ISSUE 83 25thFebruary 2000

Editor Paul Blanchard

Medical Consultant


7th Conference on Retroviruses and Opportunistic Infections
January 30th-February 2, 2000.
San Francisco, CA.

Opportunistic Infections

Pharmacokinetics, Drug-Drug Interactions and Therapeutic Drug Monitoring

7th Conference on Retroviruses and Opportunistic Infections
January 30th-February 2, 2000
San Francisco, CA.

Resistance to Antiretrovirals and Resistance Testing

Graeme Moyle, MD, Associate Director of HIV Research, Chelsea and Westminster Hospital, London, UK.

Phenotypic testing improves response to 2nd line regimens

Resistance testing is now the standard of care for persons experiencing viral rebound on antiretroviral therapy. Indeed, resistance testing is, arguably, better proven than many commonly used interventions in HIV. Not surprisingly, the recommendation to use resistance testing is now a part of new HIV management guidelines in both European and other countries. While genotypic testing is most widely used, the conference saw the first prospective clinical trial demonstrating the benefits of phenotypic tests[1]. This study, VIRA 3001, was a randomised trial evaluating the short term virologic responses to phenotypic resistance testing using the Virco Antivirogram (AVG) with 'standard of care' (SOC) (i.e. making treatment selection without assay, commonly known as educated guesswork) in 274 patients experiencing loss of suppression on first PI-containing regimen. Study entrants were NNRTI na•ve (except, for some reason 4%) and had prior therapy with 2 or more NRTIs and 1 PI, detectable viral load >2000 copies/ml on stable antiretroviral therapy for >1 month. A new regimen was nominated prior to randomisation. Patients were randomised to AVG (n=144) or SOC (n=130) although only 111 patients in the AVG arm, and 107 in the SOC arm actually took randomised therapy. It is unclear why so many patients dropped out before randomisation. Patients were predominately experienced with 3TC (85%), ZDV (83%), d4T (61%) and NFV (53%) or IDV (36%).

Despite prior exposure, at screening phenotypic resistance to PIs was relatively infrequent; to amprenavir (about 10% had 4-10 fold), to indinavir (about 15% had >10 fold resistance, about 20% had 4-10 fold resistance) to RTV, (about 15% 4-10 fold, and 20% >10 fold), and to SQV (10% 4-10 fold, and 10% >10 fold ), whereas 80% were 3TC resistant.

Doctor's are fickle creatures. Perhaps not surprisingly, after receiving the phenotypic test results, 78% in the AVG arm altered the regimen those chosen prior to receiving the test results: 63% changed their NRTIs, 49% changed their NNRTI, 57% changed their PI. However, despite not getting any new data on the day of randomisation, 42% in the SOC arm also changed their planned regimen: NRTIs 35%, PI 32%, NNRTI 24%.

A higher proportion of patients in the AVG arm received 3 or more 'active' (<4 fold resistant) drugs (77%) than in the SOC arm (62%). Resistance testing did not increase the number of drugs used. A median number of 3.4 drugs were taken in each of the two arms.

By ITT (missing=failure) analysis the proportion below 400 copies/ml at16 weeks was 38% in the AVG arm and 22% had <400 in the SOC arm. This rose to 43% and 23% respectively when baseline RNA was <10,000 copies/ml as considered. In those whose baseline VL was >100,000, 38% who were phenotypically tested but only 17% in the SOC arm achieved <400. Thus, the benefits of phenotypic testing were consistently observed across all viral load strata. Using an observed analysis 58% in the AVG arm (n=72) and 37% in the SOC arm (n=68) (p=0.001) had <400 copies/ml at week 16. The magnitude of viral load decline to week 16 was -1.27log in the AVG arm versus -0.75 log in the SOC arm (p=0.005).

Whilst this represents the third study in which resistance testing has provided benefit, the results are somewhat disappointing with a relatively low proportion of patients achieving <400 copies/ml. Results across studies are summarised below.

Key outcomes in resistance testing studies

Study Design % st
@wk 16 *(log10)
@wk 16*
Viradapt Geno vs. SOC ~40% -1.04
GART Geno+ Expert vs. SOC ~50% -1.19
VIRA3001 Pheno vs. SOC 100% -1.27

* Week 12 for GART #<200 copies/ml

Results of a second prospective randomised trial of phenotypic resistance testing, this time in 'heavily pretreated' persons were also reported in a poster[2]. 115 patients failing a PI-containing regimen were randomised SOC or AVG phenotypic testing. The mean number of sensitive drugs prescribed was significantly (p=0.0001) greater in the AVG arm compared with SOC but was still <3. Decreases in virus load and rises in CD4 count favoured the AVG arm but lost significance by week 16. Ultimately, these data imply, that for resistance testing to provide valuable guidance, the physician must have available new drugs which are active against viruses resistant to the old drugs. Thus, sometimes (see below) waiting before changing may be advantageous.


A model examining the cost-effectiveness of resistance testing suggested the costs would be offset by savings in the cost of drugs[3]. Instead of prescribing the pharmacopia in blunderbuss kitchen-sink therapies, physicians would prescribe only the active drugs, reducing the overall cost of the regimen. Treatment breaks while waiting for the resistance results, may not only benefit response to the new regimen by allowing outgrowth of more sensitive virus but may alone be enough to offset resistance test costs. Hopefully, the penny will soon drop for Luddite purchasers who have been, to date, too short-sighted to appreciate that resistance testing will save money in the long and probably the short run.

Resistance and viral fitness

Despite the presence of resistance, 3TC continues to be retained in many subsequent regimens because of the perception that it may diminish the viral replicative capacity or fitness without adding additional toxicity. Whilst this is likely to be true in theory, whether it is clinically beneficial remains unclear. It is possible that some protease inhibitors could also be used in this way as high exposures of some of these compounds (e.g. saquinavir, ABT-378) appears to have limited toxicity. Data from ViroLogic discussed an assay developed to measure the fitness of both laboratory and clinically derived viruses using recombinant vectors similar to their phenotypic assay PhenoSense. Their results in laboratory created viruses indicated fitness decreased with some mutations, such as the D30N/N88D nelfinavir-associated mutations, and G48V/L90M saquinavir-associated mutations[4]. However, fitness was often markedly restored when additional mutations (such as at L63P, V77I and gag cleavage sites) were added. Ultimately, multiply mutated viruses often had similar fitness to wild type virus. Indeed there was a significant inverse correlation between increasing fold resistance and decreasing viral fitness: the higher the resistance the lower the fitness. During a treatment interruption resistant viruses with good fitness could therefore effectively compete with wild type, delaying potentially beneficial wild type outgrowth.


Knowing what to do with the results of resistance assays remains an obstacle to their widespread use. Genotypic resistance leads to variable amounts of phenotypic resistance, in part dependent on an array of secondary or additional mutations. The clinical importance of individual mutation clusters not being established. Additionally, as viral resistance is relative and never absolute increasing drug exposures, through booster strategies like 'baby' dose ritonavir for PIs and hydroxyurea for at least ddI, may enable drug activity despite apparent resistance. Although a 4-fold increase in IC50 over a reference viral isolate is often defined as resistance, this definition may not apply with ABT-378/ritonavir (new name lopinavir) or saquinavir/ritonavir combinations. With lopinavir, drug exposures exceed the wild-type IC50 by >30-fold. In a sub-study of 61 of the 70 PI-experienced patients who received lopinavir for 2 weeks then added new NA plus nevirapine (study M97-765), neither the presence of PI-associated mutations nor a >4-fold increase in IC50 correlated with success or failure of the drug[5]. The extent of cross-resistance between PIs remains unclear. Cross-resistance to amprenavir was present in only 37% of 62 samples with complete cross-resistance to all of indinavir, saquinavir, ritonavir, and nelfinavir, 63% of samples remaining sensitive to amprenavir. Mutations at codons 84 or 46+ 54+ 90 were significantly associated with amprenavir resistance[6]. In 111 persons failed on IDV, RTV or SQV, cross-resistance to nelfinavir was highly correlated with the presence of the L90M mutation. This is not surprising, as nelfinavir has been reported to select for this mutation[7].

Resistance to tenofovir

The new and promising nucleotide tenofovir (PMPA), as with adefovir, selects for a K65R mutation in vitro. This mutation potentially causes cross resistance to 3TC, ddI and ddC. In samples patients in clinical studies who added tenofovir to stable background therapy, 35% of patients developed new ZDV-type mutations, and 4 developed the L74V mutation associated with ddI and abacavir resistance. Only 3 patients developed the K65R mutation. These patients were also taking ddI and abacavir, so several pressures selecting this mutation may have been present[8].

Regimen failure from single mutation of "brittle" agents

Resistance was investigated in persons experiencing viral rebound on triple therapy with ZDV/3TC/IDV or ZDV/3TC/abacavir in the CNA3005 study[9]. The 3TC 184V mutation was found in 18/24 study participants from the ABC arm and in 12/20 from the IDV arm. In general, no additional mutations were present, underlining that resistance to this resistance-brittle agent is enough to diminish the potency of these regimens. Similarly, resistance in nucleoside-experienced patients who received treatment with efavirenz + a protease inhibitor + 2 new nucleoside RT inhibitors commonly (11/22) involved the K103N NNRTI mutation[10]. Underlining that despite excellent pharmacology, if these agents are not adequately supported by potent co-therapies, they too are resistance brittle.

Risking higher resistance to maintain CD4 count - the discordant response

A mismatch between virus rebound and maintenance of CD4 elevation is well recognised and indeed occurred in early ZDV monotherapy and subsequent dual therapy studies, long before the idea of full suppression. It is unknown how long this 'discordant' response lasts. In a clinic-based observational cohort of 479 patients with detectable viral load on a PI-based regimen (defined as started a PI-regimen before 9/97, remained on therapy >16 weeks, with viral load >500 copies/mL after >16 weeks of treatment) the median time for the CD4 cell count to return to pre-therapy levels was 36.4 months[11]. Several biases exist within this sort of cohort. Over the median 36.7 months follow-up 58 (12%) patients were lost to follow-up and there were 49 (10%) deaths, mostly due to non-HIV-related causes. Additional patients were censored from analysis if they changed therapy to a 'salvage' regimen. The criteria for starting salvage were patient and physician decision. Interestingly, when compared with an untreated, historical control group known as the San Francisco Men's Health Study, the rate of CD4 fall, matched for viral load, was significantly (p<0.05) slower in the PI recipients. Thus in persons with few treatment options and a 'low risk' CD4 count changing therapy may be delayed while new treatment options accumulate. It is not known if the durability of the CD4 elevation differs between PIs or nucleoside backbones or with an NNRTI-based regimen. While the majority of patients experiencing incomplete viral suppression will eventually experience a CD4 decline back to baseline, viral 'failure' for the majority of individuals does not equate to CD4 or clinical 'failure'. In individuals in whom viral suppression is not feasible, maintaining a low risk CD4 seems a reasonable approach.

Additional unidentified reservoirs?

The persistence of a pool of latently infected, resting memory CD4+ T cells as a reservoir of HIV persistence is a well documented obstacle to HIV eradication. Data published in 1999 indicated that this pool was often refilled or added to by 'blips' of viral activity despite apparent viral suppression. More recent data using LTR circles have indicated that around 2/3 of persons on apparently effective therapy have evidence of recent rounds of viral replication. The clinical impact of this persistence of low level viral activity is not known. Data presented at the conference indicate that this pool of cells does not account entirely for the early rebounding plasma HIV in infected individuals in whom HAART has been discontinued, even after prolonged periods of successful suppression of plasma viraemia[12]. The HIV envelope gene, env, of the rebounding virus was found to be genetically distinct from either cell-associated HIV RNA or virus derived from CD4 memory cells suggesting the possibility of existence of other persistent HIV reservoirs.

Attempts to purge the latent CD4 memory population by recruiting them into an active phase using interleukin-2 (IL-2) sadly does not appear to delay the time to viral rebound if therapy is stopped[13].

Drug resistance may arise during transient "blips" in viral load

Studies reported last year did not identify new drug resistance mutations in HIV recovered from latent CD4 cells during suppressive therapy. While modest changes in env were reported during blips no drug resistance was detected as a result. Unfortunately, this good news only lasted a year. In virus isolated from PBMC in persons with plasma HIV RNA <50 copies/ml pol genotyping with phylogenetic analyses and Phenosense resistance assays were performed[14]. Analyses of replication-competent virus recovered from latently infected lymphocytes identified new drug resistance mutations selected during successful therapy in 4 of the 5 drug-experienced, patients following episodes of transient viraemia. However, new resistance mutations were not detected in 7 patients who did not have blips. New resistance mutations occurred in both protease and RT. The authors concluded that 'drug resistant HIV-1 can be selected in vivo in the absence of sustained rebound of viraemia during treatment of established infection with regimens currently considered optimal'. Thus, for perhaps as many as 2/3 of persons on therapy, HAART may be inadequately active. As tests for markers of on-going replication become available, the challenge remains to decide whether early treatment intensification is needed for these individuals. Alternatively, these blips, like the more deliberate viral exposures being pursued in structured treatment interruption studies, might actually stimulate maintenance of beneficial HIV-specific immune responses. The risk of drug resistance during such minor episodes of viraemia while maintaining drug pressure, however, appears real. [see further report on this data below]

1. Cohen C, Hunt S, Sension M, et al. Phenotypic resistance testing significantly improves response to therapy: a randomised trial (VIRA 3001). Programs and abstracts of the 7th CROI, 2000. Abstract 237.

2. Melnick D, Rosenthal J, Cameron M, et al. Impact of phenotypic antiretroviral drug resistance testing on the response to salvage antiretroviral therapy (ART) in heavily experienced patients. Program and abstracts of the 7th CROI, 2000. Abstract 786.

3. Simpson KN. Budget impacts of the integration of genotypic antiretroviral resistance testing into HIV treatment guidelines. Program and abstracts of the 7th CROI, 2000. Abstract 789.

4. Wrin T, Gamarnik A, Paxinos E, et al. Measuring the replicative fitness of recombinant HIV-1 vectors expressing protease and reverse transcriptase derived from patient viruses. Programs and abstracts of the 7th CROI, 2000. Abstract 233.

5. Kempf D, Xu Y, Brun S, et al. Baseline genotype and phenotype do not predict response to ABT378/ritonavir in PI-experienced patients at 24 and 48 weeks. Program and abstracts of the 7th CROI, 2000. Abstract 731.

6. Schmidt B. Cross-resistance to amprenavir in PI-treated patients. Program and abstracts of the 7th CROI, 2000. Abstract 726.

7. Dronda F, Casado JL, Moreno S. Cross-resistance to nelfinavir can be predicted by previous antiretroviral exposure in the absence of the D30N mutation. Program and abstracts of the 7th CROI, 2000. Abstract 729.

8. Miller MD, Margo NA, Robison M, Schooley R, Mills R, McGowan I. HIV-1 RT mutations in patients after 24 weeks of tenofovir disoproxil fumarate (formerly PMPA prodrug) therapy added to stable background ART. Program and abstracts of the 7th CROI, 2000. Abstract 740A.

9. Melby T, Tortell S, Thorburn D, et al. Comparison of viral resistance emerging over 48 weeks of therapy with DIV/COM vs. ABC/COM (CNA3005). Program and abstracts of the 7th CROI, 2000. Abstract 750.

10. Wallace L, Anton E, Bunville J, et al. Genotypic characterization of resistance in patients failin an EFV/IDV/NRTI regimen (Study DMP266-020). Program and abstracts of the 7th CROI, 2000. Abstract 752.

11. Deeks SG, Barbour JD, Martin JN. Grant RM. Delayed immunologic deterioration among patients who virologically fail protease inhibitor-based therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 236.

12. Chun TW, Davey R, Ostrowski M, et al. Relationship between pre-existing viral reservoirs and the re-emergence of plasma viremia after discontinuation of highly active antiretroviral therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 239.

13. H. J. Stellbrink, J. Van Lunzen, M. Westby, et al. Influence of interleukin-2 (IL-2) on productive and latent HIV infection and on viral rebound. Programs and abstracts of the 7th CROI, 2000. Abstract 240.

14. Martinez-Picado J, Depasquale MP, Kartsonis NA et al. Selection of Antiretroviral Resistance in the Latent Reservoir of Human Immunodeficiency Virus Type 1 during Successful Therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 238.

Return to Contents

Sub-optimal treatment can often be identified at 4 and 12 weeks

Simon Collins, ATP.

Presentations showing on-going replication in patients previously assumed to be maximally suppressed with PCR RNA levels undetectable on assays <50 copies/ml concluded that this strengthened the importance of aiming for maximal suppression. John Mellors presented an analysis of three ACTG studies with a view to identify if early patterns of response could identify patients that could benefit from intensification of treatment earlier than the current 24 weeks given to access efficacy.

ACTG 343, 368 and 359 were in predominantly treatment naive, nucleoside-experienced and PI-experienced patients respectively. Total number of patients from the three studies in this analysis was 1037, with 6248 HIV RNA results.

Two different HIV RNA responses to a new treatment in each of these situations were identified as either 'on-track' or 'off-track'. On-track patients experienced a bi-phasic response with an initial decline (at week 4) followed by a continued slower decline (to undetectable <200) that was sustained. Off-track patients either showed a much slower initial decline or experienced a relatively sudden rebound after a period of initial suppression.

Although a clear relationship to the rate of initial decline and risk of later rebound was not identified, this analysis defined this slope as decline from baseline to week 4. The importance of initial decline often is only found with closer monitoring (after 7-14 days) which involves more sensitive and frequent monitoring than was available for this analysis.

Significant conclusions from the study with relevance to clinical management were:

While this study showed that predicting patients likely to benefit from intensification of their antiretroviral treatment between weeks 4-12 is not straight-forward, more frequent (monthly) monitoring over the first 3 months is effective at identifying early treatment failure in both naive and experienced patients. It could also strengthen the argument for choosing more potent initial regimens using 4-5 drugs for patients changing or initiating treatment at higher baseline viral load.

Ref: W. HUANG et al - Abstract 451 - Patterns of Plasma HIV RNA Responses in

Return to Contents

Antiretroviral Treatment Success and Failure. Benefit of continued treatment for patients with CD4 <50 cells/mm3

Simon Collins, ATP.

The strategy of promoting a return to wild-type virus using a treatment break for patients with resistance to currently available treatments includes a significant risk of CD4 decline (approximate drop of 90 cells/mm3 over the 2-3 month interruption period, not restored in all patients even after 12 months of subsequent successful therapy). This risk is likely to be unacceptable for patients whose existing CD4 count fails to provide a protective buffer against such a drop.

Veronica Miller presented an analysis from the EuroSIDA Cohort of the risk of disease progression for patients with CD4 counts <50 cells/mm3. EuroSIDA collects data from 52 centres in 17 European countries and in this study data was only collected from patients while their CD4 count remained below this level.

A total of 292 clinical events from 1252 patients (637.2 patient years) yielded an overall incidence rate of 45.8 per 100 PYFU (95% CI approx 38.9-49.7). The incidence rate was significantly lower for patients on a PI-including regimen (37.5 vs 68.4, p = 0.0001). It is important to recognise that the relationship between CDC classification and CD4 count at baseline was itself found to be significant. 60% patients diagnosed as AIDS had a mean CD4 count of 23 vs 30 in the non-AIDS group (p = 0.0003).

In conclusion, it was stressed that continued vs no-treatment:

Ref: V. MILLER et al - Abstract 454 - Association of Viral Load, CD4 Cell Count, and Treatment with Clinical Progression in HIV Patients with Very Low CD4 Cell Counts: The EuroSIDA Cohort.

Return to Contents

Opportunistic Infections

Mark Nelson, MD.
Chelsea and Westminster Hospital, London, UK.

Although commonly referred to as "retrovirus" one must remember that the full title of this conference is Retroviruses AND Opportunistic Infections. This is the 7th in the series of these meetings and it is sobering to remember that when these conferences began there were few therapeutic modalities against HIV available to physicians. Indeed those medications that were available were often misused, with either sequential monotherapy or reluctant dual therapy and much of the discussion at the time of the first conference was not if patients would die, but when they would die and whether the use of effective prophylaxis and appropriate therapeutic treatments for opportunistic infections could improve quality and quantity of life.

How times change

At the 7th CROI there were no symposium on opportunistic infections and a single symposium on hepatitis C. Although there was the occasional discussion of opportunistic infection within slide sessions, those few session dedicated to such subjects were poorly attended and had very little new to say.

This is a shame. We must realise that although HAART has revolutionised HIV disease the patients may eventually fail their HAART regimes, whether due to poor choice by the clinician or poor compliance by the patients, or secondary to drug toxicity. Opportunistic infections still do occur and we have few new therapeutic options compared with six years ago to deal with them. In addition the knowledge of many physicians who have been brought up in the era of HAART is minimal for the treatment of opportunistic infections. Hepatitis B and C are now major causes of morbidity and mortalities in countries where HAART has been used. This is not surprising as both hepatitis B and hepatitis C are transmitted by similar routes to HIV, and it should not be of surprise that the HIV population is commonly co-infected with these pathogens. However, these diseases have been relatively slow to progress and have been ignored by physicians who believed the HIV patient would die relatively soon, before problems from other viral infections could develop. This is clearly not the case and complications from viral hepatitis are now one of the major, if not the major, cause of death in individuals receiving HAART.

Suspending primary PCP prophylaxis

The major discussion concerning opportunistic infections was whether it is safe to stop these preventative measures once HAART has led to immunological improvement in individuals who may previously have been expected to receive life long prophylaxis. Several abstracts reported on the safe cessation of PCP prophylaxis - both primary and secondary. The cessation of primary prophylaxis has been extensively reported beforehand and in many centres in the UK is now the standard of care when CD4 count increases and remains above 200mm3. Koletar reported the results of the ACTG study 888[1] which considered two groups of patients - those who had had a CD4 count < 100 cells per mm3 at any previous time point but without confirmed PCP (group 1), and those who had a confirmed diagnosis of PCP > 6 months prior to study entry (group 2). All patients had sustained CD4 counts above 200 cells in response to HAART.

  Group 1 Group 2
Mean follow-up (weeks) 46.1 24.3
PCP cases 0 0
95% CI per 100 patient years 0 - 2.35 0 - 5.13
Serious bacterial infections 1 2

Furrer reported similar results on stopping primary prophylaxis from the Swiss StopCox study[2]. 381 patients were enrolled with a median nadir CD4 count of 105 cells/mm3. At enrolment post HAART median CD4 count had risen to 327. However 144 individuals still had a detectable plasma viral load despite the CD4 count rise. In this cohort one patient did develop PCP (who had a CD4 count of 530 cells/mm3 and a plasma viral load undetectable). The important take home message from this study is that it appears safe to stop prophylaxis in those patients with an immunological but no virological response. However, one would expect the CD4 count to fall in the future and it is important that CD4 is closely monitored and PCP prophylaxis re-instigated as necessary.

Suspension of secondary PCP prophylaxis

In the late breaker session, Ledergerber examined discontinuation of secondary prophylaxis for PCP[3]. Subjects were patients enrolled in 8 European cohort studies. Data was presented on 246 patients, 14% of whom were female. The duration of HAART when stopping secondary prophylaxis ranged from 14-27 months with a median CD4 of 277-371 cells/mm3. There was no recurrence of PCP diagnosed during follow up, resulting in a one-sided 95% upper confidence interval for the incidence of PCP of 1.27 per 100 patients years. This data on secondary prophylaxis is important as it would suggest that the immune reconstitution seen with HAART is real ie. in patients who have developed PCP, presumably due to immunological deficits, these appear to be rejuvenated.

Suspending Toxoplasmosis prophylaxis

The results of the GESIDA 04/98-B study were presented by Miro[4]. In this study prophylaxis to prevent toxoplasmosis encephalitis was withdrawn in patients with a sustained increase in CD4 cell count above 200mm-3 and a viral load of < 5000 copies for more than 3 months. 302 individuals patients were randomised to either withdraw prophylaxis ( n = 155) or continue ( n = 147). 20 individuals had had a previous episode of toxoplasmosis encephalitis and these were equally divided between the 2 groups. At the time of enrolment median CD4 count was 336. Patients had received toxoplasmic prophylaxis for a median of 33 months. After a median follow of up 10 months there were no episodes of toxoplasmic encephalitis or death in either of the two groups. These results suggest that both primary and secondary toxoplasmic prophylaxis may be discontinued in patients with immunological response to HAART. Suspending MAC prophylaxis

The safe discontinuation of MAC primary and secondary prophylaxis in the short term have previously been described from the UK early in 1998. Further evidence to support this view was presented at the conference. Furrer et al reported on patients who were eligible for MAC prophylaxis with two CD4 counts below 50cells/mm3 who had received HAART with a demonstrated CD4 count rise to greater than 100[5]. Median duration of CD4 count below 50 cells/mm3 was 12 months. During a total follow up of 364 patient years, there was no case of disseminated MAC suggesting that in patients responding to HAART there is no risk of development of MAC despite prolonged low CD4 counts. El-Sadr presented the results of CPCRA48 in which individuals with a CD4 count increasing from < 50 to >100 cells secondary to HAART were randomised to receive azithromycin 1200mg weekly or placebo[6]. No case of MAC occurred in either group, although the rate of bacterial pneumonia was reduced from 2.8 cases per 100 person years to 1.2 cases per 100 person years in those receiving azithromycin. This did not, however, reach statistical significance.

Suspension of prophylaxis for cryptococcus

The results of stopping cryptococcal prophylaxis in six individuals with disseminated cryptococcal disease were reported by Aberg[7]. All were asymptomatic for cryptococcosis, and had a CD4 count of greater than 150 on antiretroviral therapy for at least 16 weeks. All had received fluconazole prophylaxis therapy for at least 12 months. In all six individuals CSF and blood cultures were negative, although serum and cryptococcal antigen were positive in one of the six. CSF, urine and blood cultures remain negative off therapy and no patient has relapsed. This is important data which suggests that immune reconstitution may lead to the ability to eradicate this disease. However, with cryptococcosis the danger of sanctuary sites, such as the prostate, must not be under stressed and details of whether prostatic massage looking for such a sanctuary had been performed were not presented.

Suspending secondary cytomegalovirus prophylaxis

Several authors have previously reported on the safety of stopping secondary CMV prophylaxis. Although there was additional data presented to confirm this view, this was tinged with a variety of reports where this was not the case. Casado reported the withdrawal of maintenance therapy in a group of 35 patients[8]. The risk of relapse was associated with persistence of immunosupression only. Baldanti reported on the reappearance of human cytomegalovirus DNA in the blood of two HIV positive patients despite over two and a half years of successful treatment with HAART[9]. Johnson described an individual case of recurrent CMV retinitis - with relapse of cytomegalovirus disease at CD4 counts of 97, 373, 443, 511 and 496 cells/mm3 and suggested that this may be due to incomplete immune reconstitution[10]. Some confirmation of this was presented by Komanduri who used cytokine flow cytometry to measure antigen specific T cell response in two individuals with multiple relapses of CMV retinitis despite successful response to HAART[11]. Although there was a sustained increase in CD4 count to greater than 400 cells/mm3 in these individuals CMV specific cells constituted only 0.14% and 0.05% of the total CD4 count in these two subjects. These responses were significantly lower than the CMV specific CD4 T cell response of 44 HIV infected subjects with no history of retinitis. These results suggest that limited diversity in the T cell receptor repertoire may persist despite successful HAART. In such individuals cessation of prophylaxis may not be appropriate.

Prophylaxis for candidal vaginitis

Oral candidiasis is now relatively rare, although candidal vaginitis remains extremely common among HIV infected women. Williams reviewed the effectiveness of weekly intravaginal application of the probiotic Lactobacillus acidophilus (gelatine capsules) or clotrimazole 100mg tablets as prophylaxis against candidiasis[12]. A third patient group received placebo. All subjects received pelvic examination including vaginal swabs every 6 months. A total of 34 episodes of candidal vaginitis occurred among 164 women randomised between the 3 arms. The relative risk of experiencing an episode of candidal vaginitis was 0.4 in the clotrimazole arm and 0.5 in the lactobacillus arm. The median time to first episode of candidal vaginitis was longer for both the clotrimazole and the lactobacillus arm. These results suggest that vaginal yeast infections can be prevented with appropriate local therapy with no significant differences between the two active study arms.

Simplified regimens for TB prophylaxis

A recent study from Haiti reported a bi-weekly two month regimen of rifampicin and pyrazinamide was comparable to the standard isoniazid regimen for primary prophylaxis of patients at risk of tuberculosis. Kellman reported on a small study from the USA which examined bi-weekly rifampicin or rifabutin with pyrizinamide and compared the completion rate of this regime to that of a standard isoniazid course of 3 months[13]. Using a Fischer exact test, the two month regimen had a significantly higher completion rate (31 of 33) than the isoniazid regime (34 of 52). This higher completion rate, with minimal adverse reactions, suggests that this regime may well become the prophylaxis of choice.

The close association between HIV and tuberculosis was nicely shown within a study by Stout who collected surveillance data on all reported cases of tuberculosis in North Carolina from 1993 to 1999[14]. Of 3,395 patients reported, 12.2% were HIV seropositive. However 20% were not offered HIV testing, 14.7% were offered but refused the test and 25 individuals had results unreported. Therefore, although there was a high link between HIV and tuberculosis at least 35% of patients reported with tuberculosis in North Carolina had not received HIV testing. This is a real problem as the massive impact of combination antiretroviral therapy on the risk of tuberculosis was shown in a report by Geriade who stated that the relative hazard for tuberculosis was 0.18 for double combination therapy and 0.009 for triple combination therapy[15]. Therefore HAART significantly reduced the risk of tuberculosis in HIV infected persons, and as the authors state "...widespread use of this treatment may determine a decrease in the incidence of HIV associated tuberculosis, and could contribute to a reduction in the incidence of tuberculosis in industrialised countries."

New treatment approaches for opportunistic infection

Pneumocystis carinii pneumonia (PCP)

Only 2 posters reported information on the treatment of PCP, the commonest opportunistic infection prior to the time of HAART. Both gave leads on the treatment of this condition which still occurs, especially in those who present without knowing their HIV diagnosis. Kazanjian and colleagues presented further details of studies on mutations of di-hydrodropterate synthase (DHPS) and their effect on subsequent treatment of those presenting with PCP[16]. Previous work by the same group has shown that those who failed sulpha based prophylaxis (septrin or dapsone) commonly had mutations in this enzyme. DHPS mutations occurred in 28 of 37 isolates from patients with PCP who had received sulpha based treatments compared with only 14 of 60 receiving alternative prophylaxis. This was highly statistically significant. The risk of mutation was linked to the duration of SULFA prophylaxis. Mutations significantly increased the risk of sulpha based treatment failure, but not the risk of failure to therapy with a non-sulpha agent. This would suggest that those patients who present with PCP, having failed a sulpha based prophylaxis, should be treated with other agents either clindamycin and primaquine or pentamidine. However, clindamycin and primaquine has only shown equivalence to septrin in those with mild to moderate PCP. It is unlikely that we will ever know whether clindamycin/primaquine is equally successful in those patients who present with more severe forms of PCP.

In newly diagnosed individuals with severe, life threatening PCP, it is not unusual for the treating physician to commence antiretroviral therapy. The hope is that the general improvement in the immune system may lead to improvement in their condition, although there is no evidence to back this approach. Atzori presented data suggesting that protease inhibitors themselves may result in an anti-pneumocystis effect[17]. However, in clinical practice this approach may lead to problems due to difficulties in compliance with such regimes in very sick patients and the subsequent risk of the development of antiviral resistance.

Cytomeglavirus (CMV)

Three agents are presently licensed for the treatment of cytomeglavirus (CMV) disease - ganciclovir, foscarnet and cidofovir. All three must be given intravenously. Clearly, an orally bioavailable compound would offer advantages. Martin presented data on a trial comparing valganciclovir (an orally monovalyl ester of ganciclovir) and intravenous ganciclovir for CMV retinitis[18]. Valganciclovir was given orally 900mg bd for 3 weeks then 900mg bd. Ganciclovir was administered i.v. 5mg per kilogram bd for 3 weeks then 5mg per kilogram once daily.

  i.v. ganciclovir Oral valganciclovir
Response 77% 72%
Failure at week 4 N=7 (10%) N=7 (10%)
Median time to 1st CMV progression 198 days 120 days

These promising results suggest that valganciclovir provides a convenient and effective alternative to intravenous ganciclovir for the treatment of CMV retinitis.


Although common in the southern states of the U.S.A histoplasmosis is rare in the UK. Sporadic cases do occur and treatment is usually amphotericin B followed by itraconazole secondary prophylaxis. Amphotericin must be prescribed intravenously and may be associated with significant morbidity in the patient including severe malaise, rigors and most importantly renal failure. An alternative form of amphotericin B - liposomal amphotericin B - is available. The amphotericin B is wrapped in fat drops (liposomes). This has been shown to be equally efficacious and less toxic in the treatment of several fungal infections. Johnson suggests that this may be similar with the treatment of histoplasmosis[19]. In patients who received the liposomal form of amphotericin, success was commoner and infusion related side-effects less and the nephrotoxicity defined as twice the base line creatinine, much less frequent (9% vs 53%) in the liposomal amphotericin arm.

Human Papilloma Virus (HPV)

Scattered amongst the infectious disease abstracts were two important abstracts on human papilloma virus, the common wart virus. Mosciki studied samples from adolescents aged 13 to 19 years participating in a national study of HIV infection for anal HPV DNA and cytology[20]. HPV infections were more common in males (45%) than females (26%). HIV status and CD4 count was significantly associated with HPV infection for females only (13% HIV negative vs 32% HIV positive). In females with a CD4 count of less than 200 cells 65% had evidence of anal HPV compared with 19% of those with a CD4 count of above 500. Anal squamous intra-epithelial lesions (SIL) were strongly associated with HPV status in both males and females. Therefore the presence of anal genital warts appears to be a significant risk factor for both males and females, and carries a significant risk of anal SIL. In the second report by Piketty a high prevalence of high grade SIL and anal HPV infection in HIV infected intravenous drug users was reported despite the absence of receptive anal intercourse[21]. This would be an important finding if confirmed as it suggests that all HIV positive males - not only those with a history of anal intercourse should be screened by anal cytology.


1. Koletar SL, Heald AE, Murphy RL et al. Discontinuing primary and secondary PCP prophylaxis in patients who have increased CD4 counts in response to antiretroviral therapy: Preliminary results - ACTG 888. Programs and abstracts of the 7th CROI, 2000. Abstract 243.

2. Furrer H, Opravil M, Rossi M et al. The Swiss StopCox study: Is it safe to discontinue PCP prophylaxis in patients with detectable viraemia, low nadir CD4 count, or T. gondii seropositivity? Programs and abstracts of the 7th CROI, 2000. Abstract 244.

3. Ledergerber B, Mocroft A, Reiss P et al. Is it safe to discontinue secondary prophylaxis for PCP in HIV-infected patients treated with HAART: Results from eight prospective European cohorts. Programs and abstracts of the 7th CROI, 2000. Abstract LB5.

4. Miro JM, Lopez JC, Podzamczer D et al. Discontinuation of toxoplasmic encephalitis prophylaxis is safe in HIV-1 and T. gondii coinfected patients after immunological recovery with HAART: Preliminary results of the GESIDA 04/98-B study. Programs and abstracts of the 7th CROI, 2000. Abstract 230.

5. Furrer H, Telenti A, Rossi M et al. Discontinuing or withholding primary prophylaxis against M. avium in patients on successful antiretroviral combination therapy: The Swiss HIV Cohort experience. Programs and abstracts of the 7th CROI, 2000. Abstract 246.

6. El-Sadr WM, Burman W, Grant L et al. Prophylaxis for Mycobacterium avium complex can be deferred among patients with a past CD4 count <50 cells/mm3 who responded to antiretroviral therapy: Results of a placebo-controlled trial (CPCRA 048). Programs and abstracts of the 7th CROI, 2000. Abstract 247.

7. Aberg JA, Price RW, Heeren DM et al. Discontinuation of antifungal therapy for cryptococcosis after immunological response to antiretroviral therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 250.

8. Casado JL, Arrizabalaga J, Mallolas J et al. A 2-year, multicenter study of cytomegalovirus retinitis in AIDS patients receiving protease inhibitor therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 266.

9. Baldanti F, D'Arminio Monforte A, Zavattoni M et al. Reappearance of human cytomegalovirus DNA in blood of HIV-positive patients after 2.5 years of HAART. Programs and abstracts of the 7th CROI, 2000. Abstract 271.

10. Johnson S, Benson C, Johnson D et al. Recurrent cytomegalovirus retinitis in a patient on highly active antiretroviral therapy despite apparent immune reconstitution. Programs and abstracts of the 7th CROI, 2000. Abstract 272.

11. Komanduri K, Feinberg J, Hutchins R et al. Persistent deficits in cytomegalovirus (CMV)-specific CD4+ T-cell responses in HIV-1-infected subjects with relapsing CMV retinitis following successful highly active antiretroviral therapy. Programs and abstracts of the 7th CROI, 2000. Abstract 581.

12. Williams A, Yu C, Tashima K et al. Weekly treatment for prophylaxis of candida vaginitis. Programs and abstracts of the 7th CROI, 2000. Abstract 677.

13. Kellman M, Narita M, McMillan M et al. Short-course 2-month regimen for latent M. tuberculosis infection in HIV-infected patients in Broward County, FL. Programs and abstracts of the 7th CROI, 2000. Abstract 248.

14. Stout JE, Ratard R, Southwick K et al. Epidemiology of HIV testing among tuberculosis patients in North Carolina. Programs and abstracts of the 7th CROI, 2000. Abstract 474.

15. Girardi E, Antonucci G, Vanacore P et al. Impact of combination antiretroviral therapy on the risk of tuberculosis among persons with HIV. Programs and abstracts of the 7th CROI, 2000. Abstract 255.

16. Kazanjian P, Armstrong W, Hossler PA et al. Pneumocystis carinii mutations are associated with duration of sulfa prophylaxis and with sulfa treatment failure in AIDS patients. Programs and abstracts of the 7th CROI, 2000. Abstract 229.

17. Atzori C, Angeli E, Agostoni F et al. Indinavir and other protease inhibitors decrease pneumocystis carinii in vitro growth. Programs and abstracts of the 7th CROI, 2000. Abstract 245.

18. Martin D, Sierra-Madero J, Walmsley S et al. Valganciclovir vs. IV ganciclovir as induction therapy for newly diagnosed cytomegalovirus retinitis: A randomised, controlled study. Programs and abstracts of the 7th CROI, 2000. Abstract 231.

19. Johnson P, Wheat L, Cloud G et al. A multicenter randomised trial comparing Amphotericin B and liposomal amphotericin B (AmBisome) as induction therapy for disseminated histoplasmosis in AIDS patients. Programs and abstracts of the 7th CROI, 2000. Abstract 232.

20. Moscicki AB, Houser J, Ma Y et al. HIV status and Anal human papillomavirus infection in adolescents. Programs and abstracts of the 7th CROI, 2000. Abstract 8.

21. Piketty C, Darragh TM, Da Costa M et al. High prevalence of anal squamous intraepithelial lesions related to HPV infection in HIV-infected intravenous drug users. Programs and abstracts of the 7th CROI, 2000. Abstract 276.

Return to Contents

Pharmacokinetics, Drug-Drug Interactions and Therapeutic Drug Monitoring

Simon Collins, ATP.

The pharmacokinetics and pharmacodynamics of antiretroviral agents includes not only variability of intra- and inter-patient drug plasma levels and drug-drug interactions, but also absorption, gut wall (P-gp efflux) and hepatic metabolism, intracellular clearance and uptake, protein binding, renal clearance etc. Nevertheless, of the 30 or so abstracts that touched on or directly addressed drug levels and antiviral potency, added weight was provided for increasing access to Therapeutic Drug Monitoring (TDM) in routine clinical care for both adult and paediatric care.

Return to Contents

Added Value of TDM

In a symposium session looking at the management of treatment failure, Lisa Demeter, highlighted results from the GART study which suggested that PI trough levels were an independently predictive factor of virological response.

The primary objective of GART was to look at whether there are benefits to utilising genotypic resistance tests when choosing a second-line or salvage regimen, but it also included a PK sub-study to address a possible role for therapeutic drug level monitoring. The percentage of subjects with undetectable viral load (<200 copies/ml) dependent on whether they had access or not to genotype results and optimal/sub-optimal concentrations (OC/SOC) at three and six months was as follows:

Results showing % subjects <200 copies/mL

  3 months 6 months
No geno, SOC 0% 0%
Geno, SOC 15% 8%
No geno, OC 23% 23%
Geno, OC 34% 40%

In this study the patients with access to genotypic testing and who achieved optimum concentrations of the PIs in their combination had a greater likelihood of a sustained response. Baseline resistance was found to not fully predict the likelihood of failure and, importantly, not to assure a sustained response independently of optimal PI plasma concentrations. [1]

Other studies have already reinforced the link between drug-plasma levels and efficacy (see DrFax 53, 60, 76 for previous conference reports on this subject). A study by Anderson and colleagues from University of Minnesota, primarily looked at quantifying levels of protein binding for indinavir. These were generally found to be consistent with previously reported in vitro estimates of 60% (57-74% in this study), but here antiviral effect was also found to be greatest in people who had highest concentrations both of unbound indinavir and total indinavir. [2]

Ecosta reported results from a PK/PD sub-study from the ACTG 343 induction/maintenance study (all patients started with IDV/3TC/ADV)[3]. 208 concentration timepoints were available from 90 of the original 509 patients. Indinavir trough levels were obtained 6-10 hours post dose at weeks 4, 8, 20 and peak levels were collected 30-60 minutes post dose at week 24. Efficacy variables for exploratory PD included: nadir, decay slope, % change, log decline, and normalised area-under-the-curve (NAUC). Subjects with IDV peak/troughs and/or multiple steady-state levels were analysed using a one-compartment, linear elimination model with Bayesian estimation. Pharmacokinetic (PK) parameters were compared between subjects with or without virologic suppression (<200 copies/ml at wk 24). Subjects were also assigned as responders or non-responders according to whether plasma HIV RNA was > or < 1.5 log10 drop at week 4, respectively.

Modelled IDV troughs (median; 229 vs 137 ng/ml) were significantly different between those with < 200 (n=71) and > 200 copies/ml (n=19) at 24 wks (P=0.006). A significant relationship was observed (r=0.34, P<0.05) between measured IDV troughs at week 4 (n=36) and log change in plasma HIV RNA at week 4. In addition to reporting that IDV trough levels at week 4 were significantly related to changes in plasma HIV RNA at 24 weeks, the authors suggested a minimum effective IDV trough concentration (in this treatment naive population) to be 110 ng/ml.

David Back and colleagues at Liverpool University presented further evidence of marked inter-patient variability in trough levels in patients using single-PI regimens with saquinavir, indinavir and nelfinavir[4]. Moving to dual-PI regimens with concomitant use of ritonavir resulted in trough levels greater than the IC95 being achieved by all patients using indinavir and most, but notably not all, patients using saquinavir. Although trough concentrations may not represent the true minimum concentration over a dosing interval, evidence strongly suggests that antiretroviral effect is related to maintaining plasma concentrations greater than MIC/IC95 (rather than the relationship between the Cmax and MIC). The lack of consensus for how to utilise TDM was also addressed. Although single blood samples (trough) provide useful information, these may not always provide optimal results compared to a full AUC. However, particularly given cost and other practical limitations, truncated AUCs (0-2hr or 0-3hr) provided improved accuracy over single point samples and were found to correlate well with full AUCs.

TDM and paediatric care

Several studies also showed the risks of relying of pre-determined dosing for all patients. This is a particular issue in paediatric care. Although TDM is increasingly being utilised in the UK, many centres have yet to realise the immediate benefit available. With a relatively small HIV-positive paediatric population of around 300 children, integrating TDM for attaining and maintaining optimum dosing for all children would eliminate this unnecessary variable. There is already a subsidised (free) programme provided by Roche for patients, including children, using saquinavir and/or nelfinavir, and recent capital investment at Liverpool University has allowed for an increased throughput of samples.

Several PK studies made recommendations for dose adjustment for neonates and for babies less than two years old. Maternal PK of nevirapine was found to be unaffected compared to non-pregnant women, although insufficient concentrations were measured in babies born to mothers not using NVP as a part of HAART therapy throughout their pregnancy[5]. Use of single-dose maternal nevirapine during labour, followed by a single oral 2mg/kg NVP dose to the infant at 48-72 hours resulted in NVP concentrations in the baby falling below 100ng/ml in 4 out of 10 babies by day 7. The study recommended increasing frequency of NVP dosing to infants whose mothers use NVP in the last week of pregnancy in order to maintain therapeutic levels through to day 7. Graham Taylor from Imperial College also reported a 36% reduction in NVP plasma concentrations 24hours post-birth in infants born to mothers who have used NVP-based HAART throughout second and third trimester and who are therefore at NVP steady-state throughout labour[6].

Brenson and collegues using standard dosing of nelfinavir in 10 pregnant HIV-positive women together with ZDV/3TC during gestation (range 14-34 weeks), postpartum and with IV ZDV in labour reported maternal nelfinavir plasma concentrations at delivery of 2(g/ml. However, the initial dosing of NFV at 10mg/kg bid in the infant provided inadequate concentrations leading to increasing the dosing to 40 mg/kg BID for future studies[7]. A second study presented by Brundage for the PACTG 382 study team suggested that 50mg/kg may be more appropriate in children who are younger than 8 years old[8]. An evaluation of both efavirenz and nelfinavir levels used in the PACTG 382 study (used together with at least one NRTI) also found that adopting doses for children under 2 yrs based on doses in older children may not always be appropriate. 18 children (age 2-8 years) EFV dosed at 720mg/day, allometrically scaled to body size [(wt kg/70) 0.7], produced a median AUC concentration of 158(Mxhr, with 11 AUCs remaining below the target range of 190-380(Mxhr. NFV (powder or tablet) concomitantly dosed at 35-45 mg/kg TID produced a median AUC of 19.3 mgxhr/l, with 5 children falling below the target 10 mgxhr/l. 10 children (aged 3 months to 2 years) median AUCs were 115.4(Mxhr and 9.2mgxhr/L for EFV and NFV respectively. 7 EFV and 5 NFV AUCs were below target (70% and 50%). Dosing was revised for these children in future studies to allometric scaling of 1200 QD for EFV and disintegrated tablet dose of 50 mg/kg TID for NFV.

A statistically significant relationship was also found in the above PACTG 382 study between variability of efavirenz concentrations and virologic outcome. AUCs were obtained for 50 children at weeks two and six [9]. These parameters were then used to predict routinely collected EFV concentrations (up to 12) over the first year of the study. A measured concentration within a (40% range of the predicted concentration was considered acceptable. The fraction of all observations within this acceptable region was defined as the Integrated Pharmacokinetic-Adherence Measure (IPAM) (range 0-1) for each subject. Subjects were classified as low IPAM (<33rd percentile) or high IPAM ((33rd percentile). Viral rebound (VR) was defined as having >400 copies/mL after at least two consecutive RNA <400 copies/mL or a >0.75 log10 increase from nadir otherwise.

The median IPAM was 0.20 in the low group and 0.67 in the high. 7 of 31 children (22.6%) in the high IPAM group experienced VR vs 10 of 19 children (52.6%) in the low IPAM group (P=0.037, Fisher's exact). The low IPAM group exhibited a significantly shorter time to first VR (P=0.008, log-rank test); this finding persisted in a Cox regression analysis, which controlled for baseline viral load (P=0.022).

The study chose to link low concentrations to adherence rather than PK, but in either case, identification of sub-optimal concentrations provided by therapeutic drug level monitoring, would allow for either early dose adjustment or intensified adherence support. A similar sub-population were identified in PACTG 366 (age range 0.8 - 21 yrs) who recorded sub-optimal concentrations whilst on NFV combinations, even when co-administered with ritonavir. [10]

The substantial decrease in mortality associated with increased use of PIs in paediatric care was documented by Gortmaker [11]. This has resulted in a similar use in practice of combinations using dual PIs. However, even with CYP450 inhibition effect of concomitant RTV a sub-population of children were found to fall below optimum dosing[12]. Given the higher baseline viral load of children starting and changing treatment, until more detailed PK data is provided for use of antiretrovirals in paediatric care, using TDM routinely would seem to offer greater likelihood of guaranteeing an optimal response.

References 1. L. DEMETER - Drug Resistance and Treatment Failure - Session 73 - Symposium S32 7th CROI.

2. P. L. ANDERSON, L. R. BUSHMAN, T. N. KAKUDA - Univ. of Minnesota - Abstract 101 - Quantitation and Anti-Viral Effect of Unbound Plasma Indinavir (IDV) in HIV-Infected Persons.

3. E. P. ACOSTA, D. V. HAVLIR, D. D. RICHMAN et al for the ACTG 343 TEAM - Abstract 455 - Pharmacodynamics (PD) of Indinavir (IDV) in Protease-Naive HIV-Infected Patients Receiving ZDV and 3TC.

4. C. MERRY, D. BACK, M. BARRY et al, St. James's Hosp., Dublin; and Royal Liverpool Univ. Hosp., Liverpool. - Abstract 104 - Therapeutic Drug Monitoring (TDM) of Protease Inhibitors (PIs): What to Measure and When.

5. M. MIROCHNICK et al and the PACTG 250 Protocol Team - Abstract 716 - Pharmacokinetics of Nevirapine (NVP) in Pregnant Women and in Their Infants Following In Utero Exposure.

6. G. P. TAYLOR, D. BACK et al - Imperial College, UK, Univ. of Liverpool, UK - Abstract 717 - Plasma Clearance of Nevirapine Is Increased in Neonates of Mothers Treated with Nevirapine during Pregnancy

. 7. Y. BRYSON, A. STEK, M. MIRICHNICK, et al - Abstract 715 - PACTG 353 A Phase I Study of Safety, Pharmacokinetics, and Antiviral Activity of Combination Nelfinavir (NFV), ZDV, and 3TC in HIV-Infected Pregnant Women and Their Infants.

8. R. C. BRUNDAGE et al and PACTG 382 Study Team - Abstract 719 - Efavirenz (EFV) and Nelfinavir (NFV) Pharmacokinetics (PK) in HIV-Infected Children under 2 Years of Age.

9. R. C. BRUNDAGE et al - Univ. of Minnesota, Minneapolis - abstract 699 - Variability in Efavirenz (EFV) Concentrations Predicts Virologic Outcome in HIV-Infected Children.

10. E. CAPPARELLI et al - Abstract 661 - Nelfinavir (NFV) Pharmacokinetics in Combination with Ritonavir (RTV) in Infants and Children with Advanced HIV Disease.

11. S. GORTMAKER et al - Abstract 691 - Impact of Introduction of Protease Inhibitor Therapy on Reductions in Mortality among Children and Youth Infected with HIV-1.

12. E. CAPPARELLI et al - Abstract 661 - Nelfinavir (NFV) Pharmacokinetics in Combination with Ritonavir (RTV) in Infants and Children with Advanced HIV Disease.

Return to Contents

Methadone - nelfinavir interactions

A prospective study of 14 subject on stabilised (> 1 month) fixed doses of methadone (10-140mg/day) who added nelfinavir (1250mg BID) for 8 days. Levels of methadone enantiomers and active metabolites ((+) and (-)-EDDP) were measured before and after. Nelfinavir reduced the levels of (+)-MD, (-)-MD, (+)-EDDP, and (-)-EDDP by 47%, 39%, 29%, and 34%, respectively although none of the patients showed any withdrawal symptoms during the study or required dose adjustment during and after the study. NFV and its active metabolite M8 levels were within the normal range of historical data. A retrospective study of case histories from 36 patients who were maintained with stable MD dose (20-110mg/day) were reviewed before and after addition of NFV (750mg TID or 1250mg BID from 3 to 105 weeks) reported no withdrawal or intoxication symptoms. In this retrospective study, one patient required an increase in MD dose from 70 to 80mg/day, and one patient voluntarily reduced MD dose from 30mg to 20mg/day.

Ref: P. H. HSYU, J. H. LILLIBRIDGE, B. M. KERR - Abstract 87 -Pharmacokinetic (PK) and Pharmacodynamic (PD) Interactions between Nelfinavir and Methadone.

Methadone levels reduced by nevirapine or efavirenz

Methadone levels were found to be significantly reduced in a PK study from St James' Hospital, Dublin and Liverpool University following initiation of either nevirapine or efavirenz based combinations. 25 subjects on stable methadone initiated NNRTI-based regimens (2RTIs + either nevirapine (n=10) or efavirenz (n=15)) and methadone levels were measured before and 2-3 weeks following treatment at 0, 1, 2, 3, 4, 5, 6, 7, 8 and 24 hours post dosing. When EFV was combined with methadone, there was a decrease in the mean maximum plasma concentration of methadone from 689 (range 212-1568) to 358 (205-706) ng/ml; p=0.007 [95% CI 112 to 549]. The mean AUC over 24 hours for methadone also significantly reduced in the presence of EFV from 12341 (range 3682-34147) to 5309 (2430-10349) ng.h/ml; p=0.011 [95%, CI 1921 to 12143]. Initial data on the interaction between NVP and methadone shows a mean reduction in AUC over 24 hours for methadone of 46% (range 23-80%). Seventeen patients complained of symptoms consistent with methadone withdrawal from day 8-10 onward that required a mean increase in methadone dose of 21.65% (mean increase 16 mg, range 15-30 mg). An apparent process of 'induction -detoxification' during the first 1-2 weeks of therapy means that frequent monitoring for withdrawal symptoms from day 8-10 onwards, followed by increasing methadone dose in increments of 10mgs as required is more appropriate than altering dosing when initiating these NNRTI-based regimens.

Ref: S. Clarke, F. Mulcahy, D. Back, et al - Abstract 88 - Managing Methadone and Non-Nucleoside Reverse Transcriptase Inhibitors: Guidelines for Clinical Practice.

Ritonavir may compensate for efavirenz/saquinavir interaction

Efavirenz is contraindicated for use with saquinavir in a single-PI combination as it has been shown to significantly reduce SQV levels. Nevertheless, particularly in a salvage setting, the two drugs have been used in the same combination with ritonavir concomitantly dosed in the hope that RTV will compensate for the EFV reduction.

While we are still waiting (how long do we have to wait?) for supportive PK studies on this, and other 3-way interactive combinations, Hendrix presented data from a Johns Hopkins and DuPont Pharma study on two patients using these three drugs together. Both subjects were already taking RTV/SQV (400mg/400mg Q12H) and added EFV (600mg QD). PK was recorded at days 1, 14 and 28. From day 14-28 the SQV dose was increased

The study concluded that SQV is not affected by EFV when RTV is co-administered. However, given the inter-patient variability of PI exposure reported in other studies (see earlier TDM report above), until results from a larger PK study are available, it may still remain prudent to confirm drug levels on an individual patient basis when combining these three agents.

Ref: C.W. Hendrick. Abstract 79. Pharmacokinetics of the triple combination of saquinavir, ritonavir and efavirenz in HIV-positive patients.

Return to Contents

Nevirapine reduces efavirenz levels

Dual NNRTIs are less likely to be commonly prescribed in clinical practice. In fact, whether dual NNRTIs can offer an added benefit similar to PIs if they are competing for the same pocket site still has to be demonstrated. Although nevirapine and efavirenz have been included together within the protocol of at least one mega-HAART study (Montaner), PK data was only presented for these two drugs at this conference. A Dutch study, in collaboration with Boeringer Ingelheim, studied 20 patients receiving EFV-based combination for at least 2 weeks who added nevirapine (200mg QD for 2 weeks, then 200mg Q12H) and had PK assessed at day 28. Results were available for 12 patients. Plasma HIV-1 RNA was below 400 copies/mL in 10/12 patients. NVP PK was unaffected by the co-administration of EFV. The median AUC0-24h exposure to EFV between day 1-28, decreased by approximately 22% [54.06 to 35.58 h*mg/L (p=0.002)]. Median Cmin was reduced by 36% (1.26 to 0.96 mg/L (p=0.002). The study concluded that increasing the dose of EFV to 800mg QD may be appropriate.

Ref: A. VELDKAMP, R. HOETELMANS, J. LANGE et al - Abstract 80 - DONUT: The Pharmacokinetics (PK) of Once Daily Nevirapine (NVP) and Efavirenz (EFV) When Used in Combination.

Return to Contents

Nevirapine reduces levels of oral contraceptives

Co-administering NVP with ethinyl estradiol (EE) and norethindrone (NET) to 10 HIV-positive women on stable background therapy led to a significant reduction of AUC of 29% and T1/2 of EE and of 18% median reduction in the AUC of NET. Cmax levels were not significantly altered for either EE or NET. The study compared values following single dose EE/NET at baseline to levels of a similar single dose after 30 days NVP treatment. The study, supported by Boehringer Ingelheim, concluded that oral contraceptives should not be used as the primary method of contraception for women using nevirapine-based combinations.

Ref: G. LEITZ, D. MILDVAN, M. MCDONOUGH et al - Abstract 89 - Nevirapine (VIRAMUNE, NVP) and Ethinyl Estradiol/Norethindrone [ORTHO-NOVUM 1/35 (21 Pack) EE/NET] Interaction Study in HIV-1 Infected-Women.

Return to Contents

Pregnancy and Perinatal transmission

Polly Clayden, ATP.

In addition to being the subject of a number of abstracts at this conference, care of pregnant women and prevention of perinatal transmission was also one of the dedicated state-of-the-art lectures [L3] - 'Management of the HIV-infected pregnant woman' [1] given by Dr Mary Jo O'Sullivan from the University of Miami. This talk and many of the abstracts contained few new advances and many of the same controversies that we have covered over the past year in Dr Fax and PTN. There was, however a clear overview, in line with US guidelines for treatment of pregnant women in an industrialised nation with access to antiretroviral agents so it is worth reiterating some of the issues.

In the US pregnant women divide into two categories, the known infected and (the vast majority) those whose status is identified during pregnancy. Testing varies from state to state, in some it is mandated and in most it is offered. Of great significance in the UK where screening for HIV will soon be offered routinely to all pregnant women, Dr O'Sullivan stated that women's '...willingness to accept testing is related to their perception of their healthcare providers attitude' (to HIV).

A woman's own health and her treatment (which should be as if she were a non-pregnant adult) were stressed to be of primary importance, with prevention of perinatal transmission and the health of her infant having linkage to her ongoing care. Women are offered antiretroviral treatment appropriate to their infection, according to Center Disease Control (CDC) guidelines. [2] She explained, 'Regardless of their state of pregnancy women should receive optimal treatment for their status'.

As an adjunct to their routine antiretroviral therapy women should also be screened, and treated if necessary, for toxoplasmosis and cytomegalovirus, which can be transmitted to the infant. Although the use of amniocentesis has previously been discounted in HIV-infected women, as it is invasive, this caution may be unnecessary with use of HAART, and the procedure seems anecdotally safe if a woman is BLQ. All vaginal infections should also be treated, as they can be linked to perinatal transmission and Pap smears given and followed up aggressively particularly if a woman is <200.

Dr O'Sullivan talked about the safety of antiretrovirals in the context of pregnancy and stressed the paucity of data on their effect on the foetus; she also noted that this is the first time since thalidomide that drugs have been used in pregnancy without major studies.

She explained that despite caution over the use of protease inhibitors and their possible links to pre-term birth (ie. <37 weeks) there seemed to be no greater risk with PI containing multi-drug regimens than with monotherapy. [3] And in his slide presentation Dr David Shapiro gave results of the PACTG 664 study [4] and found that women that received a multi-agent PI-containing drug regimen vs. women that received a multi-agent regimen without a PI, no difference was observed with rates of pre-term birth.

Both IDV and RTV are known to cross the placenta, but little is known about their effect on the foetus, and studies with NFV, SQV and APV are in progress. Animal studies with amprenavir have shown skeletal abnormalities in primates. And with NNRTIs they have shown teratogenic effects with EFV use, although NVP appears to be safe.

Dr Kenneth McIntosh addressed the issue of NRTI use in preventing perinatal transmission being linked to subsequent mitrochondrial toxicity in the children, in his slide presentation [5]. The uninfected children of over 20,000 HIV-infected women including 227 deaths were retrospectively analysed and abnormalities consistent with mitrochondrial toxicities investigated.

It was reassuring that this large, extensive US study designed in response to Dr Stephan Blanche's reports of 2 deaths of infants exposed to ZDV and 3TC [6] and 6 other cases of mitrochondrial toxicity [7] last year in France, failed to show evidence of fatal mitrochondrial toxicity in children perinatally-exposed to antiretrovirals.

The controversy over mode of delivery still remains. Three studies [8], [9], [10] looked at elective C-section, which was shown to be cost effective but most of the data still relates to pre-HAART and is only applicable to ZDV monotherapy settings.

Dr O'Sullivan emphasised that there was a very low risk of transmission if a woman was below the level of detection and that C-sections can carry a number of risks for the HIV-infected woman. It has also been demonstrated that infants delivered by elective C-section at 37 to 39 weeks were 120 times more likely to receive ventilatory support for respiratory disease than those born naturally at 39-41 weeks [11]. She gave a brief outline of when she would recommend each method:

  • Vaginal delivery:
  • Where viral load is undetectable
  • Where a women has this as preference
  • In labour or following ruptured membranes
  • C-section
  • Detectable viral load
  • Not on ARV treatment
  • Patient request
  • Late diagnosis in pregnancy
  • No ART
  • Although it is recommended that all women receive HIV testing - in practice this is not happening, and this is resulting in very late diagnosis for some women. A study looking at transmission rates for mothers receiving no pre-natal care [12] showed that out of 325 births to HIV-infected women, 78 (24%) received no prenatal care. Significant difference was observed between the rates of HIV transmission from women who received pre-natal care to those that did not, (24% vs. 10%, p= 0.005). There were no differences between the two groups in terms of ethnicity, stage of disease or mode of delivery but significant factors were greater maternal age and instances of substance use.

    Dr O'Sullivan looked at some possible strategies for decreasing transmission in this population. There is a need for rapid HIV testing and then possible prophylaxis use of ZDV or NVP in line with African studies. There is as yet no information on ZDV and NVP used together in this setting, although in the light of Dr Susan Eshleman's presentation [13] it would only seem prudent to use NVP as part of a drug combination.

    Finally she spoke about postpartum care and reported this disturbing figure from the Center for Disease Control and Prevention that although HIV-infected women were 9-95% adherent during pregnancy after birth this reduced to 0-60%. Support during this time is essential to encourage a woman to also concentrate on her own health as well as that of her infant.

    1. Management of the HIV-infected Pregnant Woman. MJ O'Sullivan. 7th CROI L3

    2. Public Health Service recommendations for use of antiretroviral drugs in pregnant women infected with HIV. Morbidity and Mortality Weekly Report 1998; 47(RR-20) 1-30

    3. Safety of combined therapies in pregnancy. P Lorenzi, B Masserei, B Laubereau et al. 12th World AIDS Conference 1998. Abstract 32453

    4. Antepartum Antiretroviral Therapy and Pregnancy Outcomes in 462 HIV-infected Women in 1998-1999 (PACTG 367). D. Shapiro, R.Tuomala, J Read et al 7th CROI. Abstract 664

    5. Mitrochondrial Toxicity of Perinatally Administered Zidovudine. K. McIntosh 7th CROI S14

    6. Zidovudine-lamivudine for prevention of mother-to-child HIV infection. S Blanche, C Rouziox, L Mandeldrot et al. 6th CROI Abstract 267

    7. Persistent mitrochondrial dysfunction and perinatal exposure to antiretroviral nucleoside analogues. Lancet 1999; 354:1084-1089

    8. Prevention of mother-to-child transmission of HIV by elective caesarean section: costs, outcomes and cost-effectiveness. M Halpern, J Read et al. 7th CROI. Abstract 705

    9. Increased use of maternal combination antiretroviral therapy and elective ceasarian section to reduce perinatal HIV transmission in North Carolina. S A Fiscus, AA Adimora et al 7th CROI Abstract 706

    10. Trends to reduce perinatal HIV transmission in New York City. V Peters, A Marinovich et al. 7th CROI Abstract 707

    11. Surfactant-deficient respiratory distress after elective delivery at term. J Madar, S Richmond et al. Acta Paediatric. 1999;88(11): 1244-1248

    12. Perinatal HIV transmission among women with no prenatal care. A Bardeguez, L Pompeo et al 7th CROI Abstract 711

    13. Selection of the K103N nevirapine resistance mutation in Ugandan women receiving NVP prophylaxis to prevent vertical transmission (HIVNET 006). S Eshleman, G Becker-Pergola, L Guay et al 7th CROI Abstract 658

    Important new data presented at the 7th CROI raises questions about the use of single prophylaxis dose nevirapine (NVP) to reduce perinatal transmission.

    In July 1999, results published in the Lancet from the HIVNet 012 trial demonstrated that a single 200mg oral NVP dose administered to the mother at onset of labour, followed by a single 2mg/kg dose given to her infant 72 hours after birth, significantly reduced the rate of vertical transmission. This compared favourably with mothers and infants receiving short course zidovudine (ZDV) - 600 mg oral maternal dose at onset of labour and then 300 mg every three hours until delivery and 4mg/kg oral dose to the infant for seven days thereafter. Transmission rates at weeks 14-16 were 13% in the nevirapine arm and 25% in the ZDV arm. The significance is also important in economic terms, as this strategy is relatively inexpensive - costing approximately $4 for each nevirapine-receiving pair. This is 70 times less expensive than the longer course ZDV regimen required, offering similar risk reduction. [1]

    This study, conducted in Uganda by Dr Laura Guay from the John Hopkins University, seized global attention as a feasible intervention for reducing mother-to-child transmission, particularly in the developing world. Besides the cost, NVP was seen to have a number of useful advantages:

    In the dedicated perinatal slide session, which opened with the enthusiastic, 'dramatic improvements in therapy have correlated in reduction of perinatal transmission, most excitingly with the nevirapine trial'; Dr Guay presented updated findings from the HIVNet 012. She described the results as ' least a good as their wildest dreams', but raised the question of how best to implement these findings, and was asked, most significantly, what the likelihood was to develop resistance to NVP after just a single dose.

    This question formed the basis of Dr Susan Eshleman's later presentation, a study to determine the presence of NVP resistance mutations in treatment-naive Ugandan women, following a single 200mg dose of NVP, as described above in the HIVNet 012, for prevention of HIV perinatal transmission. [2]

    After analysing 14 plasma samples, collected six weeks after dosing in labour, the K103N mutation, associated with high level NVP resistance, was detected in 20% cases - from 1/3 transmitters and 2/11 non-transmitters. Pre-dose samples were available for 2/3 of these women and the K103N mutation was absent from both of these. Other NVP mutations were absent from all of the 14 plasma samples. In addition samples were obtained one week after the dose and these will also be analysed for resistance mutations. It was speculated that due to NVP's long half-life and HIV 's low resistance threshold to the drug, it is possible that the K103N was selected in all women receiving this treatment, but that the resistant virus may have dropped below levels of detection on the genotype assay used to obtain results (PE/Biosystems Genotype).

    Although perhaps mathematically inevitable, it was concluded that the selection of the K103N was 'unexpected in the absence of prior or concurrent ZDV therapy, and suggests that mutations arising during antiretroviral therapy may be different for subtype-B vs. non-B HIV'.

    The clinical relevance for women in countries where antiretroviral therapy is widely available is clear - ie. that NVP should only be used in the setting of perinatal transmission prevention as part of a multi-drug combination.

    In settings where the drug is only available as monotherapy we were left with many questions and risk/ benefit dilemmas as to the clinical implications of these findings:

    In conversations afterwards, reactions to these trial results varied, from the alarmist - 'That is the single most disturbing piece of news to come out of this whole conference', from Dr Steven Deeks, and 'You'll end up with a whole population resistant to the only drugs that they are likely to get to help them' from Dr Karen Beckerman, to the more optimistic, 'Well you just need to support it with another drug', from activist Raffi Babakhanian. More data is urgently required.


    1. Guay LA, Musoke P, Fleming T, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNet 012 randomised trial. Lancet.1999;354;795-802

    2. Becker-Pergola G, Guay L, Mmiro F, et al. Selection of the K103N nevirapine resistance mutation in Ugandan women receiving NVP prophylaxis to prevent HIV-1 vertical transmission (HIVNet-006). Programs and abstracts of the 7th Conference on Retroviruses and Opportunistic Infections, January 30th-February 2nd, 2000; San Francisco, California. Abstract 658.

    Thymic maturation is impaired even in uninfected but perinatally exposed children

    T lymphocyte development is abnormal in HIV-infected children. A study presented by Dr Vignano from the University of Milan demonstrates that these abnormalities persist even in HIV-uninfected infants and children of HIV-infected mothers (seroreverters) who have been exposed in utero - due to HIV's ability to cross the placenta, or during labour.

    Phenotypic analyses were performed in PBMC of 61 seroreverters born to HIV-infected mothers. Of the cohort 19/61 were ZDV-exposed newborns and 41/62 were ZDV-unexposed older children, and results from both of these groups were compared to healthy HIV-unexposed children (11 newborn and 15 older). A group of 25 older HIV-infected children were also examined.

    In both the newborns and the older seroreverters the phenotypic results observed differed considerably to the results observed in the group of healthy children. These differences were as follows: 1] CD45+/62+ na•ve cells were reduced and CD45RA- cells wee augmented; 2] CD8+/38+ bright cells were augmented; 3] CD3+/4-/8-(DN) were augmented. In addition, in the group of seroreverted newborns: DN/44+ lymphocytes were augmented whereas CD8+/na•ve cells were reduced. Results obtained from the group of HIV-infected older children were generally similar to younger and older seroreverters but differed significantly from the results observed with healthy, uninfected children.

    A mechanism for these results was speculated and it was suggested that '...very long-term HIV had been managed by the immune system'. Even in the absence of overt infection, immune activation as suggested by an altered ratio of na•ve to memory cells may indicate that infection was immunologically aborted after exposure. The clinical implications of this data are unclear.

    Abstract 182 T Lymphocyte Maturation Abnormalities in Vertically HIV-Exposed Uninfected Children. A.Vigano, M.Saresella, E.Massironi, D.Bricalli, N.Principi, P.Ferrante, L.Dally and M.Clerici. Pediatrics IV, Virology and Immunology, Univ. Milano and ISS, Rome, Italy

    Return to Contents