Time to Viremia for Patients Taking their First Antiretroviral Regimen and the Subsequent Resistance Profiles

Abstract

Background: The resistance profiles for patients on first-line antiretroviral therapy (ART) regimens after viremia have not been well studied in community clinic settings in the modern treatment era.

Objective: To determine time to viremia and the ART resistance profiles of viremic patients.

Methods: HIV-positive patients aged ≥16 years initiating a three-drug regimen were retrospectively identified from 01/01/06 to 12/31/12. The regimens were a backbone of two nucleoside reverse transcriptase inhibitors (NRTIs) and a third agent: a protease inhibitor (PI), non-nucleoside reverse transcriptase inhibitor (NNRTI), or an integrase inhibitor (II). Time to viremia was compared using a proportional hazards model, adjusting for demographic and clinical factors. Resistance profiles were described in those with baseline and follow-up genotypes.

Results: For 653 patients, distribution of third-agent use and viremia was: 244 (37%) on PIs with 80 viremia, 364 (56%) on NNRTIs with 84 viremia, and 45 (7%) on II with 11 viremia. Only for NNRTIs, time to viremia was longer than PIs (p = 0.04) for patients with a CD4 count ≥200 cells/mm³. Of the 175 with viremia, 143 (82%) had baseline and 37 (21%) had follow-up genotype. Upon viremia, emerging ART resistance was rare. One new NNRTI (Y181C) mutation was identified and three patients taking PI-based regimens developed NRTI mutations (M184 V, M184I, and T215Y).

Conclusions: Time to viremia for NNRTIs was longer than PIs. With viremia, ART resistance rarely developed without PI or II mutations, but with a few NRTI mutations in those taking PI-based regimens, and NNRTI mutations in those taking NNRTI-based regimens.

Keywords:

• HIV
• first-line antiretroviral therapy
• viremia
• resistance mutations

Introduction

Antiretroviral therapy (ART) has greatly changed the course of HIV infection over the past 25 years. The evolution from the era of mono- and dual-therapy to triple combination therapy in 1996 has significantly reduced AIDS-related morbidity and mortality,¹⁻⁵ perinatal transmission,⁶ and onward HIV transmission.⁷⁻¹¹ Current guidelines for the use of antiretrovirals (ARVs) in adults and adolescents recommend the use of a backbone of two nucleoside reverse transcriptase inhibitors (NRTIs) and a third agent: either a non-nucleoside reverse transcriptase inhibitor (NNRTI), a protease inhibitor boosted with ritonavir (PI/r), or an integrase inhibitor (II).^{12, 13}

Primary concerns for clinicians and patients are viremia, the development of drug resistance, and the possible transition from the state of viremia to virologic failure. The causes of viremia for patients on first-line ART regimens include the periodic release of HIV from latently infected cells or sites inadequately accessed by ART ^{14, 15} as well as through immune activation due to immunization, for example. ^{16, 17} Other factors, although with conflicting evidence reviewed by Ryscavage et al. (2006), ¹⁸ are related to ART treatment and adherence and include suboptimal adherence,^{18, 19} treatment initiation during late stage diagnosis,^{18, 20, 21} the type of ARV regimen particularly in those with high baseline plasma viral load, ²² and the presence of baseline drug resistance.²²⁻²⁸ Screening for baseline ARV resistance is a standard of care today and a determinant of successful first-line therapies.^{12, 29−31} The association between adherence and the development of drug resistance differs dramatically across ARV classes.^{24, 25, 28, 32−34} For example, slight differences in the chemical structure of PIs lead to primary mutations conferring cross-resistance to several PIs.³⁵ Interestingly, such mutations contribute rarely to viremia resulting in virologic failure and few or no mutations are detected in patients failing their first PI-based regimen.³⁶ This high genetic barrier translates into a requirement for the accumulation of multiple mutations before their antiviral activity is negatively affected by resistance.^{37, 38} Ritonavir-boosted lopinavir requires three to four mutations for resistance and even more mutations for ritonavir-boosted darunavir.^{30, 39−43} Thus, PIs offer long-term viral suppression and in the event of poor adherence, rarely lead to the development of resistance and maintain a wide selection of options for salvage therapies.⁴⁰

In contrast to PIs, first-generation NNRTIs have a low genetic barrier to resistance such that a single-point mutation often confers a high degree of cross-resistance within the NNRTI class.^{44, 45} The long half-life of NNRTIs could result in viral replication in the presence of suboptimal plasma ARV levels, if all drugs are stopped enabling the selection of NNRTI-resistant viruses.^{35, 46−48} In standard first-line NNRTI-based regimens, the NNRTI mutations, K103 N and Y181C have been linked to 50–70% of virologic failure.⁴⁹ The K103 N mutation emerging most commonly in response to efavirenz confers cross-resistance to efavirenz, nevirapine, and delavirdine.^{50, 51} In contrast, Y181C developed in response to nevirapine and confers cross-resistance to nevirapine and delavirdine and in some cases, efavirenz.⁵²⁻⁵⁴ Another commonly emerging resistance mutation, E138 K, has been shown to confer cross-resistance to other NNRTIs (i.e. etravirine, efavirenz, and nevirapine) in half of the individuals receiving rilpivirine.⁵⁵⁻⁵⁷ Minority NNRTI-resistant variants, not detected by standard genotypic testing, were also present in treatment-experienced individuals with virologic failure.⁵⁸

II mutations have emerged with evidence of variable levels of cross-resistance, whereby raltegravir mutations have cross-resistance with elvitegravir and with some also having an impact on dolutegravir. Specifically, the common resistance-associated mutations were: Y143C/H/R, N155H, and Q148H/K/R in the absence of secondary II mutations.⁵⁹⁻⁶² The genetic barrier of the IIs, raltegravir and elvitegravir, to resistance have been shown to be lower than that of PIs and a bit more than NNRTIs, as two mutations are often required for resistance. Dolutegravir, on the other hand, has a higher genetic barrier with R263 K (clade B) or G118R (clade C) and H51Y being the only unique mutations.⁵⁹⁻⁶¹

The purpose of this retrospective cohort study was to determine whether time to viremia for patients differed between the three standard first-line ART regimens, which included a backbone of two NRTIs, with either a (1) NNRTI, (2) PI, or (3) II. This cohort study also assessed if the resistance profiles of those experiencing viremia differed.

Methods

Study design and setting

The Maple Leaf Medical Clinic (MLMC), founded in 2001, is an urban multi-specialty community clinic in Toronto, Ontario, Canada providing primary and specialty care to approximately 6,900 persons including 2,677 HIV-positive patients. A retrospective clinical chart review of HIV-positive patients was done using information collected and retained within the MLMC Electronic Medical Record system (EMR). The EMR contains demographic and clinical characteristics of MLMC patients collected since 1 July 2005. Using the MLMC EMR, HIV-positive patients were retrospectively selected for analysis.

Participants

Patients were included if they: were aged 16 or over; started a three-drug first-line ART regimen (with two NRTIs and a third agent being either a NNRTI, PI, or II) between 1 January, 2006 and 31 December 2012;and remained on ART for at least 14 days and had at least one viral load test within 6 months of starting ART. Patients were excluded if: baseline viral load was less than 40 copies/mL (as some patients could have been transferred from other clinics on ART); ARV was within a clinical trial; or having discontinued a third-agent ARV class or added a new ARV prior to day 14 on their regimens.

Variables

Demographic variables collected from the MLMC EMR were: age, gender, and HIV risk factors. Clinical variables collected were: time since HIV diagnosis, baseline viral load, baseline CD4 count, ART regimen (i.e. NRTI backbone and a third agent, NNRTI, PI, or II), hepatitis B and C serostatus, and the number of viral load test from the first day of ARV up to and including the last day of follow-up. Viral load testing is typically carried out at MLMC monthly, after the start of ART until the viral load is undetectable, then every three–four months for two years, and then every four–six months, as per guidelines.^{12, 13} Viral load testing in Ontario is done using Abbott’s m2000 RealTime™ System and RealTime™ HIV-1 assay (Abbott Molecular Inc., Mississauga, Ontario, Canada) as of 16 August 2010 and Siemens’ Versant® HIV-1 bDNA 3.0 Assay (Siemens, Oakville, Ontario, Canada) prior to this date. The primary outcome of viremia was defined for this study as either (1) no suppression by 6 months or (2) after suppression, two consecutive viral loads greater than 40 and below 200 copies/mL at least 14 days apart or one viral load greater than 200 copies/mL. Virologic suppression was defined as a viral load less than or equal to 50 copies/mL or 40 copies/mL (after 1 January 2011). The secondary outcome of interest is the resistance profile of viremic patients. Resistance testing at MLMC is carried out at the British Columbia Centre for Excellence in HIV/AIDS (Vancouver, British Columbia, Canada) using validated “in-house” methods.²⁸ Pre-treatment resistance testing has become standard for patients prior to starting ART; however, some are missed due to being transferred from another clinic. Follow-up resistance testing is conducted on any sample with a viral load greater than 200 copies/mL. The resistance profiles were described for each ARV class and further classified by drugs within the classes. Emergent mutations per patient by third agent and definition of viremia were also described.

Statistical methods

Descriptive analyses were reported as means and standard deviations or medians and interquartile ranges (IQR) for continuous variables and compared using the Kruskal–Wallis test or analysis of variance (ANOVA). Categorical variables were summarized using frequencies and proportions and Fisher’s exact test or the C test was used for comparisons. The log rank test was used to compare time to viremia between the PI, II, and NNRTI groups, and each pairwise comparison. There was no adjustment made for multiple comparisons since this is an explorative analysis. Proportional hazards models were used to evaluate if treatment group was associated with time to viremia. A multivariable proportional hazards model was adjusted for age, sex, baseline viral load, and count of viral load tests during follow-up. Variables with an unadjusted p-value < 0.05 were included in the multivariable model. Patients were censored if they: (1) were viremic by 31 December 2013; and if between 1 January 2006 to 31 December 2012 they: (2) died; (3) were lost to follow-up; (4) discontinued or switched third agent ARV class after 14 days of initiating treatment; or (5) did not have a second viral load available to determine if viremia occurred. Resistance profiles for NRTIs and third agents, NNRTI, PI, or II, were described in patients with viremia who had both baseline and follow-up genotypes. Variables in each analysis were stratified by CD4 count less than and greater or equal to 200 cells/mm³_, where the former was suggestive of advanced HIV disease. Stratification by CD4 cell count was done to address channeling bias with patients with more advanced HIV disease more frequently starting a PI-based regimen. SAS version 9.3 (SAS Institute Inc., Cary, NC) was used for all analyses.

Missing data consisting of no dates and viral load were excluded from the analysis. Data from individuals who were loss to follow-up were included in the analysis up to the date of their last visit.

Results

Patient characteristics

There were 10,784 patients in the MLMC database with 3,801 being HIV positive. Of the 2,560 active patients living with HIV (i.e. not transferred, not dead, or loss to follow-up), 716 patients started on a standard first-line ART regimen containing two backbone NRTIs with an NNRTI, PI and II between 1 January 2006 and 31 December 2012 and stayed on it for more than 14 days. Fifty-six patients did not meet the inclusion criteria of having at least one follow-up viral load. Seven additional patients were excluded because they did not have a baseline CD4 cell count.

Of the 653 patients who met the study inclusion criteria, 91.5% with a CD4 count less than 200 cells/mm³ and 93.7% with a CD4 count greater or equal to 200 cells/mm³ were male with a mean age of 39.7 (SD = 8.8) and 38.8 (SD = 9.3), respectively, and a median follow-up period of 35.3 (IQR: 32.2–39.3) months. Distribution of third-agent use was: NNRTI 40.1% (n = 71), PI 57.6% (n = 102), and II 2.3% (n = 4) for patients with a CD4 count less than 200 cells/mm³. Distribution was: NNRTI 61.6% (n = 293), PI 29.8% (n = 142), and II 8.6% (n = 41) for those with a CD4 count greater or equal to 200 cells/mm³. Frequency of viral tests per year of follow-up (p = .004) and hepatitis B serostatus (p = .01) for patients with a CD4 cell count greater or equal to 200 cells/mm³ was significantly different between third-agent groups (Table 1). The initial backbone medications were: emtricitabine with tenofovir, lamivudine with abacavir or zidovudine or tenofovir, or stavudine, and abacavir with tenofovir. Patient characteristics are summarized in Table 1.

Table 1 Baseline demographic and clinical characteristics stratified by CD4 counts less than and equal or greater than 200 cells/mm³

	All	PI	NNRTI	II	P
	(n = 653)	(n = 244)	(n = 364)	(n = 45)
Frequency, n (%)
<200 cells/mm^3	177 (100)	102 (57.6)	71 (40.1)	4 (2.3)
≥200 cells/mm^3	476 (100)	142 (29.8)	293 (61.6)	41 (8.6)
Age, Mean (SD)
<200 cells/mm^3	39.7 (8.8)	40.2 (9.0)	39.2 (8.8)	37.5 (3.1)	0.67
≥200 cells/mm^3	38.8 (9.3)	37.9 (9.1)	39.4 (9.5)	37.4 (8.7)	0.20
Male, n (%)
<200 cells/mm^3	162 (91.5)	90 (88.2)	68 (95.8)	4 (100)	0.18
≥200 cells/mm^3	446 (93.7)	122 (85.9)	285 (97.3)	39 (95.1)	<0.0001
Time since diagnosis (months), Mean (SD)
<200 cells/mm^3	46.7 (60.2)	50.1 (68.2)	42.7 (48.1)	29.9 (29.3)	0.63
≥200 cells/mm^3	46.4 (51.7)	51.7 (63.2)	44.5 (46.0)	41 (45.5)	0.32
Baseline HIV viral load, copies/mL, Mean (SD)
<200 cells/mm^3	206233.2 (229756.1)	214140.6 (243314.2)	197657.7 (214980.8)	156809.3 (135083.8)	0.82
≥200 cells/mm^3	88654.5 (122773.1)	105650.1 (138562.0)	83940.3 (120073.9)	63480.5 (65033.5)	0.09
Baseline HIV viral load, n (%)
<200 cells/mm^3
<100'000 copies/mL	79 (44.6)	45 (44.1)	32 (45.1)	2 (50.0)	1
≥100'000 copies/mL	98 (55.4)	57 (55.9)	39 (54.9)	2 (50.0)
≥200 cells/mm^3
<100'000 copies/mL	348 (73.1)	98 (69.0)	219 (74.7)	31 (75.6)	0.41
≥100'000 copies/mL	128 (26.9)	44 (31.0)	74 (25.3)	10 (24.4)
Count of viral load tests during follow-up, Mean (SD)
<200 cells/mm^3	11.3 (7.3)	11.1 (7.4)	11.7 (7.2)	9.5 (3.7)	0.77
≥200 cells/mm^3	10 (6.3)	9.8 (6.7)	10.1 (6.1)	9.7 (6.5)	0.86
Frequency of viral load tests/year of follow-up, Mean (SD)
<200 cells/mm^3	4.8 (4.6)	5.3 (3.4)	4 (6.0)	5.2 (1.8)	0.23
≥200 cells/mm^3	4.3 (3.1)	5 (3.7)	4.1 (2.7)	3.7 (2.5)	0.004
Baseline CD4 count, n (%)
<200 cells/mm^3	120.3 (53.4)	113.4 (55.5)	129.2 (50.1)	140.3 (37.4)	0.12
≥200 cells/mm^3	361.6 (156)	350.2 (144.5)	362.5 (161.1)	394.2 (155.8)	0.28
Hepatitis B, n (%)
<200 cells/mm^3
Yes	2 (1.1)	1 (1.0)	1 (1.4)	-	1
No	122 (68.9)	73 (71.6)	47 (66.2)	2 (50.0)
Unknown	53 (29.9)	28 (27.4)	23 (32.4)	2 (50.0)
≥200 cells/mm^3
Yes	14 (2.9)	8 (5.6)	4 (1.4)	2 (4.9)	0.01
No	328 (68.9)	88 (62.0)	216 (73.7)	24 (58.5)
Unknown	134 (28.2)	46 (32.4)	73 (24.9)	15 (36.6)
Hepatitis C, n (%)
<200 cells/mm^3
Yes	9 (5.1)	6 (5.9)	3 (4.2)	-	0.79
No	168 (94.9)	96 (94.1)	68 (95.8)	4 (100.0)
≥200 cells/mm^3
Yes	24 (5.0)	9 (6.3)	12 (4.1)	3 (7.3)	0.41
No	452 (95.0)	133 (93.7)	281 (95.9)	38 (92.7)
Risk factors, n (%)
<200 cells/mm^3
Country of high HIV prevalence	10 (5.6)	8 (7.8)	2 (2.8)	-
Heterosexual	26 (14.7)	15 (14.7)	11 (15.5)	-
IDU	8 (4.5)	4 (3.9)	3 (4.2)	1 (25.0)
MSM	132 (74.6)	71 (69.6)	57 (80.3)	4 (100.0)
Other	34 (19.2)	21 (20.6)	12 (16.9)	1 (25.0)
≥200 cells/mm^3
Country of high HIV prevalence	13 (2.7)	10 (7.0)	3 (1.0)	-
Heterosexual	34 (7.1)	16 (11.3)	16 (5.5)	2 (4.9)
IDU	9 (1.9)	5 (3.5)	3 (1.0)	1 (2.4)
MSM	361 (75.8)	97 (68.3)	232 (79.2)	32 (78.0)
Other	72 (15.1)	19 (13.4)	49 (16.7)	4 (9.8)
Initial backbone medications, n (%)
<200 Cells/mm^3
3TC+ABC	39 (21.5)	27 (26.5)	10 (14.1)	1 (25.0)
3TC+AZT	9 (5.1)	7 (6.9)	2 (2.8)	-
FTC+TDF	127 (71.8)	65 (63.7)	59 (83.1)	3 (75.0)
3TC+d4T	2 (1.1)	2 (2.0)	-	-
ABC+TDF	1 ( 0.6)	1 ( 1.0)	-	-
≥200 cells/mm^3
3TC+ABC	63 (13.2)	33 (23.2)	29 (9.9)	1 (2.4)
3TC+AZT	12 (2.5)	8 (5.6)	4 (1.4)	-
FTC+TDF	399 (83.8)	101 (71.1)	258 (8.0)	40 (97.6)
3TC+d4T	2 (0.4)	-	2 (0.7)	-
ABC+TDF	-	-	-	-

Note:

Continuous variables summarized using means and standard deviations (SD) and categorical variables as frequency and proportions, n (%). Risk factors are not mutually exclusive such that ‘Other’ includes drug use (not IDU), accidental and occupational exposure, no known risk factors, sex trade worker and sex with women. PI = protease inhibitor; NNRTI = non-nucleoside reverse transcriptase inhibitor; II = integrase inhibitor; IDU = injection drug use; MSM = men who have sex with men; 3TC = lamivudine; ABC = abacavir; AZT = zidovudine; FTC = emtricitabine; TDF = tenofovir; d4T = stavudine.

The majority of the patients received an efavirenz-based regimen (45.6%), next a ritonavir-boosted lopinavir-based regimen (15.0%), and thirdly, an atazanavir-based regimen (11.6%) (Table 2). Third ARV agents and viremia distribution are summarized in Table 2. In brief, raltegravir was the only II investigated as a third agent; first-and-second generation NNRTIs (i.e. efavirenz, etravirine, nevirapine, and rilpivirine) were among the third agents studied and lastly, six PIs (i.e. atazanavir- and ritonavir-boosted atazanavir, ritonavir-boosted darunavir, ritonavir-boosted saquinavir, ritonavir-boosted lopinavir, nelfinavir,and ritonavir-boosted fosamprenavir) (Table 2).

Table 2 Third antiretroviral agents and viremia distributions

				CD4 <200 cells/mm^3			CD4 ≥200 cells/mm^3
Class	ARV	Total (n = 653)	Viremia (n = 175)	VL > 50 × 2 (n = 7)	VL > 200 (n = 21)	Never VS (n = 21)	VL > 50 × 2 (n = 18)	VL > 200 (n = 32)	Never VS (n = 54)
II, n (%)
	RAL	45 (100)	11 (24.4)	0 (0.0)	1 (100)	0 (0.0)	4 (40.0)	4 (40.0)	2 (20.0)
NNRTI, n (%)
	EFV	298 (81.8)	66 (22.1)	2 (10.5)	6 (31.6)	11 (57.9)	8 (17.0)	10 (21.3)	29 (61.7)
	ETV	15 (4.1)	7 (46.7)	-	-	-	3 (42.9)	0 (0.0)	4 (57.1)
	NVP	35 (9.6)	11 (31.4)	0 (0.0)	1 (33.3)	2 (66.7)	2 (25.0)	2 (25.0)	4 (50.0)
	RLP	16 (4.4)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
PI, n (%)
	ATV	4 (1.6)	1 (25.0)	-	-	-	0 (0.0)	1 (100)	0 (0.0)
	ATV/r	76 (31.1)	27 (35.5)	0 (0.0)	6 (46.2)	7 (53.8)	1 (7.1)	6 42.9)	7 (50.0)
	DRV/r	54 (22.1)	12 (22.2)	1 (50.0)	0 (0.0)	1 (50.0)	0 (0.0)	2 (20.0)	8 (80.0)
	SQV/r	9 (3.7)	1 (11.1)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	1 (100)	0 (0.0)
	LPV/r	98 (40.2)	38 (38.8)	4 (18.2)	7 (31.8)	11 (50.0)	0 (0.0)	6 (37.5)	10 (62.5)
	NFV	1 (0.4)	1 (100)	-	-	-	0 (0.0)	0 (0.0)	1 (100)
	fAMP/r	2 (0.8)	0 (0.0)	-	-	-	0 (0.0)	0 (0.0)	0 (0.0)

Note:

Categorical variables summarized using frequency and proportions, n (%). VL = viral load; VS = virologic suppresion; ARV = antiretroviral; II = integrase inhibitor; RAL = raltegravir; NNRTI = non-nucleoside reverse transcriptase inhibitor; EFV = efavirenz; ETV = etravirine; NVP = nevirapine; RLP = rilpivirine; PI = protease inhibitor; ATV = atazanavir; /r = ritonavir boosted; DRV = darunavir; SQV = saquinavir; LPV = lopinavir; NFV = nelfinavir; fAMP = fosamprenavir.

Viremia and resistance analysis

The viremia rate was 23% (n = 84/364) for patients taking an NNRTI-based regimen, 33% (n = 80/244) for patients taking a PI-based regimen, and 24% (n = 11/45) for the II-based regimen (Table 2). In the patients who experienced viremia, median time to viremia (months) was longer with NNRTIs (point estimate [PE]: 37.4, 95% CI: 32.5–43.4) compared to PIs (PE: 33.1, 95% CI: 25.1–39.3) and IIs (PE: 35.4, 95% CI: 25.1–42.5). There was no evidence that time to viremia with NNRTIs was different from IIs and IIs from PIs (data not shown). There was no difference in time to viremia between NNRTI and PI for patients with a CD4 cell count less than 200 cells/mm³ (p = .08) (Figure 1A), whereas there was a significant difference in time to viremia between NNRTI and PI (p = .016) (Figure 1B) and between NNRTI, PI, and II (p = .04) (Figure 1C) for patients with a CD4 cell count greater or equal to 200 cells/mm³. Using PI treatment as the reference group and adjusting for age, sex, baseline CD4 count, hepatitis B serostatus, baseline viral load, and count of viral load test during follow-up, the significant aHRs for viremia were: 2.11 [95% CI: 1.32–3.37] for baseline viral (Log10) for patients with a CD4 count less than 200 cells/mm³; and for patients with a CD4 count greater or equal to 200 cells/mm³, 1.43 [95% CI: 1.06–1.93] for baseline viral load (Log10) and 0.61 [95% CI: 0.41–0.91] for NNRTI. The unadjusted and adjusted Cox proportional hazards models are presented in Table 3.

Figure 1 Time to viremia compared by antiretroviral classes; integrase inhibitor (II), non-nucleoside reverse transcriptase inhibitor (NNRTI), and protease inhibitor (PI) for patients with a CD4 count less than 200 cells/mm³ (A) and greater or equal to 200 cells/mm³ (B and C).

Table 3 Unadjusted and adjusted Cox proportional hazard models for viremia

	CD4<200 copies/mL		CD4≥200 copies/mL
	Unadjusted hazard ratio (95% CI)	Adjusted hazard ratio (95% CI)	Unadjusted hazard ratio (95% CI)	Adjusted hazard ratio (95% CI)
Age	0.97 (0.94-1.00)	0.97 (0.94-1.00)	1.00 (0.98-1.02)	1.00 (0.98-1.02)
Sex
Male	1	1	1	1
Female	1.51 (0.64-3.53)	1.31 (0.54-3.22)	0.66 (0.24-1.78)	0.56 (0.20-1.55)
Baseline viral load Log10	1.76 (1.13-2.74)	2.11 (1.32-3.37)	1.37 (1.03-1.83)	1.43 (1.06-1.93)
HIV class
PI	1	1	1	1
NNRTI	0.64 (0.38-1.09)	0.57 (0.33-1.00)	0.63 (0.42-0.92)	0.61 (0.41-0.91)
II			1.00 (0.98-1.02)	0.61 (0.30-1.22)
Baseline hepatitis B serostatus	1.03 (0.97-1.09)	1.06 (1.00-1.13)	1.04 (0.99-1.08)	1.03 (0.99-1.08)
Count of viral load tests during follow-up	0.96 (0.92-1.00)	0.94 (0.90-0.98)	0.97 (0.94-1.00)	0.96 (0.93-0.99)

Note:

PI = protease inhibitor; NNRTI = non-nucleoside reverse transcriptase inhibitor; II = integrase inhibitor.

Of the 653 patients, 86% (n = 565) had a baseline genotype. Among these patients, 175 had viremia of which 82% (n = 143) had a baseline genotype and 21% (n = 37) had a follow-up genotype (Table 4). Among those with a follow-up genotype, 46% (n = 17) were on a PI-based regimen, 49% (n = 18) on an NNRTI-based regimen, and 5% (n = 2) on an II-based regimen (Table 4). The NNRTI resistance mutation, Y181C, was detected in one person on an NNRTI-based regimen who experienced no virologic suppression at 6 months and had CD4 count greater or equal to 200 cells/mm³. Three persons on a PI-based regimen had the following NRTI mutation patterns each; M184 V + T215Y after viremia with one viral load greater than 200 copies/mL and a CD4 count greater or equal to 200 cells/mm³ and M184 V and M184 V/I with no virologic suppression at 6 months in patients with a CD4 count less than or greater or equal to 200 cells/mm³, respectively (Table 4).

Table 4 New mutations per patient experiencing viremia by class and type of viremia

	PI	NNRTI	II
BL genotype, n (%)	63 (78.8)	70 (83.3)	10 (90.9)
BL & F/U genotype, n (%)	17 (21.2)	18 (21.4)	2 (18.2)
	CD4 <200 cells/mm^3
Viremia type
VL >50 x 2 consecutive	-	-	-
VL >200 x 1	PI: +35DPI: +77IRT: +62V ; PI: +37N+64V	PI: +13V+15V+37N+77IRT: +236L
No VS @ 6 months	PI: +62VRT: +184V ; PI: +10I+63SRT: +335D+371V+399D ; PI: +60E+62V+72V+93L	-	-
	CD4 ≥200 cells/mm^3
VL >50 x 2 consecutive	-	PI: +37NRT: +189I ; PI: +16E+62V+63PRT: +98S	-
VL >200 x 1	PI: +15V	PI: +64LPI: +64V+69QRT: +238T	PI: +10V+35D+37N+75IPI: +62V+82I
	PI: +63P
	RT: +184V+215Y ; PI: +10V+11I
	RT: +62V ; PI: +37N+63S+93L
No VS @ 6 months	RT: +179D+184I+184V+210F ; PI: +37N+64V	PI: +15V+62V+64L+72TRT: +181CRT: +365I+399D ; PI: +10I+64L	-

Note:

Categorical variables summarized using frequency and proportions, n (%). PI = protease inhibitor; NNRTI = non-nucleoside reverse transcriptase inhibitor; II = integrase inhibitor; BL = baseline; F/U = follow-up; VL = viral load; VS = virologic suppression; RT = reverse transcriptase.

Discussion

Among patients accessing care in an urban community clinic setting, time to viremia was significantly delayed among persons who started a first-line NNRTI-based regimen compared to those who started a first-line PI-based regimen. In those experiencing viremia, ART resistance rarely developed with no PI mutations, but a few NRTI mutations in those taking PI-based regimens, and NNRTI mutations in those taking an NNRTI-based regimen.

Our results show that time to viremia was significantly longer among persons with a CD4 cell count greater or equal to 200 cells/mm³_, following an NNRTI-based regimen compared to those on a PI-based regimen. Other studies have similarly shown longer time to viremia and specifically virologic failure for persons on an NNRTI-based regimen compared to a ritonavir-boosted PI-based regimen and that low level rebound occurred less frequently in patients on an NNRTI-based regimen compared to other regimens.⁶³⁻⁷⁰ Also, 45.8% of the patients in our study were on an efavirenz-based regimen and the literature has shown that efavirenz-based regimen exhibits improved immunological and virological outcomes as well as greater adherence compared to nevirapine, boosted PI, and single PI-based regimens.⁷⁰ Other reasons may also include that NNRTIs have a longer half-life than PIs and in the event of less than perfect adherence, levels of NNRTI may still be adequate.^{35, 46−48, 71} An example of this has been demonstrated in the FOTO pilot study where a short cycle treatment interruption strategy (i.e. five-days-on (FO) and two-days-off (TO)) demonstrated that virologic suppression was durable for patients on an efavirenz- and nevirapine-based regimens.^{72, 73} Also, patients with lower CD4 count presumed poorer adherence and higher viral load was placed on a PI-based regimen due to the high genetic barrier and the multiple mutations that are required for resistance. As such, we stratified our findings by baseline CD4 counts less than 200 cells/mm³ as suggestive of advanced HIV disease and greater than or equal to 200 cells/mm³ to account for this channeling bias. In our study, only a few patients were on an II-based regimen (n = 45) and all of them restricted to raltegravir precluding us from drawing comparisons regarding time to viremia with patients on either a PI- or NNRTI-based regimen.

Several factors have previously been described as leading to viremia and include as previously mentioned release of HIV from latently infected cells or sites not readily accessed by ART, immune activation, suboptimal adherence, type of ARV regimen, stage of infection prior to initiating ART, and baseline drug resistance.^{14, 15, 18−28} We showed an association between viremia and increasing baseline viral load for both patients with a CD4 count less than 200 cells/mm³ and patients with CD4 counts greater or equal to 200 cells/mm³. Careful monitoring of viremia is required since several studies have also demonstrated persistent low-level viremia is associated with increased risk of virologic failure.^{63, 74, 75}

In our study, major mutations were not reported in HIV-positive patients following an II-based regimen. In contrast, in patients following NNRTI- and PI-based regimens, NRTI resistance mutations were described. The discriminatory NRTI resistance mutations, M184 V and M184 V/I, were each identified in two patients following a PI-based regimen. The NRTIs mutations, M184 V/I, are the most common NRTI mutations and are selected by lamivudine and emtricitabine, which confer cross-resistance to abacavir and didanosine and susceptibility to tenofovir and zidovudine.^{35, 37, 76−84} Lamivudine in the MLMC cohort is part of four of the five backbone regimens. The NRTI thymidine analog mutation T215Y was also identified among a HIV-positive patient taking a PI-based regimen who also had a M184 V mutation. T125Y is selected only by zidovudine- and stavudine-based regimens and confer cross-resistance to tenofovir, abacavir, and didanosine.⁸²⁻⁸⁴ Few patients in the MLMC cohort are on a regimen, which includes stavudine (n = 4) and zidovudine (n = 21). The mutational pattern of M184 V and T215Y is one of the most common combinations and is associated with a high level of resistance against NRTIs.^{35, 37, 76−84} Also in this cohort, a HIV-positive patient who experienced viremia after taking an NNRTI-based regimen had the NNRTI Y181C mutation found. The latter mutation develops in response to nevirapine with cross-resistance to delavirdine and in some cases, efavirenz.^{35, 54, 84} Three out of the four mutations described were present in patients with a CD4 count greater or equal to 200 cells/mm³ and on a PI-based regimen. A previous study in British Columbia with treatment of naïve patients without baseline resistance showed that 19% (38/196) of patients had a least one drug resistance mutation at their first low-level viremia affecting NRTIs and NNRTIs ⁸⁵ and other studies have similarly showed low risk of PI-resistant mutations at low-level viremia^{86, 87} and important major NRTI (e.g. M184 V and T215Y) and NNRTI (e.g. Y181C) resistance mutations at the first low-level viremia of individuals primarily on PI-based regimens.⁸⁸

Virologic suppression at six months was not attained in a person taking the NNRTI-based regimen with an Y181C mutation and in two patients following PI-based regimens exhibiting M184 V/I mutations. As such, major resistance mutations and viremia in these patients are suggestive of a failure of the ART regimen to suppress HIV. In a patient with a viral load greater or equal to 200 copies/mL, the mutational pattern of M184 V and T215Y was described. Viremia and major resistance mutations in these patients may be indicative of low-level viremia, requiring further monitoring to prevent transition to virologic failure. There were few major mutations present among the patients. Despite this, viremia was reported for 2 (18.2%) patients on an II-based regimen, 18 (20.7%) patients on an NNRTI-based regimen, and 17 (21.0%) patients following a PI-based regimen. Identified causes for viremia such as the existence of minority variants not detected by genotypic testing, poor adherence, or other variables were not characterized in our study.

A previous cohort study in British Columbia (1996–1999) detected ARV resistance mutations in approximately 25% (n = 298/1191) of their cohort, where the majority of patients were receiving nevirapine (n = 288, 96.6%) followed by efavirenz (n = 8, 2.7%).²⁸ Factors associated with drug resistance mutations included high baseline viral load and high (80–89%) to perfect (≥95%) refill prescriptions.²⁸ In our study, the majority of the patients were taking an efavirenz-based regimen (45.8%), followed by a ritonavir-boosted lopinavir-based regimen (15.0%), and thirdly an atazanavir-based regimen (12.1%). Both nevirapine (Y181C) and efavirenz (K103 N) are associated with major resistance mutations. The lower rate of resistance mutations in our study may reflect the fact that in contrast to the BC cohort, which took place at the beginning of the new treatment era consisted of both nevirapine and efavirenz mostly versus the current era with a shift toward efavirenz. In addition, modern day regimens as in our study may be more resistant to the development of resistance mutations.

A major study limitation is that the data were derived from a convenience sample. One possible source of selection bias is that patients are required to have a viral load within six months and to be engaged in care. The higher frequency of men in the analysis is reflective of the population in care at MLMC, which consists primarily of a gay male patient population. The results may also only be generalizable to viremia and viral suppression in persons living with HIV who are accessing care in specialized HIV clinics in urban areas. Possible unobserved confounding variables shown in other studies to influence viremia and viral suppression are substance use and ART adherence which were not reported for our cohort.^{89, 90} Other possible unobserved confounders are factors involving the selection of a regimen by the patient and physician. Persons initiating PI-based regimens may have different characteristics from those initiating an NNRTI-based regimen prior to the start of treatment. Also, channeling bias may be an issue since patients with presumed poorer adherence and higher viral load are placed on a PI-based regimen which we addressed by stratifying our population by CD4 cell counts less than and greater or equal to 200 cells/mm³. Most patients with a CD4 cell count less than 200 cells/mm³ in this study were on a PI-based regimen (n = 102, 57.6%), followed by a NNRTI-based regimen (n = 71, 40.1%), and lastly only a few were on an II-based regimen (n = 4, 2.3%). In contrast, most patients with a CD4 count greater or equal to 200 cells/mm³ were on a NNRTI-based regimen (n = 293, 61.6%), followed by a PI-based regimen (n = 142, 29.8%), and lastly an II-based regimen (n = 41, 8.6%). Findings regarding patients on an II-based regimen were restricted to those on raltegravir. Another important study limitation is that the resistance profile is limited to a small number of patients experiencing viremia with both baseline and follow-up genotype testing. Lastly, our study did not differentiate between intermittent versus persistent low-level viremia. This is important to further understand the clinical significance of our findings since the evidence implicates persistent low-level viremia among those on treatment to have an increased risk of virologic failure.^{21, 91, 92}

Evidence from our retrospective chart review suggests that among HIV-positive patients accessing specialized HIV care in an urban setting, ART resistance rarely develops after starting first-line therapy. Despite only few cases of ART resistance, it is important to consider the clinical significance of resistance on viremia and salvage therapies. Future research needs to investigate the newer ART regimens and agents and an assessment of their virologic and resistance outcomes in cases where viremia occurs in the clinic setting.