J Antimicrob Chemother 2014; 69: 2826 – 2834 doi:10.1093/jac/dku190 Advance Access publication 16 June 2014
Factors associated with mortality among persistently viraemic triple-antiretroviral-class-experienced patients receiving antiretroviral therapy in the HIV Outpatient Study (HOPS) Frank J. Palella Jr1*, Carl Armon2, Kate Buchacz3, Joan S. Chmiel1, Richard M. Novak4, Richard T. D’Aquila1 and John T. Brooks3 on behalf of the HOPS Investigators† 1
Northwestern University, Chicago, IL, USA; 2Cerner Corporation, Vienna, VA, USA; 3Centers for Disease Control and Prevention, Atlanta, GA, USA; 4University of Illinois at Chicago, Chicago, IL, USA
Received 22 October 2013; returned 16 December 2013; revised 17 April 2014; accepted 3 May 2014 Background: Identifying factors associated with mortality for HIV-infected patients with persistent viraemia despite antiretroviral (ARV) therapy may inform diagnostic and treatment strategies. Methods: We analysed data from viraemic triple-ARV-class-experienced HIV Outpatient Study patients seen during 1 January 1999 to 31 December 2012 who, despite treatment that included ARVs from three major drug classes [nucleoside analogue reverse transcriptase inhibitors, non-nucleoside analogue reverse transcriptase inhibitors and protease inhibitors (PIs)], had plasma HIV RNA levels [viral load (VL)] .1000 copies/mL [‘triple ARV class failure’ (TCF)]. The baseline was defined as the date of meeting the TCF criteria during 1999 – 2008. We identified factors associated with mortality using Cox regression. Results: Of 597 patients who met the TCF criteria (median follow-up after baseline 4.9 years), 115 (19.3%) died. Baseline factors associated with mortality were age per 10 years [hazard ratio (HR) 1.61, 95% CI 1.28 –2.02], risk of HIV from use of injection drugs (HR 1.81, 95% CI 1.10 –2.98), CD4+ T cell count ,200 cells/mm3 (HR 3.68, 95% CI 2.41–5.62), VL ≥5.0 log10 copies/mL (HR 2.91, 95% CI 1.88–4.49) and receiving a first combination ARV therapy regimen that was PI-based (HR 2.44, 95% CI 1.47 –4.06); receiving a novel ARV agent during follow-up (HR 0.45, 95% CI 0.22 –0.93) was protective. Genotypic resistance testing results were available for 274 (45.9%) of the TCF patients, of whom 47 (17.2%) died. In this group, factors associated with death were increasing age (HR 1.94, 95% CI 1.36 – 2.78, per 10 year increment), risk of HIV from use of injection drugs (HR 2.71, 95% CI 1.37 –5.39), baseline VL ≥5.0 log10 copies/mL (HR 5.35, 95% CI 2.82–10.1) and receiving PI-based first combination ARV therapy regimen (HR 3.20, 95% CI 1.25–8.17). No HIV mutations or combinations of mutations were significantly associated with survival. Conclusions: Factors significantly associated with mortality risk among TCF patients who received ongoing ARV therapy included traditional clinical predictors but not the presence, type or number of HIV genetic mutations. The use of novel ARV drugs by these ARV therapy-experienced patients was associated with an improved survival. Keywords: ARV, antiretroviral resistance, viraemia
Introduction The optimal management of extensively antiretroviral (ARV) therapy-experienced HIV-infected persons who remain viraemic despite continued ARV therapy remains a clinical challenge.1 – 4 Identifying factors associated with improved clinical outcomes among such patients can inform ARV treatment strategies. Existing data indicate that ARV resistance among both ARVtreated and ARV-naive persons is associated with an increased
risk of death,5,6 suggesting that strategies aimed at reducing the induction of resistance may improve survival. As new ARV agents from diverse drug classes become available, routine ARV resistance testing among viraemic treatment-experienced patients has become a standard of care to guide ARV selection.1,3,7 – 9 In this report, using data from the HIV Outpatient Study (HOPS), we evaluated the associations between mortality risk, patient factors and ARV resistance mutations in a diverse group of highly ARV
# The Author 2014. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
2826
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
*Corresponding author. Tel: +1-312-695-5053; E-mail: [email protected] †The HOPS Investigators are listed in the Acknowledgements section.
Mortality factors among viraemic ARV-experienced HIV patients
treatment-experienced persons who remained viraemic despite ongoing ARV exposure.
Methods The HOPS
Study population and definitions We analysed data from participants with at least two clinical encounters who met the criteria for triple ARV class failure (TCF; defined below) between 1 January 1999 and 31 December 2008 and contributed observation time during 1 January 1999 to 31 December 2012, using the dataset updated 31 December 2013. We defined triple-class-experienced (TCE) as having received a prescription for ≥4 continuous months of an ARV from each of the three major drug classes [nucleoside analogue reverse transcriptase inhibitors (NRTIs), non-nucleoside analogue reverse transcriptase inhibitors (NNRTIs) and protease inhibitors (PIs), although not necessarily simultaneously]. We limited our analyses to patients who had a plasma HIV RNA level [viral load (VL)] .1000 copies/mL (VL .3.0 log10 copies/mL) after becoming TCE and while receiving ARV therapy. We defined such patients as TCF. The baseline date was defined as the earliest date at which a VL .1000 copies/mL was seen after meeting the TCE criteria during 1999 –2008. Patients could have met the TCF definition prior to 1999 if they contributed observation time after 1 January 1999. We conducted further analyses for the subset of TCF patients for whom results of HIV genotypic testing performed on or after the baseline date were available, incorporating resistance results from this first postbaseline genotypic testing as well as results from prior resistance testing for these patients. Thymidine analogue NRTI-associated resistance mutations (TAMs) were M41L, D67N, K70E/R, L210W, T215F/Y and K219E/Q. Novel ARV agents were fusion inhibitors, integrase strand transfer inhibitors, entry inhibitors and CCR5 attachment inhibitors.
Statistical analyses We used Cox proportional hazards regression models to examine risk factors for mortality among TCF patients. Observation began at baseline, continued until death (if documented within 180 days of last patient contact) and was censored (the patient was assumed to be alive) at the last patient contact plus 180 days if the patient was not known to be deceased by then, or at 31 December 2012, whichever occurred first. Among a subgroup of TCF patients with genotypic resistance testing, we analysed cumulative mutations at first genotypic testing following TCF as well as mutations detected by prior genotypic testing studies. We evaluated the associations between mortality and both the total number and
individual types of mutations, after adjusting for other clinical and demographic factors. For these subgroup analyses, the start of observation was reset to the date of the first post-baseline genotypic testing. For both unadjusted and adjusted Cox regression models, the results are reported using hazard ratios (HRs) and 95% CIs. To develop multivariable regression models for mortality after the start of observation, variables with P values ,0.20 in univariate analyses were considered for inclusion in multivariable models; manual backwards selection was then used to eliminate non-significant variables sequentially in order of highest P value. We estimated the proportion of the observation time during which patients were viraemic (defined as VL ≥200 copies/mL) by summing the duration of observation times around a given detectable VL; a patient was considered to be viraemic from the mid-point in time between the prior VL test and the ‘current’ time, to the mid-point in time between the ‘current’ time and the next VL test. We summed the time spent as viraemic across patients included in the analysis. All the descriptive summaries, survival plots and regression analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA). Statistical significance was defined as a two-sided P value ,0.05 for statistical tests and, for an HR, when the associated 95% CI excluded 1.0.
Results Of 3840 HOPS patients with at least two clinical encounters between 1 January 1999 and 31 December 2008 and whose ARV prescription history was known, 1536 patients met the definition of TCE; of these, 597 (38.9%) met the definition of TCF. Of the 597 patients, 158 (26.5%) first met the TCF criteria in 1998 or earlier and 439 (73.5%) met them during 1999 – 2008. Among the TCF patients, the median baseline age was 41 years, 77.2% were male, 79.9% had a prior diagnosis of AIDS and 79.9% had been exposed to mono or dual ARV therapy prior to their first combination ARV therapy (cART). The median baseline and nadir CD4+ T cell counts were 283 and 105 cells/mm3 , respectively. Most (71.4%) patients had been prescribed a first cART regimen that was PI-based. The median observation time after baseline was 6.2 years (IQR 3.1–10.5) (Table 1). Of the 597 TCF patients, 115 (19.3%) died (Table 1), giving a mortality rate of 2.77 deaths per 100 person-years of follow-up. In the final multivariable Cox regression model, age per 10 years (HR 1.61, 95% CI 1.28 – 2.02), risk of HIV from use of injection drugs (HR 1.81, 95% CI 1.10 – 2.98), baseline CD4+ T cell count ,200 cells/mm3 (HR 3.68, 95% CI 2.41 – 5.62), baseline plasma HIV RNA ≥5.0 log10 copies/mL (HR 2.91, 95% CI 1.88 –4.49) and having received a first cART regimen that was PI-based (HR 2.44, 95% CI 1.47 – 4.06) were associated with an increased risk of death, while receiving a novel ARV agent during follow-up (HR 0.45, 95% CI 0.22 – 0.93) was associated with an decreased risk of death (Table 2). Among the 597 TCF patients, 556 (93.1%) had a VL measurement that was ≥200 copies/mL after meeting the TCF criteria; these patients remained viraemic above this level during 68% of their post-baseline follow-up. A total of 13/597 (2.2%) of patients had subsequent measurements that were all ,200 copies/mL and 28/597 (4.7%) had no subsequent VL recorded. The 556 viraemic patients had a median of 17 (IQR 8 –29) VL measurements after meeting the TCF criteria, with median and mean durations of viraemia after baseline of 2.7 (IQR 1.1 – 4.9) and 3.4 years, respectively. Among the 597 TCF patients, 156 patients received a ‘novel’ ARV agent (i.e. one other than an NRTI, NNRTI or PI
2827
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
The HOPS is an ongoing prospective cohort study of HIV-infected adults receiving care at participating HIV clinics. For this analysis, data were included from 12 clinics (university-based, private and public) in eight US cities (Chicago, IL, USA; Denver, CO, USA; Stony Brook, NY, USA; Oakland/ San Leandro, CA, USA; Walnut Creek, CA, USA; Philadelphia, PA, USA; Tampa, FL, USA; and Washington, DC, USA) since 1993.10 Patient data, including sociodemographic characteristics, diagnoses, treatments and laboratory values, are abstracted from medical charts and entered into an electronic database by trained staff. These data are reviewed for quality and analysed centrally. HOPS investigators have collected information on more than 10 000 patients seen at over 400 000 clinical encounters. The HOPS protocol has been reviewed annually and approved by the CDC (Atlanta, GA, USA), the Cerner Corporation (Vienna, VA, USA) and the institutional review board at each participating site. The study protocol conforms to the guidelines of the US Department of Health and Human Services for the protection of human subjects in research.
JAC
Palella et al.
Table 1. Characteristics of ARV-experienced patients who experienced TCF, the HOPS 1999 –2008 All TCF patients (N ¼597)
TCF with genotype testa (n ¼274)
TCF without genotype test (n¼323)
P
Characteristics as of baseline dateb Male, n (%) Age (years), median (IQR) Age .40 years, n (%)
461 (77.2) 41 (36 –46) 339 (56.8)
201 (73.4) 41 (36 –46) 154 (56.2)
260 (80.5) 41 (36 –46) 185 (57.3)
Race or ethnicity, n (%) non-Hispanic white non-Hispanic black Hispanic other/unknown race/ethnicity
291 (48.7) 230 (38.5) 62 (10.4) 14 (2.4)
131 (47.8) 104 (38.0) 34 (12.4) 5 (1.8)
160 (49.5) 126 (39.0) 28 (8.7) 9 (2.8)
HIV risk category (non-exclusive categories), n (%) men who have sex with men injection drug use heterosexual other/unknown HIV risk
324 (54.3) 67 (11.2) 172 (28.8) 34 (5.7)
138 (50.4) 30 (11.0) 88 (32.1) 18 (6.6)
186 (57.6) 37 (11.5) 84 (26.0) 16 (5.0)
Insurance/payer at date of TCF, n (%) private public, other or unknown
291 (48.7) 306 (51.3)
127 (46.4) 147 (53.7)
164 (50.8) 159 (49.2)
Calendar year patient first fulfilled criteria for TCF, n (%) 1998 or earlier 1999 –2001 2002 –2004 2005 –2008
158 (26.5) 226 (37.9) 128 (21.4) 85 (14.2)
65 (23.7) 101 (36.9) 61 (22.3) 47 (17.2)
93 (28.8) 125 (38.7) 67 (20.7) 38 (11.8)
477 (79.9) 105 (31 –219) 215 (93 –361)
217 (79.2) 111 (36 –227) 226 (103– 368)
260 (80.5) 99 (26 –211) 206 (88 –350)
0.77 0.22 0.17
283 (145– 417) 4.1 (3.4 –4.8) 7.4 (4.6 –11.1) 4.4 (2.7 –6.5) 89 (14.9) 365 (61.1) 425 (71.2) 477 (79.9) 4.3 (2.5 –6.6) 0.8 (0.5 –1.6) 1.9 (1.1 –3.2)
289 (171– 467) 4.1 (3.4– 4.8) 7.1 (4.4– 11.1) 4.2 (2.5– 6.2) 33 (12.0) 151 (55.1) 183 (66.8) 209 (76.3) 4.2 (2.5– 6.2) 0.9 (0.5– 1.5) 1.9 (1.2– 3.6)
267 (127– 395) 4.2 (3.4– 4.9) 7.5 (4.7– 11.1) 4.5 (2.8– 6.6) 56 (17.3) 214 (66.3) 242 (74.9) 268 (83.0) 4.5 (2.5– 6.7) 0.8 (0.5– 1.7) 1.8 (1.1– 3.0)
0.030 0.38 0.59 0.26 0.09 0.007 0.036 0.053 0.52 0.90 0.18
51 (8.5) 115 (19.3) 426 (71.4) 5 (0.8)
15 (5.5) 55 (20.1) 200 (73.0) 4 (1.5)
36 (11.2) 60 (18.6) 226 (70.0) 1 (0.3)
5 (0.8)
3 (1.1)
2 (0.6)
0.048 0.80 0.86 0.44
First cART regimen type, n (%) NNRTI+PI NNRTI PI three or more NRTIs
0.32
0.19
0.035
Additional first cART regimen novel agentsc, n (%)
0.67
b
Characteristics after baseline date Years follow-up post-baseline, median (IQR) Persons who died within 6 months of baseline, n (%) Persons who died within 12 months of baseline, n (%)
6.2 (3.1 –10.5) 7 (1.2) 18 (3.0)
8.1 (5.3– 12.1) 0 (0.0) 3 (1.1)
4.5 (1.8– 8.7) 7 (2.2) 15 (4.6)
,0.001 0.017 0.022 Continued
2828
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
Prior AIDS diagnosis, n (%) Nadir CD4+ T cell count (cells/mm3), median (IQR) CD4+ T cell count at cART initiation (cells/mm3), median (IQR) CD4+ T cell count (cells/mm3), median (IQR) HIV VL (log10 copies/mL), median (IQR) Years since HIV diagnosis, median (IQR) Years since AIDS diagnosis (n¼477), median (IQR) Co-infected with hepatitis C, n (%) Exposed to mono ARV therapy, n (%) Exposed to dual ARV therapy, n (%) Exposed to mono ARV or dual ARV therapy, n (%) Years of prior NRTI use, median (IQR) Years of prior NNRTI use, median (IQR) Years of prior PI use, median (IQR)
0.26
JAC
Mortality factors among viraemic ARV-experienced HIV patients
Table 1. Continued All TCF patients (N ¼597) Persons who died during follow-up, n (%)d cause of death AIDS non-AIDS unknown cause of death
Percentage of time since TCF with HIV RNA ≥200 copies/ mL, median (IQR)
TCF without genotype test (n¼323)
115 (19.3)
47 (17.2)
68 (21.1)
56 (48.7) 40 (34.8) 19 (16.5)
20 (42.6) 17 (36.2) 10 (21.3)
36 (52.9) 23 (33.8) 9 (13.2)
595 (99.7) 499 (83.6) 546 (91.5) 156 (26.1)
274 (100.0) 233 (85.0) 262 (95.6) 119 (43.4)
321 (99.4) 266 (82.4) 284 (87.9) 37 (11.5)
74 (38 –100)
65 (38 –97)
88 (35 –100)
P 0.27 0.42
0.50 0.44 0.001 ,0.001 0.002
a
First genotype test at or after baseline date. Baseline date: the earliest date during 1999– 2008 when a triple-ARV-class-experienced patient had an HIV VL .1000 copies/mL (log10 VL .3.0). c Using the HOPS dataset updated 31 December 2013, novel agents include fusion inhibitors, entry inhibitors and integrase inhibitors. d Deaths are included only if they occurred within 6 months of the last HOPS contact and before 31 December 2012. b
such as raltegravir, maraviroc or enfuvirtide); the mortality rate in these patients after baseline and after starting the novel agent was 0.63 and 1.56 deaths/100 person-years of observation respectively, compared with 3.00 deaths/100 person-years for patients who did not receive a novel agent after baseline. There were 274 patients who underwent genotypic testing after baseline; 47 died (17.2%) within 180 days of their last contact and their mortality rate over the remaining observation was 2.53 deaths/100 person-years. The median time from baseline to the first post-baseline genotypic testing was 12.1 (IQR 3.8 –32.5) months. Among these 274 patients, the median age at baseline was 41 years and most were male (73.4%), had a prior AIDS diagnosis (79.2%) and had been exposed to mono or dual ARV therapy (76.3%) prior to first cART (Table 1); their median baseline and nadir CD4+ T cell counts were 289 and 111 cells/mm3, respectively. Most (73.0%) had also received a PI-based initial cART regimen. The median observation time after the date of the first post-baseline genotypic testing was 8.1 years (compared with a median observation time of 6.2 years after the baseline date used for the analyses of the entire TCF cohort). In univariate analyses, older age, HIV risk from use of injection drugs, baseline CD4+ cell count ,200 copies/mm3, baseline VL ≥5.0 log10 copies/mL, having exposure to mono or dual ARV therapy before cART therapy and having a PI-based first cART were each associated with an increased mortality risk, while a later year of meeting the TCF criteria (in 1999–2008 versus 1997–1998) was associated with a decreased mortality risk; however, neither the number nor the type of HIV mutations, the number of ARV drug classes against which mutations were present, or number of mutations per ARV class was significantly associated with mortality in univariate analyses (Table 3). In the final multivariable regression model, factors independently associated with increased risk of death were increasing age (HR 1.94, 95% CI 1.36– 2.78, per 10 year increment), HIV risk from use of injection drugs (HR 2.71, 95% CI 1.37 – 5.39), baseline plasma HIV RNA
≥5.0 log10 copies/mL (HR 5.35, 95% CI 2.82 –10.1) and having a PI-based first cART regimen (HR 3.20, 95% CI 1.25–8.17) (Table 3).
Discussion In the HOPS, among substantially treatment-experienced patients who remained viraemic despite continuous ARV therapy, risk of death was associated with established clinical factors, including a lower CD4+ T cell count and higher plasma HIV RNA levels. For these patients the use of any ‘novel’ ARV, including an integrase strand transfer inhibitor, a CCR5 receptor antagonist or a fusion inhibitor, was associated with a reduction in mortality, while the historical use of PI-based initial cART regimens was associated with a greater mortality risk. Among the subset of TCF patients with available HIV genotypic resistance data, contrary to our hypothesis and to some preliminary data (not shown), no genotypic patterns of mutations were independently associated with risk of death. This was true despite the fact that many of the patients whom we studied who were receiving nonsuppressive ARV had had a prior extensive exposure to drugs that are no longer recommended first-line therapies, such as thymidine analogue NRTIs,11,12 and had viral isolates demonstrating mutations associated with resistance to some of these older drugs. Our observation that no increased (or decreased) mortality risk was associated with the presence of genotypic resistance, regardless of the number and types of mutations and whether single or multiple ARVs or classes of ARVs were affected, is in contrast to some recent reports5,6 that have identified associations between ARV resistance and risk of death, but it is consistent with one study.13 Although extensive genotypic resistance may be a marker for patient or virus characteristics that make viral suppression with ARV more difficult, and specific mutational patterns can be associated with increased viral fitness or replication competence, we did not ascertain any associations between the
2829
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
ARV exposure, n (%) NRTI-exposed NNRTI-exposed PI-exposed novel-agent-exposedc
TCF with genotype testa (n ¼274)
Palella et al.
Table 2. Cox proportional hazards results for mortality risk after baseline among all TCF participants, HOPS 1999– 2008 (N¼597) Univariateb Baselinea variable
HR (95% CI)
Multivariablec
Full multivariable P
P
HR (95% CI)
P
1.61 (1.28– 2.02)
,0.001
1.81 (1.10– 2.98) referent referent referent
0.019
Male Age, per 10 year increment
1.15 (0.73–1.81) 1.29 (1.04–1.60)
0.56 0.020
1.15 (0.71– 1.86) 1.56 (1.24– 1.97)
0.57 ,0.001
Race/ethnicity non-Hispanic white non-Hispanic black Hispanic other race
referent 0.93 (0.63–1.37) 0.44 (0.19–1.02) 0.92 (0.29–2.92)
0.70 0.054 0.88
referent 1.02 (0.66– 1.57) 0.47 (0.19– 1.15) 0.69 (0.21– 2.24)
0.94 0.10 0.54
HIV risk injection drug use men who have sex with men heterosexual other risk
1.73 (1.03–2.90) referent 1.00 (0.64–1.55) 1.19 (0.54–2.60)
0.039
Public or no health insurance yes no
1.71 (1.17–2.49) referent
0.006
Year first met TCF definition 1998 or earlier 1999 –2001 2002 –2004 2005 –2008
referent 0.68 (0.45–1.05) 0.58 (0.34–0.99) 0.40 (0.18–0.89)
CD4+ T cell count (cells/mm3) ,200 ≥200
4.33 (2.97–6.32) referent
,0.001
3.36 (2.18– 5.17) referent
,0.001
3.68 (2.41– 5.62) referent
,0.001
HIV VL (log10 copies/mL) ≥5.0 ,5.0
4.24 (2.90–6.20) referent
,0.001
2.90 (1.86– 4.51) referent
,0.001
2.91 (1.88– 4.49) referent
,0.001
Prior mono or dual ART exposure yes no
2.04 (1.09–3.80) referent
0.025
1.59 (0.83– 3.04) referent
0.16
First cART regimen PI-based yes no
2.24 (1.36–3.71) referent
0.002
2.43 (1.44– 4.09) referent
,0.001
2.44 (1.47– 4.06) referent
,0.001
0.46 (0.22– 0.97) referent
0.041
0.45 (0.22– 0.93) referent
0.031
1.00 0.67
0.08 0.048 0.025
1.79 (1.04– 3.07) referent referent referent
0.034
1.52 (1.01– 2.30) referent
0.045
referent 0.84 (0.53– 1.31) 0.91 (0.52– 1.62) 0.84 (0.36– 1.98)
0.44 0.76 0.70
Received genotype test during follow-up (time-dependent variable) yes 0.93 (0.63–1.38) 0.73 no referent Novel agent use during follow-up (time-dependent variable) yes 0.58 (0.29–1.18) no referent
0.13
ART, ARV therapy. a The baseline date was defined as the earliest date during 1999 –2008 when a triple-ARV-class-experienced patient had an HIV VL .1000 copies/mL (log10 VL .3.0). b Additional baseline variables (listed in Table 1) were explored in univariate analyses and were not associated with mortality risk; for brevity, only the key demographic characteristics are shown in the table. c The multivariable model contains variables with P,0.20 in univariate analyses, and manual backwards selection was used to eliminate non-significant variables in order of highest P value.
2830
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
HR (95% CI)
JAC
Mortality factors among viraemic ARV-experienced HIV patients
Table 3. Cox proportional hazards models for mortality risk among the subset of TCF participants who had a genotype test at or after TCF date, HOPS 1999– 2013 (N ¼274) Variablea
Univariate HR (95% CI)
P
201 (73.4)
113 (41.2) 178 (65.0) 152 (55.5) 56 (20.4)
1.87 (0.87–4.01) 1.57 (1.16–2.13) 1.69 (0.94–3.04) 2.90 (1.47–5.72) 1.47 (0.82–2.65) 0.53 (0.29–0.96) 2.30 (1.30–4.09) 4.39 (2.39–8.07) 3.17 (1.14–8.85) 0.65 (0.29–1.47) 2.94 (1.16–7.44) 1.00 (0.99–1.01) 1.00 (0.98–1.01) 0.96 (0.89–1.03) 0.99 (0.97–1.02) 1.13 (0.94–1.35) 1.08 (0.61–1.92) 0.94 (0.51–1.74) 1.11 (0.62–2.01) 1.70 (0.91–3.19)
0.11 0.003 0.08 0.002 0.20 0.037 0.004 ,0.001 0.027 0.30 0.023 0.64 0.63 0.25 0.60 0.20 0.79 0.85 0.72 0.10
Any NRTI mutation M41L D67N K70E/R L74V M184I M184V L210W T215F T215Y T215F/Y K219E/Q
210 (76.6) 89 (32.5) 86 (31.4) 59 (21.5) 38 (13.9) 2 (0.7) 125 (45.6) 56 (20.4) 28 (10.2) 86 (31.4) 111 (40.5) 41 (15.0)
1.17 (0.57–2.43) 1.42 (0.79–2.56) 1.21 (0.67–2.19) 0.94 (0.47–1.86) 1.32 (0.62–2.83) 5.52 (0.75–40.7) 0.80 (0.44–1.43) 1.39 (0.73–2.64) 1.41 (0.60–3.33) 1.61 (0.64–2.09) 1.39 (0.78–2.46) 0.79 (0.34–1.87)
0.67 0.24 0.53 0.85 0.48 0.09 0.45 0.32 0.44 0.62 0.26 0.60
Any NNRTI mutation L100I K103N Y181C Y181I Y188L G190A/S
202 (73.7) 20 (7.3) 129 (47.1) 61 (22.3) 2 (0.7) 13 (4.7) 47 (17.2)
0.74 (0.39–1.39) 1.18 (0.42–3.28) 0.89 (0.50–1.59) 0.89 (0.44–1.79) 2.62 (0.36–19.0) 2.51 (0.99–6.34) 1.09 (0.54–2.20)
0.35 0.76 0.69 0.74 0.34 0.053 0.80
Any PI mutation V32I L33F M46I M46L I47V G48V I50V I54M
135 (49.3) 18 (6.6) 0 (0.0) 53 (19.3) 13 (4.7) 11 (4.0) 9 (3.3) 5 (1.8) 6 (2.2)
1.09 (0.61–1.95) 1.01 (0.31–3.26)
0.76 0.98
0.48 (0.20–1.12) 0.73 (0.18–3.00) 0.60 (0.08–4.32) 2.07 (0.64–6.66) 1.06 (0.15–7.70) 1.45 (0.20–10.5)
0.09 0.66 0.61 0.22 0.95 0.72
Male Age in years, per 10 year increment Non-Hispanic white race/ethnicity HIV risk from injection drug use Public/no health insurance TCF year 1999 –2008 (versus 1997 –1998) Baseline CD4+ T cell count (,200 cells/mm3) Baseline HIV VL ≥5.0 log10 copies/mL Prior mono or dual ART exposure Novel agent use during follow-up (time-dependent variable) First cART PI-based Total number of IAS major mutations Total number of IAS major NRTI mutations Total number of IAS major NNRTI mutations Total number of IAS major PI mutations Total number of TAMs Mutations in all three main classes ≥3 any mutations Any TAM ≥4 TAMs
131 (47.8) 30 (11.0) 147 (53.7) 209 (76.3) 87 (31.8) 36 (13.1) 209 (76.3) 119 (43.4) 200 (73.0)
Multivariableb HR (95% CI)
P
1.94 (1.36–2.78)
,0.001
2.71 (1.37–5.39)
0.004
5.35 (2.82–10.1)
,0.001
3.20 (1.25–8.17)
0.015
Continued
2831
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
n (%)
Palella et al.
Table 3. Continued Variablea V82A/F/L/S/T L90M
n (%)
Univariate HR (95% CI)
P
53 (19.3) 81 (29.6)
1.36 (0.71–2.58) 1.72 (0.97–3.07)
0.35 0.07
Multivariableb HR (95% CI)
P
ART, ARV therapy; IAS, International Antiviral Society. Baseline date: the earliest date during 1999 –2008 when a triple-ARV-class-experienced patient had HIV VL .1000 copies/mL (log10 VL .3.0). All variables in the model were measured as of baseline, except for variables related to ARV resistance mutations (assessed as of first post-baseline genotypic testing). b The multivariable model contains variables with P,0.20 in univariate analysis, and manual backwards selection was used to eliminate non-significant variables in order of highest P value. a
2832
associated with improved survival are consistent with this evolution. Our analysis has several limitations. Approximately half of the TCF patients had evaluable HIV genotype data. While we cannot readily explain the lack of association with mortality risk (improved or worsened) among persons with extensive resistance mutations, unmeasured confounding (e.g. resistant virus reflecting poor adherence, which was not assessed in this cohort) rather than some direct effect of specific mutational patterns may have been present. Finally, it is possible that we failed to detect associations between specific mutational combinations and mortality due to the relative rarity of these mutations or mutational patterns (a type 2 error). It is reassuring, however, that we did identify positive associations between mortality risk and known validated clinical variables in the overall group of TCF patients, i.e. a lower baseline CD4+ T cell count and higher plasma HIV RNA. In conclusion, among substantially ARV-experienced HIVinfected participants in the HOPS who remained viraemic despite continued therapy, mortality was associated with established clinical factors and ARV therapy factors (with PI-based initial cART associated with an increased mortality risk and the use of novel agents associated with improved survival), but not the presence of specific HIV drug resistance mutations or mutational patterns. Whether or not our findings are generalizable will require further evaluation in a larger population that includes more deaths, more genotypic HIV resistance tests and perhaps more person-years of observation. Ample data exist supporting the importance of routine HIV resistance testing among virally nonsuppressed ARV treatment-experienced patients11 in order to characterize the extent and type of ARV resistance present and to optimize ARV drug selection.29
Acknowledgements HOPS Investigators The HIV Outpatient Study (HOPS) Investigators include the following persons and sites: John T. Brooks, Kate Buchacz, Marcus D. Durham, Division of HIV/AIDS Prevention, National Center for HIV, STD, and TB Prevention (NCHSTP), Centers for Disease Control and Prevention (CDC), Atlanta, GA; Kathleen C. Wood, Darlene Hankerson, Rachel Debes, Thila Subramanian, Cheryl Akridge, Harlan Hayes, Carl Armon, Bonnie Dean, Jeff Binkley and Sam Bozzette, Cerner Corporation, Vienna, VA; Frank J. Palella, Joan S. Chmiel, Carolyn Studney, Saira Jahangir, Feinberg School
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
pattern of mutations and survival even among persons whose viral isolates demonstrated the most extensive resistance. Among our TCF patients and the subset of these with available HIV genotypic data, having received a first cART regimen that contained a PI (versus an NNRTI) was associated with a greater risk of death. While PI-containing initial cART, particularly regimens that include older PIs, can be associated with greater treatment-limiting toxicity, pill burden and lower rates of HIV suppression compared with NNRTI therapies, the association with a subsequent increased mortality among such persons at a timepoint distant from the initial cART, often after multiple cART regimens, is notable. The use of novel ARV agents among TCF patients was associated with a survival benefit. This was not surprising since the use of such therapies among ARV therapy-experienced patients in randomized prospective clinical trials has been shown to be associated with improved virological and immunological outcomes,14 – 17 yet this cohort is among the first in which a survival benefit associated with the use of these drugs among ARV therapy-experienced patients is apparent. Greater HIV replication capacity has been associated with mutations known to confer resistance to thymidine analogue NRTIs, notably with ≥4 TAMs.18 – 20 Other data exist indicating that viruses containing fewer TAMs20,21 can demonstrate decreased replication capacity and/or RT processivity. Other point mutations selected by ARVs reduce viral fitness in potentially clinically relevant ways that are exploitable among ARV recipients who experience incomplete HIV suppression.22 – 26 However, among our TCF patients, we observed no survival disadvantage associated with having ≥4 TAMS and no survival advantages associated with HIV mutations known to reduce viral fitness, such as M184V.27,28 While we did not identify any discrete care-related factors that could reasonably have impacted on mortality risk, and that improved in general over time, persons who became TCF in earlier time periods were more likely to have had exposure to mono or dual NRTI therapy (although such exposure was not associated with an increased risk of death in multivariable models). Moreover, the prescribed ARV regimens substantially evolved over time (for both the initial and the salvage regimen), with more persons receiving NNRTI-based first-line therapy and more effective, less toxic background ARVs during later time periods; our findings that a PI-based regimen as first cART was a risk factor for mortality and that the use of novel ARV agents during follow-up was
Mortality factors among viraemic ARV-experienced HIV patients
of Medicine, Northwestern University, Chicago, IL; Kenneth A. Lichtenstein and Cheryl Stewart, National Jewish Medical and Research Center Denver, CO; John Hammer, Kenneth S. Greenberg, Barbara Widick, and Rosa Franklin, Rose Medical Center, Denver, CO; Bienvenido G. Yangco and Kalliope Chagaris, Infectious Disease Research Institute, Tampa, FL; Doug Ward, Troy Thomas, and Fletcher Neale, Dupont Circle Physicians Group, Washington, DC; Jack Fuhrer, Linda Ording-Bauer, Rita Kelly and Jane Esteves, State University of New York (SUNY), Stony Brook, NY; Ellen M. Tedaldi, Ramona A. Christian, Faye Ruley, Dania Beadle and Princess Graham, Temple University School of Medicine, Philadelphia, PA; Richard M. Novak and Andrea Wendrow and Renata Smith, University of Illinois at Chicago, Chicago, IL; Benjamin Young, Barbara Widick, Mia Scott, APEX Family Medicine, Denver, CO.
This work was supported by contracts 200-2001-00133, 200-2006-18797 and 200-2011-41872 from the Centers for Disease Control and Prevention.
Transparency declarations F. J. P. is on speakers’ bureaus for Gilead Sciences, Merck, Janssen Pharmaceuticals and Bristol-Myers Squibb. C. A.’s time in working on the analysis, writing and editing, and preparing this article was funded by a contract with Cerner Corporation, which contracts with the CDC for the HOPS. J. S. C.’s time in working on the analysis, writing and editing, and preparing this article was funded under a subcontract to Northwestern University from Cerner Corporation, which contracts with the CDC for the HOPS. K. B., R. M. N., R. T. D’A. and J. T. B.: none to declare.
Disclaimer The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
References 1 Lorenzen T, Walther I, Stoehr A et al. Effective treatment of patients in a deep salvage situation with “non-active HAART”: experiences with the expert advice system RADATA. Infection 2009; 37: 528–33. 2 Napravnik S, Keys JR, Quinlivan EB et al. Triple-class antiretroviral drug resistance: risk and predictors among HIV-1-infected patients. AIDS 2007; 21: 825–34. 3 Steigbigel RT, Cooper DA, Kumar PN et al. Raltegravir with optimized background therapy for resistant HIV-1 infection. N Engl J Med 2008; 359: 339– 54. 4 Montaner J, Guimaraes D, Chung J et al. Prognostic staging of extensively pretreated patients with advanced HIV-1 disease. HIV Clin Trials 2005; 6: 281–90. 5 Lohse N, Jorgensen LB, Kronborg G et al. Genotypic drug resistance and long-term mortality in patients with triple-class antiretroviral drug failure. Antivir Ther 2007; 12: 909– 17. 6 Hogg RS, Bangsberg DR, Lima VD et al. Emergence of drug resistance is associated with an increased risk of death among patients first starting HAART. PLoS Med 2006; 3: e356. 7 Hardy WD, Gulick RM, Mayer H et al. Two-year safety and virologic efficacy of maraviroc in treatment-experienced patients with CCR5tropic HIV-1 infection: 96-week combined analysis of MOTIVATE 1 and 2. J Acquir Immune Defic Syndr 2010; 55: 558–64. 8 Nelson M, Fisher M, Gonzalez-Garcia J et al. Impact of baseline antiretroviral resistance status on efficacy outcomes among patients
receiving maraviroc plus optimized background therapy in the MOTIVATE 1 and 2 trials. HIV Clin Trials 2010; 11: 145–55. 9 Buchacz K, Baker RK, Young B et al. Changes in the use of HIV antiretroviral resistance testing in a large cohort of U.S. patients, 1999 to 2006. J Acquir Immune Defic Syndr 2010; 53: 625– 32. 10 Moorman AC, Holmberg SD, Marlowe SI et al. Changing conditions and treatments in a dynamic cohort of ambulatory HIV patients: the HIV outpatient study (HOPS). Ann Epidemiol 1999; 9: 349–57. 11 Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Updated February 12, 2013. http://aidsinfo.nih.gov/ guidelines (16 April 2014, date last accessed). 12 Williams I, Churchill D, Anderson J et al. British HIV Association guidelines for the treatment of HIV-1-positive adults with antiretroviral therapy 2012. HIV Med 2012; 13 Suppl 2: 1 –85. 13 Swindells S, Jiang H, Mukherjee AL et al. Lower CD4 cell count and higher virus load, but not antiretroviral drug resistance, are associated with AIDS-defining events and mortality: an ACTG Longitudinal Linked Randomized Trials (ALLRT) analysis. HIV Clin Trials 2011; 12: 79– 88. 14 Eron JJ, Cooper DA, Steigbigel RT et al. Efficacy and safety of raltegravir for treatment of HIV for 5 years in the BENCHMRK studies: final results of two randomised, placebo-controlled trials. Lancet Infect Dis 2013; 13: 587–96. 15 Anderson D, DeMasi R, DeLaitsch L et al. Week 96 outcomes of patients with less treatment experience versus more treatment experience receiving etravirine in the DUET trials. Curr HIV Res 2012; 10: 256–61. 16 Reynes J, Araste´h K, Clotet B et al. TORO: ninety-six-week virologic and immunologic response and safety evaluation of enfuvirtide with an optimized background of antiretrovirals. AIDS Patient Care STDS 2007; 21: 533– 43. 17 Taiwo B, Murphy RL, Katlama C. Novel antiretroviral combinations in treatment-experienced patients with HIV infection: rationale and results. Drugs 2010; 70: 1629 –42. 18 D’Aquila RT, Johnson VA, Welles SL et al. Zidovudine resistance and HIV-1 disease progression during antiretroviral therapy. AIDS Clinical Trials Group Protocol 116B/117 Team and the Virology Committee Resistance Working Group. Ann Intern Med 1995; 122: 401–8. 19 Caliendo AM, Savara A, An D et al. Effects of zidovudine-selected human immunodeficiency virus type 1 reverse transcriptase amino acid substitutions on processive DNA synthesis and viral replication. J Virol 1996; 70: 2146– 53. 20 Hu Z, Giguel F, Hatano H et al. Fitness comparison of thymidine analog resistance pathways in human immunodeficiency virus type 1. J Virol 2006; 80: 7020– 7. 21 Cong ME, Heneine W, Garcia-Lerma JG. The fitness cost of mutations associated with human immunodeficiency virus type 1 drug resistance is modulated by mutational interactions. J Virol 2007; 81: 3037–41. 22 White KL, Margot NA, Wrin T et al. Molecular mechanisms of resistance to human immunodeficiency virus type 1 with reverse transcriptase mutations K65R and K65R+M184V and their effects on enzyme function and viral replication capacity. Antimicrob Agents Chemother 2002; 46: 3437– 46. 23 Deval J, White KL, Miller MD et al. Mechanistic basis for reduced viral and enzymatic fitness of HIV-1 reverse transcriptase containing both K65R and M184V mutations. J Biol Chem 2004; 279: 509–16. 24 Tupinambas U, Aleixo A, Greco D. HIV-1 genotypes related to failure of nelfinavir as the first protease inhibitor treatment. Braz J Infect Dis 2005; 9: 324– 9.
2833
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
Funding
JAC
Palella et al.
25 Martinez-Picado J, Savara AV, Sutton L et al. Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1. J Virol 1999; 73: 3744– 52. 26 Shafer R. Stanford University HIV Drug Resistance Database. http:// hivdb.stanford.edu/DR/NRTIResiNote.html (16 April 2014, date last accessed). 27 von Wyl V, Ehteshami M, Demeter LM et al. HIV-1 reverse transcriptase connection domain mutations: dynamics of emergence and implications
for success of combination antiretroviral therapy. Clin Infect Dis 2010; 51: 620– 8. 28 Kuritzkes DR, Bassett RL, Hazelwood JD et al. Rate of thymidine analogue resistance mutation accumulation with zidovudine- or stavudine-based regimens. J Acquir Immune Defic Syndr 2004; 36: 600–3. 29 Palella FJ Jr, Armon C, Buchacz K et al. The association of HIV susceptibility testing with survival among HIV-infected patients receiving antiretroviral therapy: a cohort study. Ann Intern Med 2009; 151: 73– 84.
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
2834
Factors associated with mortality among persistently viraemic triple-antiretroviral-class-experienced patients receiving antiretroviral therapy in the HIV Outpatient Study (HOPS) Frank J. Palella Jr1*, Carl Armon2, Kate Buchacz3, Joan S. Chmiel1, Richard M. Novak4, Richard T. D’Aquila1 and John T. Brooks3 on behalf of the HOPS Investigators† 1
Northwestern University, Chicago, IL, USA; 2Cerner Corporation, Vienna, VA, USA; 3Centers for Disease Control and Prevention, Atlanta, GA, USA; 4University of Illinois at Chicago, Chicago, IL, USA
Received 22 October 2013; returned 16 December 2013; revised 17 April 2014; accepted 3 May 2014 Background: Identifying factors associated with mortality for HIV-infected patients with persistent viraemia despite antiretroviral (ARV) therapy may inform diagnostic and treatment strategies. Methods: We analysed data from viraemic triple-ARV-class-experienced HIV Outpatient Study patients seen during 1 January 1999 to 31 December 2012 who, despite treatment that included ARVs from three major drug classes [nucleoside analogue reverse transcriptase inhibitors, non-nucleoside analogue reverse transcriptase inhibitors and protease inhibitors (PIs)], had plasma HIV RNA levels [viral load (VL)] .1000 copies/mL [‘triple ARV class failure’ (TCF)]. The baseline was defined as the date of meeting the TCF criteria during 1999 – 2008. We identified factors associated with mortality using Cox regression. Results: Of 597 patients who met the TCF criteria (median follow-up after baseline 4.9 years), 115 (19.3%) died. Baseline factors associated with mortality were age per 10 years [hazard ratio (HR) 1.61, 95% CI 1.28 –2.02], risk of HIV from use of injection drugs (HR 1.81, 95% CI 1.10 –2.98), CD4+ T cell count ,200 cells/mm3 (HR 3.68, 95% CI 2.41–5.62), VL ≥5.0 log10 copies/mL (HR 2.91, 95% CI 1.88–4.49) and receiving a first combination ARV therapy regimen that was PI-based (HR 2.44, 95% CI 1.47 –4.06); receiving a novel ARV agent during follow-up (HR 0.45, 95% CI 0.22 –0.93) was protective. Genotypic resistance testing results were available for 274 (45.9%) of the TCF patients, of whom 47 (17.2%) died. In this group, factors associated with death were increasing age (HR 1.94, 95% CI 1.36 – 2.78, per 10 year increment), risk of HIV from use of injection drugs (HR 2.71, 95% CI 1.37 –5.39), baseline VL ≥5.0 log10 copies/mL (HR 5.35, 95% CI 2.82–10.1) and receiving PI-based first combination ARV therapy regimen (HR 3.20, 95% CI 1.25–8.17). No HIV mutations or combinations of mutations were significantly associated with survival. Conclusions: Factors significantly associated with mortality risk among TCF patients who received ongoing ARV therapy included traditional clinical predictors but not the presence, type or number of HIV genetic mutations. The use of novel ARV drugs by these ARV therapy-experienced patients was associated with an improved survival. Keywords: ARV, antiretroviral resistance, viraemia
Introduction The optimal management of extensively antiretroviral (ARV) therapy-experienced HIV-infected persons who remain viraemic despite continued ARV therapy remains a clinical challenge.1 – 4 Identifying factors associated with improved clinical outcomes among such patients can inform ARV treatment strategies. Existing data indicate that ARV resistance among both ARVtreated and ARV-naive persons is associated with an increased
risk of death,5,6 suggesting that strategies aimed at reducing the induction of resistance may improve survival. As new ARV agents from diverse drug classes become available, routine ARV resistance testing among viraemic treatment-experienced patients has become a standard of care to guide ARV selection.1,3,7 – 9 In this report, using data from the HIV Outpatient Study (HOPS), we evaluated the associations between mortality risk, patient factors and ARV resistance mutations in a diverse group of highly ARV
# The Author 2014. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please e-mail: [email protected]
2826
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
*Corresponding author. Tel: +1-312-695-5053; E-mail: [email protected] †The HOPS Investigators are listed in the Acknowledgements section.
Mortality factors among viraemic ARV-experienced HIV patients
treatment-experienced persons who remained viraemic despite ongoing ARV exposure.
Methods The HOPS
Study population and definitions We analysed data from participants with at least two clinical encounters who met the criteria for triple ARV class failure (TCF; defined below) between 1 January 1999 and 31 December 2008 and contributed observation time during 1 January 1999 to 31 December 2012, using the dataset updated 31 December 2013. We defined triple-class-experienced (TCE) as having received a prescription for ≥4 continuous months of an ARV from each of the three major drug classes [nucleoside analogue reverse transcriptase inhibitors (NRTIs), non-nucleoside analogue reverse transcriptase inhibitors (NNRTIs) and protease inhibitors (PIs), although not necessarily simultaneously]. We limited our analyses to patients who had a plasma HIV RNA level [viral load (VL)] .1000 copies/mL (VL .3.0 log10 copies/mL) after becoming TCE and while receiving ARV therapy. We defined such patients as TCF. The baseline date was defined as the earliest date at which a VL .1000 copies/mL was seen after meeting the TCE criteria during 1999 –2008. Patients could have met the TCF definition prior to 1999 if they contributed observation time after 1 January 1999. We conducted further analyses for the subset of TCF patients for whom results of HIV genotypic testing performed on or after the baseline date were available, incorporating resistance results from this first postbaseline genotypic testing as well as results from prior resistance testing for these patients. Thymidine analogue NRTI-associated resistance mutations (TAMs) were M41L, D67N, K70E/R, L210W, T215F/Y and K219E/Q. Novel ARV agents were fusion inhibitors, integrase strand transfer inhibitors, entry inhibitors and CCR5 attachment inhibitors.
Statistical analyses We used Cox proportional hazards regression models to examine risk factors for mortality among TCF patients. Observation began at baseline, continued until death (if documented within 180 days of last patient contact) and was censored (the patient was assumed to be alive) at the last patient contact plus 180 days if the patient was not known to be deceased by then, or at 31 December 2012, whichever occurred first. Among a subgroup of TCF patients with genotypic resistance testing, we analysed cumulative mutations at first genotypic testing following TCF as well as mutations detected by prior genotypic testing studies. We evaluated the associations between mortality and both the total number and
individual types of mutations, after adjusting for other clinical and demographic factors. For these subgroup analyses, the start of observation was reset to the date of the first post-baseline genotypic testing. For both unadjusted and adjusted Cox regression models, the results are reported using hazard ratios (HRs) and 95% CIs. To develop multivariable regression models for mortality after the start of observation, variables with P values ,0.20 in univariate analyses were considered for inclusion in multivariable models; manual backwards selection was then used to eliminate non-significant variables sequentially in order of highest P value. We estimated the proportion of the observation time during which patients were viraemic (defined as VL ≥200 copies/mL) by summing the duration of observation times around a given detectable VL; a patient was considered to be viraemic from the mid-point in time between the prior VL test and the ‘current’ time, to the mid-point in time between the ‘current’ time and the next VL test. We summed the time spent as viraemic across patients included in the analysis. All the descriptive summaries, survival plots and regression analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA). Statistical significance was defined as a two-sided P value ,0.05 for statistical tests and, for an HR, when the associated 95% CI excluded 1.0.
Results Of 3840 HOPS patients with at least two clinical encounters between 1 January 1999 and 31 December 2008 and whose ARV prescription history was known, 1536 patients met the definition of TCE; of these, 597 (38.9%) met the definition of TCF. Of the 597 patients, 158 (26.5%) first met the TCF criteria in 1998 or earlier and 439 (73.5%) met them during 1999 – 2008. Among the TCF patients, the median baseline age was 41 years, 77.2% were male, 79.9% had a prior diagnosis of AIDS and 79.9% had been exposed to mono or dual ARV therapy prior to their first combination ARV therapy (cART). The median baseline and nadir CD4+ T cell counts were 283 and 105 cells/mm3 , respectively. Most (71.4%) patients had been prescribed a first cART regimen that was PI-based. The median observation time after baseline was 6.2 years (IQR 3.1–10.5) (Table 1). Of the 597 TCF patients, 115 (19.3%) died (Table 1), giving a mortality rate of 2.77 deaths per 100 person-years of follow-up. In the final multivariable Cox regression model, age per 10 years (HR 1.61, 95% CI 1.28 – 2.02), risk of HIV from use of injection drugs (HR 1.81, 95% CI 1.10 – 2.98), baseline CD4+ T cell count ,200 cells/mm3 (HR 3.68, 95% CI 2.41 – 5.62), baseline plasma HIV RNA ≥5.0 log10 copies/mL (HR 2.91, 95% CI 1.88 –4.49) and having received a first cART regimen that was PI-based (HR 2.44, 95% CI 1.47 – 4.06) were associated with an increased risk of death, while receiving a novel ARV agent during follow-up (HR 0.45, 95% CI 0.22 – 0.93) was associated with an decreased risk of death (Table 2). Among the 597 TCF patients, 556 (93.1%) had a VL measurement that was ≥200 copies/mL after meeting the TCF criteria; these patients remained viraemic above this level during 68% of their post-baseline follow-up. A total of 13/597 (2.2%) of patients had subsequent measurements that were all ,200 copies/mL and 28/597 (4.7%) had no subsequent VL recorded. The 556 viraemic patients had a median of 17 (IQR 8 –29) VL measurements after meeting the TCF criteria, with median and mean durations of viraemia after baseline of 2.7 (IQR 1.1 – 4.9) and 3.4 years, respectively. Among the 597 TCF patients, 156 patients received a ‘novel’ ARV agent (i.e. one other than an NRTI, NNRTI or PI
2827
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
The HOPS is an ongoing prospective cohort study of HIV-infected adults receiving care at participating HIV clinics. For this analysis, data were included from 12 clinics (university-based, private and public) in eight US cities (Chicago, IL, USA; Denver, CO, USA; Stony Brook, NY, USA; Oakland/ San Leandro, CA, USA; Walnut Creek, CA, USA; Philadelphia, PA, USA; Tampa, FL, USA; and Washington, DC, USA) since 1993.10 Patient data, including sociodemographic characteristics, diagnoses, treatments and laboratory values, are abstracted from medical charts and entered into an electronic database by trained staff. These data are reviewed for quality and analysed centrally. HOPS investigators have collected information on more than 10 000 patients seen at over 400 000 clinical encounters. The HOPS protocol has been reviewed annually and approved by the CDC (Atlanta, GA, USA), the Cerner Corporation (Vienna, VA, USA) and the institutional review board at each participating site. The study protocol conforms to the guidelines of the US Department of Health and Human Services for the protection of human subjects in research.
JAC
Palella et al.
Table 1. Characteristics of ARV-experienced patients who experienced TCF, the HOPS 1999 –2008 All TCF patients (N ¼597)
TCF with genotype testa (n ¼274)
TCF without genotype test (n¼323)
P
Characteristics as of baseline dateb Male, n (%) Age (years), median (IQR) Age .40 years, n (%)
461 (77.2) 41 (36 –46) 339 (56.8)
201 (73.4) 41 (36 –46) 154 (56.2)
260 (80.5) 41 (36 –46) 185 (57.3)
Race or ethnicity, n (%) non-Hispanic white non-Hispanic black Hispanic other/unknown race/ethnicity
291 (48.7) 230 (38.5) 62 (10.4) 14 (2.4)
131 (47.8) 104 (38.0) 34 (12.4) 5 (1.8)
160 (49.5) 126 (39.0) 28 (8.7) 9 (2.8)
HIV risk category (non-exclusive categories), n (%) men who have sex with men injection drug use heterosexual other/unknown HIV risk
324 (54.3) 67 (11.2) 172 (28.8) 34 (5.7)
138 (50.4) 30 (11.0) 88 (32.1) 18 (6.6)
186 (57.6) 37 (11.5) 84 (26.0) 16 (5.0)
Insurance/payer at date of TCF, n (%) private public, other or unknown
291 (48.7) 306 (51.3)
127 (46.4) 147 (53.7)
164 (50.8) 159 (49.2)
Calendar year patient first fulfilled criteria for TCF, n (%) 1998 or earlier 1999 –2001 2002 –2004 2005 –2008
158 (26.5) 226 (37.9) 128 (21.4) 85 (14.2)
65 (23.7) 101 (36.9) 61 (22.3) 47 (17.2)
93 (28.8) 125 (38.7) 67 (20.7) 38 (11.8)
477 (79.9) 105 (31 –219) 215 (93 –361)
217 (79.2) 111 (36 –227) 226 (103– 368)
260 (80.5) 99 (26 –211) 206 (88 –350)
0.77 0.22 0.17
283 (145– 417) 4.1 (3.4 –4.8) 7.4 (4.6 –11.1) 4.4 (2.7 –6.5) 89 (14.9) 365 (61.1) 425 (71.2) 477 (79.9) 4.3 (2.5 –6.6) 0.8 (0.5 –1.6) 1.9 (1.1 –3.2)
289 (171– 467) 4.1 (3.4– 4.8) 7.1 (4.4– 11.1) 4.2 (2.5– 6.2) 33 (12.0) 151 (55.1) 183 (66.8) 209 (76.3) 4.2 (2.5– 6.2) 0.9 (0.5– 1.5) 1.9 (1.2– 3.6)
267 (127– 395) 4.2 (3.4– 4.9) 7.5 (4.7– 11.1) 4.5 (2.8– 6.6) 56 (17.3) 214 (66.3) 242 (74.9) 268 (83.0) 4.5 (2.5– 6.7) 0.8 (0.5– 1.7) 1.8 (1.1– 3.0)
0.030 0.38 0.59 0.26 0.09 0.007 0.036 0.053 0.52 0.90 0.18
51 (8.5) 115 (19.3) 426 (71.4) 5 (0.8)
15 (5.5) 55 (20.1) 200 (73.0) 4 (1.5)
36 (11.2) 60 (18.6) 226 (70.0) 1 (0.3)
5 (0.8)
3 (1.1)
2 (0.6)
0.048 0.80 0.86 0.44
First cART regimen type, n (%) NNRTI+PI NNRTI PI three or more NRTIs
0.32
0.19
0.035
Additional first cART regimen novel agentsc, n (%)
0.67
b
Characteristics after baseline date Years follow-up post-baseline, median (IQR) Persons who died within 6 months of baseline, n (%) Persons who died within 12 months of baseline, n (%)
6.2 (3.1 –10.5) 7 (1.2) 18 (3.0)
8.1 (5.3– 12.1) 0 (0.0) 3 (1.1)
4.5 (1.8– 8.7) 7 (2.2) 15 (4.6)
,0.001 0.017 0.022 Continued
2828
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
Prior AIDS diagnosis, n (%) Nadir CD4+ T cell count (cells/mm3), median (IQR) CD4+ T cell count at cART initiation (cells/mm3), median (IQR) CD4+ T cell count (cells/mm3), median (IQR) HIV VL (log10 copies/mL), median (IQR) Years since HIV diagnosis, median (IQR) Years since AIDS diagnosis (n¼477), median (IQR) Co-infected with hepatitis C, n (%) Exposed to mono ARV therapy, n (%) Exposed to dual ARV therapy, n (%) Exposed to mono ARV or dual ARV therapy, n (%) Years of prior NRTI use, median (IQR) Years of prior NNRTI use, median (IQR) Years of prior PI use, median (IQR)
0.26
JAC
Mortality factors among viraemic ARV-experienced HIV patients
Table 1. Continued All TCF patients (N ¼597) Persons who died during follow-up, n (%)d cause of death AIDS non-AIDS unknown cause of death
Percentage of time since TCF with HIV RNA ≥200 copies/ mL, median (IQR)
TCF without genotype test (n¼323)
115 (19.3)
47 (17.2)
68 (21.1)
56 (48.7) 40 (34.8) 19 (16.5)
20 (42.6) 17 (36.2) 10 (21.3)
36 (52.9) 23 (33.8) 9 (13.2)
595 (99.7) 499 (83.6) 546 (91.5) 156 (26.1)
274 (100.0) 233 (85.0) 262 (95.6) 119 (43.4)
321 (99.4) 266 (82.4) 284 (87.9) 37 (11.5)
74 (38 –100)
65 (38 –97)
88 (35 –100)
P 0.27 0.42
0.50 0.44 0.001 ,0.001 0.002
a
First genotype test at or after baseline date. Baseline date: the earliest date during 1999– 2008 when a triple-ARV-class-experienced patient had an HIV VL .1000 copies/mL (log10 VL .3.0). c Using the HOPS dataset updated 31 December 2013, novel agents include fusion inhibitors, entry inhibitors and integrase inhibitors. d Deaths are included only if they occurred within 6 months of the last HOPS contact and before 31 December 2012. b
such as raltegravir, maraviroc or enfuvirtide); the mortality rate in these patients after baseline and after starting the novel agent was 0.63 and 1.56 deaths/100 person-years of observation respectively, compared with 3.00 deaths/100 person-years for patients who did not receive a novel agent after baseline. There were 274 patients who underwent genotypic testing after baseline; 47 died (17.2%) within 180 days of their last contact and their mortality rate over the remaining observation was 2.53 deaths/100 person-years. The median time from baseline to the first post-baseline genotypic testing was 12.1 (IQR 3.8 –32.5) months. Among these 274 patients, the median age at baseline was 41 years and most were male (73.4%), had a prior AIDS diagnosis (79.2%) and had been exposed to mono or dual ARV therapy (76.3%) prior to first cART (Table 1); their median baseline and nadir CD4+ T cell counts were 289 and 111 cells/mm3, respectively. Most (73.0%) had also received a PI-based initial cART regimen. The median observation time after the date of the first post-baseline genotypic testing was 8.1 years (compared with a median observation time of 6.2 years after the baseline date used for the analyses of the entire TCF cohort). In univariate analyses, older age, HIV risk from use of injection drugs, baseline CD4+ cell count ,200 copies/mm3, baseline VL ≥5.0 log10 copies/mL, having exposure to mono or dual ARV therapy before cART therapy and having a PI-based first cART were each associated with an increased mortality risk, while a later year of meeting the TCF criteria (in 1999–2008 versus 1997–1998) was associated with a decreased mortality risk; however, neither the number nor the type of HIV mutations, the number of ARV drug classes against which mutations were present, or number of mutations per ARV class was significantly associated with mortality in univariate analyses (Table 3). In the final multivariable regression model, factors independently associated with increased risk of death were increasing age (HR 1.94, 95% CI 1.36– 2.78, per 10 year increment), HIV risk from use of injection drugs (HR 2.71, 95% CI 1.37 – 5.39), baseline plasma HIV RNA
≥5.0 log10 copies/mL (HR 5.35, 95% CI 2.82 –10.1) and having a PI-based first cART regimen (HR 3.20, 95% CI 1.25–8.17) (Table 3).
Discussion In the HOPS, among substantially treatment-experienced patients who remained viraemic despite continuous ARV therapy, risk of death was associated with established clinical factors, including a lower CD4+ T cell count and higher plasma HIV RNA levels. For these patients the use of any ‘novel’ ARV, including an integrase strand transfer inhibitor, a CCR5 receptor antagonist or a fusion inhibitor, was associated with a reduction in mortality, while the historical use of PI-based initial cART regimens was associated with a greater mortality risk. Among the subset of TCF patients with available HIV genotypic resistance data, contrary to our hypothesis and to some preliminary data (not shown), no genotypic patterns of mutations were independently associated with risk of death. This was true despite the fact that many of the patients whom we studied who were receiving nonsuppressive ARV had had a prior extensive exposure to drugs that are no longer recommended first-line therapies, such as thymidine analogue NRTIs,11,12 and had viral isolates demonstrating mutations associated with resistance to some of these older drugs. Our observation that no increased (or decreased) mortality risk was associated with the presence of genotypic resistance, regardless of the number and types of mutations and whether single or multiple ARVs or classes of ARVs were affected, is in contrast to some recent reports5,6 that have identified associations between ARV resistance and risk of death, but it is consistent with one study.13 Although extensive genotypic resistance may be a marker for patient or virus characteristics that make viral suppression with ARV more difficult, and specific mutational patterns can be associated with increased viral fitness or replication competence, we did not ascertain any associations between the
2829
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
ARV exposure, n (%) NRTI-exposed NNRTI-exposed PI-exposed novel-agent-exposedc
TCF with genotype testa (n ¼274)
Palella et al.
Table 2. Cox proportional hazards results for mortality risk after baseline among all TCF participants, HOPS 1999– 2008 (N¼597) Univariateb Baselinea variable
HR (95% CI)
Multivariablec
Full multivariable P
P
HR (95% CI)
P
1.61 (1.28– 2.02)
,0.001
1.81 (1.10– 2.98) referent referent referent
0.019
Male Age, per 10 year increment
1.15 (0.73–1.81) 1.29 (1.04–1.60)
0.56 0.020
1.15 (0.71– 1.86) 1.56 (1.24– 1.97)
0.57 ,0.001
Race/ethnicity non-Hispanic white non-Hispanic black Hispanic other race
referent 0.93 (0.63–1.37) 0.44 (0.19–1.02) 0.92 (0.29–2.92)
0.70 0.054 0.88
referent 1.02 (0.66– 1.57) 0.47 (0.19– 1.15) 0.69 (0.21– 2.24)
0.94 0.10 0.54
HIV risk injection drug use men who have sex with men heterosexual other risk
1.73 (1.03–2.90) referent 1.00 (0.64–1.55) 1.19 (0.54–2.60)
0.039
Public or no health insurance yes no
1.71 (1.17–2.49) referent
0.006
Year first met TCF definition 1998 or earlier 1999 –2001 2002 –2004 2005 –2008
referent 0.68 (0.45–1.05) 0.58 (0.34–0.99) 0.40 (0.18–0.89)
CD4+ T cell count (cells/mm3) ,200 ≥200
4.33 (2.97–6.32) referent
,0.001
3.36 (2.18– 5.17) referent
,0.001
3.68 (2.41– 5.62) referent
,0.001
HIV VL (log10 copies/mL) ≥5.0 ,5.0
4.24 (2.90–6.20) referent
,0.001
2.90 (1.86– 4.51) referent
,0.001
2.91 (1.88– 4.49) referent
,0.001
Prior mono or dual ART exposure yes no
2.04 (1.09–3.80) referent
0.025
1.59 (0.83– 3.04) referent
0.16
First cART regimen PI-based yes no
2.24 (1.36–3.71) referent
0.002
2.43 (1.44– 4.09) referent
,0.001
2.44 (1.47– 4.06) referent
,0.001
0.46 (0.22– 0.97) referent
0.041
0.45 (0.22– 0.93) referent
0.031
1.00 0.67
0.08 0.048 0.025
1.79 (1.04– 3.07) referent referent referent
0.034
1.52 (1.01– 2.30) referent
0.045
referent 0.84 (0.53– 1.31) 0.91 (0.52– 1.62) 0.84 (0.36– 1.98)
0.44 0.76 0.70
Received genotype test during follow-up (time-dependent variable) yes 0.93 (0.63–1.38) 0.73 no referent Novel agent use during follow-up (time-dependent variable) yes 0.58 (0.29–1.18) no referent
0.13
ART, ARV therapy. a The baseline date was defined as the earliest date during 1999 –2008 when a triple-ARV-class-experienced patient had an HIV VL .1000 copies/mL (log10 VL .3.0). b Additional baseline variables (listed in Table 1) were explored in univariate analyses and were not associated with mortality risk; for brevity, only the key demographic characteristics are shown in the table. c The multivariable model contains variables with P,0.20 in univariate analyses, and manual backwards selection was used to eliminate non-significant variables in order of highest P value.
2830
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
HR (95% CI)
JAC
Mortality factors among viraemic ARV-experienced HIV patients
Table 3. Cox proportional hazards models for mortality risk among the subset of TCF participants who had a genotype test at or after TCF date, HOPS 1999– 2013 (N ¼274) Variablea
Univariate HR (95% CI)
P
201 (73.4)
113 (41.2) 178 (65.0) 152 (55.5) 56 (20.4)
1.87 (0.87–4.01) 1.57 (1.16–2.13) 1.69 (0.94–3.04) 2.90 (1.47–5.72) 1.47 (0.82–2.65) 0.53 (0.29–0.96) 2.30 (1.30–4.09) 4.39 (2.39–8.07) 3.17 (1.14–8.85) 0.65 (0.29–1.47) 2.94 (1.16–7.44) 1.00 (0.99–1.01) 1.00 (0.98–1.01) 0.96 (0.89–1.03) 0.99 (0.97–1.02) 1.13 (0.94–1.35) 1.08 (0.61–1.92) 0.94 (0.51–1.74) 1.11 (0.62–2.01) 1.70 (0.91–3.19)
0.11 0.003 0.08 0.002 0.20 0.037 0.004 ,0.001 0.027 0.30 0.023 0.64 0.63 0.25 0.60 0.20 0.79 0.85 0.72 0.10
Any NRTI mutation M41L D67N K70E/R L74V M184I M184V L210W T215F T215Y T215F/Y K219E/Q
210 (76.6) 89 (32.5) 86 (31.4) 59 (21.5) 38 (13.9) 2 (0.7) 125 (45.6) 56 (20.4) 28 (10.2) 86 (31.4) 111 (40.5) 41 (15.0)
1.17 (0.57–2.43) 1.42 (0.79–2.56) 1.21 (0.67–2.19) 0.94 (0.47–1.86) 1.32 (0.62–2.83) 5.52 (0.75–40.7) 0.80 (0.44–1.43) 1.39 (0.73–2.64) 1.41 (0.60–3.33) 1.61 (0.64–2.09) 1.39 (0.78–2.46) 0.79 (0.34–1.87)
0.67 0.24 0.53 0.85 0.48 0.09 0.45 0.32 0.44 0.62 0.26 0.60
Any NNRTI mutation L100I K103N Y181C Y181I Y188L G190A/S
202 (73.7) 20 (7.3) 129 (47.1) 61 (22.3) 2 (0.7) 13 (4.7) 47 (17.2)
0.74 (0.39–1.39) 1.18 (0.42–3.28) 0.89 (0.50–1.59) 0.89 (0.44–1.79) 2.62 (0.36–19.0) 2.51 (0.99–6.34) 1.09 (0.54–2.20)
0.35 0.76 0.69 0.74 0.34 0.053 0.80
Any PI mutation V32I L33F M46I M46L I47V G48V I50V I54M
135 (49.3) 18 (6.6) 0 (0.0) 53 (19.3) 13 (4.7) 11 (4.0) 9 (3.3) 5 (1.8) 6 (2.2)
1.09 (0.61–1.95) 1.01 (0.31–3.26)
0.76 0.98
0.48 (0.20–1.12) 0.73 (0.18–3.00) 0.60 (0.08–4.32) 2.07 (0.64–6.66) 1.06 (0.15–7.70) 1.45 (0.20–10.5)
0.09 0.66 0.61 0.22 0.95 0.72
Male Age in years, per 10 year increment Non-Hispanic white race/ethnicity HIV risk from injection drug use Public/no health insurance TCF year 1999 –2008 (versus 1997 –1998) Baseline CD4+ T cell count (,200 cells/mm3) Baseline HIV VL ≥5.0 log10 copies/mL Prior mono or dual ART exposure Novel agent use during follow-up (time-dependent variable) First cART PI-based Total number of IAS major mutations Total number of IAS major NRTI mutations Total number of IAS major NNRTI mutations Total number of IAS major PI mutations Total number of TAMs Mutations in all three main classes ≥3 any mutations Any TAM ≥4 TAMs
131 (47.8) 30 (11.0) 147 (53.7) 209 (76.3) 87 (31.8) 36 (13.1) 209 (76.3) 119 (43.4) 200 (73.0)
Multivariableb HR (95% CI)
P
1.94 (1.36–2.78)
,0.001
2.71 (1.37–5.39)
0.004
5.35 (2.82–10.1)
,0.001
3.20 (1.25–8.17)
0.015
Continued
2831
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
n (%)
Palella et al.
Table 3. Continued Variablea V82A/F/L/S/T L90M
n (%)
Univariate HR (95% CI)
P
53 (19.3) 81 (29.6)
1.36 (0.71–2.58) 1.72 (0.97–3.07)
0.35 0.07
Multivariableb HR (95% CI)
P
ART, ARV therapy; IAS, International Antiviral Society. Baseline date: the earliest date during 1999 –2008 when a triple-ARV-class-experienced patient had HIV VL .1000 copies/mL (log10 VL .3.0). All variables in the model were measured as of baseline, except for variables related to ARV resistance mutations (assessed as of first post-baseline genotypic testing). b The multivariable model contains variables with P,0.20 in univariate analysis, and manual backwards selection was used to eliminate non-significant variables in order of highest P value. a
2832
associated with improved survival are consistent with this evolution. Our analysis has several limitations. Approximately half of the TCF patients had evaluable HIV genotype data. While we cannot readily explain the lack of association with mortality risk (improved or worsened) among persons with extensive resistance mutations, unmeasured confounding (e.g. resistant virus reflecting poor adherence, which was not assessed in this cohort) rather than some direct effect of specific mutational patterns may have been present. Finally, it is possible that we failed to detect associations between specific mutational combinations and mortality due to the relative rarity of these mutations or mutational patterns (a type 2 error). It is reassuring, however, that we did identify positive associations between mortality risk and known validated clinical variables in the overall group of TCF patients, i.e. a lower baseline CD4+ T cell count and higher plasma HIV RNA. In conclusion, among substantially ARV-experienced HIVinfected participants in the HOPS who remained viraemic despite continued therapy, mortality was associated with established clinical factors and ARV therapy factors (with PI-based initial cART associated with an increased mortality risk and the use of novel agents associated with improved survival), but not the presence of specific HIV drug resistance mutations or mutational patterns. Whether or not our findings are generalizable will require further evaluation in a larger population that includes more deaths, more genotypic HIV resistance tests and perhaps more person-years of observation. Ample data exist supporting the importance of routine HIV resistance testing among virally nonsuppressed ARV treatment-experienced patients11 in order to characterize the extent and type of ARV resistance present and to optimize ARV drug selection.29
Acknowledgements HOPS Investigators The HIV Outpatient Study (HOPS) Investigators include the following persons and sites: John T. Brooks, Kate Buchacz, Marcus D. Durham, Division of HIV/AIDS Prevention, National Center for HIV, STD, and TB Prevention (NCHSTP), Centers for Disease Control and Prevention (CDC), Atlanta, GA; Kathleen C. Wood, Darlene Hankerson, Rachel Debes, Thila Subramanian, Cheryl Akridge, Harlan Hayes, Carl Armon, Bonnie Dean, Jeff Binkley and Sam Bozzette, Cerner Corporation, Vienna, VA; Frank J. Palella, Joan S. Chmiel, Carolyn Studney, Saira Jahangir, Feinberg School
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
pattern of mutations and survival even among persons whose viral isolates demonstrated the most extensive resistance. Among our TCF patients and the subset of these with available HIV genotypic data, having received a first cART regimen that contained a PI (versus an NNRTI) was associated with a greater risk of death. While PI-containing initial cART, particularly regimens that include older PIs, can be associated with greater treatment-limiting toxicity, pill burden and lower rates of HIV suppression compared with NNRTI therapies, the association with a subsequent increased mortality among such persons at a timepoint distant from the initial cART, often after multiple cART regimens, is notable. The use of novel ARV agents among TCF patients was associated with a survival benefit. This was not surprising since the use of such therapies among ARV therapy-experienced patients in randomized prospective clinical trials has been shown to be associated with improved virological and immunological outcomes,14 – 17 yet this cohort is among the first in which a survival benefit associated with the use of these drugs among ARV therapy-experienced patients is apparent. Greater HIV replication capacity has been associated with mutations known to confer resistance to thymidine analogue NRTIs, notably with ≥4 TAMs.18 – 20 Other data exist indicating that viruses containing fewer TAMs20,21 can demonstrate decreased replication capacity and/or RT processivity. Other point mutations selected by ARVs reduce viral fitness in potentially clinically relevant ways that are exploitable among ARV recipients who experience incomplete HIV suppression.22 – 26 However, among our TCF patients, we observed no survival disadvantage associated with having ≥4 TAMS and no survival advantages associated with HIV mutations known to reduce viral fitness, such as M184V.27,28 While we did not identify any discrete care-related factors that could reasonably have impacted on mortality risk, and that improved in general over time, persons who became TCF in earlier time periods were more likely to have had exposure to mono or dual NRTI therapy (although such exposure was not associated with an increased risk of death in multivariable models). Moreover, the prescribed ARV regimens substantially evolved over time (for both the initial and the salvage regimen), with more persons receiving NNRTI-based first-line therapy and more effective, less toxic background ARVs during later time periods; our findings that a PI-based regimen as first cART was a risk factor for mortality and that the use of novel ARV agents during follow-up was
Mortality factors among viraemic ARV-experienced HIV patients
of Medicine, Northwestern University, Chicago, IL; Kenneth A. Lichtenstein and Cheryl Stewart, National Jewish Medical and Research Center Denver, CO; John Hammer, Kenneth S. Greenberg, Barbara Widick, and Rosa Franklin, Rose Medical Center, Denver, CO; Bienvenido G. Yangco and Kalliope Chagaris, Infectious Disease Research Institute, Tampa, FL; Doug Ward, Troy Thomas, and Fletcher Neale, Dupont Circle Physicians Group, Washington, DC; Jack Fuhrer, Linda Ording-Bauer, Rita Kelly and Jane Esteves, State University of New York (SUNY), Stony Brook, NY; Ellen M. Tedaldi, Ramona A. Christian, Faye Ruley, Dania Beadle and Princess Graham, Temple University School of Medicine, Philadelphia, PA; Richard M. Novak and Andrea Wendrow and Renata Smith, University of Illinois at Chicago, Chicago, IL; Benjamin Young, Barbara Widick, Mia Scott, APEX Family Medicine, Denver, CO.
This work was supported by contracts 200-2001-00133, 200-2006-18797 and 200-2011-41872 from the Centers for Disease Control and Prevention.
Transparency declarations F. J. P. is on speakers’ bureaus for Gilead Sciences, Merck, Janssen Pharmaceuticals and Bristol-Myers Squibb. C. A.’s time in working on the analysis, writing and editing, and preparing this article was funded by a contract with Cerner Corporation, which contracts with the CDC for the HOPS. J. S. C.’s time in working on the analysis, writing and editing, and preparing this article was funded under a subcontract to Northwestern University from Cerner Corporation, which contracts with the CDC for the HOPS. K. B., R. M. N., R. T. D’A. and J. T. B.: none to declare.
Disclaimer The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
References 1 Lorenzen T, Walther I, Stoehr A et al. Effective treatment of patients in a deep salvage situation with “non-active HAART”: experiences with the expert advice system RADATA. Infection 2009; 37: 528–33. 2 Napravnik S, Keys JR, Quinlivan EB et al. Triple-class antiretroviral drug resistance: risk and predictors among HIV-1-infected patients. AIDS 2007; 21: 825–34. 3 Steigbigel RT, Cooper DA, Kumar PN et al. Raltegravir with optimized background therapy for resistant HIV-1 infection. N Engl J Med 2008; 359: 339– 54. 4 Montaner J, Guimaraes D, Chung J et al. Prognostic staging of extensively pretreated patients with advanced HIV-1 disease. HIV Clin Trials 2005; 6: 281–90. 5 Lohse N, Jorgensen LB, Kronborg G et al. Genotypic drug resistance and long-term mortality in patients with triple-class antiretroviral drug failure. Antivir Ther 2007; 12: 909– 17. 6 Hogg RS, Bangsberg DR, Lima VD et al. Emergence of drug resistance is associated with an increased risk of death among patients first starting HAART. PLoS Med 2006; 3: e356. 7 Hardy WD, Gulick RM, Mayer H et al. Two-year safety and virologic efficacy of maraviroc in treatment-experienced patients with CCR5tropic HIV-1 infection: 96-week combined analysis of MOTIVATE 1 and 2. J Acquir Immune Defic Syndr 2010; 55: 558–64. 8 Nelson M, Fisher M, Gonzalez-Garcia J et al. Impact of baseline antiretroviral resistance status on efficacy outcomes among patients
receiving maraviroc plus optimized background therapy in the MOTIVATE 1 and 2 trials. HIV Clin Trials 2010; 11: 145–55. 9 Buchacz K, Baker RK, Young B et al. Changes in the use of HIV antiretroviral resistance testing in a large cohort of U.S. patients, 1999 to 2006. J Acquir Immune Defic Syndr 2010; 53: 625– 32. 10 Moorman AC, Holmberg SD, Marlowe SI et al. Changing conditions and treatments in a dynamic cohort of ambulatory HIV patients: the HIV outpatient study (HOPS). Ann Epidemiol 1999; 9: 349–57. 11 Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Updated February 12, 2013. http://aidsinfo.nih.gov/ guidelines (16 April 2014, date last accessed). 12 Williams I, Churchill D, Anderson J et al. British HIV Association guidelines for the treatment of HIV-1-positive adults with antiretroviral therapy 2012. HIV Med 2012; 13 Suppl 2: 1 –85. 13 Swindells S, Jiang H, Mukherjee AL et al. Lower CD4 cell count and higher virus load, but not antiretroviral drug resistance, are associated with AIDS-defining events and mortality: an ACTG Longitudinal Linked Randomized Trials (ALLRT) analysis. HIV Clin Trials 2011; 12: 79– 88. 14 Eron JJ, Cooper DA, Steigbigel RT et al. Efficacy and safety of raltegravir for treatment of HIV for 5 years in the BENCHMRK studies: final results of two randomised, placebo-controlled trials. Lancet Infect Dis 2013; 13: 587–96. 15 Anderson D, DeMasi R, DeLaitsch L et al. Week 96 outcomes of patients with less treatment experience versus more treatment experience receiving etravirine in the DUET trials. Curr HIV Res 2012; 10: 256–61. 16 Reynes J, Araste´h K, Clotet B et al. TORO: ninety-six-week virologic and immunologic response and safety evaluation of enfuvirtide with an optimized background of antiretrovirals. AIDS Patient Care STDS 2007; 21: 533– 43. 17 Taiwo B, Murphy RL, Katlama C. Novel antiretroviral combinations in treatment-experienced patients with HIV infection: rationale and results. Drugs 2010; 70: 1629 –42. 18 D’Aquila RT, Johnson VA, Welles SL et al. Zidovudine resistance and HIV-1 disease progression during antiretroviral therapy. AIDS Clinical Trials Group Protocol 116B/117 Team and the Virology Committee Resistance Working Group. Ann Intern Med 1995; 122: 401–8. 19 Caliendo AM, Savara A, An D et al. Effects of zidovudine-selected human immunodeficiency virus type 1 reverse transcriptase amino acid substitutions on processive DNA synthesis and viral replication. J Virol 1996; 70: 2146– 53. 20 Hu Z, Giguel F, Hatano H et al. Fitness comparison of thymidine analog resistance pathways in human immunodeficiency virus type 1. J Virol 2006; 80: 7020– 7. 21 Cong ME, Heneine W, Garcia-Lerma JG. The fitness cost of mutations associated with human immunodeficiency virus type 1 drug resistance is modulated by mutational interactions. J Virol 2007; 81: 3037–41. 22 White KL, Margot NA, Wrin T et al. Molecular mechanisms of resistance to human immunodeficiency virus type 1 with reverse transcriptase mutations K65R and K65R+M184V and their effects on enzyme function and viral replication capacity. Antimicrob Agents Chemother 2002; 46: 3437– 46. 23 Deval J, White KL, Miller MD et al. Mechanistic basis for reduced viral and enzymatic fitness of HIV-1 reverse transcriptase containing both K65R and M184V mutations. J Biol Chem 2004; 279: 509–16. 24 Tupinambas U, Aleixo A, Greco D. HIV-1 genotypes related to failure of nelfinavir as the first protease inhibitor treatment. Braz J Infect Dis 2005; 9: 324– 9.
2833
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
Funding
JAC
Palella et al.
25 Martinez-Picado J, Savara AV, Sutton L et al. Replicative fitness of protease inhibitor-resistant mutants of human immunodeficiency virus type 1. J Virol 1999; 73: 3744– 52. 26 Shafer R. Stanford University HIV Drug Resistance Database. http:// hivdb.stanford.edu/DR/NRTIResiNote.html (16 April 2014, date last accessed). 27 von Wyl V, Ehteshami M, Demeter LM et al. HIV-1 reverse transcriptase connection domain mutations: dynamics of emergence and implications
for success of combination antiretroviral therapy. Clin Infect Dis 2010; 51: 620– 8. 28 Kuritzkes DR, Bassett RL, Hazelwood JD et al. Rate of thymidine analogue resistance mutation accumulation with zidovudine- or stavudine-based regimens. J Acquir Immune Defic Syndr 2004; 36: 600–3. 29 Palella FJ Jr, Armon C, Buchacz K et al. The association of HIV susceptibility testing with survival among HIV-infected patients receiving antiretroviral therapy: a cohort study. Ann Intern Med 2009; 151: 73– 84.
Downloaded from http://jac.oxfordjournals.org/ at Mount Allison University on June 23, 2015
2834
