VH, Guimaraes MD, Rocha GM, et al. Prevalence and risk factors associated with anal intraepithelial neoplasia among HIV-positive men in Brazil. J Low Genit Tract Dis. Apr 2014;18(2):128-135.

22. Alvarez J, de Pokomandy A, Rouleau D, et al. Episomal and integrated human papillomavirus type 16 loads and anal intraepithelial neoplasia in HIV-seropositive men. AIDS. Sep 24 2010;24(15):2355-2363.

23. Libois A, Feoli F, Nkuize M, et al. Prolonged antiretroviral therapy is associated with fewer anal high-grade squamous intraepithelial lesions in HIV-positive MSM in a cross-sectional study. Sex Transm Infect. Feb 2017;93(1):15-17.

24. Kreimer AR, Brennan P, Lang Kuhs KA, et al. Human papillomavirus antibodies and future risk of anogenital cancer: a nested case-control study in the European prospective investigation into cancer and nutrition study. J Clin Oncol. Mar 10 2015;33(8):877-884.

25. Kreimer AR, Johansson M, Waterboer T, et al. Evaluation of human papillomavirus antibodies and risk of subsequent head and neck cancer. J Clin Oncol. Jul 20 2013;31(21):2708-2715.

26. Kreimer AR, Johansson M, Yanik EL, et al. Kinetics of the Human Papillomavirus Type 16 E6 Antibody Response Prior to Oropharyngeal Cancer. J Natl Cancer Inst. Aug 1 2017;109(8).

27. van Sighem AI, van de Wiel MA, Ghani AC, et al. Mortality and progression to AIDS after starting highly active antiretroviral therapy. AIDS. Oct 17 2003;17(15):2227-2236.

28. Molijn A, Kleter B, Quint W, van Doorn LJ. Molecular diagnosis of human papillomavirus (HPV) infections. J Clin Virol. Mar 2005;32 Suppl 1:S43-51.

29. Van der Weele P, Breeuwsma M, Van Logchem E, et al. Bivalent HPV vaccines lead to reduced (vaccine type) incident and persistent HPV infections and lower viral load in young Dutch females. Submitted. 2018.

30. van der Weele P, van Logchem E, Wolffs P, et al. Correlation between viral load, multiplicity of infection, and persistence of HPV16 and HPV18 infection in a Dutch cohort of young women. J Clin Virol. Oct 2016;83:6-11.

31. Waterboer T, Sehr P, Michael KM, et al. Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem. Oct 2005;51(10):1845-1853.

32. Richel O, de Vries HJ, van Noesel CJ, Dijkgraaf MG, Prins JM. Comparison of imiquimod, topical fluorouracil, and electrocautery for the treatment of anal intraepithelial neoplasia in HIV-positive men who have sex with men: an open-label, randomised controlled trial. Lancet Oncol. Apr 2013;14(4):346-353.

33. Richel O, Quint KD, Lindeman J, et al. One lesion, one virus: individual components of high-grade anal intraepithelial neoplasia in HIV-positive men contain a single HPV type. J Infect Dis. Jul 01 2014;210(1):111-120.

34. Quint W, Jenkins D, Molijn A, et al. One virus, one lesion--individual components of CIN lesions contain a specific HPV type. J Pathol. May 2012;227(1):62-71.

35. Schiffman M, Kjaer SK. Chapter 2: Natural history of anogenital human papillomavirus infection and neoplasia. JNCI Monographs. 2003;2003(31):14-19.

36. Moscicki AB, Schiffman M, Burchell A, et al. Updating the natural history of human papillomavirus and anogenital cancers. Vaccine. Nov 20 2012;30 Suppl 5:F24-33.

37. Jin F, Roberts JM, Grulich AE, et al. The performance of human papillomavirus biomarkers in predicting anal high-grade squamous intraepithelial lesions in gay and bisexual men. AIDS. Jun 1 2017;31(9):1303-1311.

38. Tong WW, Jin F, McHugh LC, et al. Progression to and spontaneous regression of high-grade anal squamous intraepithelial lesions in HIV-infected and uninfected men. AIDS. Sep 10 2013;27(14):2233-2243.

39. Siegenbeek van Heukelom ML, Richel O, de Vries HJ, et al. Low- and high-risk human papillomavirus genotype infections in intra-anal warts in HIV-positive men who have sex with men. Br J Dermatol. Mar 19 2016.

40. Guimera N, Lloveras B, Lindeman J, et al. The occasional role of low-risk human papillomaviruses 6, 11, 42, 44, and 70 in anogenital carcinoma defined by laser capture microdissection/PCR methodology: results from a global study. Am J Surg Pathol. Sep 2013;37(9):1299-1310.

41. Bruni L, Barrionuevo-Rosas L, Albero G, et al. Human Papillomavirus and Related Diseases in the World. Summary Report 27 July 2017. 2017; http://www.hpvcentre.net/statistics/reports/XWX.pdf. Accessed 24-8, 2017.

42. Schroeder L, Wichmann G, Willner M, et al. Antibodies against human papillomaviruses as diagnostic and prognostic biomarker in patients with neck squamous cell carcinoma from unknown primary tumor. Int J Cancer. Nov 21 2017.

144

Predictors of anal HSIL

Supplementary table 1 - Cut-offs (in MFI) used to determine seropositivity for HPV antibody detection of

L1, E1, E2, E6, E7 in the H2M2-study*. HPV type L1 E1 E2 E6 E7

6 571

500 364 11 500

260 200

16 422 200 679 484 548 18 394 200 600 243 789 31 712

890 200

33 515

253 500 45 368

249 200

52 547

271 200 58 371

250 200

* Antibody specific seropositivity for all HPV antigens was calculated based on standardized serology cutoff values based

on the mean plus 5 standard deviations excluding positive outliers of 371 DNA negative, female self-reported virgins

(L1 antigens), or the mean plus 5 standard deviations of 117 DNA negative, female self-reported virgins (E antigens) [25, 31] Abbreviations: HPV – human papillomavirus; MFI - median

fluorescence intensity; DNA - deoxyribonucleic acid.

Supp

lem

enta

ry ta

ble

2. F

requ

enci

es o

f HPV

type

-spe

cific

bio

mar

kers

and

AIN

dia

gnos

is a

mon

g HI

V-po

sitiv

e M

SM in

Am

ster

dam

, the

Net

herla

nds (

the

H2M

2-st

udy)

- in

clud

ing

HPV

type

s 35,

39,

51,

56,

59.

HPV3

5 HP

V39

HPV5

1 HP

V56

HPV5

9

N

n %

n

%

n %

n

%

n %

An

al H

PV p

ersis

tenc

e

At le

ast 4

out

of 5

visi

ts p

ositi

ve

155

10

6%

5 3%

25

14

%

10

6%

0 0%

3

posit

ive

visit

s *

162

16

9%

8 4%

37

20

%

22

12%

2

1%

2 co

nsec

utiv

e po

sitiv

e vi

sits

167

19

10%

15

8%

46

24

%

28

15%

6

3%

At le

ast o

nce

posit

ive

169

43

23%

37

19

%

68

36%

53

28

%

15

8%

Posit

ive

in la

st v

isit

169

18

9%

20

10%

38

20

%

29

15%

6

3%

Anal

HPV

vira

l loa

d in

copi

es/h

uman

cell*

*# -

-

-

-

-

-

-

-

-

-

HP

V se

ropo

sitiv

ity**

#

L1

- -

- -

- -

- -

- -

E1

-

- -

- -

- -

- -

- E2

- -

- -

- -

- -

- -

E6

-

- -

- -

- -

- -

- E7

- -

- -

- -

- -

- -

HPV

type

in A

IN le

sion*

**

AI

N2

1

0.5%

1

0.5%

2

1%

1 0.

5%

1 0.

5%

AIN

3

3 2%

0

0%

0 0%

0

0%

0 0%

Ab

brev

iatio

ns: H

PV -

hum

an p

apill

omav

irus;

IQR

- int

erqu

artil

e ra

nge;

AIN

- an

al in

trae

pith

elia

l neo

plas

ia

- : n

ot m

easu

red;

* 3

pos

itive

visi

ts w

ith m

ax. 1

neg

ativ

e vi

sit in

bet

wee

n; *

* M

easu

red

at la

st H

2M v

isit:

for m

ost o

f the

men

this

was

visi

t #5,

bu

t for

som

e th

is w

as v

isit #

4, o

r #3,

or #

2; #

Not

mea

sure

d fo

r HPV

type

s 35,

39,

51,

56,

59;

***

as e

stab

lishe

d by

lase

r cap

ture

mic

rodi

ssec

tion.

145

Predictors of anal HSIL

6

Supplementary table 1 - Cut-offs (in MFI) used to determine seropositivity for HPV antibody detection of

L1, E1, E2, E6, E7 in the H2M2-study*. HPV type L1 E1 E2 E6 E7

6 571

500 364 11 500

260 200

16 422 200 679 484 548 18 394 200 600 243 789 31 712

890 200

33 515

253 500 45 368

249 200

52 547

271 200 58 371

250 200

* Antibody specific seropositivity for all HPV antigens was calculated based on standardized serology cutoff values based

on the mean plus 5 standard deviations excluding positive outliers of 371 DNA negative, female self-reported virgins

(L1 antigens), or the mean plus 5 standard deviations of 117 DNA negative, female self-reported virgins (E antigens) [25, 31] Abbreviations: HPV – human papillomavirus; MFI - median

fluorescence intensity; DNA - deoxyribonucleic acid.

Supp

lem

enta

ry ta

ble

2. F

requ

enci

es o

f HPV

type

-spe

cific

bio

mar

kers

and

AIN

dia

gnos

is a

mon

g HI

V-po

sitiv

e M

SM in

Am

ster

dam

, the

Net

herla

nds (

the

H2M

2-st

udy)

- in

clud

ing

HPV

type

s 35,

39,

51,

56,

59.

HPV3

5 HP

V39

HPV5

1 HP

V56

HPV5

9

N

n %

n

%

n %

n

%

n %

An

al H

PV p

ersis

tenc

e

At le

ast 4

out

of 5

visi

ts p

ositi

ve

155

10

6%

5 3%

25

14

%

10

6%

0 0%

3

posit

ive

visit

s *

162

16

9%

8 4%

37

20

%

22

12%

2

1%

2 co

nsec

utiv

e po

sitiv

e vi

sits

167

19

10%

15

8%

46

24

%

28

15%

6

3%

At le

ast o

nce

posit

ive

169

43

23%

37

19

%

68

36%

53

28

%

15

8%

Posit

ive

in la

st v

isit

169

18

9%

20

10%

38

20

%

29

15%

6

3%

Anal

HPV

vira

l loa

d in

copi

es/h

uman

cell*

*# -

-

-

-

-

-

-

-

-

-

HP

V se

ropo

sitiv

ity**

#

L1

- -

- -

- -

- -

- -

E1

-

- -

- -

- -

- -

- E2

- -

- -

- -

- -

- -

E6

-

- -

- -

- -

- -

- E7

- -

- -

- -

- -

- -

HPV

type

in A

IN le

sion*

**

AI

N2

1

0.5%

1

0.5%

2

1%

1 0.

5%

1 0.

5%

AIN

3

3 2%

0

0%

0 0%

0

0%

0 0%

Ab

brev

iatio

ns: H

PV -

hum

an p

apill

omav

irus;

IQR

- int

erqu

artil

e ra

nge;

AIN

- an

al in

trae

pith

elia

l neo

plas

ia

- : n

ot m

easu

red;

* 3

pos

itive

visi

ts w

ith m

ax. 1

neg

ativ

e vi

sit in

bet

wee

n; *

* M

easu

red

at la

st H

2M v

isit:

for m

ost o

f the

men

this

was

visi

t #5,

bu

t for

som

e th

is w

as v

isit #

4, o

r #3,

or #

2; #

Not

mea

sure

d fo

r HPV

type

s 35,

39,

51,

56,

59;

***

as e

stab

lishe

d by

lase

r cap

ture

mic

rodi

ssec

tion.

146

Predictors of anal HSIL Su

pple

men

tary

tabl

e 3.

Cha

ract

eris

tics o

f the

stud

y po

pula

tion

of th

e H2

M2

(HIV

& H

PV in

MSM

2) b

y an

al H

SIL

diag

nosi

s with

kno

wn

caus

ativ

e HP

V ty

pe, b

ased

on

HPV

type

-spe

cific

re-a

sses

smen

t of b

iops

y m

ater

ial.

To

tal (

N=1

69)

No

HSIL

** (N

=143

)

HSIL

link

ed to

sp

ecifi

c HP

V ty

pe

(N=2

6)

No.

%

N

o.

%

No.

%

P

valu

e# De

mog

raph

ic an

d be

havi

oura

l var

iabl

es

Ag

e in

yea

rs a

t mom

ent o

f HRA

, mea

n(SD

) 50

(9

) 50

(1

0)

48

(8)

0.3

Smok

ing

stat

us a

t mom

ent o

f HRA

(a)

0.9

Nev

er sm

oked

60

38

%

52

38%

8

40%

Prev

ious

ly sm

oked

52

33

%

46

34%

6

30%

Curr

ently

smok

ing

44

28%

38

28

%

6 30

%

Hi

gher

edu

catio

n (h

ighe

r pro

fess

iona

l edu

catio

n or

uni

vers

ity)*

0.1

No

62

37%

56

39

%

6 23

%

Ye

s 10

7 63

%

87

61%

20

77

%

Li

ving

situ

atio

n* (b

)

1.

0 Li

ving

alo

ne

84

50%

71

50

%

13

52%

Livi

ng to

geth

er w

ith a

stea

dy p

artn

er

78

47%

66

46

%

12

48%

Oth

er

5 3%

5

4%

0 0%

Sexu

al b

ehav

iour

al v

aria

bles

Life

time

num

ber o

f sex

par

tner

s, m

edia

n(IQ

R)*

(a)

500

(150

-100

0)

400

(150

-100

0)

500

(150

-100

0)

0.7

Life

time

num

ber o

f sex

par

tner

s* (a

)

0.

5 <2

5 7

5%

6 5%

1

4%

25

-99

16

10%

13

10

%

3 13

%

10

0-49

9 53

34

%

48

36%

5

22%

≥500

79

51

%

65

49%

14

61

%

Co

ndom

use

dur

ing

anal

sex

in th

e pr

eced

ing

6 m

onth

s* (c

)

0.

9 N

ever

17

13

%

14

12%

3

14%

Som

etim

es

74

55%

62

54

%

12

57%

Alw

ays

44

33%

38

33

%

6 29

%

HI

V-re

late

d va

riabl

es (a

ll ar

ound

tim

e of

HRA

)

Curr

ently

usin

g cA

RT

0.5

No

4 2%

3

2%

1 4%

Yes

165

98%

14

0 98

%

25

96%

Dura

tion

of c

ART

use

in y

ears

(med

ian/

IQR)

(d)

8.8

(5.0

-15.

9)

9.7

(5.0

-16.

1)

6.8

(3.4

-11.

6)

0.1

Year

s liv

ing

with

vira

l sup

pres

sion

at m

omen

t of H

RA (e

)

0.

7 <3

yea

rs

26

16%

21

15

%

5 20

%

3-

10 y

ears

75

45

%

63

44%

12

48

%

>1

0 ye

ars

66

40%

58

41

%

8 32

%

N

adir

CD4

T-ce

ll co

unt c

ells/

µl (m

ean/

SD) (

f) 24

4 (1

36)

241

(140

) 25

9 (1

11)

0.5

CD4

T-ce

ll co

unt c

ells/

µl (m

ean/

SD) (

g)

676

(247

) 68

1 (2

57)

651

(197

) 0.

6 HI

V pl

asm

a vi

ral l

oad

copi

es/m

l

0.

6 <5

0 16

1 95

%

136

95%

25

96

%

≥5

0 8

5%

7 5%

1

4%

Ti

me

betw

een

last

H2M

visi

t and

HRA

in y

ears

(med

ian/

IQR)

1.

3 (0

.9-1

.8)

1.3

(0.9

-1.8

) 1.

2 (0

.9-1

.6)

0.6

* Va

riabl

e fr

om th

e ba

selin

e vi

sit o

f H2M

stud

y **

For

the

HPV

type

-leve

l ana

lyse

s a d

atas

et w

as c

reat

ed o

f one

reco

rd p

er H

PV ty

pe p

er p

atie

nt. E

ach

reco

rd in

clud

ed a

val

ue fo

r HSI

L; th

is w

as 1

if th

e pa

tient

had

an

HSI

L ca

used

by

that

par

ticul

ar ty

pe a

nd 0

if th

e pa

tient

did

not

hav

e HS

IL. O

f som

e le

sions

the

HSIL

dia

gnos

is co

uld

not b

e co

nfirm

ed, t

here

fore

thes

e re

cord

s are

ex

clud

ed. T

his m

ay b

e be

caus

e in

the

left-

over

mat

eria

l of t

he le

sion

no H

SIL

was

rem

aini

ng, o

r bec

ause

of a

n in

itial

mis-

diag

nosis

of H

SIL.

As t

hese

HSI

L di

agno

ses

coul

d no

t be

assig

ned

to a

n HP

V ty

pe, t

hey

wer

e ex

clud

ed fr

om th

e an

alys

es.

# Si

gnifi

canc

e of

diff

eren

ces b

etw

een

part

icip

ants

with

and

with

out a

nal H

SIL

was

ass

esse

d by

Chi

-squ

are

test

s or F

isher

exa

ct te

st fo

r cat

egor

ical

var

iabl

es a

nd

Wilc

oxon

rank

-sum

test

s for

con

tinue

s var

iabl

es.

(a) 1

7 m

issin

gs; (

b) 2

miss

ings

; (c)

38

miss

ing;

(d) 5

miss

ings

; (e)

2 m

issin

gs; (

f) 1

miss

ing;

(g) 2

8 m

issin

gs; (

h) 4

miss

ings

Ab

brev

iatio

ns: H

PV -

hum

an p

apill

omav

irus;

HIV

- hu

man

imm

unod

efic

ienc

y vi

rus;

cAR

T - c

ombi

natio

n an

tiret

rovi

ral t

hera

py; H

RA -

high

reso

lutio

n an

osco

py; A

IN -

anal

intr

aepi

thel

ial n

eopl

asia

; HSI

L - h

igh

grad

e sq

uam

ous i

ntra

epith

elia

l les

ion;

SD

- sta

ndar

d de

viat

ion;

IQR

- int

erqu

artil

e ra

nge.

147

Predictors of anal HSIL

6

Supp

lem

enta

ry ta

ble

3. C

hara

cter

istic

s of t

he st

udy

popu

latio

n of

the

H2M

2 (H

IV &

HPV

in M

SM 2

) by

anal

HSI

L di

agno

sis w

ith k

now

n ca

usat

ive

HPV

type

, bas

ed o

n HP

V ty

pe-s

peci

fic re

-ass

essm

ent o

f bio

psy

mat

eria

l.

To

tal (

N=1

69)

No

HSIL

** (N

=143

)

HSIL

link

ed to

sp

ecifi

c HP

V ty

pe

(N=2

6)

No.

%

N

o.

%

No.

%

P

valu

e# De

mog

raph

ic an

d be

havi

oura

l var

iabl

es

Ag

e in

yea

rs a

t mom

ent o

f HRA

, mea

n(SD

) 50

(9

) 50

(1

0)

48

(8)

0.3

Smok

ing

stat

us a

t mom

ent o

f HRA

(a)

0.9

Nev

er sm

oked

60

38

%

52

38%

8

40%

Prev

ious

ly sm

oked

52

33

%

46

34%

6

30%

Curr

ently

smok

ing

44

28%

38

28

%

6 30

%

Hi

gher

edu

catio

n (h

ighe

r pro

fess

iona

l edu

catio

n or

uni

vers

ity)*

0.1

No

62

37%

56

39

%

6 23

%

Ye

s 10

7 63

%

87

61%

20

77

%

Li

ving

situ

atio

n* (b

)

1.

0 Li

ving

alo

ne

84

50%

71

50

%

13

52%

Livi

ng to

geth

er w

ith a

stea

dy p

artn

er

78

47%

66

46

%

12

48%

Oth

er

5 3%

5

4%

0 0%

Sexu

al b

ehav

iour

al v

aria

bles

Life

time

num

ber o

f sex

par

tner

s, m

edia

n(IQ

R)*

(a)

500

(150

-100

0)

400

(150

-100

0)

500

(150

-100

0)

0.7

Life

time

num

ber o

f sex

par

tner

s* (a

)

0.

5 <2

5 7

5%

6 5%

1

4%

25

-99

16

10%

13

10

%

3 13

%

10

0-49

9 53

34

%

48

36%

5

22%

≥500

79

51

%

65

49%

14

61

%

Co

ndom

use

dur

ing

anal

sex

in th

e pr

eced

ing

6 m

onth

s* (c

)

0.

9 N

ever

17

13

%

14

12%

3

14%

Som

etim

es

74

55%

62

54

%

12

57%

Alw

ays

44

33%

38

33

%

6 29

%

HI

V-re

late

d va

riabl

es (a

ll ar

ound

tim

e of

HRA

)

Curr

ently

usin

g cA

RT

0.5

No

4 2%

3

2%

1 4%

Yes

165

98%

14

0 98

%

25

96%

Dura

tion

of c

ART

use

in y

ears

(med

ian/

IQR)

(d)

8.8

(5.0

-15.

9)

9.7

(5.0

-16.

1)

6.8

(3.4

-11.

6)

0.1

Year

s liv

ing

with

vira

l sup

pres

sion

at m

omen

t of H

RA (e

)

0.

7 <3

yea

rs

26

16%

21

15

%

5 20

%

3-

10 y

ears

75

45

%

63

44%

12

48

%

>1

0 ye

ars

66

40%

58

41

%

8 32

%

N

adir

CD4

T-ce

ll co

unt c

ells/

µl (m

ean/

SD) (

f) 24

4 (1

36)

241

(140

) 25

9 (1

11)

0.5

CD4

T-ce

ll co

unt c

ells/

µl (m

ean/

SD) (

g)

676

(247

) 68

1 (2

57)

651

(197

) 0.

6 HI

V pl

asm

a vi

ral l

oad

copi

es/m

l

0.

6 <5

0 16

1 95

%

136

95%

25

96

%

≥5

0 8

5%

7 5%

1

4%

Ti

me

betw

een

last

H2M

visi

t and

HRA

in y

ears

(med

ian/

IQR)

1.

3 (0

.9-1

.8)

1.3

(0.9

-1.8

) 1.

2 (0

.9-1

.6)

0.6

* Va

riabl

e fr

om th

e ba

selin

e vi

sit o

f H2M

stud

y **

For

the

HPV

type

-leve

l ana

lyse

s a d

atas

et w

as c

reat

ed o

f one

reco

rd p

er H

PV ty

pe p

er p

atie

nt. E

ach

reco

rd in

clud

ed a

val

ue fo

r HSI

L; th

is w

as 1

if th

e pa

tient

had

an

HSI

L ca

used

by

that

par

ticul

ar ty

pe a

nd 0

if th

e pa

tient

did

not

hav

e HS

IL. O

f som

e le

sions

the

HSIL

dia

gnos

is co

uld

not b

e co

nfirm

ed, t

here

fore

thes

e re

cord

s are

ex

clud

ed. T

his m

ay b

e be

caus

e in

the

left-

over

mat

eria

l of t

he le

sion

no H

SIL

was

rem

aini

ng, o

r bec

ause

of a

n in

itial

mis-

diag

nosis

of H

SIL.

As t

hese

HSI

L di

agno

ses

coul

d no

t be

assig

ned

to a

n HP

V ty

pe, t

hey

wer

e ex

clud

ed fr

om th

e an

alys

es.

# Si

gnifi

canc

e of

diff

eren

ces b

etw

een

part

icip

ants

with

and

with

out a

nal H

SIL

was

ass

esse

d by

Chi

-squ

are

test

s or F

isher

exa

ct te

st fo

r cat

egor

ical

var

iabl

es a

nd

Wilc

oxon

rank

-sum

test

s for

con

tinue

s var

iabl

es.

(a) 1

7 m

issin

gs; (

b) 2

miss

ings

; (c)

38

miss

ing;

(d) 5

miss

ings

; (e)

2 m

issin

gs; (

f) 1

miss

ing;

(g) 2

8 m

issin

gs; (

h) 4

miss

ings

Ab

brev

iatio

ns: H

PV -

hum

an p

apill

omav

irus;

HIV

- hu

man

imm

unod

efic

ienc

y vi

rus;

cAR

T - c

ombi

natio

n an

tiret

rovi

ral t

hera

py; H

RA -

high

reso

lutio

n an

osco

py; A

IN -

anal

intr

aepi

thel

ial n

eopl

asia

; HSI

L - h

igh

grad

e sq

uam

ous i

ntra

epith

elia

l les

ion;

SD

- sta

ndar

d de

viat

ion;

IQR

- int

erqu

artil

e ra

nge.

148

Predictors of anal HSIL

Supplementary table 4 - Univariable logistic regression of possible determinants of anal HSIL, results of logistic regression using GEE including HPV types 6, 11, 16, 18, 31, 33, 45, 52, and 58.

Univariable logistic regression -

HSIL vs. no HSIL (N=191) OR (95% CI) P-value Demographic variables

Age

0.2 ≤44 years ref

45-54 years 0.87 (0.35-2.20) ≥55 years 0.29 (0.08-1.05) HIV-related variables

Years living with viral suppression §

0.2 < 3 year Ϯ ref

3-10 years 1.22 (0.40-3.72) >10 years 0.49 (0.14-1.71) Time variables

Time between last H2M visit and HRA in years

0.9 <1 year ref

1-2 years 1.23 (0.44-3.46) ≥2 years 1.53 (0.32-7.26) Abbreviations: CI - confidence Interval, HSIL - High grade Squamous Intraepithelial Lesion, HIV -

Human Immunodeficiency Virus, n.a. = not assessable Ϯ participants who never had an undetectable viral load are included in the category <1 year undetectable viral load § viral suppression was defined as having a viral load of <200 copies/ml in tests from 1-8-1999 onwards allowing for a onetime blip in viral load between 200 and 400 copies/ml. For samples tested prior to 1-8-1999 the cut-off of detectability of the laboratory assay that was used for that sample is the cut-off for viral suppression. This varies by time period (sensitivity of the assays increased over time) and hospital (based on the used assay).

Supplementary figure 1. Associations between biomarkers and different definitions of anal HSIL with known causative HPV type among 183 HIV-positive MSM in Amsterdam, the Netherlands (H2M2 study); results from univariable logistic regression using GEE, including HPV16 and HPV18. Odds ratios are displayed on a logarithmic scale. If no odds ratio and 95% confidence intervals are shown for a specific biomarker, these biomarkers were unassessable due to small numbers. * Persistence is defined as: at least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Seropositivity is defined as seropositive above the standard cut-off at the last H2M study visit. Abbreviations: OR – odds ratio, aHSIL – anal high-grade squamous intraepithelial lesion, HPV – human papillomavirus.

149

Predictors of anal HSIL

6

Supplementary table 4 - Univariable logistic regression of possible determinants of anal HSIL, results of logistic regression using GEE including HPV types 6, 11, 16, 18, 31, 33, 45, 52, and 58.

Univariable logistic regression -

HSIL vs. no HSIL (N=191) OR (95% CI) P-value Demographic variables

Age

0.2 ≤44 years ref

45-54 years 0.87 (0.35-2.20) ≥55 years 0.29 (0.08-1.05) HIV-related variables

Years living with viral suppression §

0.2 < 3 year Ϯ ref

3-10 years 1.22 (0.40-3.72) >10 years 0.49 (0.14-1.71) Time variables

Time between last H2M visit and HRA in years

0.9 <1 year ref

1-2 years 1.23 (0.44-3.46) ≥2 years 1.53 (0.32-7.26) Abbreviations: CI - confidence Interval, HSIL - High grade Squamous Intraepithelial Lesion, HIV -

Human Immunodeficiency Virus, n.a. = not assessable Ϯ participants who never had an undetectable viral load are included in the category <1 year undetectable viral load § viral suppression was defined as having a viral load of <200 copies/ml in tests from 1-8-1999 onwards allowing for a onetime blip in viral load between 200 and 400 copies/ml. For samples tested prior to 1-8-1999 the cut-off of detectability of the laboratory assay that was used for that sample is the cut-off for viral suppression. This varies by time period (sensitivity of the assays increased over time) and hospital (based on the used assay).

Supplementary figure 1. Associations between biomarkers and different definitions of anal HSIL with known causative HPV type among 183 HIV-positive MSM in Amsterdam, the Netherlands (H2M2 study); results from univariable logistic regression using GEE, including HPV16 and HPV18. Odds ratios are displayed on a logarithmic scale. If no odds ratio and 95% confidence intervals are shown for a specific biomarker, these biomarkers were unassessable due to small numbers. * Persistence is defined as: at least three anal samples positive for the same HPV type. We allowed one negative sample between two positive sample per participant as long as there were at least three positive anal samples. Seropositivity is defined as seropositive above the standard cut-off at the last H2M study visit. Abbreviations: OR – odds ratio, aHSIL – anal high-grade squamous intraepithelial lesion, HPV – human papillomavirus.

150

Predictors of anal HSIL

Supplementary figure 2: Association between anal HPV viral load of HPV types 16,18, 31, 33, 45, 52, 58 and anal HSIL using restricted cubic splines. Odds ratios for having anal HSIL by standardized Z-score of the natural log transformed anal HPV viral load of each hrHPV type.

.11

1010

010

0010

000

Odd

s R

atio

-2 -1 0 1 2Standardized Z-score of the natural log transformed anal HPV viral load

Odds Ratio95% Confidence Interval

.1

110

100

1000

1000

0O

dds

Rat

io

-2 -1 0 1 2Standardized Z-score of the natural log transformed anal HPV viral load

Odds Ratio95% Confidence Interval