Research article.

Transcription

1 Reserch rticle Lte presenttion for HIV cre cross Europe: updte from the Collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) study, 2010 to 2013 The lte presenters working group in COHERE in EuroCoord 1 1. The members of the group re listed t the end of the rticle Correspondence: Amnd Mocroft (.mocroft@ucl.c.uk) Cittion style for this rticle: The lte presenters working group in COHERE in EuroCoord. Lte presenttion for HIV cre cross Europe: updte from the Collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) study, 2010 to Euro Surveill. 2015;20(47):pii= DOI: ES Article submitted on 01 September 2015 / ccepted on 18 November 2015 / published on 26 November 2015 Lte presenttion (LP) for HIV cre cross Europe remins significnt issue. We provide cross-europen updte from 34 countries on the prevlence nd risk fctors of LP for People ged 16 presenting for HIV cre (erliest of HIV-dignosis, first clinic visit or cohort enrolment) fter 1 Jnury 2010 with vilble CD4 count within six months of presenttion were included. LP ws defined s presenttion with CD4 count < 350/mm 3 or n AIDS defining event (t ny CD4), in the six months following HIV dignosis. Logistic regression investigted chnges in LP over time. A totl of 30,454 people were included. The medin CD4 count t presenttion ws 368/ mm 3 (interqurtile rnge (IQR) /mm 3 ), with no chnge over time (p = 0.70). In 2010, 4,775/10,766 (47.5%) were LP wheres in 2013, 1,642/3,375 (48.7%) were LP (p = 0.63). LP ws most common in centrl Europe (4,791/9,625, 49.8%), followed by northern (5,704/11,692; 48.8%), southern (3,550/7,760; 45.8%) nd estern Europe (541/1,377; 38.3%; p < ). There ws significnt increse in LP in mle nd femle people who inject drugs (PWID) (djusted odds rtio (OR)/yer lter 1.16; 95% confidence intervl (CI): ), nd significnt decline in LP in northern Europe (OR/yer lter 0.89; 95% CI: ). Further improvements in effective HIV testing strtegies, with focus on vulnerble groups, re required cross the Europen continent. Introduction The United Ntions Progrmme on HIV/AIDS (UNAIDS) recently relesed n mbitious strtegy tht clls for 90% of HIV infections to be dignosed by 2020 [1]. Mny people however remin unwre of their HIV sttus nd cscdes for cre for HIV vry widely from country to country within Europe, with 20 70% of people infected with HIV remining undignosed [2-4]. These estimtes rely on estimtes of the popultion with HIV, which itself is estimted using vriety of different methods [5]. In ddition, c 40 60% of HIVpositive people re dignosed with HIV t lte stge of infection [6,7], defined s people presenting for HIV cre with CD4 count of less thn 350/mm 3 or n AIDS defining illness [8]. Individuls t gretest risk of lte dignosis nd/or lte entry into cre hve poorer outcomes nd higher resource use once dignosed. Those who re unwre of their HIV sttus re lso less likely to tke steps to prevent onwrd trnsmission to others [9-11]. However, mny who present lte, do so becuse they perceive their risk for HIV s low, s they hve few sex prtners for exmple. The rtes of lte presenttion (LP) mong newly dignosed HIV positive people in ny setting serves s proxy of effective HIV testing strtegies. Such strtegies should ensure people enter pproprite cre to strt ntiretrovirl therpy (ART) [12]. However, the extent to which these recommendtions re implemented cross Europe is vrible [13]. Following the recent publiction of findings from the Strtegic Timing of Antiretrovirl Tretment (START) [14], tretment is now recommended for ll people infected with HIV [13]. The World Helth Orgniztion (WHO) recognises erly HIV dignosis s crucil first step in the successful cre of HIV [15]. There re number of progrmmes nd inititives to increse HIV testing; these include indictor-condition-guided HIV testing nd ntionl HIV testing strtegies, linkge nd retention in cre of those lredy dignosed. Furthermore, there re inititives specificlly imed t reducing HIV trnsmission, such s hrm-reduction, condom use, initition of ntiretrovirl therpy nd pre-exposure prophylxis [16-18]. 1

2 Figure 1 Adjusted odds of hving no CD4 count informtion within six months following HIV dignosis, COHERE study, Dte HIV test/yer lter Age/10 yers older Region of origin Unknown Other Afric Europe Region of cre in Europe Estern Northern Centrl Southern HIV exposure group Femle other Mle other Femle PWID Mle PWID Femle heterosexul Mle heterosexul MSM Adjusted odds rtio (95% confidence intervl) COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe; MSM: men who hve sex with men; PWID: people who inject drugs. Logrithmic scle. The Collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) study provides unique opportunity to describe the epidemiology of those dignosed HIV-positive t lte stge of HIV infection compred with those dignosed erlier, nd to look t geogrphicl differences within HIV exposure groups. COHERE is collbortion of 39 cohorts cross Europe nd is prt of the EuroCoord network ( net). COHERE ws estblished in 2005 with the im of conducting epidemiologicl reserch on the prognosis nd outcome of HIV-positive people, which the individul contributing cohorts cnnot ddress themselves becuse of insufficient smple size or heterogeneity of specific subgroups of HIV-positive people. Locl ethics committee nd/or other regultory pprovls were obtined s pplicble ccording to locl nd/or ntionl regultions in ll prticipting cohorts unless no such requirement pplied to observtionl studies. Ech cohort submits dt using the stndrdised HIV Collbortion Dt Exchnge Protocol (HICDEP) [19], including informtion on ptient demogrphics, use of combintion ntiretrovirl therpy (cart), CD4 counts, AIDS, nd deths. Further detils cn be found on EuroCoord website [20]. Previous work showed n overll 4% decrese in LP per yer of lte presenttion between 2000 nd 2010 cross Europe, lbeit with n increse over time in people who inject drugs (PWID) [7]. The ims of this updte were to determine if the downwrd trend in LP observed between 2000 nd 2010 continued, nd whether there were ny groups of individuls in which LP continues to increse. Methods Ptients Twenty-four cohorts including dt from 34 Europen countries provided dt for the present nlysis. All people ged 16 yers, who presented for cre (defined s erliest dte of HIV dignosis, first clinic visit, or enrolment into the prticipting cohort, referred to s bseline ) for the first time fter 1 Jnury 2010 were included to provide n updte to the report from 2013 which included people dignosed to the end of 2010 [7]. People were excluded if informtion on sex or dte of HIV dignosis ws missing, or where there ws evidence of n erlier HIV dignosis (CD4 count, AIDS dignosis, or hving strted ntiretrovirl therpy 2

3 Figure 2 Chnges over time in stges of lte presenttion nd CD4 count t HIV dignosis, COHERE study, (N=30,454) Percentge Medin CD4 t HIV dignosis (interqurtile rnge) (n=10,766) 2011 (n=10,054) 2012 (n=6,259) 2013 (n=3,375) Yer of dignosis Lte presenttion Lte presenttion with very dvnced disese CD4 t dignosis (interqurtile rnge) Lte presenttion with dvnced disese AIDS COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe. Lte presenttion: dignosed with HIV with CD4 count below 350/mm 3 or n AIDS defining event regrdless of the CD4 count, in the six months following HIV dignosis. Lte presenttion with dvnced disese: dignosed with HIV with CD4 count below 200/mm 3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis. Lte presenttion with very dvnced disese: dignosed with HIV with CD4 count below 50/mm 3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis (ART)) more thn one month before first clinic visit, s were people from Argentinen centres in EuroSIDA [21]. People from the seroconverter cohorts in COHERE were lso excluded, s in our previous work [7]. By definition, such people re dignosed soon fter HIV infection, even if they hve low CD4 count t HIV-1 dignosis. Definitions of lte presenttion LP ws defined s n individul dignosed with HIV with CD4 count below 350/mm 3 or n AIDS-defining event regrdless of the CD4 count, in the six months following HIV dignosis. LP with dvnced disese ws defined s n individul dignosed with HIV with CD4 count below 200/mm 3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis. LP with very dvnced disese ws defined s n individul dignosed with HIV with CD4 count below 50/mm 3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis. The proportion presenting with AIDS, regrdless of the CD4 count t which it occurred, ws lso presented. Delyed entry into cre ws defined s more thn three months between HIV dignosis nd first clinic visit, in those where both dtes were recorded. All people were required to hve t lest one CD4 count mesured in the six months following dignosis. Sttisticl methods Bseline chrcteristics of lte presenters were compred with those of non-lte presenters nd logistic regression ws used to identify fctors ssocited with lte presenttion nd lte presenttion with dvnced disese. Fctors investigted were ge, HIV exposure group (men who hve sex with men (MSM), heterosexul men, heterosexul femle, mle PWID, femle PWID, other (including ptients with unknown HIV exposure group)), continent of origin (Europe, Afric, other (including ptients from Centrl/Southern 3

4 Figure 3 Chnges over time in lte presenttion by HIV exposure groups, COHERE study, (n=30,454) Percentge with LP MSM M Het F Het M/F PWID M/F Other HIV exposure group Odds of lte presenttion: per yer lter of presenttion. Univrite Multivrite 0.99 ( ) 0.99 ( ) 1.04 ( ) 1.16 ( ) 1.10 ( ) 0.98 ( ) 0.97 ( ) 0.99 ( ) 1.16 ( ) 1.08 ( ) 2010 (n=10,766) 2011 (n=10,054) 2012 (n=6,259) 2013 (n=3,375) COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe; F: femle; Het: heterosexul; LP: lte presenttion; M: mle; MSM: men who hve sex with men; PWID: people who inject drugs. Adjusted for ge, region of cre in Europe, continent of origin, delyed ccess to cre. Lte presenttion: dignosed with HIV with CD4 count below 350/mm 3 or n AIDS defining event regrdless of the CD4 count, in the six months following HIV dignosis. Americ nd Asi), unknown), region of HIV dignosis in Europe, bsed on the cohort loction nd defined similrly to the EuroSIDA study [22] (Tble 1), clendr yer of dignosis, nd delyed entry into cre. A priori, we were interested in compring chnges over time within region of HIV dignosis in Europe nd HIV exposure groups. Simple descriptive dt were used to present the proportions of LP, dvnced LP, very dvnced LP nd presenttion with AIDS by country; countries were grouped into regions nd nonymised, those countries with less thn 50 people included were combined. Liner regression ws used to ssess chnge over time in CD4 counts overll nd mong LP, nd Cox proportionl hzrds models were used to compre the risk of development of new clinicl event i.e. new AIDS defining illness occurring more thn one month fter the first if the person hd AIDS within six months of bseline or deth), whether this hs chnged over time, nd t similr rte for LP nd non-lp. Descriptive nlyses were used to investigte whether someone clssified s LP bsed on CD4 count > 350/ mm 3 (but no AIDS dignosis) would not be LP if the next CD4 count ws used (misclssifiction). This nlysis ws limited to the smll subset with CD4 counts mesured fter HIV dignosis nd before strting ART. Logistic regression ws used to determine whether potentil misclssifiction of people s LP hd chnged over time, fter djusting for ge, HIV exposure group, region of origin, region of HIV dignosis in Europe, 4

5 Figure 4 Chnges over time in lte presenttion by region of cre in Europe, COHERE study, (N=30,454) Percentge with LP Southern Centrl Northern Estern Region of cre in Europe Odds of lte presenttion: per yer lter of presenttion. Univrite Multivrite b 1.04 ( ) 1.01 ( ) 0.93 ( ) 1.12 ( ) 1.02 ( ) 1.04 ( ) 0.89 ( ) 1.05 ( ) 2010 (n=10,766) 2011 (n=10,054) 2012 (n=6,259) 2013 (n=3,375) COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe; LP: lte presenttion. Adjusted for ge, HIV exposure group, region of origin, delyed ccess to cre. b No dt vilble for Lte presenttion: dignosed with HIV with CD4 count below 350/mm 3 or n AIDS defining event regrdless of the CD4 count, in the six months following HIV dignosis. clendr yer of dignosis, delyed entry into cre, nd CD4 count nd HIV virl lod t HIV dignosis. All nlyses were performed using Sttisticl Anlysis Softwre Version 9.3 (Sttisticl Anlysis Softwre). Results Of 37,859 people with HIV-1 test fter 1 Jnury 2010, 4,197 were excluded becuse they were ged < 16, from seroconverter cohorts, or where there ws evidence tht the person hd strted ART, hd CD4 count or n AIDS dignosis more thn 28 dys before the first reported HIV-1 test. A further 3,208 of 33,662 (9.5%) were excluded due to missing CD4 counts, 9.6% from south Europe nd 8.6%, 9.8% nd 13.5% from centrl, northern nd estern Europe respectively. Compred with the 30,454 included, those excluded due to missing CD4 counts were more likely to be PWID, from other (including unknown) HIV exposure groups, nd to be in cre in northern, centrl or estern Europe compred with southern Europe (Figure 1). Older people were less likely to be excluded, s were those with more recent test for HIV. Lte presenttion nd chnges over time Tble 2 summrises the chrcteristics of the 30,454 people included, strtified by LP sttus; 14,586 (47.9%) were LP, rnging from > 60% of heterosexul men or people originting from Afric to c 39% MSM nd femle PWID. 5

6 Tble 1 Geogrphicl ctegoristion of Europen countries to regions, COHERE study, Region in Europe Northern Southern Centrl Estern Countries included Denmrk, Finlnd, Irelnd, the Netherlnds, Norwy, Sweden, United Kingdom Greece, Isrel, Itly, Portugl, Spin Austri, Belgium, Frnce, Germny, Luxemburg, Switzerlnd Belrus, Bosni nd Herzegovin, Bulgri, Croti, Czech Republic, Estoni, Hungry, Ltvi, Lithuni, Polnd, Romni, Russi, Serbi, Slovki, Sloveni, Ukrine COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe. Defined similrly to the EuroSIDA study [22]. Among people in whom both first visit nd HIV test dte were known, 1,247/27,818 (4.5%) hd delyed entry into cre. Figure 2 shows the nnul proportion of people with LP, LP with dvnced or very dvnced disese, nd with n AIDS dignosis, regrdless of the CD4 count t presenttion. In ,775/10,766 (47.5%) were LP, compred with 1,642/3,375 (48.7%) in 2013 or lter (p = 0.63). The proportion of people with LP, dvnced disese, very dvnced disese, or AIDS did not chnge significntly over time (p = 0.63, 0.090, 0.16, nd respectively). The proportion of those presenting who would be eligible for strting cart with CD4 count of <500/ mm 3 ws 69.0% in 2010, 68.8% in 2011, 68.3% in 2012 nd 69.0 in 2013 or lter (p = 0.77). In multivrite nlyses, there ws no evidence of chnge over time in LP (djusted odds rtio (OR) 0.99/ yer lter; 95% confidence intervl (CI): ; p = 0.60), or in LP with dvnced disese (OR 0.99/ yer lter; 95% CI: ; p = 0.65). This finding ws consistent cross wide rnge of sensitivity nlyses, such s including those with AIDS but without CD4 count mesured s LP, including deths within the first six months s LP, when the window required for CD4 count fter HIV dignosis to three months, nd defining LP s CD4 count < 350/mm 3 or n AIDS dignosis within three months of HIV dignosis. There ws some evidence tht presenttion with very dvnced disese hd decresed over time by 3% per yer lter (OR 0.97/yer lter; 95% CI: ; p = 0.035), nd tht LP bsed on n AIDS dignosis lone, regrdless of the CD4 count t which it ws dignosed, hd decresed by 7% per yer (OR 0.93; 95% CI: ; p = ). Chnges in CD4 count t presenttion The medin CD4 count t presenttion ws 368/mm 3 (interqurtile rnge (IQR) /mm 3 ). There ws no evidence of chnge over time in the medin CD4 count t presenttion (p=0.70), suggesting tht overll, the level of immunodeficiency t which HIV ws dignosed hs not chnged over time (djusted chnge/yer 1.2/ mm 3 ; 95% CI: -0.8 to 3.3/mm 3 ; p = 0.89). Similr results were seen in n nlysis limited to LP (djusted chnge/ yer -1.1/mm 3 ; 95% CI: -3.1 to 0.8/mm 3, p = 0.31), demonstrting tht LP re dignosed with HIV t similr level of immunodeficiency between 2010 nd Of 14,586 LP, 3,984 (27.3%) did not hve AIDS s prt of the LP definition nd hd t lest one CD4 count during follow-up before strting ART. Among these, for 1,067 (26.8%) the next CD4 count ws > 350/mm 3, suggesting they my not be LP or they my be seroconverters; this proportion ws highest for MSM (698/2,154; 32.4%), nd ws c 20% in ll other HIV exposure groups (p < ). There ws some evidence tht the proportion tht my be incorrectly clssified s LP hd incresed over time (OR 1.14/yer lter; 95% CI: ; p = ). This proportion of potentilly misclssified LP ws lower using confirmed CD4 count > 350/ mm 3 (317/1,279 (24.8%) with more thn two CD4 counts fter HIV dignosis nd before ART strted). Chnges in lte presenttion in HIV exposure groups nd regions of Europe Figure 3 summrises the chnge over time in LP mong HIV exposure groups. Mle nd femle PWID were combined due to smller numbers s were men nd women belonging to the other risk groups. There ws strong evidence to suggest tht the rte of chnge in LP differed between HIV exposure groups (p < , test for interction). After djustment, there ws no chnge over time in LP mong MSM, or mle or femle heterosexuls, but there ws significnt increse in LP mong PWID (both men nd women combined) (OR 1.16/yer lter; 95% CI: ; p = 0.024) nd in the other exposure groups (OR 1.08/yer lter; 95% CI: , p = 0.040). LP ws most common in centrl Europe (4,791/9,625, 49.8%), followed by northern (5,704/11,692; 48.8%), southern (3,550/7,760; 45.8%) nd estern Europe (541/1,377; 38.3%; p<0.0001). There were considerble differences in LP in countries within regions of cre in Europe (Tble 3), prticulrly within estern Europe. Figure 4 presents similr dt to Tble 3, strtified by region of cre in Europe, with evidence to suggest the rte of chnge in LP differed between regions (p < ; test for interction). There ws mrginlly significnt increse in LP over time in centrl Europe 6

7 Tble 2 Chrcteristics of included ptients, COHERE study, (n=30,454) All Lte presenters Ptients chrcteristics Percentge of lte N Percentge (%) N presenters (%) b All 30, , Delyed entry to No 26, , cre c Yes 1, HIV exposure group MSM 15, , Heterosexul men 4, , Heterosexul femles 5, , PWID (mle) PWID (femle) Mle other 2, , Femle other 1, Region of cre in Europe Southern 11, , Centrl 9, , Northern 7, , Estern 1, Europe 20, , Continent of origin Afric 2, , Other 2, , Unknown 4, , Medin IQR Medin IQR Age Yers CD4 /mm Bseline d Month/yer 6/2011 9/2010 4/2012 6/2011 9/2010 4/2012 COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe; IQR: interqurtile rnge; MSM: men who hve sex with men; PWID: people who inject drugs. % represents percentge of totl; for exmple, 15,371/30,454 (50.5%) of the popultion included were MSM. b % represents the percentge of lte presenters; for exmple, 5,993/15,371 (39.0%) of MSM were lte presenters. c Delyed entry into cre ws defined s more thn three months between HIV dignosis nd first visit to clinic, in people with both dtes recorded (n = 27,998). d Bseline ws defined s the erliest of HIV test, first study visit or cohort enrolment. Lte presenttion: dignosed with HIV with CD4 count below 350/mm 3 or n AIDS defining event regrdless of the CD4 count, in the six months following HIV-dignosis [8]. (OR 1.04/yer lter; 95% CI: , p = 0.084), nd significnt decrese in LP over time in northern Europe (OR 0.89; 95% CI: ; p < ). Clinicl disese progression During 39,790 person-yers of follow-up (PYFU) 886 (2.9%) people developed new AIDS defining illness or died, giving n incidence of clinicl progression of 22.3/1,000 PYFU (95% CI: ). A totl of 409 disese progression events were deth, 486 were new AIDS event, nd nine ptients hd both types of events on the sme dte. There were no differences in the proportion of events tht were ttributble to AIDS (63/125 (50.4%) vs 423/761 (55.6%); p=0.28), deths (63/125 (50.4%) vs 346/761 (45.5%); p=0.31), or in the specific AIDS events dignosed (p=0.053) compring LP nd non-lp. The incidence of clinicl progression ws 6.5-fold higher mong LP (761 events, 39.6/1,000 PYFU; 95% CI: ) compred with non-lp (125 events, 6.1/1,000 PYFU; 95% CI: ). There ws no evidence of ny chnge over time in the risk of developing new AIDS event or deth within the first six months of presenttion or fter this time (Tble 4). For exmple, fter djustment, there ws no chnge over time in the risk of developing new clinicl event per yer lter of presenttion fter six months of follow-up (djusted hzrd rtio (HR) 1.04/yer lter; 95% CI: , p = 0.40). There ws no evidence tht this reltionship differed in LP vs non-lp (test for interction > 0.2), nd the results re lso shown in Tble 4. These findings were lso consistent for 7

8 Tble 3 Lte presenttion, lte presenttion with dvnced disese, lte presenttion with very dvnced disese nd presenttion with n AIDS defining event t ny CD4 count strtified by country, COHERE study, Lte presenttion Lte presenttion with dvnced disese Lte presenttion with very dvnced disese Region of cre in Country N % 95% CI % 95% CI % 95% CI % 95% CI Europe Southern Southern 1 1, Southern 2 4, Southern 3 5, Centrl Centrl Centrl Centrl Centrl 4 2, Centrl 5 4, Northern Northern Northern 2 3, Northern 3 3, Estern Estern Estern Estern 3 1, AIDS COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe. Represents > 1 country within specific regions where countries with < 50 people hve been combined. Lte presenttion: dignosed with HIV with CD4 count below 350/mm3 or n AIDS defining event regrdless of the CD4 count, in the six months following HIV dignosis. Lte presenttion with dvnced disese: dignosed with HIV with CD4 count below 200/mm3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis. Lte presenttion with very dvnced disese: dignosed with HIV with CD4 count below 50/mm3 or n AIDS defining event, regrdless of CD4 cell count, in the six months following HIV dignosis. AIDS: presenttion with AIDS regrdless of CD4 count. different HIV exposure ctegories nd cross regions of cre in Europe nd when using deth lone s the clinicl endpoint. Discussion This study, which included more thn 30,000 people from cross 34 Europen countries, demonstrted no overll chnge in the proportion of LP cross Europe since LP incresed significntly in PWID s presumed HIV exposure. This lck of improvement in dignosing HIV erlier ws consistent cross wide rnge of nlyses; there ws no chnge over time in LP with dvnced disese, in the verge CD4 count t presenttion, or in progression to new AIDS event/deth. The proportion of LP decresed significntly cross Europe between 2000 nd 2010 [7], but despite numerous interventions nd inititives in recent yers to optimise testing for HIV, we found no overll chnge between 2010 nd 2013 in the proportion of LP cross Europe. LP is impcted by the underlying incidence of HIV which itself is difficult to estimte [5]. If HIV incidence increses nd HIV testing does not chnge, the proportion of LP will decline s more re dignosed erly. Conversely, if incidence is declining nd HIV testing does not chnge, the proportion of LP increses. Assuming the overll incidence of HIV is not decresing in Europe, s there ppers to hve been no decline in HIV dignoses per 100,000 popultion over the lst decde [4], our findings of no decrese in LP overll suggests there re res for further interventions for reducing LP on Europen level. A more detiled nlysis by region showed smll decrese over time in LP from northern Europe, but not from other regions. Combining countries into these regions ws decided priori nd used the strtifiction previously used by EuroSIDA [23]. Such brod grouping my not be idel for number of resons, including history, politics nd economy nd the rtes of LP within regions vried considerbly, reflecting this heterogeneity. HIV surveillnce in Europe for the Europen Union (EU)/Europen Economic Are (EEA) is coordinted by the Europen 8

9 Tble 4 Reltive hzrd of new AIDS defining event or deth following HIV dignosis per yer lter of presenttion, COHERE study, All Non-LP LP In first six months fter presenttion More thn six months fter presenttion HR 95% CI P HR 95% CI p Univrite Multivrite Univrite Multivrite Univrite Multivrite CI: confidence intervl; COHERE: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe; HR: hzrd rtio; LP: lte presenttion. Multivrite models were djusted for ge, HIV exposure group, region of cre in Europe, continent of origin, nd whether n AIDS dignosis ws present t bseline. The model in ll ptients djusts dditionlly for lte presenttion. All HR re per yer lter of presenting for cre. Centre for Disese Prevention nd Control (ECDC) in collbortion with the WHO Regionl Office for Europe nd the ECDC report displys LP in centrl-est nd estern Europe seprtely [4]; in this study, they were combined due to smll numbers from the centrl-est region. Further, combining southern, centrl nd northern Europe into one region in our study to compre to ECDC dt my hide importnt findings within these regions. Despite these differences in clssifiction, the proportion with LP in this study ws very similr to tht recently reported by the ECDC [4], who reported eight countries with > 50% s LP, including Greece nd Itly, both of which contributed significnt numbers to our nlyses. Possible ction points for incresing HIV wreness nd HIV testing, nd therefore minimising LP include combintion of both community-bsed nd providerinitited models for HIV testing, removl of stigmtistion, s well s working towrds cceptnce of verbl informed consent for testing [16,18,24,25]. Communitybsed testing should be priority, s should trgeting key popultions. Previous nlyses from COHERE showed n increse in LP mong PWID from southern nd estern Europe [7], trend tht continued in these nlyses. It is worth noting gin tht these dt re difficult to interpret; the number of new dignoses of HIV is declining in PWID in Europe [4] nd if HIV testing is stble this could led to n increse in the proportion of PIWD presenting lte. Although PWID ccount for comprtively smll proportion of new HIV infections in western Europe, this route of trnsmission is more common in estern Europe [4,7], nd issues continue to exist round needle exchnge, opite substitution therpy, s well s ccess to ART nd retention in cre once HIV hs been dignosed [26,27]. Recent dt suggest tht those in prison nd migrnts were mong those lest likely to be trgeted for HIV testing; with chllenges being providing HIV services nd support, lthough the ECDC-funded report by Deblonde et l. cknowledges excellence in some countries [28]. Further, PWID re more likely to fce greter brriers to ccessing helthcre nd to belong to lower socioeconomic groups nd hve lower levels of eduction, ll fctors known to be ssocited with poorer medicl outcomes [29,30]. Thus while there is evidence for brriers for PWID to ccess cre nd be retined in cre, there is much less evidence tht HIV is not dignosed, or tht dignosis of HIV or ccess to cre is even worsening. We found smll decline in the proportion of LP with very dvnced disese or in the proportion presenting with AIDS over time. The fct tht the CD4 count t presenttion hs remined stble over time my suggest tht helth systems re better ble to recognise nd cpture people with symptomtic HIV disese occurring t higher CD4 counts nd tht symptomtic ptients re not routinely dignosed with HIV, especilly in groups t low risk of HIV infection. Evidence from other studies concerning chnges over time in CD4 count t presenttion in recent times hve shown mixed results [31-34]; some hve limited dt from 2011 nd others hve not been ble to djust for confounding vribles. Other studies hve described decrese in the proportion presenting with AIDS [35,36], lthough the extent to which this is due to the under-reporting of AIDS is unknown. We found no evidence of chnge over time in shortterm clinicl progression (within six months) or fter tht time in ll people or in LP nd non-lp considered seprtely, in different regions of Europe or HIV exposure ctegories, lthough medin follow-up ws limited by only including people dignosed with HIV-1 since The gretest risk in clinicl progression for LP hs been observed in the yers immeditely following LP [7,10,37,38], nd in this study, LP hd pproximtely six-fold higher incidence of clinicl progression. A lck of chnge in clinicl outcomes in LP over clendr time suggests tht, once people hve ccessed cre, tretment nd outcomes re uniform cross vriety of settings. Given the poor outcomes fter LP, work is needed to reduce the proportion of those presenting 9

10 lte to reduce morbidity nd mortlity ssocited with HIV, s well s to reduce the finncil impct on helth systems nd onwrd trnsmission of HIV [9-11]. It is possible tht some people presenting with symptomtic seroconversion for HIV re misclssified s LP due to the trnsient drop in CD4 count occurring t this time [39]. Misclssifiction of LP my be highest in MSM [40]. In our study, there ws smll subset with CD4 counts fter HIV dignosis but before strting ntiretrovirl therpy. Approximtely 20% did not hve CD4 count of < 350/mm 3 t the next CD4 count, nd this proportion ws highest for MSM. This proportion ws similr when using confirmed CD4 count of > 350/ mm 3, suggesting this is not lrgely due to lbortory vrition, lthough it is worth noting tht only smll subset of people hd one or two CD4 counts fter HIV dignosis nd before strting ART. In ddition, only 5% hd CD4 count >500/mm 3 t the second mesurement, which is higher thn the currently recommended threshold for initition of ntiretrovirl therpy [12]. In ddition, we found no chnges over time in the proportion of LP, presenttion with dvnced or very dvnced disese, suggesting tht n incresing proportion of primry HIV infections is unlikely to explin the lck of chnge in LP in recent yers. There re number of limittions which should be considered. We re likely underestimting LP s people who do not survive long enough to hve CD4 count mesured were excluded [41]. Our dt suggest this ws more likely in PWID, other HIV exposure groups, those under cre in northern nd estern Europe, nd ffected c 10% of those in the COHERE cohorts. This is considerbly lower thn reported by surveillnce studies [42], highlighting tht cohort studies such s COHERE cn supplement informtion vilble from the WHO or ECDC. Furthermore, cohorts prticipting in COHERE tend to be receiving helthcre t centres of excellence nd clinic-bsed cohorts rther thn nonclinic outptient settings, where LP my be higher. We excluded seroconverter cohorts prticipting in COHERE s in our previous work, where inclusion of these cohorts did not lter our findings [7]. Even in collbortion s lrge s COHERE we were not ble to consider LP for mle nd femle PWID seprtely, lthough it is worth noting tht there ws no sttisticlly significnt rise in LP in ny one region, suggesting tht the problem of LP in IDUs is not limited to one region of Europe, but potentil problem on wider scle. In conclusion, LP cross Europe ccount for lmost 50% of HIV dignoses with no evidence of chnge since Incresed HIV testing, with focus on vulnerble groups, will reduce the hrm for the individul nd it my s well reduce onwrd trnsmission. Erlier dignosis for HIV is n importnt component of chieving the UNAIDS trget of ending the AIDS epidemic by 2030 [1]. The Collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) group Anlysis nd Writing Committee: Amnd Mocroft, Jens Lundgren, Andre Antinori, Antonell d Arminio Monforte, Johnn Brännström, Fbrice Bonnet, Norbert Brockmeyer, Jordi Csbon, Antonell Cstgn, Dominique Costgliol, Stéphne De Wit, Gerd Fätkenheuer, Hnsjkob Furrer, Corinne Jdnd, Anne Johnson, Mrio Lzns, Ctherine Leport, Sntigo Moreno, Christin Mussini, Niels Obel, Frnk Post, Peter Reiss, Croline Sbin, Adrine Skletz-Rorowski, Igncio Surez-Lono, Crlo Torti, Josine Wrszwski, Lind Wittkop, Robert Zngerle, Genevieve Chene, Dorthe Rben, Ole Kirk. Conflict of interest None declred. Authors contributions Amnd Mocroft, Ole Kirk nd Jens Lundgren proposed nd developed the project. Amnd Mocroft performed the nlyses nd wrote the first drft of the mnuscript. All other uthors contributed to discussions of results, interprettion of dt nd contributed to writing the mnuscript. Acknowledgements Members of the collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) group: COHERE Steering Committee Contributing Cohorts: Robert Zngerle (AHIVCOS),Giot Touloumi (AMACS), Josine Wrszwski (ANRS CO1 EPF/ANRS CO11 OBSERVATOIRE EPF), Lurence Meyer (ANRS CO2 SEROCO), Frnçois Dbis (ANRS CO3 AQUITAINE), Murielle Mry Kruse (ANRS CO4 FHDH), Jde Ghosn (ANRS CO6 PRIMO), Ctherine Leport (ANRS CO8 COPILOTE), Lind Wittkop (ANRS CO13 HEPAVIH), Peter Reiss (ATHENA), Ferdinnd Wit (ATHENA), Mri Prins (CASCADE), Heiner Bucher (CASCADE), Din Gibb (CHIPS), Gerd Fätkenheuer (Cologne-Bonn), Juli Del Amo (CoRIS), Niels Obel (Dnish HIV Cohort), Clire Thorne (ECS), Amnd Mocroft (EuroSIDA), Ole Kirk (EuroSIDA), Christoph Stephn (Frnkfurt), Sntigo Pérez-Hoyos (GEMES-Hemo), Osmh Hmoud (Germn ClinSurv), Brbr Brtmeyer (Germn ClinSurv), Nikoloz Chkhrtishvili (Georgin Ntionl HIV/ AIDS), Antoni Noguer-Julin (CORISPE-ct), Andre Antinori (ICC), Antonell d Arminio Monforte (ICONA), Norbert Brockmeyer (KOMPNET), Luis Prieto (Mdrid PMTCT Cohort), Pblo Rojo Conejo (CORISPES-Mdrid), Antoni Sorino- Arndes (NENEXP), Mnuel Bttegy (SHCS), Roger Kouyos (SHCS), Cristin Mussini (Moden Cohort), Pt Tookey (NSHPC), Jordi Csbon (PISCIS), Jose M. Miró (PISCIS), Antonell Cstgn (Sn Rffele), Deborh Konopnick (St. Pierre Cohort), Tess Goetghebuer (St Pierre Peditric Cohort), Anders Sönnerborg (Swedish InfCre), Crlo Torti (The Itlin Mster Cohort), Croline Sbin (UK CHIC), Rmon Teir (VACH), Myrim Grrido (VACH). Dvid Herry (Europen AIDS Tretment Group) Executive Committee: Stéphne de Wit (Chir, St. Pierre University Hospitl), Jose M. Miró (PISCIS), Dominique Costgliol (FHDH), Antonell d Arminio-Monforte (ICONA), Antonell Cstgn (Sn Rffele), Juli del Amo (CoRIS), Amnd Mocroft (EuroSid), Dorthe Rben (Hed, Copenhgen Regionl Coordinting Centre), Geneviève Chêne (Hed, Bordeux Regionl Coordinting Centre). Peditric Cohort Representtives: Ali Judd, Pblo Rojo Conejo. Regionl Coordinting Centres: Bordeux RCC: Din Brger, Christine Schwimmer, Monique Termote, Lind Wittkop; 10

11 Copenhgen RCC: Mri Cmpbell, Csper M. Frederiksen, Nin Friis-Moller, Jesper Kjer, Dorthe Rben, Rikke Slbøl Brndt. Project Leders nd Sttisticins: Jun Berenguer, Juli Bohlius, Vincent Bouteloup, Heiner Bucher, Alessndro Cozzi-Lepri, Frnçois Dbis, Antonell d Arminio Monforte, Mry-Anne Dvies, Juli del Amo, Mri Dorrucci, Dvid Dunn, Mtthis Egger, Hnsjkob Furrer, Mrguerite Guiguet, Sophie Grbr, Ali Judd, Ole Kirk, Olivier Lmbotte, Vlérine Leroy, Sr Lodi, Sophie Mtheron, Lurence Meyer, Jose M. Miró, Amnd Mocroft, Susn Monge, Fumiyo Nkgw, Roger Predes, Andrew Phillips, Mssimo Puoti, Michel Schomker, Colette Smit, Jonthn Sterne, Rodolphe Thiebut, Clire Thorne, Crlo Torti,, Mrc vn der Vlk, Lind Wittkop, Ntsh Wyss. The COHERE cknowledgement ppendix my be found t: Funding: The COHERE study group hs received unrestricted funding from: Agence Ntionle de Recherches sur le SIDA et les Héptites Virles (ANRS), Frnce; HIV Monitoring Foundtion, The Netherlnds; the Augustinus Foundtion, Denmrk nd the Dnish Ntionl Reserch Foundtion (grnt number DNRF126). COHERE receives funding from the Europen Union Seventh Frmework Progrmme (FP7/ ) under EuroCoord grnt greement n A list of the funders of the prticipting cohorts cn be found t The study sponsors hd no role in the design of the study, the collection, nlysis nd interprettion of dt, the writing of the report or the decision to submit the pper for publiction. References 1. UNAIDS. Fst Trck: Ending the AIDS epidemic by Genev: UNAIDS Avilble from: JC2686_WAD2014report_en.pdf 2. Rymond A, Hill A, Poznik A. Lrge disprities in HIV tretment cscdes between eight Europen nd high-income countries - nlysis of brek points.j Int AIDS Soc. 2014;17(4) Suppl 3;19507.PMID: Chkhrtishvili N, Shrvdze L, Chokoshvili O, DeHovitz JA, del Rio C, Tsertsvdze T. The cscde of cre in the Estern Europen country of Georgi.HIV Med. 2015;16(1):62-6. DOI: /hiv PMID: Europen Centre for Disese Prevention nd Control (ECDC). HIV/AIDS surveillnce in Europe Stockholm: ECDC Avilble from: Publictions/hiv-ids-surveillnce-report-Europe-2013.pdf 5. Working Group on Estimtion of HIV Prevlence in Europe,. HIV in hiding: methods nd dt requirements for the estimtion of the number of people living with undignosed HIV.AIDS. 2011;25(8): PMID: ICoNA Behviourl Epidemiology Study Group,Girrdi E, Aloisi MS, Arici C, Pezzotti P, Serrino D, Blzno R, et l.. Delyed presenttion nd lte testing for HIV: demogrphic nd behviorl risk fctors in multicenter study in Itly.J Acquir Immune Defic Syndr. 2004;36(4): DOI: / PMID: Collbortion of Observtionl HIV Epidemiologicl Reserch Europe (COHERE) study in EuroCoord,Mocroft A, Lundgren JD, Sbin ML, Monforte A, Brockmeyer N, Csbon J, et l.. Risk fctors nd outcomes for lte presenttion for HIVpositive persons in Europe: results from the Collbortion of Observtionl HIV Epidemiologicl Reserch Europe Study (COHERE).PLoS Med. 2013;10(9):e DOI: / journl.pmed PMID: Europen Lte Presenter Consensus Working Group,Antinori A, Coenen T, Costgiol D, Dedes N, Ellefson M, Gtell J, et l.. Lte presenttion of HIV infection: consensus definition.hiv Med. 2011;12(1):61-4. DOI: /j x PMID: HIV Reserch Network,Fleishmn JA, Yehi BR, Moore RD, Gebo KA. The economic burden of lte entry into medicl cre for ptients with HIV infection.med Cre. 2010;48(12): DOI: /MLR.0b013e3181f81c4 PMID: ANRS C004 French Hospitl Dtbse on HIV Clinicl Epidemiologicl Group,Lnoy E, Mry-Kruse M, Tttevin P, Perbost I, Poizot-Mrtin I, Dupont C, et l.. Frequency, determinnts nd consequences of delyed ccess to cre for HIV infection in Frnce.Antivir Ther. 2007;12(1):89-96.PMID: Rki Project Study Group,Quinn TC, Wwer MJ, Sewnkmbo N, Serwdd D, Li C, Wbwire-Mngen F, et l.. Virl lod nd heterosexul trnsmission of humn immunodeficiency virus type 1.N Engl J Med. 2000;342(13): DOI: / NEJM PMID: World Helth Orgniztion (WHO). Guideline on when to strt ntiretrovirl therpy nd on pre-exposure prophylxis for HIV. Sep Genev: WHO. Avilble from: iris/bitstrem/10665/186275/1/ _eng.pdf 13. Eurosid in Eurocoord,Lut KG, Mocroft A, Lzrus J, Reiss P, Rockstroh J, Krpov I, et l.. Regionl differences in selfreported HIV cre nd mngement in the EuroSIDA study.j Int AIDS Soc. 2014;17(4) Suppl 3;19504.PMID: INSIGHT START Study Group,Lundgren JD, Bbiker AG, Gordin F, Emery S, Grund B, Shrm S, et l.. Initition of Antiretrovirl Therpy in Erly Asymptomtic HIV Infection.N Engl J Med. 2015;373(9): DOI: /NEJMo PMID: World Helth Orgniztion (WHO). Globl updte on HIV tretment 2013: Results, impct nd opportunities. Jun Genev: WHO. Avilble from: bitstrem/10665/85326/1/ _eng.pdf?u=1 16. Europen Centre for Disese Prevention nd Control (ECDC). HIV testing: incresing uptke nd effectiveness in the Europen Union. Stockholm: ECDC Avilble from: GUI_HIV_testing.pdf 17. Sullivn AK, Rben D, Reekie J, Ryment M, Mocroft A, Esser S, et l. Fesibility nd effectiveness of indictor conditionguided testing for HIV: results from HIDES I (HIV indictor diseses cross Europe study). PLoS ONE. 2013;8(1):e DOI: /journl.pone PMID: World Helth Orgniztion (WHO) Regionl Office for Europe. Scling up HIV testing nd counseling in the WHO Europen Region - s n essentil component of efforts to chieve universl ccess to HIV prevention, tretment, cre nd support. Policy frmework. Copenhgen: WHO Europe Avilble from: dt/ssets/pdf_ file/0007/85489/e93715.pdf?u=1 19. Kjer J, Ledergerber B. HIV cohort collbortions: proposl for hrmoniztion of dt exchnge.antivir Ther. 2004;9(4): PMID: Europen Network of HIV/AIDS cohort studies to coordinte t Europen nd Interntionl level clinicl reserch on HIV/ AIDS (EuroCoord). COHERE - Collbortion of Observtionl HIV Epidemiologicl Reserch in Europe. London: EuroCoord. [Accessed Nov 2015]. Avilble from: net/prtners/founding_networks/cohere.spx 21. Centre for Helth nd Infectious Disese Reserch (CHIP). EuroSIDA. Copenhgen: CHIP. [Accessed 23 Nov 2015]. Avilble from: EuroSIDA/About 22. EuroSIDA Study Group,Bnnister WP, Kirk O, Gtell JM, Knysz B, Vird JP, Mens H, et l.. Regionl chnges over time in initil virologic response rtes to combintion ntiretrovirl therpy cross Europe.J Acquir Immune Defic Syndr. 2006;42(2): DOI: /01.qi PMID: EuroSIDA Study Group,Bnnister WP, Ruiz L, Lovedy C, Vell S, Zilmer K, Kjer J, et l.. HIV-1 subtypes nd response to combintion ntiretrovirl therpy in Europe.Antivir Ther. 2006;11(6): PMID: Gökengin D, Geretti AM, Begovc J, Plfreemn A, Stevnovic M, Trsenko O, et l Europen Guideline on HIV testing. Int J STD AIDS. 2014;25(10): DOI: / PMID: World Helth Orgniztion (WHO). Guidnce on providerinitited HIV testing nd counselling in helth fcilities. Genev: WHO Avilble from: bitstrem/10665/43688/1/ _eng.pdf 26. Degenhrdt L, Mthers BM, Wirtz AL, Wolfe D, Kmrulzmn A, Crrieri MP, et l. Wht hs been chieved in HIV prevention, tretment nd cre for people who inject drugs, ? A review of the six highest burden countries. Int J Drug Policy. 2014;25(1): DOI: /j. drugpo PMID:

12 27. Kztchkine M. Russi s bn on methdone for drug users in Crime will worsen the HIV/AIDS epidemic nd risk public helth. BMJ. 2014;348(my08 1):g Deblonde J, Meulemns H, Cllens S, Luchters S, Temmermn M, Hmers FF. HIV testing in Europe: mpping policies. Helth Policy. 2011;103(2-3): DOI: /j. helthpol PMID: Dry-Spir R, Lert F. Socil helth inequlities during the course of chronic HIV disese in the er of highly ctive ntiretrovirl therpy.aids. 2003;17(3): DOI: / PMID: Socio-economic Inequlities nd HIV Writing Group for Collbortion of Observtionl HIV Epidemiologicl Reserch in Europe (COHERE) in EuroCoord,Lodi S, Dry-Spir R, Touloumi G, Brun D, Teir R, D Arminio Monforte A, et l.. Delyed HIV dignosis nd initition of ntiretrovirl therpy: inequlities by eductionl level, COHERE in EuroCoord.AIDS. 2014;28(15): PMID: Helth Protection Agency (HPA),. HIV in the United Kingdom: 2010 report.hpr. 2010;4(47): UK Collbortive HIV Cohort (CHIC) Study Steering Committee,. HIV dignosis t CD4 count bove 500 cells/mm3 nd progression to below 350 cells/mm3 without ntiretrovirl therpy.j Acquir Immune Defic Syndr. 2007;46(3): DOI: /QAI.0b013e PMID: HIV Reserch Network,Hines CF, Fleishmn JA, Yehi BR, Berry SA, Moore RD, Bmford LP, et l.. Increse in CD4 count mong new enrollees in HIV cre in the modern ntiretrovirl therpy er.j Acquir Immune Defic Syndr. 2014;67(1): DOI: /QAI PMID: Lesko CR, Cole SR, Zinski A, Poole C, Mugvero MJ. A systemtic review nd met-regression of temporl trends in dult CD4(+) cell count t presenttion to HIV cre, Clin Infect Dis. 2013;57(7): DOI: /cid/cit421 PMID: Mirnd AC, Moneti V, Brogueir P, Peres S, Bptist T, Aldir I, et l. Evolution trends over three decdes of HIV infection lte dignosis: the experience of Portuguese cohort of 705 HIVinfected ptients. J Int AIDS Soc. 2014;17(4) Suppl 3; PMID: Jkopnec I, Grjibovski AM, Nilsen Ø, Blystd H, Avitslnd P. Trends in HIV infection surveillnce dt mong men who hve sex with men in Norwy, BMC Public Helth. 2013;13(1):144. DOI: / PMID: NADIS Group,Delpierre C, Luwers-Cnces V, Pugliese P, Poizot-Mrtin I, Billud E, Duvivier C, et l.. Chrcteristics trends, mortlity nd morbidity in persons newly dignosed HIV positive during the lst decde: the profile of new HIV dignosed people.eur J Public Helth. 2008;18(3): DOI: /eurpub/ckm117 PMID: Helleberg M, Engsig FN, Kronborg G, Lursen AL, Pedersen G, Lrsen O, et l. Lte presenters, repeted testing, nd missed opportunities in Dnish ntionwide HIV cohort. Scnd J Infect Dis. 2012;44(4): DOI: / PMID: Pntleo G, Grziosi C, Fuci AS. New concepts in the immunopthogenesis of humn immunodeficiency virus infection.n Engl J Med. 1993;328(5): DOI: / NEJM PMID: Ssse A, Florence E, Phrris A, De Wit S, Lcor P, Vn Beckhoven D, et l. Belgin Reserch AIDS & HIV Consortium (BREACH). Lte presenttion to HIV testing is overestimted when bsed on the consensus definition. HIV Med. 2015; 41. Mocroft A. Lte presenttion to HIV/AIDS testing, tretment or continued cre: clrifying the use of CD4 evlution in the consensus definition.hiv Med. 2014;15(3):129. DOI: / hiv PMID: Liktvicius G, vn de Lr MJ. HIV infection nd AIDS in the Europen Union nd Europen Economic Are, Euro Surveill. 2011;16(48):16(48):