Important New Findings on HIV Control in the Brain, Neurocognitive Dysfunction, ART CNS-Penetration
14th Conference on Retroviruses
February 25-28, 2007, Los Angeles
Newer, convenient, and relatively nontoxic antiretroviral combinations that swiftly shut down HIV replication in blood have swept aside more cumbersome, side-effect-laden regimens popular only a few years ago. But a clutch of compelling reports at the 14th Conference on Retroviruses reminded prescribers that stopping HIV in blood is not the only goal of antiretroviral therapy. Rooting out replicating virus from the central nervous system (CNS) also figures strongly in returning HIV-infected people to health. These studies traced ties between viral suppression in blood and brain, pinpointed newer antiretrovirals that do a better job in the CNS, reckoned whether CNS-penetrating combos prevent rebounds in blood, and sized up the antiviral potential of antidepressants and statins.
Links between HIV suppression in blood and brain
An analysis of 101 people starting a new antiretroviral regimen traced some intriguing links between plasma and cerebrospinal fluid (CSF) loads--as well as CD4 counts and neuropsychological function . Christina Marra (University of Washington, Seattle) and colleagues found higher neuropsychological performance scores in people who began antiretroviral therapy or switched to a new regimen with a higher CD4 count in AIDS Clinical Trials Group (ACTG) protocol 736, an observational study. People who took antiretrovirals that penetrate the CNS well had a significantly better chance of reaching an undetectable viral load in both plasma and CSF.
ACTG researchers analyzed RNA loads, CD4 counts, neuropsychological performance Z score on 4-item or 8-item tests (NPZ4 or NPZ8), and other variables in 101 people embarking on their first antiretroviral course or shifting to a new regimen with a CD4 count under 200 and a plasma load over 2000 copies or with a load above 50,000 copies and any CD4 count. They ranked AZT, d4T, 3TC, abacavir, nevirapine, efavirenz, and indinavir as good CNS penetrators. Regimens including two or more of these drugs rated as good CNS-penetrating combinations.
The study group was 17% female and 48% nonwhite. Median age stood at 39 years (interquartile range [IQR] 34 to 44 years) and median years of education at 13 (IQR 12 to 15). Median starting CD4 count measured 108 cells (IQR 49 to 224), median CSF load 3.17 log (IQR 2.51 to 3.83 log), and median plasma load 4.88 log (IQR 4.55 to 5.55 log).
NPZ4 scores during treatment were better in people with a higher baseline NPZ4 and in those with a higher baseline CD4 count. People with higher baseline NPZ8 scores and higher baseline CD4 counts scored higher NPZ8 marks during treatment.
Marra and coworkers never detected HIV RNA in CSF when they failed to measure it in plasma. After 52 weeks of treatment, HIV RNA rarely turned up in CSF if the plasma load stayed below 1000 copies. Among 39 people whose plasma load did climb above 1000 copies, only one factor correlated with an undetectable CSF load: A CNS-penetrating regimen raised the chances of an undetectable CSF load 5 times in these people (P = 0.05). Baseline CSF RNA, baseline plasma RNA, baseline CD4 count, number of antiretrovirals, and taking a first-line regimen did not correlate with undetectable CSF load in this analysis.
Two variables independently predicted an undetectable load in both CSF and plasma: Taking a CNS-penetrating regimen boosted chances of doubly undetectable loads 2.6 times (P = 0.03), and taking one's first regimen upped the odds 2.9 times (P = 0.02).
Marra and colleagues believe their results "suggest that an antiretroviral regimen with good CNS penetration is important in achieving suppression of HIV RNA in both CSF and plasma." They caution that it will take controlled trials to confirm causality for the correlations they found between CSF and plasma loads, higher baseline NPZ scores, higher baseline CD4 count, and taking a CNS-penetrating regimen.
A small case-control study by Barcelona clinicians suggested that a lower pretreatment CD4 count raises the risk of neuropsychological impairment . Measuring neurocognitive function with a battery of tests in people who began treatment with CD4 counts above or below 200, 250, 300, or 350, Jose Munoz-Moreno and coworkers (Germans Trias i Pujol Hospital) found that more people starting therapy below versus above the lower three cutoffs had evidence of neurocognitive impairment.
But these differences all lacked statistical significance: Whereas 53% of 38 people starting with more than 200 CD4s had lower marks on neurocognitive tests, 73% of 26 people starting with fewer than 200 CD4s had evidence of such problems. Respective percentages with neurocognitive impairment were 53% versus 67% for more versus fewer than 250 CD4s, and 57% versus 64% for more versus fewer than 300 CD4s. Slightly more people starting above versus below 350 CD4s had lower test marks (62.5% versus 57%). Each comparison involved about 60 people.
Individual neurocognitive tests recorded significantly worse attention/working memory among people starting with fewer than 200 CD4s (P < 0.05) and significantly worse executive function for those who began therapy with a sub-200 count (P < 0.05). These and other individual test results did not vary significantly when the Barcelona team compared people above or below CD4 cutoffs of 250, 300, or 350 cells.
Does CNS penetration alter risk of plasma rebound?
Italian investigators explored the potential impact of CNS-penetrating regimens on rebound risk in people who reach a plasma load below 80 copies with their first antiretrovirals . Andrea Antinori (National Institute of Infectious Diseases, Rome) and coworkers across Italy studied 2785 previously untreated people, only 12 of whom (0.4%) had a recorded neurologic disorder. Everyone reached a plasma load below 80 copies with their first regimen.
To see whether antiretroviral CNS penetration affects rebound risk in plasma, Antinori and coworkers used a CSF penetration gauge that sums scores for individual drugs in a regimen [see note 4]. Individual antiretrovirals score 0 for low penetration, 0.5 for intermediate penetration, and 1 for high penetration. Defining rebound as two consecutive plasma loads above 400 copies, the researchers charted a rebound rate of 6% yearly.
A statistical analysis that did not account for other rebound risk factors determined that every 1 unit higher penetration score lowered the rebound risk by one third (crude relative rate 0.67, 95% confidence interval [CI] 0.56 to 0.81, P = 0.0001). That correlation disappeared, however, when statisticians adjusted the analysis for antiretrovirals in regimens being taken at the time of rebound. This adjusted analysis did find that each unit higher penetration score sliced the rebound risk 96% in people with a current CD4 count under 200. But that correlation was weak because the 95% CI crossed 1.0 (adjusted relative rate 0.04, 95% CI 0.01 to 1.61). And the adjusted analysis linked a higher penetration score to a higher rebound risk in people with a current CD4 count above 200.
Antinori's findings show how much remains to be learned about whether CSF-penetrating antiretrovirals--not to mention proposed CSF penetration scores--have any consistent impact on sustained virologic response.
Brain-breaching potential of lopinavir/ritonavir
Published research indicates that lopinavir/ritonavir ranks as a reliable CNS penetrator [5,6], and a Retrovirus Conference study confirmed that finding in previously untreated people starting these protease inhibitors (PIs) as monotherapy . Joseph Gathe (Therapeutic Concepts, Houston) measured CSF loads in people who reached a plasma load below 75 copies after 24 weeks of solo standard-dose lopinavir/ritonavir. The report did not mention how many people failed to control plasma viremia with this regimen. Among 22 people who did get their plasma load under 75 copies, 12 agreed to lumbar puncture.
Gathe and coworkers excluded CSF results in one person with a traumatic lumbar puncture. Of the remaining 11, 6 were men, 5 women, and 5 African American. They had a median pretreatment CD4 count of 264 (range 200 to 491) and a median pretreatment plasma load of 4.28 log (range 3.72 to 4.78 log). Plasma loads stayed below 75 copies in all 11 people when they had their lumbar puncture, but 1 of the 11 had detectable HIV RNA in CSF (251 copies on the first reading and 747 copies 4 months later).
Median CSF lopinavir levels in these 11 study participants stood far below those measured in plasma (24.3 versus 9679 ng/mL) to yield a CSF-to-plasma ratio of 0.24%. But CSF lopinavir concentrations in all patients exceeded the 99th percentile for 50% inhibitory concentration recorded by Monogram Biosciences (8.1 nmol/L or 5.1 ng/mL) for wild-type (nonmutant) virus. CSF LPV concentrations ranged from 7.1 to 85.3 ng/mL in the study group.
A separate study found better brain penetration with lopinavir/ritonavir than with atazanavir (see next section).
Surprising correlates of HIV hidden in CSF
Countless studies demonstrate that HIV may lurk at low levels in plasma when it goes undetected by a 50-copy assay. At the Retrovirus Conference Scott Letendre (University of California, San Diego) showed that HIV can also hide out in CSF at levels below 50 copies/mL . With CHARTER cohort colleagues, Letendre also investigated potential correlates of HIV hidden in CSF, including use of certain popular antiretrovirals.
Letendre measured plasma and CSF loads in 125 people (group 1), 113 of them (90%) taking antiretroviral therapy (ART). Forty of these 125 (32%) had plasma loads under 50 copies (group 2), all of them taking ART. Everyone had a CSF load under 50 copies. Groups 1 and 2 were 55% and 58% non-Caucasian, had median CD4 counts of 421 and 538, and median CD4 nadirs of 138 and 131. Equivalent proportions in groups 1 and 2 (71% and 70%) had AIDS.
In group 1, the overall study group of 125 people, 62 (50%) had HIV RNA detected in CSF with a sub-50-copy assay. Compared with people who had no sign of HIV in CSF, those with low-level CSF virus had higher plasma loads (median 218 versus 133 copies, P = 0.01), higher CD4 nadirs (median 180 versus 101 cells, P = 0.01), higher current CD4 counts (425 versus 408 cells, P = 0.059), and less advanced HIV disease (36% versus 21% without AIDS, P = 0.04).
In the 90% of group 1 members taking antiretrovirals, people with detectable HIV in CSF had a lower median CNS Penetration-Effectiveness (CPE) score --1.5 versus 2.0, a difference just short of statistical significance (P = 0.09). The 48 people taking tenofovir had more than twice the risk of detectable virus in CSF than did the 38 taking AZT or abacavir (60% versus 38%, odds ratio [OR] 2.46, P = 0.04). Detection of CSF load did not differ significantly between people taking nevirapine versus efavirenz or between those taking lopinavir/ritonavir versus atazanavir.
Among 40 people with an undetectable plasma load, 17 (42%) harbored between 2.5 and 50 copies of HIV RNA in CSF. As in the overall group, people with detectable CSF HIV in this subgroup had a higher nadir CD4 count (215 versus 101 cells, P = 0.04), higher current CD4 count (580 versus 481 cells, P = 0.06), and earlier stage disease (50% versus 13% non-AIDS, P = 0.03). All 17 people with detectable virus in CSF had global impairment measured on neuropsychological testing, compared with 59% of those without detectable CSF HIV, a difference that fell shy of statistical significance (P = 0.07). Worse neuropsychological impairment did not correlate with detectable HIV in CSF of the overall group. The study also failed to link HCV coinfection or duration of the current regimen to CSF load in the overall group or the subgroup.
Why would higher CD4 nadirs and higher current CD4s favor detectable CSF HIV in both the overall group and the subgroup? Fielding that question via e-mail, Letendre suggests two possibilities:
"First, we (and others) have seen that people with higher CD4 lymphocyte counts are more likely to have lymphocytes in their CSF. Individuals who have more lymphocytes in CSF tend to have higher viral loads in CSF, possibly because this reflects trafficking of replication-competent cells across the blood-CSF barrier. Therefore people who have higher CD4 counts may be more likely to have ongoing low-level replication in CSF because they have more trafficking of lymphocytes and the HIV they produce into CSF.
"Second, people with detectable HIV in CSF also had higher CD4 nadirs, but overall the CD4 nadirs were low. A low CD4 nadir with a higher current CD4 count may reflect relatively greater immune reconstitution. When you consider that the individuals who had detectable HIV in CSF were taking antiretroviral regimens with worse penetration characteristics, this suggests a hypothesis: Potent but poorly penetrating antiretroviral regimens lead to the undesirable combination of incomplete suppression of HIV in the central nervous system along with greater immune reconstitution, possibly leading to low-level IRIS [immune reconstitution inflammatory syndrome] in the CNS."
Letendre cautions that these two theories are not mutually exclusive: "Some combination of these factors (poor penetration, immune reconstitution, lymphocyte trafficking) may be responsible for our observations. The findings should be confirmed by independent analyses and we should analyze more specimens to better understand the correlates and consequences."
In the 40-person subgroup with undetectable plasma loads, those with a detectable CSF load had a lower median CPE score (1.5 versus 2.0, P = 0.01). While 67% of 13 people taking tenofovir had a detectable CSF load, only 21% of 14 taking AZT or abacavir did (P = 0.02). Taking atazanavir (12 people) rather than lopinavir/ritonavir (8 people) also favored detectable HIV in CSF (75% versus 25%, P = 0.04). Again, nevirapine and efavirenz did not differ in CSF-penetrating power.
Letendre and colleagues measured CSF load twice in 12 people. Three of them (25%) had no detectable CSF HIV on the first reading but low-level viral quotients on the second. These three had lower CPE scores than the other 9 (median 0.75 versus 2.0, P = 0.02).
Together these CNS penetration scores remind prescribers not to lose sight of this antiretroviral selection criterion as simpler regimens compete for attention.
Four-class regimens slow lethal HIV brain disease
A four-class regimen engineered for ART-naive or -experienced people greatly prolonged survival--at least in the short term--among people with progressive multifocal leukoencephalopathy (PML), according to preliminary results from the ANRS 125 trial presented by Jacques Gasnault (Bicetre Hospital, Paris) and coworkers . All regimens include the fusion inhibitor enfuvirtide for 6 months; after that point everyone will try a different new antiretroviral.
So far the trial involves 28 people with clinical and/or laboratory evidence of PML, a diagnosis confirmed by an independent panel. Twelve people with no antiretroviral experience began tenofovir/emtricitabine (as Truvada), efavirenz, lopinavir/ritonavir, and enfuvirtide. Sixteen ART-experienced people switched to enfuvirtide plus three or more other antiretrovirals picked on the basis of resistance testing and treatment history. (According to a US-proposed antiretroviral brain-penetration score, lopinavir/ritonavir has the highest score possible, 1.0, while efavirenz and emtricitabine score 0.5, and tenofovir and enfuvirtide 0 .) Median baseline CD4 count in the PML study stood at 54 cells (range 2 to 345 cells) and median plasma and CSF loads at 4.1 (1.6 to 5.2) and 2.3 (1.6 to 5.1) log copies/mL. Twenty-two people (79%) had JC virus DNA detectable in CSF.
In January 2007 the cumulative 6-month probability of survival measured 77% (95% CI 63% to 95%), more than 50% better than the 50% 6-month survival generally expected for PML patients taking ART. Sixteen of 19 people treated for 6 months (84%) had an HIV load below 40 copies in plasma, and 10 of 13 (77%) had undetectable JC virus DNA in CSF. Both CD4 and CD8 counts continued rising through 12 months of therapy. The percentage of people with a positive anti-JC virus CD4-cell proliferative assay climbed from 4% at baseline to 23% at week 6, 32% at month 3 (P = 0.03 versus baseline), and 47% at month 6 (P = 0.008 versus baseline).
The ANRS team did not differentiate between results in antiretroviral-experienced and previously untreated people. Nor did they note side effects or dropouts with this heavy-duty regimen. ACTG trials of lopinavir/ritonavir plus efavirenz--with no other drugs--found tolerability a problem [10,11].
Distinct neurodisease pattern if CSF load tops plasma load
Plasma HIV loads usually exceed CSF loads. But people with more HIV swarming through spinal fluid than blood have much worse neuropsychological profiles, even in the absence of overt clinical problems. That conclusion emerged from a 153-person analysis by veteran HIV neurologist Gabriele Arendt (University Hospital of Duesseldorf), who split study participants into 33 with bigger CSF loads than plasma loads (a high proportion, she noted) and 120 with higher plasma loads . Arendt further divided each of these two groups into six others depending on HIV disease stage, ART status, and drug abuse:
Group 1: Early CDC stage, no ART: 29 higher plasma load, 7 higher CSF load
Group 2: Early CDC stage with ART: 27 higher plasma load, 7 higher CSF load
Group 3: Late CDC stage, no ART: 6 higher plasma load, 4 higher CSF load
Group 4: Late CDC stage with ART: 48 higher plasma load, 12 higher CSF load
Group 5: Drug abuse, no ART: 4 higher plasma load
Group 6: Drug abuse with ART: 6 higher plasma load
No study participant had overt neurologic disease, and everyone took an array of neuropsychological performance tests, including pure motor tests, motor/cognitive tests, a motor/cognitive/visual test (grooved pegboard), and pure cognitive tests.
Arendt found that people with higher CSF viral loads had strikingly pathologic neuropsych test profiles when compared with the high-plasma-load group. People with higher plasma loads showed signs of mild to moderate motor dysfunction, but not cognitive deficits.
Everyone in group 3 (late CDC stage without ART) with a higher CSF load had signs of severe dysfunction on all tests, a clear indication of the neurologic value of antiretroviral therapy in later-stage HIV infection. In contrast, group 4 (antiretroviral-treated) people with higher CSF loads showed moderate signs of impairment on pure motor and motor/cognitive tests, and only mild impairment on the grooved pegboard test.
Most people with higher plasma loads had only mild or moderate neuropsychological impairment. The exceptions were group 5 members (drug abusers), who had severe impairment on the pure motor test, and group 3 members (late CDC stage and no ART), who showed severe impairment on one motor/cognitive test.
Antiretroviral activity of SSRIs and statins?
Earlier research offered conflicting evidence on whether statins have anti-HIV activity. A Retrovirus Conference report by Scott Letendre (University of California, San Diego) and CHARTER cohort colleagues suggested statins help lower HIV levels in CSF, but mainly in people already taking antiretrovirals . On the other hand, selective serotonin reuptake inhibitors (SSRIs, the popular antidepressants) seemed to have antiretroviral activity in CSF mainly in people not taking antiretrovirals.
Among 659 CHARTER members studied, 467 (71%) used antiretrovirals, 195 (30%) SSRIs, and 63 (10%) statins when they made their study visit. CHARTER clinicians measured HIV RNA in plasma of 383 people (58%) and in CSF of 230 (35%). The group had a median CD4 count of 402, and 64% had AIDS.
Not surprisingly, people taking SSRIs reported more severe depression symptoms than people not taking these antidepressants (P < 0.001). SSRI users also were more likely to be taking antiretrovirals (77% versus 70%, P = 0.06) and had higher CD4 counts (median 433 versus 389 cells, P = 0.07) and lower viral loads (median 1.9 versus 2.4 log, P = 0.07).
An unadjusted analysis determined that people taking SSRIs were about 30% less likely to have detectable HIV in CSF (29% versus 37%, odds ratio 0.69, P = 0.05). Three of seven SSRIs used appeared to have the biggest impact on HIV in CSF: citalopram (Celexa), trazodone (Desyrel), and sertraline (Zoloft). Combining these three antidepressants into a single "antiretroviral" category strengthened their apparent anti-HIV activity in CSF.
People who took one of these three SSRIs also had better global deficit scores on neuropsychological testing (median 0.37 versus 0.47, P = 0.04), and a lower proportion had evidence of neuropsychological impairment (55% versus 64%, P = 0.05). Logistic regression analysis determined that these three "antiretroviral" SSRIs correlated with lower HIV RNA levels in CSF (odds ratio 0.72, P < 0.05), as did antiretroviral use, lower CD4 counts, undetectable HIV RNA in plasma, and Caucasian race.
One might speculate that SSRI use correlates with lower viral loads in plasma and CSF because less depressed people taking these drugs are more likely to adhere to antiretroviral therapy. But Letendre found that "antiretroviral" SSRI users not taking licensed antiretrovirals had the best chance of reaching an undetectable HIV load in CSF. Compared with people not taking SSRIs, this group had about a 70% lower chance of detectable HIV in CSF (odds ratio 0.31, P = 0.01).
Statins, on the other hand, exerted their greatest anti-HIV activity in people already taking antiretrovirals. Cohort members combining statins with antiretrovirals had about a 90% lower chance of detectable HIV in CSF than non-statin users (odds ratio 0.09, P < 0.001). Among all statin users, regardless of antiretroviral use, 16% had detectable HIV in CSF compared with 37% of people not taking statins (odds ratio 0.32, P < 0.001). The apparently stronger anti-HIV activity of statins with antiretrovirals led Letendre and colleagues to speculate that statins' apparent anti-HIV activity "may be mediated by modifying antiretroviral metabolism or distribution."
Next the CHARTER researchers compared HIV CSF levels in four groups--people taking neither SSRIs nor statins, people taking only SSRIs, people taking only statins, and people taking both SSRIs and statins. SSRIs and statins appeared to have additive antiretroviral activity in this analysis. Rates of detectable HIV in CSF measured 39% in the no-SSRI/no-statin group, 27% in the SSRI-only group, 17% in the statin-only group, and 12% in the SSRI-plus-statin group (P = 0.001).
Unlike SSRIs, statins had no neurocognitive benefit in this cohort. Such benefits, Letendre and coworkers surmised, may be negated by concomitant PI therapy, statins themselves, or the lipodystrophy syndrome suffered by many people taking statins. The researchers also caution that some evidence ties statins to brain injury, such as memory loss.
HIV risk in people with acute meningitis
HIV swirling through CSF may cause another problem, one that should alert clinicians to a possible diagnosis of acute HIV infection, according to a report by Kimberly Hanson and colleagues at Duke University in Durham, North Carolina . She found that undiagnosed HIV infection probably explained aseptic meningitis (meningeal inflammation with negative cultures) in 3 of 57 people (5%) with clinical and laboratory signs of meningitis. High CSF HIV load and clinical symptoms suggested that 2 of these 3 people had acute HIV infection when they sought medical care in an emergency room. Both got discharged without an HIV diagnosis.
The study involved 288 stored CSF specimens, 72 of them (25%) from people with abnormal CSF findings (more than 5 nucleated cells/mm3 and protein levels above 50 mg/dL). Fifty-seven of these 72 people (79%) had clinical signs of meningitis or encephalitis. HIV RNA could be detected in CSF of 3 of these 57 samples, all from people who had hospital evaluations for clinical and lab markers of CNS infection. Two of these three people were discharged from the emergency department, and one with bacterial pneumonia died in the hospital.
Hanson and colleague urged clinicians to look for HIV with HIV RNA testing and an HIV ELISA in anyone at risk of infection with the virus who has otherwise unexplainable aseptic meningitis. Because so many of these people turn up in the emergency room, the Duke team called for HIV testing and counseling programs in urgent-care facilities.
1. Marra C, Sinha S, Clifford D, et al. Suppression of HIV replication in plasma requires suppression of HIV replication in CSF. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 382b.
2. Munoz-Moreno J, Rodriguez C, Prats A, et al. Recommended earlier initiation of ART based on nadir CD4 cell count as a risk factor for HIV-related neurocognitive impairment. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 383.
3. Antinori A,Cozzi-Lepri A, Cinque P, et al. Could cerebrospinal fluid ART penetration scores predict viral rebound in ART-naive individuals who achieve plasma viral load suppression on their first therapy? 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 382a.
4. Letendre S, Capparelli E, Best B, et al. Better antiretroviral penetration into the central nervous system is associated with lower CSF viral load. 13th Conference on Retroviruses and Opportunistic Infections. February 5-8, 2006. Denver. Abstract 74. Higher scores indicate better brain penetration. For nucleosides/nucleotides, AZT and abacavir score 1, d4T, 3TC, and FTC 0.5, and ddI, ddC, and tenofovir 0. For nonnucleosides, nevirapine and delavirdine score 1 and efavirenz 0.5. For protease inhibitors, ritonavir-boosted amprenavir, fosamprenavir, atazanavir, indinavir, and lopinavir score 1, unboosted amprenavir, fosamprenavir, atazanavir, and indinavir score 0.5, and nelfinavir, saquinavir, ritonavir-boosted saquinavir, ritonavir, and ritonavir-boosted tipranavir score 0. Darunavir is not yet rated. Enfuvirtide scores 0.
5. Capparelli EV, Holland D, Okamoto C, et al. Lopinavir concentrations in cerebrospinal fluid exceed the 50% inhibitory concentration for HIV. AIDS. 2005;19:949-952.
6. Yilmaz A, Stahle L, Hagberg L, et al. Cerebrospinal fluid and plasma HIV-1 RNA levels and lopinavir concentrations following lopinavir/ritonavir regimen. Scand J Infect Dis. 2004;36:823-828.
7. Yeh R, Letendre S, Novak I, et al. Single-agent therapy with lopinavir/ritonavir controls HIV-1 viral replication in the central nervous system. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 381.
8. Letendre S, McClernon D, Benjamin R, et al. Presence of HIV RNA in cerebrospinal fluid that is undetectable with the ultrasensitive assay. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 369.
9. Gasnault J, H Hendel Chavez H, Dorofeev E, et al. Acceleration of Immune Recovery on Intensified ART Improves Survival in Patients with AIDS-related Progressive Multifocal Leukoencephalopathy: Preliminary Reports of the ANRS 125 Trial. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 379.
10. Fischl MA, Collier AC, Mukherjee AL, et al. Randomized open-label trial of two simplified, class-sparing regimens following a first suppressive three or four-drug regimen. AIDS. 2007;21:325-333.
11. Riddler SA, Haubrich R, DiRienzo G, et al. A prospective, randomized, phase III trial of NRTI-, PI-, and NNRTI-sparing regimens for initial treatment of HIV infection--ACTG 5142. XVI International AIDS Conference. August 13-18, 2006. Toronto. Abstract THLB0204.
12. Arendt G, Nolting T, Husstedt IW, et al. Distinct neuropsychological deficit pattern in HIV-1+ patients with higher CSF than plasma viral load. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 385.
13. Letendre S, Marquie-Beck J, Clifford D, et al. Clinical evidence of antiviral activity of serotonin reuptake inhibitors and HMG-CoA reductase inhibitors in the central nervous system. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 384.
14. Hanson K, Reckleff J, Hicks L, et al. Unsuspected HIV infection in patients presenting with acute meningitis. 14th Conference on Retroviruses and Opportunistic Infections. February 25-28, 2007. Los Angeles. Abstract 962.