Carl Armon

Incidence of and risk factors for lipoatrophy (abnormal fat loss) in ambulatory HIV-1-infected patients.

To identify clinical factors associated with the incidence of HIV-1–associated lipoatrophy, HIV-1–infected patients in the HIV Outpatient Study (HOPS) were prospectively evaluated for clinical signs of lipoatrophy at two visits about 21 months apart. Development of lipoatrophy was analyzed in stratified and multivariate analyses for its relationship to immunologic, virologic, clinical, and drug treatment information for each patient. Of 337 patients with no lipoatrophy at Survey 1, 44 (13.1%) developed moderate or severe lipoatrophy between the two surveys. In multivariate analyses, significant risk factors for incident lipoatrophy were white race (OR = 5.2; 95% CI: 1.9–17.1;p = .003), CD4 T-lymphocyte count at Survey 2 less than 100 cells/mm3 (OR = 4.2; 95% CI: 1.3–13.1;p = .013), and body mass index (BMI) less than 24 kg/m2 (OR = 2.4; 95% CI: 1.1–5.4;p = .024). Analyses that controlled for the severity of HIV illness demonstrated no significant association with use of or time on any antiretroviral agent or class of agents and the development of lipoatrophy. Some host factors and factors associated with previous or current severity of HIV infection, especially CD4 T-lymphocyte cell count, appeared to have the strongest association with incidence of lipoatrophy.

Low CD4+ T Cell Count Is a Risk Factor for Cardiovascular Disease Events in the HIV Outpatient Study

BACKGROUND Traditional cardiovascular disease (CVD) risk factors, human immunodeficiency virus (HIV) infection, and antiretroviral (ARV) agents have been associated with CVD events in HIV-infected patients. We investigated the association of low CD4(+) T lymphocyte cell count with incident CVD in a cohort of outpatients treated in 10 HIV specialty clinics in the United States. METHODS We studied patients who were under observation from 1 January 2002 (baseline), categorized them according to National Cholesterol Education Program guidelines into 10-year cardiovascular risk score (10-y CVR) groups , and observed them until CVD event, death, last HIV Outpatient Study contact, or 30 September 2009. We calculated rates of incident CVD events and identified associated baseline risk factors using Cox proportional hazard models. We also performed a nested case-control study to examine the association of latest CD4(+) cell count with CVD events. RESULTS Among 2005 patients, 148 experienced incident CVD events. CVD incidence increased steadily from 0.4 to 3.0 events per 100 person-years from lowest to highest 10-y CVR group (P < .001). In multivariable Cox analyses adjusted for 10-y CVR, CD4(+) cell count <350 cells/mm(3) was associated with incident CVD events (hazard ratio, 1.58 [95% confidence interval, 1.09-2.30], compared with >500 cells/mm(3)), suggesting an attributable risk of approximately 20%. In the multivariable case-control analyses, traditional CVD risk factors and latest CD4(+) cell count <500 cells/mm(3), but not cumulative use of ARV class or individual drugs, were associated with higher odds of experiencing CVD events. CONCLUSION CD4(+) count <500 cells/mm(3) is an independent risk factor for incident CVD, comparable in attributable risk to several traditional CVD risk factors in the HIV Outpatient Study cohort.

Modification of the Incidence of Drug-Associated Symmetrical Peripheral Neuropathy by Host and Disease Factors in the HIV Outpatient Study Cohort

BACKGROUND We sought to identify factors associated with the clinical diagnosis of symmetrical peripheral neuropathy (SPN) during the era of highly active antiretroviral therapy (HAART) in a retrospective, longitudinal cohort analysis. METHODS Patients infected with human immunodeficiency virus type 1 were evaluated for clinical signs of SPN and its association with immunologic, virologic, clinical, and drug treatment factors by means of univariate and multivariate logistic regression analyses. RESULTS Of 2515 patients, 329 (13.1%) received a diagnosis of SPN. In the logistic regression analysis, statistically significant non-drug-based risk factors for SPN were age >40 years (adjusted odds ratio [aOR], 1.17), diabetes mellitus (aOR, 1.79), white race (aOR, 1.33), nadir CD4(+) T lymphocyte count <50 cells/mm(3) (aOR, 1.64), CD4(+) T lymphocyte count 50-199 cells/mm(3) (aOR, 1.40), and viral load >10,000 copies/mL at first measurement (aOR, 1.44). Although initial use of didanosine, stavudine (40 mg b.i.d.), nevirapine, or 4 protease inhibitors was associated with SPN (ORs for all 4 treatments, >1.41), the strength of association decreased with continued use of all medications studied. CONCLUSION Since HAART was introduced, the incidence of SPN has decreased. Host factors and signs of increased disease severity were associated with an increased risk of developing SPN during the initial period of exposure to drug therapy. Immunity improved and the risk of SPN decreased with continued use of HAART. Delaying the initiation of therapy may select those individuals who will be more likely to develop SPN, and earlier initiation of HAART may decrease the risk of developing this common problem, as well as increase the therapeutic effects and decrease the toxic effects of the drugs.

Initiation of antiretroviral therapy at CD4 cell counts >/=350 cells/mm3 does not increase incidence or risk of peripheral neuropathy, anemia, or renal insufficiency.

Background:US guidelines recommend deferring initiation of highly active antiretroviral therapy (HAART) for most patients with CD4 counts >350 cells/mm3 in part because of concerns about antiretroviral toxicity. Methods:Incidence rates of peripheral neuropathy, anemia, and renal insufficiency in a cohort of 2165 patients followed more than 3 years (mean) were analyzed in multivariate Cox proportional hazards models by CD4 cell counts at initiation of HAART. A nested cohort of 895 patients restricted to study participants who did or did not start HAART within a CD4 cell count stratum were also compared. Results:Incidence and risks of all 3 comorbidities decreased with initiation of HAART at CD4 counts >200 cells/mm3 versus <200 cells/mm3. Incidence and risks of renal insufficiency were similar with HAART initiation at CD4 counts ≥350 cells/mm3 versus 200 to 349 cells/mm3, but risk of peripheral neuropathy and anemia were further decreased in persons starting HAART at a CD4 count ≥350 cells/mm3. The incidence of these conditions was highest during the first 6 months of treatment at any CD4 cell count and declined up to 19-fold with further therapy. Discussion:Initiating HAART at CD4 cell counts ≥200 cells/mm3 reduced the incidence and risk of the 3 comorbid conditions and for anemia and peripheral neuropathy as well by starting at CD4 counts ≥350 cells/mm3. The incidence of each condition decreased rapidly and remained low with increasing time on HAART.

Initiation of HAART at higher CD4 cell counts is associated with a lower frequency of antiretroviral drug resistance mutations at virologic failure

Background:There are limited data on the risk of developing HIV drug resistance based on the CD4 cell count at which highly active antiretroviral therapy (HAART) is initiated. Methods:We examined data from participants in the HIV Outpatient Study who initiated antiretroviral therapy with HAART in 1999 or later (when genotypic resistance testing became more commonly used in clinical practice and in the HIV Outpatient Study), achieved virologic suppression, and subsequently experienced virologic failure and received a genotypic assay for antiretroviral resistance mutations. We assessed the frequency of resistance mutations at virologic failure and the differences in the frequencies of mutations by the CD4 stratum at which HAART was initiated using the Cochran-Armitage exact test. Results:Of 683 patients who achieved virologic suppression on a first HAART regimen, 243 had virologic failure and 78 of these had a genotype resistance test done. Among these patients, the frequency of any HIV resistance mutations was 50% among patients who started HAART at 0-199 CD4 cells per cubic millimeter or 200-349 CD4 cells per cubic millimeter compared with 22% among patients who started HAART at ≥350 CD4 cells per cubic millimeter (P = 0.062). The frequency of nucleoside reverse transcriptase inhibitor-associated mutations was 48%, 31%, and 11% among persons who initiated nucleoside reverse transcriptase inhibitor-containing HAART within these respective CD4 cell count strata (P = 0.005). We observed similar trends for nonnucleoside reverse transcriptase inhibitor-associated (P = 0.040) and protease inhibitor-associated (P = 0.063) mutations among persons initiating HAART containing these agents. Conclusions:Patients failing HAART that was initiated at <350 CD4 cells per cubic millimeter had higher frequencies of resistance mutations to the classes of antiretrovirals to which they had been exposed than failing patients who initiated at ≥350 CD4 cells per cubic millimeter. Initiating HAART at higher CD4 cell counts may decrease the risk of developing treatment-limiting antiretroviral resistance.

The association of HIV susceptibility testing with survival among HIV-infected patients receiving antiretroviral therapy: A cohort study

Context Although HIV-1 susceptibility testing has become standard practice in some settings to guide the selection of antiretroviral therapy regimens, we do not know whether testing is associated with longer survival. Contribution These authors evaluated data on 2699 patients in 10 U.S. HIV clinics and found that, over a median follow-up of about 3 years, patients who had HIV-1 susceptibility testing were less likely to die than were patients who did not have testing. Implication HIV susceptibility testing is associated with longer survival. However, this observational study cannot tell us whether the association was due to better antiretroviral treatment selection in tested patients or due to confounding. The Editors Use of HIV genotypic or phenotypic susceptibility testing (GPT) has become the recommended standard of care to guide antiretroviral treatment selection for HIV-infected persons, both those who are antiretroviral-naive and those who are already receiving highly active antiretroviral therapy (HAART) but experience suboptimal viral suppression with their current treatment regimens (15). These recommendations are based on findings from studies that demonstrate short-term improvements in viral suppression associated with use of GPT among antiretroviral-experienced patients, as well as anticipated clinical benefit for antiretroviral-naive patientsparticularly in populations and geographic areas where the prevalence of transmitted drug-resistant HIV exceeds 5% (615). It remains unclear whether GPT use is associated with a reduction in AIDS-related morbidity or improved survival. In our analyses, we sought to evaluate the effect of GPT on survival among a diverse sample of HIV-infected patients with varied antiretroviral treatment histories and therapeutic responses who have participated in the HOPS (HIV Outpatient Study). Methods The HOPS is an ongoing, prospective, observational cohort study of HIV-infected adults who have been receiving care at 10 participating HIV clinics (1 public, 4 university, and 4 private clinics) in 8 U.S. cities (Chicago, Illinois; Denver, Colorado; Long Island, New York; Oakland/San Leandro, California; Philadelphia, Pennsylvania; Tampa, Florida; and Washington, DC) since 1993 (16). Patient data, including sociodemographic characteristics, diagnoses, treatments, and laboratory values, are abstracted from medical charts and entered into an electronic database (Clinical Practice Analyst, Cerner Corporation, Vienna, Virginia) by trained staff. These data are reviewed for quality and analyzed centrally. To date, HOPS has collected information on more than 8200 patients from more than 280000 clinical encounters. Since its inception, the HOPS protocol has been reviewed annually and approved by the Centers for Disease Control and Prevention (CDC), the Cerner Corporation, and each local sites institutional review board. The study protocol conforms to the guidelines of the U.S. Department of Health and Human Services for the protection of human subjects in research. Study Population We analyzed data from antiretroviral-naive and HAART-experienced participants who had at least 1 HOPS clinic visit between 1 January 1999 and 31 December 2005 and at least 1 plasma HIV RNA determination greater than 1000 copies/mL. For this analysis, baseline observation time was the date at which an HIV RNA level greater than 1000 copies/mL was first noted after 1 January 1999. Observation continued until death (if documented within 180 days of last patient contact), censoring at last patient contact plus 180 days (patients not known to be deceased were assumed to be alive), or 31 December 2005, whichever occurred first. We excluded patients who had GPT before baseline or for whom a CD4+ cell count was not measured within 6 months before or 90 days after baseline. For subgroup analyses, we first defined 2 mutually exclusive groups of patients: those who were antiretroviral-naive at baseline and whose first regimen during follow-up was HAART (therapy-naive patients who started HAART) and those who were already treated with HAART at baseline (HAART experienced) (Figure 1). We also excluded patients whose antiretroviral therapy consisted only of monotherapy or dual therapy with nucleoside reverse transcriptase inhibitors at baseline. Further, we defined a subset of HAART-experienced patients who had received at least 180 cumulative days of therapy with antiretroviral drugs in all 3 major classes (nucleoside analogue reverse transcriptase inhibitors, nonnucleoside analogue reverse transcriptase inhibitors, and protease inhibitorsnot necessarily simultaneously) yet remained viremic with a plasma HIV RNA level above 500 copies/mL (triple class experienced). Figure 1. Study flow diagram. GPT = genotypic and phenotypic susceptibility testing; HAART = highly active antiretroviral therapy; HOPS = HIV Outpatient Study. * Defined as date of first determination of HIV RNA >1000 copies/mL after 1 January 1999. Statistical Analysis We calculated crude mortality rates per 100 person-years of observation for patients who did and did not have GPT, overall and stratified by patient characteristics and potential risk factors for death. We calculated mortality rate ratios (RRs) on the basis of the stratified mortality rates for each variable analyzed and statistically assessed for differences from 1.0. We used univariable and multivariable Cox proportional hazards regression models to identify risk factors for death; we have reported our results in terms of hazard ratios (HRs) and 95% CIs. Because of the many clinical and sociodemographic variables we considered in the univariable analyses (some of which were highly correlated, such as injection drug use and hepatitis C virus infection), we evaluated each of these variables as a potential confounder in the association between GPT and death in exploratory analyses. We defined confounding as altering the HR for the association between GPT and death by more than 10% in exploratory multivariable Cox models. In the final multivariable models, we also included those factors known to be of clinical importance or that seemed to confound the relationship between GPT and death. In the Cox models, GPT was treated as a time-dependent variable; during follow-up, patients contributed time to the unexposed category until they had GPT, after which they contributed time to the exposed category. We used Cox proportional hazards models with time-varying GPT as a covariate to account for possible survival biases that could occur if some patients who did not have GPT died before having the opportunity for GPT. We evaluated the proportional hazards assumption by using loglog survival curves for select baseline variables. In addition, we used marginal structural models (1719) to control for potential time-varying confounders, such as CD4+ cell count and plasma HIV RNA, that may have been affected by the primary exposure of interest (receipt of GPT) and are themselves determinants of survival. Accounting for study site did not appreciably alter the findings in exploratory marginal structural models; therefore, the final models we present do not account for clustering by site. Appendix 2 provides further details of the methodology. Supplement. Appendix 2 We used StatCalc Epi Info, version 3.2.2 (CDC, Atlanta, Georgia), to compare crude mortality and GPT use rates. We used SAS, version 9.1 (SAS Institute, Cary, North Carolina), for all other statistical analyses. We considered statistical associations with P values less than 0.05 to be significant. Role of the Funding Source The CDC supported the HOPS data collection through a contract with the Cerner Corporation. The co-authors from the CDC participated in the design of the study and the analysis and interpretation of the data and approved the final version of the manuscript. Results Of the 2699 patients who met inclusion criteria, 915 (34%) had GPT and 1784 (66%) did not (Table 1). Overall, 2107 (78%) were HAART experienced. Median follow-up was 3.3 years for the entire cohort, 4.8 years for patients who had GPT, and 2.7 years for those who did not have GPT (Table 1). Appendix 3 and Appendix 4 provide further details on the frequency of clinic visits, laboratory measurements (including GPT), and plasma HIV RNA levels. We considered only the first GPT measurement in all analyses. Eight hundred seventy-two (95%) patients who had GPT had an antiretroviral regimen change compared with 1341 (75%) patients who did not have GPT. Median time from first plasma viral load determination greater than 1000 copies/mL (baseline) to antiretroviral regimen change was 157 days for patients who had GPT compared with 98 days for patients who did not have GPT (P< 0.001). The median time between GPT and antiretroviral regimen change for patients who had GPT was 63 days. Supplement. Appendix 3 Supplement. Appendix 4 Table 1. Study Population Characteristics, by GPT Status Patients who had GPT were more likely (P< 0.05) to be white; to be men who have sex with men; to have private medical insurance, a history of AIDS-defining illness, lower nadir CD4+ cell counts, and longer follow-up since HIV diagnosis (median, 6.9 vs. 5.9 years); to have received antiretroviral therapy before the start of observation and mono- or dual-drug antiretroviral therapy for some portion of their treatment; and to have had more plasma HIV RNA determinations (median, 4.0 vs. 3.3 determinations per person-year) (Table 1). Patients who had GPT were less likely at baseline to have a history of injection drug use or to have been co-infected with hepatitis C virus and were less likely in follow-up to remain antiretroviral-naive or to die within the first 6 or 12 months after baseline. The mortality rate for the overall cohort was 2.5 deaths per 100 person-years. Patients who had GPT had lower mortality rates than those who did not (2.0 vs. 2.7 deaths per 100 person-years; RR, 0.74; P= 0.054) (Table

Provider compliance with guidelines for management of cardiovascular risk in HIV-infected patients.

Introduction Compliance with National Cholesterol Education Program Adult Treatment Panel III (NCEP) guidelines has been shown to significantly reduce incident cardiovascular events. We investigated physicians’ compliance with NCEP guidelines to reduce cardiovascular disease (CVD) risk in a population infected with HIV. Methods We analyzed HIV Outpatient Study (HOPS) data, following eligible patients from January 1, 2002, or first HOPS visit thereafter to calculate 10-year cardiovascular risk (10yCVR), until September 30, 2009, death, or last office visit. We categorized participants into four 10yCVR strata, according to guidelines determined by NCEP, the Infectious Disease Society of America, and the Adult AIDS Clinical Trials Group. We calculated percentages of patients treated for dyslipidemia and hypertension, calculated percentages of patients who achieved recommended goals, and categorized them by 10yCVR stratum. Results Of 2,005 patients analyzed, 33.7% had fewer than 2 CVD risk factors. For patients who had 2 or more risk factors, 10yCVR was less than 10% for 28.2%, 10% to 20% for 18.2%, and higher than 20% for 20.0% of patients. Of patients eligible for treatment, 81% to 87% were treated for elevated low-density lipoprotein cholesterol/non–high-density lipoprotein cholesterol (LDL-C/non–HDL-C), 2% to 11% were treated for low HDL-C, 56% to 91% were treated for high triglycerides, and 46% to 69% were treated for hypertension. Patients in higher 10yCVR categories were less likely to meet treatment goals than patients in lower 10yCVR categories. Conclusion At least one-fifth of contemporary HOPS patients have a 10yCVR higher than 20%, yet a large percentage of at-risk patients who were eligible for pharmacologic treatment did not receive recommended interventions and did not reach recommended treatment goals. Opportunities exist for CVD prevention in the HIV-infected population.

CD4 Cell Counts at HIV Diagnosis among HIV Outpatient Study Participants, 2000–2009

Background. It is unclear if CD4 cell counts at HIV diagnosis have improved over a 10-year period of expanded HIV testing in the USA. Methods. We studied HOPS participants diagnosed with HIV infection ≤6 months prior to entry into care during 2000–2009. We assessed the correlates of CD4 count <200 cells/mm3 at HIV diagnosis (late HIV diagnosis) by logistic regression. Results. Of 1,203 eligible patients, 936 (78%) had a CD4 count within 3 months after HIV diagnosis. Median CD4 count at HIV diagnosis was 299 cells/mm3 and did not significantly improve over time (P = 0.13). Comparing periods 2000-2001 versus 2008-2009, respectively, 39% and 35% of patients had a late HIV diagnosis (P = 0.34). Independent correlates of late HIV diagnosis were having an HIV risk other than being MSM, age ≥35 years at diagnosis, and being of nonwhite race/ethnicity. Conclusions. There is need for routine universal HIV testing to reduce the frequency of late HIV diagnosis and increase opportunity for patient- and potentially population-level benefits associated with early antiretroviral treatment.

Ongoing Pain Despite Aggressive Opioid Pain Management Among Persons With HIV

BackgroundChronic pain is a common problem among persons living with HIV and opioids are frequently used in its treatment. However, data on the variables associated with opioids use and the efficacy of this practice are lacking. MethodsWe performed a cross-sectional cohort study of self-reported pain during the year 2005 in our clinic. Patients were grouped into 3 cohorts: those receiving daily opioid therapy for chronic pain (cohort 1, n=115), those with a chronic pain diagnosis but not on daily opioid therapy (cohort 2, n=209), and those without a chronic pain diagnosis (cohort 3, n=796). ResultsIn multivariate analysis comparing cohorts 1 and 2, patients in cohort 1 were significantly more likely to be on a benzodiazepine or gamma-aminobutyric receptor agonist [odds ratio (OR)=15.2], have injection drug use as a HIV risk factor (OR=4.27), lack private insurance (OR=3.51), have been abused (OR=3.08), have a history of AIDS (OR=2.21), and be seen more frequently (OR=1.18). Patients in cohort 1 reported significantly more pain [mean pain scores (0 to 10): 4.3 cohort 1; 1.9 cohort 2; 0.7 cohort 3], and were more likely to have pain that was of moderate or greater severity (58.6% cohort 1; 15.5% cohort 2; 4.9% cohort 3). ConclusionsPsychosocial variables and a history of AIDS were associated with opioid use in our clinic. Persons on opioids continued to experience significantly more pain than other patients in our clinic.

Low bone mineral density and risk of incident fracture in HIV-infected adults

BACKGROUND Prevalence rates of low bone mineral density (BMD) and bone fractures are higher among HIV-infected adults compared with the general United States (US) population, but the relationship between BMD and incident fractures in HIV-infected persons has not been well described. METHODS Dual energy X-ray absorptiometry (DXA) results of the femoral neck of the hip and clinical data were obtained prospectively during 2004-2012 from participants in two HIV cohort studies. Low BMD was defined by a T-score in the interval >-2.5 to <-1.0 (osteopenia) or ≤-2.5 (osteoporosis). We analysed the association of low BMD with risk of subsequent incident fractures, adjusted for sociodemographics, other risk factors and covariables, using multivariable proportional hazards regression. RESULTS Among 1,006 participants analysed (median age 43 years [IQR 36-49], 83% male, 67% non-Hispanic white, median CD4(+) T-cell count 461 cells/mm(3) [IQR 311-658]), 36% (n=358) had osteopenia and 4% (n=37) osteoporosis; 67 had a prior fracture documented. During 4,068 person-years of observation after DXA scanning, 85 incident fractures occurred, predominantly rib/sternum (n=18), hand (n=14), foot (n=13) and wrist (n=11). In multivariable analyses, osteoporosis (adjusted hazard ratio [aHR] 4.02, 95% CI 2.02, 8.01) and current/prior tobacco use (aHR 1.59, 95% CI 1.02, 2.50) were associated with incident fracture. CONCLUSIONS In this large sample of HIV-infected adults in the US, low baseline BMD was significantly associated with elevated risk of incident fracture. There is potential value of DXA screening in this population.