Important note: Information in this article was accurate in May 2000. The state of the art may have changed since the publication date.| Trouble with Indinavir Troughs Can TDM Spot Poor Adherence? Will ATHENA Enlighten? Quality Control Woes Efflux-influx Flip-flops Failure's Fault Lines Table of Abstracts References and Notes |
| Can you all reconcile our current hope-based strategy of one dose for all with what must be our ultimate aim--to prevent the postantiretroviral era? --Concepta Merry, MB, PhD |
Does a one-dose-fits-all strategy work for antiretrovirals? Or, as Northwestern University's Concepta Merry, MB, PhD, asked a roomful of pharmacologists at a rain-swept North Sea resort, will that monolithic tactic so speed the pace of treatment failure that HIV will once again go unchallenged?
The alternative to one-dose-fits-all, therapeutic drug monitoring (TDM), got a two-day hearing at the March 30-31 First International Workshop on Clinical Pharmacology of HIV Therapy in Noordwijk, The Netherlands. And although most presentations addressing that issue positioned TDM as a fourth leg in HIV disease monitoring (along with tests for RNA, CD4, and resistance), more than one attendee saw TDM not as a fourth leg but as a fifth wheel, a gratuitous bit of pharmacologic filigree.
Arguments favoring TDM are phar-flung. One workshop organizer, the University of Liverpool's David Back, PhD, deftly summed them up in reviewing this contentious issue for the Journal of IAPAC in February.1 Those arguments can be boiled down to three:
So far, the most clinically persuasive work in this field involves protease inhibitors (PIs). Plasma levels of nucleosides (NRTIs) are meaningless because they don't correlate with intracellular levels. And since nukes must graduate to their triphosphate form inside cells, only intracellular triphosphate levels count. Although David Back showed the workshop his approach to a more dependable intracellular NRTI assay [abstract 1.1], he emphasized that the method's complexity precludes its clinical use. Nonnucleoside (NNRTI) levels can be measured in blood, but their concentrations are so often adequate that measuring them seems superfluous to some.
So PI plasma levels are the playing field for the TDM tug-of-war. But, as Charles Flexner, MD, argues in the interview in the May JIAPAC, that contest may already be over because drug development and treatment practice evolve so rapidly. Dual PIs have become the standard of care, Flexner notes, precisely because piggybacking protease drugs boosts treacherously low troughs and smoothes toxic peaks. Newer agents will have even comelier trough and peak profiles, says Flexner. Nonetheless, several Noordwijk reports suggested why TDM proponents see drug level assays as the next technology HIV docs must master.
Giorgio Gatti, MD (University of Genoa) laid the groundwork for the primary TDM rationale--that PI concentration correlates with antiviral effect--in a study of standard-dose indinavir given with two NRTIs [abstract 5.4]. Comparing area under the concentration-time curve (AUC) with six-month viral load changes in 35 persons, he found that indinavir AUC predicted virologic response. People with an AUC above 25 mg/Lh had a median 2.53-log six-month drop in viral load, whereas those with an AUC below 25 mg/Lh had a median 1.08-log drop in viremia (P < 0.01). Gatti proposed that the 25-mg mark may be a reasonable target for indinavir exposure but added that this recommendation should be confirmed in bigger studies. Indeed, lack of verified target concentrations is one impediment to clinical adoption of TDM.
David Burger, PharmD, PhD (University Medical Center St. Radboud, Nijmegen) thinks he may have pinned down an indinavir target concentration for kids--20 mg/Lh. Burger's multicenter study tracked indinavir concentrations in 41 children treated with one of three doses: 100, 150, or 200 mg/kg metabolic weight/day (100 mg/kg metabolic weight/day is equivalent to 1250 mg/m2/day) [abstract 5.3].
Shooting for an AUC of 20 mg/Lh, Burger found that the 150-mg dose hit the bull's eye most consistently. Among 11 children whose indinavir AUC lagged below 20 mg/Lh, six (55 percent) had a viral load below 500 copies/mL after six months of treatment. But all 11 children whose AUCs exceeded the 20-mg benchmark attained a sub-500 six-month viral load.
The study did turn up some problems, Burger noted. Six children with an average indinavir AUC of 40.6 mg/Lh suffered renal side effects. So he cautioned that clinicians must monitor closely for renal toxicity if indinavir's AUC tops 30 to 35 mg/Lh. Also, the 150-mg dose yielded the lowest indinavir troughs in the study, raising the risk of virologic breakthrough if a dose is missed. One should also note that Burger picked the 20 mg/Lh AUC target for kids because his experience suggested that's the best indinavir target in adults. So Burger's experience varies from Gatti's indinavir AUC recommendation for adults.
A principal US proponent of TDM, Courtney Fletcher, PharmD (University of Minnesota), offered six-month results of another indinavir study, this one, like Gatti's, in adults [abstract 6.2]. Besides adjusting indinavir concentrations, though, Fletcher also monitored levels of zidovudine (ZDV) and lamivudine (3TC), then boosted them if low. The Minnesotans aimed for an indinavir trough of 0.15 mg/L and steady-state intracellular triphosphate concentrations of 0.19 mg/L for ZDV and 0.44 mg/L for 3TC. They randomized 11 treatment-naive people to concentration-controlled (CC) therapy and 13 to standard of care treatment.
Among the 11 individuals in the CC arm, the median time to reach a viral load below 50 copies/mL measured 110 days, compared with 176 days in the standard of care arm (P = 0.05). Because studies have linked time to undetectability with more durable responses, that result suggests tinkering with antiretroviral doses may improve treatment. Indinavir appeared to be the weakest link in this combination insofar as target concentrations are concerned: Whereas nine of 11 in the CC arm achieved the target indinavir concentration, only three of 13 in the standard of care arm did so (P < 0.05). But four people in the standard treatment arm also failed to hit the ZDV target, compared with none in the CC arm (P < 0.05).
If one accepts that the quicker virologic response in the CC arm will translate into a clinical benefit, the question becomes how practical this strategy will prove clinically. Fletcher found no difference between arms in rates of toxicity or adherence (measured by pill count). But two of 11 people in the CC arm had to switch from twice- to thrice-daily 3TC, and six jumped from three-times-daily indinavir to four-times-daily dosing. Outside of a small well-conducted trial, few clinicians may be willing to impose such burdens on the people they treat. And since indinavir proved the prime culprit in missing the concentration target, many would argue that adding ritonavir and cutting the PI dosing to twice daily makes lots more sense than tracking blood levels and pushing indinavir doses to 600 or 800 mg every six hours. Finally, as noted, measuring intracellular nucleoside triphosphate levels remains a specialized skill.
Two groups studied TDM as a tool to improve adherence. P.W.H. Hugen, PharmD, and colleagues (St. Radboud, Nijmegen) devised a thoughtful scheme to spot wobbly adherence to indinavir, nelfinavir, or ritonavir/saquinavir [abstract 3.1]. First they constructed eight-hour pharmacokinetic (PK) curves based on 12 blood samples per person after observed ingestion of the PI with the correct food requirements. Then they set cutoff points at the extreme ranges of those curves and figured that values outside those cutoffs may signal poor compliance. They determined the cutoffs by calculating the ratios of the concentration at each time point against the median population value (Table 1).
Hugen reported results using a range from 5 percent of the median population concentration (P05 in Table 1) to 95 percent of the median (P95 in Table 1). Yes, yes, it sounds sort of complicated. But if someone (Hugen et al, perhaps) were to establish and verify what might be called a concentration confidence interval for every PI, much of the hard work would be done.
The Nijmegen team's results at least suggest that people with concentrations outside the 5- to 95-percent range are doing something wrong. First the treating physician or pharmacist determined that interacting comedications, compromised liver function, or severe diarrhea or vomiting would not account for low drug levels in study participants suspected of poor adherence. Next they collected random blood samples and measured PI concentrations. Finally they compared PI levels that fell outside the 5- to 95-percent limit in the random samples with those that fell outside the limits during observed treatment. Percentages outside the target range for indinavir, nelfinavir, ritonavir, and saquinavir during observed treatment were 3, 8, 5, and 4 percent, respectively. But in random samples from people suspected of poor adherence, respective percentages outside the target range soared to 31, 55, 46, and 46 percent.
Hugen concluded that this scheme will not identify all weak compliers. Because blood samples have to be collected randomly--for the sake of convenience and to prevent cheating--some iffy pill-takers may get lucky and take their meds just before sample collection. But Hugen proposed that concentrations outside the 5- to 95-percent range should send a red flag snapping above the heads of hit-or-miss adherers.
Some workshop attendees remained unswayed. Ruling out other reasons for low drug levels would be harder in practice than in a small trial, they said. And collecting and measuring random blood samples may be no more effective than simpler techniques to gauge adherence. Still, because good adherence remains a critical ingredient in treatment success, any reasonable attempt to uncover bad pill-taking habits deserves close attention.
Let's say the Nijmegen formula seems a tad too rococo to you. Then how about IPAM? That's the catchy acronym crafted by Richard Brundage, PharmD, in Courtney Fletcher's group at the University of Minnesota. It stands for integrated pharmacokinetic-adherence measure. Brundage used IPAM to track adherence by 50 children in Pediatric AIDS Clinical Trials Group (PACTG) protocol 382, an open-label study of efavirenz, nelfinavir, and at least one nucleoside [abstract 3.2]. Here's how it works.
First Brundage measured AUCs for efavirenz during study weeks two to six. Then, throughout the first year of study, he collected up to 12 blood samples to check efavirenz concentrations. His acceptable concentration range was a little narrower than the Nijmegen range, plus or minus 40 percent of the predicted concentration based on the week two to six samples. Brundage defined the fraction of all observations inside this range as that child's IPAM, scored from 0 to 1. Children below the 33rd percentile for all IPAMs got labeled "low IPAM," and those at or above the 33rd percentile won a "high IPAM" star.
The low-IPAM group had a median IPAM of 0.20, compared with 0.67 for the high-IPAM group. In the high-IPAM cadre, seven of 31 (23 percent) had a viral load rebound (>400 copies/mL after at least two consecutive viral loads <400 copies/mL, or a >0.75-log bounce from the nadir viral load), but 10 of 19 low-IPAMers (53 percent) endured a rebound (P = 0.037). The low-IPAM group had a significantly shorter time to first rebound (P = 0.008), and that difference from the high-IPAM group remained significant after statisticians controlled the results for baseline viral load (P = 0.022). In other words, IPAM predicted rebound independently of viral load.
Brundage suggested that IPAM may ferret out faulty adherence "and could allow interventions to minimize therapeutic failure." But this preliminary study is small, and Brundage was not certain how many blood levels would be needed to establish and track a person's IPAM in practice.
Still, even these early findings are interesting because they show that close to 40 percent of children in this study (19 of 50) had marginal efavirenz concentrations. That finding runs against the grain of other workshop studies that found consistently high NNRTI concentration in adults. For example, Kees Groen, PharmD (Kinesis, Breda, The Netherlands) measured NNRTI and PI concentrations in 770 samples and found subtherapeutic NNRTI values in only 3 percent [abstract 5.5]. In contrast, 16 percent of samples had subtherapeutic PI levels. In the randomized ATHENA study of TDM (see next section), the only nonnucleoside studied, nevirapine, yielded consistently higher concentrations than PIs. Perhaps children differ substantially from adults in adherence to NNRTIs or in their ability to attain effective plasma concentrations.
What clinicians would like to know about TDM is whether its clinical use will cut morbidity and mortality. Dutch investigators David Burger, PharmD, PhD (St. Radboud, Nijmegen) and Richard Hoetelmans, PharmD, PhD (Slotervaart Hospital, Amsterdam) get credit for tackling that question as part of a large trial, ATHENA [abstract 6.6]. But in reporting the earliest results, Burger cautioned that ATHENA may not impart her legendary practical wisdom on the question of TDM.
The big difficulty in mounting the ATHENA TDM study is that no one had ever tried such a thing before. So no one knew how often plasma levels should be gauged or in which groups of patients. No one knew what target concentrations to aim for. Burger and Hoetelmans decided to draw random samples and to shoot for a concentration between 0.75 and 2.0 of the population value for each agent. They began with an unselected group of 600 individuals, half of whom had just started their first antiretroviral regimen. Because all study participants had to fill out a questionnaire, Burger noted, the sample necessarily excludes immigrants not literate in Dutch.
Investigators randomize participants to the TDM arm, in which drug concentrations and advice are shared with treating clinicians, or a control arm, in which drug concentrations are measured but not shared. Clinicians in the TDM arm are encouraged to adjust doses that fall outside the target concentration range. By December 1999, 391 people had enrolled in the TDM study and Burger had 1828 drug levels on file.
So far drug levels have been lowest with saquinavir (41 percent of saquinavir samples below 0.75 of the population concentration), followed by a tight cluster of the other PIs: indinavir, 28 percent; ritonavir, 27 percent; and nelfinavir, 26 percent. Only 10 percent of nevirapine samples fell short of the 0.75 mark. Fewer drug levels exceeded the upper limit of twice the population concentration. Saquinavir once more led the pack with 11.5 percent of samples above the 2.0 rung, followed by ritonavir (9.9 percent), indinavir (5.9 percent), nelfinavir (5.0 percent), and, best again, nevirapine (3.5 percent).
Even though the ATHENA TDM study is in its earliest stages, Burger worried that it may prove nothing. Because the trial began before dual PI prescribing became common, final results may not reflect standard practice when the study ends. Another problem already encountered is that clinicians in the control arm may measure drug levels on the sly. Burger noted that a few control arm docs have called him begging to know the concentration for a particular person taking a failing regimen. (That very possibility helped torpedo a British trial of TDM planned by Liverpool's David Back. After 16 months of "fast track" consideration, Back added, Britain's Medical Research Council also complained that it didn't like the trial's name, OPIUM.)
Burger's list of potential ATHENA confounders didn't end there. He observed that it's difficult to measure whether clinicians in the TDM arm heed the advice that comes with the drug level. Indeed, the study already includes cases of treatment failure among individuals in the TDM group. Finally, Burger noted, more stringent (rather than random) sampling may have yielded more robust results. He concluded that no TDM trial will give quick answers. And even if the answer supports clinical use of TDM, a drug concentration will remain a solitary datum that must be considered along with viral load, CD4+ count, resistance, adherence, and, of course, clinical developments (Figure 1).
David Burger's lowering of expectations about ATHENA's TDM study did little to stoke enthusiasm for this treatment strategy. A few other reports at the Noordwijk workshop did even less to rally the troops. Two of those studies made it plain that much work remains before clinicians can feel confident that assay results closely match actual drug levels in a patient.
Rob Aarnoutse, PharmD (St. Radboud, Nijmegen) reported findings of an international quality control survey involving labs from nine European countries and one in North America [abstract 1.4]. His conclusions weren't cheery.
All labs got blinded samples that contained some mix of low, medium, and high concentrations of saquinavir, ritonavir, indinavir, and nelfinavir. Aarnoutse decided that any reported level not within 20 percent of the actual PI concentration would be called inaccurate. He had tabulated results from nine labs by the time of his presentation.
The median percentages of inaccurate reports were 22.9 percent for ritonavir, 21.4 percent for nelfinavir, 14.3 percent for saquinavir, and 9.9 percent for indinavir. Four labs came within the 20 percent target range with more than 75 percent of samples, three labs did so with 50 to 75 percent of samples, and two labs got the answer right with fewer than half the samples. Two labs reported lower limits "that were above trough levels for some PIs," Aarnoutse noted, a finding "indicating that their methods can not be used universally in clinical practice."
Results from a 27-lab French study proved no more reassuring. C. Palette, PharmD (Corentin Celton Hospital, Issy les Moulineaux) dispatched blinded samples spiked with varying dollops of saquinavir, ritonavir, indinavir, nelfinavir, amprenavir, nevirapine, and efavirenz [abstract 1.5]. Like the Dutch team, Palette and colleagues scored a concentration report as accurate if it came within 20 percent of the actual drug level. For some drugs, Palette had results for two separate samples.
The 27 labs did worst gauging levels of the nonnucleosides. For nevirapine, only 50 percent of reports fell within the 20 percent target, and for efavirenz a woeful 28.6 percent did so. Results were somewhat better with the PIs, but not great. Accurate calls for ritonavir came in at 53.3 percent, for saquinavir at 75 percent, for two nelfinavir samples at 73.3 and 26.7 percent, for two indinavir samples at 60.2 and 69.2 percent, and for two amprenavir samples at 63.6 and 81.8 percent. Palette concluded that standardization between labs is essential.
As gloomy as these two studies' results are, readers should remember that labs running resistance assays haven't done a whole lot better. Rob Schuurman, PhD (University of Utrecht) has reported lots of bad calls in two multilab studies, usually with samples that contain mixed mutant and wild-type virus.2 Those findings apparently have not dissuaded clinicians from routinely testing for resistance, if the patient or an insurer can foot the bill. But Aarnoutse and Palette have spotlighted a troubling deficiency that needs fixing before TDM can be considered a reliable tool.
And the two studies raise another question: Even if you decide that TDM may help with a particular person, where can you get it done? The answer appears to be almost nowhere, unless you've got a tie-in with a lab, or unless you're close enough to send samples to David Back's outfit in Liverpool. As noted at the University Web site,3 Liverpool technicians will reckon PI or NNRTI levels at £25 a pop. Kees Groen described an assay that can size up levels of all current PIs and NNRTIs in a single 100-µL plasma sample [abstract 5.5]. But the Kinesis-Virco collaborators developing this test are far from selling a kit. On the other hand, since no one has reported a dwindling of the profit motive in this world, one must assume that demonstration (or even strong hints) of clinically effective TDM will speed the commercialization of drug level assays remarkably.
Although several studies correlate PI plasma concentrations with virologic response,4-7 levels of PIs sloshing through veins don't necessarily reflect PI levels inside cells. That's not necessarily bad news, according to results of a study by Martina Hennessy, PharmD, and colleagues at St. James's Hospital in Dublin and the University of Liverpool [abstract 1.3]. They measured indinavir levels in plasma and peripheral blood mononuclear cells (PBMCs) from 10 people; eight had plasma viral loads below 50 copies/mL, and the other two had measurable but durably suppressed viremia.
Plasma and PBMC levels of indinavir correlated inconsistently in the 10 people studied. Sometimes high PBMC levels belied apparent deficits in plasma. Three people who had indinavir plasma trough levels below the accepted minimum effective concentration had just fine intracellular concentrations. And all three kept their viral loads under 50 copies/mL. Another person had plenty of indinavir in plasma, but scanty amounts in PBMCs. That person suffered a virologic rebound.
Although this is a tiny study, it raises the possibility that one can't religiously rely on plasma levels, even with PIs, to tell if a person is getting enough drug to the scene of the crime. And what would happen, one wonders, if a clinician got a low plasma reading on a patient who actually had dandy intracellular drug levels? Would that clinician push the dose of the first PI, add a second PI, switch PIs, deliver the adherence lecture?
Hennessy observed that science can explain some inconsistent plasma-versus-PBMC readings. Drug transporters such as P-glycoprotein (P-gp) stud the blood-brain barrier, dot both liver and intestines, and sit atop T lymphocytes. If P-gp pumps click as smoothly in humans as they do in cell cultures, it could explain why some cells stay PI-starved. With such possibilities in mind, Hennessy suggested her early results may mean intracellular indinavir levels, not just plasma levels, are needed to predict virologic response.
That would certainly further dim the prospects for clean-and-easy TDM. And another of David Back's colleagues, Saye Khoo, MB, larded on several layers of complication [abstract 4.1]. Besides P-gp pumping drugs from cells, he reminded attendees, one must contend with the multidrug-resistance protein (MRP) pump, the avidity of drug binding to 
1-acid glycoprotein (AGP) and other plasma proteins, and the efficiency of transporters that pipe drugs into cells. Unlike Hennessy, Khoo's early in vivo results suggest "little significant intracellular accumulation" of indinavir, even though less indinavir (50 percent) latched onto plasma proteins than saquinavir (76 percent) or ritonavir (81 percent). Saquinavir did best in settling down in cells, followed by ritonavir, then indinavir.
At this stage of efflux-influx research, it's tough to guess what all this may mean clinically. Perhaps the only sound conclusion one can draw right now is the one drawn by Khoo and Hennessy: In Khoo's measured words, "This may complicate therapeutic drug monitoring."
Anyone who needs further convincing could consult the results of two other efflux studies summarized in Noordwijk. R. Garraffo, PharmD (University Hospital, Nice) measured PI levels in cells that do not express MDR-1 (the gene that programs P-gp) and in cells that do [abstract 4.5]. PIs quickly pile into MDR-1-negative cells, led by nelfinavir, then, in decreasing order, by saquinavir, ritonavir, and (tied for last) amprenavir and indinavir. Even in these cells, though, concentrations of all the PIs except nelfinavir and saquinavir plummet to unmeasurable iotas in minutes. The story gets grimmer with cells that do express MDR-1. Only nelfinavir breached these cells as well as it did the MDR-1-negative cells. When Garraffo added ritonavir and a second PI to cell cultures, he found that ritonavir stanched MDR-1-mediated efflux in a dose-dependent manner.
Garraffo concluded that researchers must pin down not only the dose of ritonavir needed to inhibit CYP 3A4-dependent metabolism of other PIs, but also the dose needed to plug the P-gp pump. These cell line studies led him to suggest that 100 mg of ritonavir will not be enough to thwart P-gp-induced efflux. But Garraffo cautioned that this analysis may not mirror P-gp activity in T cells of infected people, because the MDR-1-expressing cells he used evict drug with the remorseless efficiency of a slum landlord.
Mice, at any rate, may credit that caveat, because high-dose ritonavir did little to gum up P-gp gears in cells from their brains or placentas [abstract 4.6]. Maarten Huisman, MD (Netherlands Cancer Institute, Amsterdam) gave a 50 mg/kg oral dose of ritonavir, followed by a radiolabeled 5 mg/kg oral dose of saquinavir, to P-gp-deficient mice and to mice with normal P-gp function. Even at this ritonavir dose, high for a mouse, P-gp still squelched saquinavir's bioavailability and limited its penetration into brain and fetus.
"It is therefore unlikely," Huisman concluded, "that coadministration of ritonavir with saquinavir to patients will substantially affect the contribution of P-gp to potential pharmacological sanctuaries such a brain, testis, and fetus."
At this point in the study of therapeutic drug monitoring, hurdles to its clinical use seem Himalayan. As Charles Flexner points out in the interview beginning on page 136, one precipitous obstacle is that target concentration ranges remain undefined. On top of that, no one knows which concentration value to measure--troughs, troughs plus peaks, area under the curve (AUC), or 50 percent or 90 percent inhibitory concentration (IC50 or IC90)? And when should blood levels be sampled--soon after a new regimen begins, regularly after that, at early hints of drug failure?
Of course, even Himalayan peaks can be scaled, and routinely are. But if clinicians today could order a drug level as easily as a pizza pie, or a genotype, what would they do with the results, wondered Jonathan Schapiro, MD (Tel Aviv University). Clearly, he said, simply bumping up a drug's dose may not be the right answer every time. Schapiro called current studies linking low blood levels with poor virologic responses "sterile," because they don't factor in the messy complications that will make interpreting TDM tricky in the clinic.
But one of Schapiro's mentors, Thomas Merigan, MD (Stanford University) encouraged colleagues to continue refining TDM trial strategies in hopes of yielding data showing that drug level monitoring is worthwhile and, just as important, showing "how to do it" in clinical practice. The finding of Stephen Piscitelli, PharmD, that St. John's wort can dangerously lower levels of indinavir8 [abstract 2.1] may have come to light sooner, Merigan suggested, if TDM were a routine clinical procedure. That very possibility arose at the workshop when workers at Slotervaart Hospital in Amsterdam charted a 19 percent dip in nevirapine levels after a person started taking St. John's wort [abstract 2.8, see Note 9].
In arguing for more diligent consideration of TDM, Concepta Merry agreed that still-limited understanding of antiretroviral pharmacokinetics and pharmacodynamics makes it "really difficult to determine the optimal time of sampling for TDM and how to interpret these results" [abstract 6.1]. But, she argued, similar limitations have not discouraged a warm embrace of clinical resistance testing. Merry pointed to antiretroviral guidelines published earlier this year,10 noting that the International AIDS Society-USA Panel recommends resistance testing despite several reservations.
"They're concerned about the cost," Merry observed. "They're concerned about quality assurance. They're concerned about optimal use and interpretation of results. Now, does that sound familiar?" As a clinician, she confessed, "I find this a little perplexing, because I maintain that we're maintaining a double standard in HIV therapeutics today."
That double standard in considering resistance testing versus TDM, Merry said, creates a familiar scenario. "We start a patient on a drug. We monitor their CD4+ count. We monitor their viral load. And when all else fails we rush in with resistance testing." A more reasonable approach, she proposed, may be to start treatment then check drug concentrations once or twice to make sure everything is on track, and to adjust doses if everything's not (Figure 1).
Merry concluded with the story of a patient who had just decided to spend $1400 of his own money for resistance testing because his treatment appeared to be faltering. She regretted not checking his plasma levels early in the course of treatment. "You know," she lamented, "for $1400 I could have done a lot of drug assays for this man."
But what are the chances that low drug levels explain failing treatment in this man, or in anyone taking a regimen faithfully? The chances are slim indeed, contended John Gerber, MD (University of Colorado). Today's antiretrovirals do just fine getting where they have to go, he argued. Anyone who needs convincing, he said, should refer to Margaret Fischl's study of directly observed anti-HIV therapy versus standard care.11 Forty-two of 42 prisoners watched while taking their drugs had viral loads below 400 copies/mL 48 weeks after starting their first PI or NNRTI regimen, compared with 68 percent of those in the standard-of-care arm (P < 0.01). Better adherence, not higher drug levels, should be the prime target for monitoring, Gerber proposed. The fault lies less in the drugs, perhaps, than in their taking.
Therapeutic drug monitoring, although the most contentious issue pondered at the Noordwijk pharmacology workshop, accounted for only 11 of 47 abstracts. Several studies focused on critical antiretroviral drug interactions, usually involving ritonavir combined with another protease inhibitor, with or without a nonnucleoside. This table outlines six such studies.
| Presenting author, site(s), type of study [abstract number] | Drugs (doses), study participants (n) | Methods | Results | Conclusions or comments |
|---|---|---|---|---|
| R. Aarnoutse, Nijmegen and Amsterdam, PKs in healthy volunteers [2.3] | IDV/RTV/EFV: IDV (800 mg bid) + RTV (100 mg bid), then EFV (600 mg qd); n = 18; 14 completed study | IDV + RTV with food for 14 d; steady-state PKs for IDV and RTV for 12 h on d 15; add EFV d 15 to 29; PKs repeated on d 29 | Adding EFV TMDN IDV AUC 19% (P=0.001), TMDN IDV Cmin 48% (P=0.001), and TMDN IDV Cmax 13% (P=0.024); but d 29 IDV Cmin >0.10 mg/L therapeutic threshold value; EFV PKs comparable to historical data | No need to adjust IDV dose when adding EFV to IDV/RTV; IDV/RTV/EFV may be simple, NRTI-sparing combination; but some question whether 0.10 mg/L threshold for IDV is a reliable indicator of efficacy |
| J. Mallolas, Barcelona, PKs in ART-naive adults [2.14] | IDV/RTV/CBV: IDV (800 mg bid) + RTV liquid* (100 mg bid) + CBV (1 pill bid) for 4 w; then IDV (1000 mg qd) + RTV (100 mg qd) + CBV; then IDV (800 mg qd) + RTV (200 mg qd) + CBV#; n = 27, 11 completed study; mean BL CD4+ count 107 cells/µL | IDV/RTV liquid* (800/100 mg bid) + CBV for 4 wk; then IDV/RTV (1000/100 mg qd) + CBV; then IDV/RTV (800/200 mg qd) + CBV#; PKs for each dosage level | Three D/C IDV/RTV within 4 wk because of clinical progression; 11 D/C within 4 wk because of RTV-related side effects; mean IDV Cmin 1.4 µg/mL with IDV/RTV 800/100 bid (n = 16), 0.18 µg/mL with 1000/100 qd (n = 9), 0.2 µg/mL with 800/200 qd (group A# n = 9), and 0.4 µg/mL with 800/200 qd (group B† n = 7); VL <5 copies/mL | 15 of 27 (56%) ART-naive adults with advanced disease respond to qd IDV/RTV plus CBV bid; high early D/C rate for progression and side effects, 11 of 27 (41%) in 15/16 who completed 32 wk |
| D. Burger, Thailand, PK substudy of IDV/RTV trial [2.5] | IDV/RTV + ZDV/3TC: IDV (800 mg bid) + RTV (100 mg bid); n = 5 Thai women, 4 Thai men | 4 wk after starting IDV/RTV, 8- or 12-h PKs in 20 Thai men and women; results available for first 9 | IDV AUC and Cmax higher in 9 Thais than in Caucasians (historical data), but body weight does not correlate with IDV AUC, Cmax, or Cmin; nonsignificant trend to higher AUC and Cmax in Thai women than Thai men | IDV exposure appears not to depend on body weight in Thai adults but may depend on gender or some other anthropometric measure; data needed from more patients |
| A. Hsu, USA, 3 PK studies of LPV/RTV + other PIs [2.4] | LPV/RTV + other PIs: LPV/RTV (400/100 mg bid) + SQV (800 mg, n = 11) or IDV (600 mg, n = 11) or NFV (750 mg, n = 5) or APV (450 or 750 mg, n = 10) | After 10 d of LPV/RTV, add single dose of SQV, or IDV, or NFV; or, after 17 d of LPV/RTV, add APV for 5 d | SQV and IDV: AUC, Cmax, Cmin modestly affected by LPV/RTV NFV: LPV/RTV TMDN AUC 23%, TMDN Cmin42% APV: LPV/RTV TMDN AUC 15%, TMDN Cmax 11%, TMDN Cmin19% | After multiple doses, IDV may TMUP LPV trough and NFV may TMDN LPV trough; APV and SQV not expected to have significant effect on LPV; suggested doses when combining PIs with LPV/RTV (400/100 mg bid): SQV, 800 mg bid; IDV, 600 mg bid; NFV, 750 mg bid; APV, 750 mg bid |
| C. Lamotte, Paris, APV Cmin in patients taking APV + low-dose RTV ± EFV [2.7] | APV/RTV ± EFV: EFV (600 mg qd) + APV/RTV (450/100 mg bid [n = 15] or 600/100 mg bid [n = 12] or 600/200 mg bid [n = 12]) or APV (1200 mg bid [n = 39]) without RTV; or each APV/RTV combination without EFV (n = 59) | APV Cmin determined at steady state >15 d after beginning regimen | Cmin <100 ng/mL (2.5 x IC90) in 23 of 39 (59%) taking APV (1200 mg bid) + EFV (600 mg qd) without RTV; Cmins significantly higher (P < 0.001) in all RTV-containing arms, regardless of use of EFV | Low-dose RTV TMUP Cminof APV even when given with EFV; tolerance of APV/RTV/EFV regimen "acceptable"; further study needed to assess efficacy |
| O. Degen, Hamburg and Berlin, APV PKs in patients taking APV + low- dose RTV ± EFV or NVP [2.12] | APV/RTV ± EFV or NVP: APV (450 mg bid) + RTV (200 mg bid) + EFV (600 mg qd [n = 7]) or NVP (200 mg qd [n = 2]) or with no NNRTI (n = 3) | After 4 to 7 wk of APV/RTV + >2 NRTIs ± 1 NNRTI, serial blood samples collected over 8 h, with the first sample collected 12 h after evening dose and just before next dose | AUC, Cmax, and Cminof APV high and stable in 12 of 12, regardless of use of EFV or NFV; mean APV Cmin= 3200 ng/mL with EFV, 1368 ng/mL with NVP, 1387 ng/mL without an NNRTI | Low-dose RTV TMUP concentrations of APV given with or without EFV or NVP; efficacy of combinations should be determined in further study |
| APV = amprenavir; ART = antiretroviral; AUC = area under the concentration-time curve; bid = twice daily; BL = baseline; CBV = Combivir (ZDV + 3TC); Cmax = maximum (peak) concentration; Cmin= minimum (trough) concentration; D/C = discontinue; EFV = efavirenz; IC90 = concentration needed to inhibit viral replication by 90%; IDV = indinavir; LPV = lopinavir (ABT-378); NFV = nelfinavir; NNRTI = nonnucleoside reverse transcriptase inhibitor; NRTI = nucleoside reverse transcriptase inhibitor; NVP = nevirapine; PKs = pharmacokinetics; qd = once daily; RTV = ritonavir; SQV = saquinavir; VL = plasma viral load.
*RTV liquid used throughout study. |
||||
Mark Mascolini writes about HIV infection (mailmark@ptd.net).
1. Back DJ, Gibbons SE, Khoo SH, et al. Therapeutic drug monitoring of antiretrovirals: ready for the clinic? J IAPAC 2000;6:34-37.
2. Mascolini M. Drug-resistant HIV: science concocts counterattacks. J IAPAC 1999;5(9):38.
3. The University of Liverpool offers therapeutic drug monitoring (TDM) for saquinavir, ritonavir, indinavir, nelfinavir, nevirapine, delavirdine, efavirenz, sildenafil, and methadone. The assay for amprenavir is being validated. The cost is £25 per drug, and the turnaround time is two weeks. For details, go to: http://www.hiv-druginteractions.org/tdm/ tdm_ind.htm (accessed April 23, 2000) or contact Sara Gibbons at 44-151-794-5553 (phone), 44-151-794-5540 (fax), or hivgroup@liv.ac.uk.
4. Hoetelmans RMW, Reijers MHE, Weverling GJ, et al. The effect of plasma drug concentrations on HIV-1 clearance rate during quadruple therapy. AIDS 1998 Jul 30;12(11):F111-5.
5. Gieschke R, Fotteler B, Buss N, Steiner J-L. Relationship between exposure to saquinavir monotherapy and antiviral response in HIV-positive patients. Clin Pharmacokinet 1999 Jul;37(1):75-86.
6. Descamps D, Flandre P, Calvez V, et al. Mechanisms of virologic failure in previously untreated HIV-infected patients from a trial of induction-maintenance therapy. JAMA 2000 Jan 12;283(2):205-11. Available at: http://jama.ama-assn.org/issues/v283n2/full/joc90679.html. Accessed April 24, 2000.
7. Burger DM, Hoetelmans RMW, Hugen PWH, et al. Low plasma concentrations of indinavir are related to virological treatment failure in HIV-1 infected patients on indinavir-containing triple therapy. Antivir Ther 1998;3(4):215-20.
8. Piscitelli SC, Burstein AH, Chaitt D, et al. Indinavir concentrations and St John's wort. Lancet 2000 Feb 12;355(9203):547-8.
9. A case report by M.M.R. de Maat and colleagues at Slotervaart Hospital and the University of Utrecht [abstract 2.8] told the story of a man being treated with ZDV, 3TC, and nevirapine whose plasma concentrations of nevirapine were checked every three months. On the first two measures, nevirapine levels were within the reported population concentration range. A third measure, after the man began taking St. John's wort, logged a 19 percent drop in nevirapine level. "The concomitant use of St. John's wort could explain the low exposure to nevirapine in this patient," de Maat concluded, "as nevirapine is extensively metabolized via the CYP450 system." The investigators suggested further study.
10. Carpenter CC, Cooper DA, Fischl MA, et al. Antiretroviral therapy in adults: updated recommendations of the International AIDS Society-USA Panel. JAMA 2000;283:381-390. Available at: http://jama.ama-assn.org/issues/v283n3/full/jst90023.html. Accessed April 24, 2000.
11. Fischl M, Rodriguez A, Scerpella E, et al. Impact of directly observed therapy on outcome in HIV clinical trials. Presented at: 7th Conference on Retroviruses and Opportunistic Infections; January 30-February 2, 2000; San Francisco. Abstract 71.
000510
IA000500
Copyright © 2000 - International Association of Physicians in AIDS Care. All rights reserved. http://www.iapac.org
ÆGiS is made possible through unrestricted grants from Roxane Laboratories, Inc., iMetrikus, Inc., the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2000. This material is designed to support, not replace, the relationship that exists between you and your doctor.
ÆGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.
Copyright ©1990, 2000. ÆGiS & the Sisters of Saint Elizabeth of Hungary. All materials appearing on ÆGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of ÆGIS and the Sisters of Saint. Elizabeth of Hungary, or the party credited as the provider of the content.