Stepped-Dose Versus Full-Dose Efavirenz for HIV Infection and Neuropsychiatric Adverse Events A Randomized Trial

BACKGROUND
More than 50% of patients who start efavirenz treatment develop limiting neuropsychiatric adverse events (NPAEs).


OBJECTIVE
To assess whether stepwise dosing of efavirenz decreases the incidence and severity of NPAEs while maintaining virologic efficacy.


DESIGN
Randomized, double-blind, controlled trial.


SETTING
7 HIV clinics in Spain.


PATIENTS
114 HIV-infected patients eligible for efavirenz treatment plus 2 nucleoside or nucleotide reverse transcriptase inhibitors.


INTERVENTION
Random assignment (by computer-generated sequence) to receive efavirenz, 200 mg/d on days 1 through 6, 400 mg/d on days 7 through 13, and 600 mg/d on day 14 and after, or efavirenz, 600 mg/d, from day 1. Both groups received 2 nucleoside or nucleotide reverse transcriptase inhibitors chosen by the patient's physician.


MEASUREMENTS
Neuropsychiatric symptoms and sleep quality were assessed by questionnaires at 0, 7, 14, and 30 days. The primary outcome was efavirenz-related NPAEs during the first 2 weeks, and the secondary outcome was plasma HIV RNA level at 24 weeks.


RESULTS
Compared with the stepped-dose group, the full-dose group had higher incidence and severity of dizziness (66.0% vs. 32.8%; P = 0.001), hangover (45.8% vs. 20.7%; P = 0.008), impaired concentration (22.9% vs. 8.9%; P = 0.038), and hallucinations (6.1% vs. 0%; P = 0.056) during the first week. From week 2, the incidence of efavirenz-related NPAEs was similar in both groups, although the severity was greater in the full-dose group. Virologic and immunologic efficacy seemed similar in both groups.


LIMITATIONS
The sample size was calculated on the basis of a high absolute difference in rates of efavirenz-related NPAEs between the groups. A lower absolute difference and a larger sample size could have made the differences between groups reach statistical significance beyond the first week. In addition, the sample size does not allow confirmation of similar efficacy between treatment groups.


CONCLUSION
Stepwise dose escalation of efavirenz over 2 weeks reduces the incidence and intensity of efavirenz-related NPAEs while maintaining efficacy.


PRIMARY FUNDING SOURCE
Consejería de Salud, Junta de Andalucía, Spain.

of treatment decreases the incidence and severity of efavirenz-related NPAEs compared with the standard full dose. Participants were enrolled at 7 HIV clinics in Andalusia, Spain, from April 2006 to January 2008. The study was approved by the Regional Ethics Committee for Clinical Research of the Community of Andalusia and by the Spanish Agency for Medicines and Healthcare Products and was conducted according to the principles in the Declaration of Helsinki.

Participants
We recruited participants through the investigators' clinical practices. Patients with HIV-1 infection were eligible if they were older than 18 years and were scheduled to receive an efavirenz-containing treatment plus 2 nucleoside or nucleotide reverse transcriptase inhibitors. Exclusion criteria were pregnancy, previous antiretroviral treatments or genotypic resistance test results suggesting nonnucleoside reverse transcriptase inhibitor (NNRTI) resistance, presence of major psychiatric disorders, initiation of psychotropic drug treatment in the 4 weeks before inclusion, methadone treatment, concomitant use of drugs that have potential adverse interactions with efavirenz pharmacokinetics, and cirrhosis with clinical or analytic data of liver failure. All patients gave dated and written informed consent.

Randomization and Intervention
The treatment packages were manufactured by the pharmacy department at Hospital Universitario Virgen del Rocío, Seville, Spain, where randomization was performed by using a computer-generated random number sequence. Randomization was stratified according to center in blocks of 20, so that each block comprised 10 patients randomly assigned to efavirenz at either a stepped dose (200 mg/d from day 1 to 6, 400 mg/d from day 7 to 13, and 600 mg/d from day 14 and after) or the usual full dose (standard dose of 600 mg/d from the first day). Each patient was assigned the next sequential number for the particular center. Treating physicians did not know the participants' group assignments until the trial was completed.
Both treatment groups received 2 nucleoside or nucleotide reverse transcriptase inhibitors, chosen by the treating physician on the basis of the patient's clinical history or previous genotypic resistance test results, plus 3 identical capsules containing either 200 mg/d of efavirenz or placebo, taken at night, during the first 13 days of treatment. From day 14 and after, patients received the efavirenz standard tablet of 600 mg/d at night and were followed until week 24.

Measurements, Outcomes, and Follow-up Procedures
We collected information on demographic characteristics and clinical data regarding HIV history, as well as current and previous antiretroviral regimens. After randomization, patients were assessed at baseline and weeks 1, 2, 4, 12, and 24. At each visit and after an overnight fast, clinical data and blood samples were collected for determination of plasma HIV-1 RNA (Amplicor HIV Monitor, Roche Diagnostic Systems, Basel, Switzerland; lower limit of detection, 50 copies/mL); CD4 ϩ count (standard flow cytometry); complete blood count; glucose measurement; lipid profile; and tests of liver, renal, and pancreatic function.
The primary outcome was the frequency and intensity of efavirenz-related NPAEs during the first 2 weeks of treatment. For that purpose, patients were interviewed by means of 2 questionnaires at days 0, 7, 14, and 30. The first questionnaire included 11 questions on common efavirenz-related NPAEs (dizziness, feeling of drunkenness or hangover, headache, impaired concentration, mood disorders, anxiety, and depression) graded from 0 to 6 (0 ϭ no adverse events, 2 ϭ mild adverse events [without alterations in normal daily activities], 4 ϭ moderate adverse events [causing difficulties at work or in daily activities but not disrupting them entirely], and 6 ϭ severe adverse events [having to stop doing some tasks at work or in daily activities]). Hallucinations and disorientation were evaluated as qualitative variables and always considered severe. The second questionnaire was a 13-item instrument validated by Oviedo University, Oviedo, Spain (20), that measured subjective sleep quality, somnolence, insomnia, and nightmares. The score assigned to each item in the questionnaires at weeks 1, 2, and 4 after treatment was initiated was the result of subtracting the base score from the score obtained every week. At each time point of the analysis of efavirenz-related NPAEs (days 0, 7, 14, and 30), we analyzed efavirenz-related NPAEs only for patients who were receiving treatment.
The secondary outcome was virologic efficacy analyzed by intention-to-treat, with treatment failure considered to

Context
Efavirenz is a powerful component of effective highly active antiretroviral therapy. Unfortunately, about half of patients develop disturbing neuropsychiatric side effects. Although side effects subside for many patients as they continue to take the drug, other patients have to stop taking efavirenz.

Contribution
This randomized trial compared full-dose efavirenz with stepped-dose efavirenz during the first 2 weeks of treatment and found fewer side effects during the first week in patients receiving stepped-dose therapy. After 24 weeks, HIV RNA levels and CD4 ϩ cell counts were similar in both groups.

Caution
The study had too few patients to detect small differences in effectiveness.

-The Editors
Article Efavirenz for HIV Infection and Neuropsychiatric Adverse Events be treatment interruption for any reason or virologic failure, defined as inability to suppress plasma HIV RNA to less than 200 copies/mL after 24 weeks of treatment or a confirmed viral load greater than 200 copies/mL in patients who had previously achieved viral suppression or had an undetectable viral load at inclusion. We evaluated the HIV genotypic pattern of resistance by sequencing the HIV protease and retrotranscriptase (1-335 amino acids) genes by using the ViroSeq HIV-1 Genotyping system (Celera Diagnostics, Alameda, California) whenever viral loads allowed. We defined mutations conferring resistance to antiretroviral drugs according to the 2006 International AIDS Society. We considered patients who missed 2 consecutive scheduled visits as lost to follow-up. We evaluated adherence by personal interview at each follow-up visit.
Blood samples for measurement of plasma efavirenz were obtained 12 hours after efavirenz intake (EFV-C 12 ) on days 7, 14, and 28 from patients at Hospital Universitario Virgen del Rocío. Plasma samples were stored frozen at Ϫ80°C for determination of efavirenz levels by highperformance liquid chromatographic assay according to a validated method (21).

Statistical Analysis
We designed the study to have a statistical power of 80% (with a 1-sided ␣ level of 0.05) to detect an absolute difference of 50% in the rates of efavirenz-related NPAEs (50% in the standard full-dose group vs. 25% in the stepped-dose group) at a 5% level of significance. A withdrawal rate of 15% was estimated. Therefore, we expected to include a total of 106 patients. We analyzed differences in the incidence and severity of adverse events, virologic failures, and absolute CD4 ϩ counts in each study group by using the chi-square, t, and Mann-Whitney-Wilcoxon tests as appropriate.

Role of the Funding Source
This investigator-initiated study was supported by the Consejería de Salud, Junta de Andalucía, Spain. The fund- ing source had no role in the design, conduct, or reporting of the study or in the decision to submit the manuscript for publication.

RESULTS
We enrolled 114 participants: 60 in the stepped-dose group and 54 in the full-dose group. We excluded 2 patients in the stepped-dose group and 4 patients in the fulldose group because of protocol violations, wrong doses, or concomitant use of rifampin. Therefore, we analyzed results from 108 patients. Figure 1 shows the flow of study participants. Baseline characteristics were similar in both groups ( Table 1). Genotypic resistance test results were available before randomization in 39 patients (18 in the stepped-dose group and 21 in the full-dose group), none of whom had NNRTI resistance-associated mutations. No recent genotypic resistance test results were available in the remaining patients for several reasons: a very low or undetectable viral load (n ϭ 15); long treatment interruption before inclusion, which meant that the test results were not expected to add relevant data (n ϭ 14); or test not requested (n ϭ 40). Nine patients (4 in the stepped-dose group and 5 in the full-dose group) were lost to follow-up between weeks 4 and 20 of treatment; those patients had had an undetectable viral load in the previous evaluation. Twelve patients discontinued the study because of adverse events: 5 (8.6%) in the stepped-dose group and 7 (14%) in the full-dose group (P ϭ 0.28) (Figure 1). Table 2 shows the incidence and severity of each efavirenz-related NPAE and sleep disorders in both groups at weeks 1, 2, and 4. During the first week of treatment, 60 of 108 patients (55.5%) developed efavirenz-related NPAEs; these were more frequent in the full-dose group (66%) than in the stepped-dose group (46.5%) (P ϭ 0.040). The incidence of dizziness (66.0% vs. 32.8%; P ϭ 0.001), feeling of drunkenness or hangover (45.8% vs. 20.7%; P ϭ 0.008), impaired concentration (22.9% vs. 8.9%; P ϭ 0.038), and hallucinations (6.1% vs. 0%; P ϭ 0.056) was higher and more severe in the full-dose group than in the stepped-dose group. Throughout the second week, the incidence of efavirenz-related NPAEs was similar in both groups (57.8% vs. 48.9%), although more severe NPAEs occurred in the full-dose group. Hallucinations occurred in 4 patients in the full-dose group during the first 2 weeks and in 1 patient in the stepped-dose group when the efavirenz dose was increased to 600 mg/d. No differences were observed between the 2 groups regarding other efavirenz-related NPAEs, although a trend toward increased incidence of sleep disorders and nightmares was observed in the full-dose group. At week 4, 51 patients (51.5%) still reported some efavirenz-related NPAEs. Throughout the first 2 weeks, only 1 patient from the full-dose group discontinued efavirenz treatment because of NPAEs, and 6 other patients discontinued later (weeks 5 to 20). Overall, the withdrawal rate due to efavirenzrelated NPAEs was 6.5%.

Immunovirologic Efficacy
No apparent differences were observed between the 2 groups regarding immunovirologic efficacy (Figure 2). The percentages of patients receiving treatment and with undetectable viral load at weeks 0, 1, 2, 4, 12, and 24 in the stepped-dose and full-dose groups were 12.1% and 10% (P ϭ 0.49), 14.8% and 14.6% (P ϭ 0.60), 26.9% and 17% (P ϭ 0.173), 32.1% and 27.7% (P ϭ 0.40), 71.2% and 59% (P ϭ 0.161), and 95.7% and 86.5% (P ϭ 0.137), respectively. Among patients with a baseline viral load greater than 100 000 copies/mL (23 in the stepped-dose group and 20 in the full-dose group), these percentages at weeks 4, 12, and 24 were 5.6% and Six cases of virologic failure were recorded: 2 in the stepped-dose group and 4 in the full-dose group. Among these, treatment was considered to have failed because the viral load was greater than 200 copies/mL (219 to 319 copies/mL) at week 24, although further testing yielded undetectable levels. In contrast, 1 patient in each group had a rebounding viral load at 12 weeks, with genotypic resistance test results showing mutations in M184V, L100I, K103N, and M230L (in 1 patient) and K103N (in the other patient).
Five patients self-reported in at least 1 study visit that they neglected to take efavirenz doses. The 2 patients with virologic failure and NNRTI mutations were among them.

Article
Efavirenz for HIV Infection and Neuropsychiatric Adverse Events www.annals.org We found a significant association between virologic failure and incomplete adherence (odds ratio, 33 [Mantel-Haenszel test CI, 3.5 to 236]; P ϭ 0.001).

Efavirenz Plasma Levels
Samples for EFV-C 12 measurement were available from 63 patients (33 in the stepped-dose group and 30 in the full-dose group; 56, 52, and 54 patients on treatment days 7, 14, and 28, respectively). The baseline characteristics of the patients in the stepped-dose and full-dose groups were similar in terms of sex (men, 78.8% vs. 73.3%; P ϭ 0.39), age (mean, 37.7 years [SD, 12.4] vs. 41 years [SD, 9.6]; P ϭ 0.20), body weight (mean, 69.9 kg [SD, 13.5] vs. 68.4 kg [SD, 14.7]; P ϭ 0.38), and prevalence of chronic viral hepatitis (21.2% vs. 26.7%; P ϭ 0.53). The interindividual variability in EFV-C 12 levels was large in both groups. As expected, EFV-C 12 levels were lower during the first 2 weeks in the stepped-dose group (P ϭ 0.047) than in the full-dose group (P ϭ 0.023), whereas no differences were observed at day 30 ( Table 3). A trend toward a higher EFV-C 12 levels at day 7 was observed in patients reporting dizziness than in those who did not, but the trend did not reach statistical significance. Nonsignificant differences were observed in EFV-C 12 levels regarding efavirenz-related NPAEs at any time point, and regarding sex, weight, or presence of chronic hepatitis.

DISCUSSION
In this randomized, multicenter, double-blind clinical trial, we found that the incidence of efavirenz-related NPAEs in the full-dose group was similar to that reported in previous studies (22)(23)(24)(25). However, both the incidence and severity of efavirenz-related NPAEs in the steppeddose group were lower throughout the first 2 weeks of treatment, although the differences observed were statistically significant only for the first week. Likewise, during the first 2 weeks, 6 patients (10.3%) versus 10 patients (20%), respectively, experienced severe efavirenz-related NPAEs in the stepped-dose and full-dose groups. Although this represents a 2-fold greater incidence in the full-dose group than in the stepped-dose group, the difference did not reach statistical significance. Hallucinations were among the more severe efavirenz-related NPAEs. They occurred in 4 patients in the full-dose group only; no patient from the stepped-dose group reported this event throughout the first 2 weeks. From the third week on, 1 more patient in the stepped-dose group reported hallucinations, and hallucinations recurred in 1 patient in the full-dose group. The incidence and severity of other efavirenzrelated NPAEs were similar in both groups, which was expected because all patients were taking the same doses.
The virologic efficacy of both regimens seems similar, with 1 case of virologic failure associated with NNRTIresistant mutations in each group and 4 patients (1 in the stepped-dose group and 3 in the full-dose group) with very low, but still detectable, viral loads at week 24, although further determination showed undetectable plasma HIV RNA and 2 of the patients had very high viral loads (Ͼ7 log 10 copies/mL) at baseline. Likewise, the increases in CD4 ϩ counts were similar in both groups.
To our knowledge, this is the first randomized clinical trial directly comparing the incidence of efavirenz-related NPAEs of efavirenz treatment given in stepwise doses with that of standard, full-dose efavirenz treatment given from the first day. In the efavirenz phase 2 dose-finding study (3), a similar 16 weeks of virologic efficacy was observed with efavirenz doses ranging from 200 to 600 mg/d, whereas the incidence of efavirenz-related NPAEs increased with efavirenz dose. Also, in a pilot study that included 41 patients (26) and had a stepped efavirenz administration  Article Efavirenz for HIV Infection and Neuropsychiatric Adverse Events schedule similar to ours, a better tolerance profile was observed than that reported in the literature. Although the exact mechanism by which efavirenz causes NPAEs is not well known yet, the incidence of NPAEs is higher during the first days of treatment and their intensity is more severe in the hours immediately after drug intake, gradually decreasing and finally diminishing by weeks 4 to 6 in most patients (3,4,22,23). In contrast, efavirenz is metabolized by the liver P450 enzymatic system, mainly by CYP2B6 and CYP3A4 isoenzymes, whose activity is induced by itself, thus leading to increased plasma clearance of the drug and lower plasma levels after an average of 7 days of treatment (3). Taking this time profile into account, the incidence of efavirenz-related NPAEs, especially during the first days or weeks, could be related to the maximum plasma concentration (C max ) of efavirenz, which occurs between 2 and 6 hours after taking the drug (27,28). Therefore, administration of efavirenz in a stepwise schedule might alleviate NPAEs.
As expected, EFV-C 12 levels were statistically significantly lower in the stepped-dose group than in the fulldose group during the first 2 weeks and similar at week 4 when all patients were taking the same efavirenz dose. The absence of relationships between EFV-C 12 concentrations and the incidence or severity of the efavirenz-related NPAEs was not surprising. Various studies (7)(8)(9)(11)(12)(13)(14)(15) have not reliably demonstrated a relationship between efavirenz pharmacokinetics and the occurrence of NPAEs. Regardless of individual susceptibility to efavirenz-related NPAEs, several factors could justify the relationship between efavirenz pharmacokinetics and NPAEs. First, efavirenz is usually taken at bedtime to ameliorate NPAEs, so obtaining the efavirenz C max (the responsible factor according to our hypothesis) or a complete 24-hour profile is difficult. Therefore, research studies on this issue were based on population pharmacokinetics; estimations of concentrations based on a theoretical average elimination halflife (although elimination half-life carries a high interindividual variability); or, as in our study, samples obtained at different times (usually 8 to 20 hours after taking efavirenz). However, neither efavirenz C max nor other pharmacokinetic variables can be precisely estimated from samples obtained 8 or more hours after drug intake (28). Second, the correlation between efavirenz plasma levels and the occurrence of efavirenz-related NPAEs after prolonged drug intake may be inherently erroneous, because NPAEs are more frequent and severe at the beginning of therapy. Finally, efavirenz-related NPAEs could probably be more closely related to the drug levels in the central nervous system than to plasma levels, and it is currently unknown whether a relationship between drug levels in both systems actually exists.
Because efavirenz-related NPAEs are subjective in nature, a strength of our study is its randomized, doubleblind design. Most previous studies of efavirenz-related NPAEs were limited to observational or open-label clinical trials. Likewise, exclusion of patients who were receiving methadone treatment was essential, because efavirenz increases metabolism of methadone and may precipitate symptoms of withdrawal that could be mistakenly attributed to efavirenz (29).
The sample size might also be a limitation of our study. A very high absolute difference in the rates of efavirenz-related NPAEs between the study groups (50%) was chosen when calculating the sample size. If we had considered a lower absolute difference and, therefore, had a larger sample, the observed differences may have reached statistical significance beyond the first week as well. Likewise, the sample size does not have enough statistical power for us to draw definitive conclusions about similar virologic efficacy in both treatment groups. A much larger clinical noninferiority trial would be required to appropriately evaluate this issue.
In summary, stepped-dose administration of efavirenz over 2 weeks significantly decreases the incidence and severity of NPAEs while apparently maintaining the same efficacy as the standard schedule.