Share this article on:

Efavirenz Versus Nevirapine in Current Clinical Practice: A Prospective, Open-Label Observational Study

Manfredi, Roberto MD; Calza, Leonardo MD; Chiodo, Francesco MD

JAIDS Journal of Acquired Immune Deficiency Syndromes: April 15th, 2004 - Volume 35 - Issue 5 - p 492-502
Clinical Science

Abstract: An open-label, observational, prospective 18-month survey was conducted to compare the efficacy and tolerability of the 2 available nonnucleoside reverse transcriptase inhibitors (NNRTIs) in all possible indications of current clinical practice. A broad range of clinical and laboratory variables accounting for drug efficacy and tolerability (with special emphasis on metabolic and hepatic toxicity) were measured in 287 evaluable patients treated with efavirenz, compared with 258 subjects taking nevirapine for 18 months. A separate efficacy analysis was performed in 154 antiretroviral-naive subjects, 298 patients experienced with 2–7 prior anti-HIV lines who abandoned protease inhibitors (PIs), and 103 subjects entering a salvage regimen containing at least 4 drugs, including PIs. Antiretroviral-naive patients experienced greater efavirenz activity at 3–12 months (maximum HIV RNA drop = −2.4 log10 copies/mL), associated with a significantly higher rate of complete viral suppression, while immunologic results proved significant only after 6–9 months. When assessing experienced patients and those on rescue regimens, a similar and progressively blunted laboratory response was achieved, on the ground of a worse baseline virologic and immunologic profile, and duration of prior anti-HIV therapy. Both first-month (4.2 and 4.3% for efavirenz and nevirapine, respectively) and overall discontinuation rates (11.5 and 12%, respectively) proved similar, but a profound difference emerged as to the different spectrum of untoward events: central nervous system (CNS) disturbances, persisting metabolic abnormalities, and possibly gynecomastia and laboratory pancreatic abnormalities for efavirenz vs. immediate allergy and increased hepatotoxicity (regardless of chronic infection with hepatitis B or C virus and methadone use) for nevirapine. A limited virologic and immunologic advantage of efavirenz was observed in the first 12-month assessment of antiretroviral-naive patients, whereas all other examined situations did not disclose relevant efficacy differences between efavirenz and nevirapine throughout the 18-month comparison. Although the short- and long-term toxicity and withdrawal rates of the 2 drugs were comparable, the different pathways prompting allergic, metabolic, liver, and CNS disturbances observed with NNRTIs deserve careful investigation, to prevent toxicity of these relevant antiretroviral compounds.

From the Department of Clinical and Experimental Medicine, Division of Infectious Diseases, University of Bologna “Alma Mater Studiorum,” S. Orsola Hospital, Bologna, Italy.

Received for publication July 28, 2003 accepted January 13, 2004.

Reprints: Roberto Manfredi, Infectious Diseases, S. Orsola Hospital, Via Massarenti 11, I-40138 Bologna, Italy (e-mail:

Only 2 nonnucleoside reverse transcriptase inhibitors (NNRTIs) are presently available for clinical use in the treatment of HIV disease. Recent recommendations for anti-HIV therapy indicated efavirenz among first-choice drugs and pointed out nevirapine among alternative agents in the setting of initial antiretroviral treatment 1; however, recent investigations (the so-called 2NN study) underlined a similar antiviral potency of efavirenz and nevirapine when they are used in antiretroviral therapy–naive subjects. Both efavirenz and nevirapine are widely used in either initial, subsequent, and salvage antiretroviral regimens, as suggested by clinical trials, international recommendations, and common clinical practice. 1–8 Although these 2 NNRTI drugs were introduced approximately 4 years ago, comprehensive, randomized, and nonrandomized studies of both efficacy and tolerability of efavirenz vs. nevirapine in all possible therapeutic backgrounds are still quite limited, all suffer from a proportionally short follow-up period (never >12 months), and sometimes gave contradictory results as to potency and safety. 4,5,9–11 As a consequence, most present recommendations are extrapolated from trials conducted in antiretroviral-naive patients, based on either efavirenz or nevirapine, and on multiple studies that focused on NNRTI use as a part of a combination anti-HIV regimen in already experienced patients, especially when a switch to an NNRTI-based treatment was suggested on the ground of a toxic or a failing protease inhibitor (PI)-based highly active antiretroviral therapy (HAART). 2,3,8,12–14

The aim of our open-label nonrandomized observational study is to perform a head-to-head prospective 18-month comparison of both efficacy and tolerability of the 2 available NNRTIs in all possible indications available in current clinical practice.

Back to Top | Article Outline


The activity of efavirenz vs. nevirapine was compared in 3 different clinical scenarios. The first one included patients naive to all antiretrovirals, who started an NNRTI-based regimen in association with 2 nucleoside analogues (NAs), selected among recommended associations. 1 The second situation analyzed a patient subgroup including a broad spectrum of pretreated but still NNRTI-naive subjects, who switched their antiretroviral therapy due to laboratory failure or emerging toxicity: all patients at their 2nd–7th therapeutic change were included in this second group, except subjects resorting to a PI-NNRTI association. Finally, patients adding for the first time an NNRTI while undergoing salvage regimens including ≥4 drugs (belonging to all available classes of anti-HIV agents, such as nucleoside/nucleotide analogues (NAs), PIs, NNRTIs, and fusion inhibitors), represented the 3rd analyzed group. In the 2nd and 3rd patient groups, low (200-mg/d) ritonavir doses used to make a booster of a concurrent PIs were not considered as a further antiretroviral drug adjunct, whereas the selection of associated NAs or tenofovir was decided by the physician in charge according to individual patients’ needs, on the basis of updated therapeutic recommendations for treatment of HIV infection. 1

Since October 2001, 662 overall patients, after giving their written consent, have undergone treatment with either efavirenz or nevirapine, and 287 evaluable patients treated with efavirenz were compared with 258 individuals taking nevirapine in our open-label, single-center prospective observational survey.

Back to Top | Article Outline

Baseline Assessment

For both efficacy and safety evaluation, a number of demographic and epidemiologic variables were compared between the efavirenz and the nevirapine group. HIV disease stage was evaluated considering class C disease or AIDS diagnosis, carried out according to the 1993 Centers for Disease Control (CDC) classification (Table 1). Laboratory markers of HIV disease progression included virology studies (with plasma viral load determined by an ultrasensitive branched- DNA technique with minimum detectable levels at 50 copies/mL), and immunologic assessment (as performed by peripheral CD4 lymphocyte count carried out by a flow cytometry assay). Underlying diseases (in particular, chronic hepatitis B or C virus (HBV or HCV) infection, as diagnosed by virologic, ultrasonographic, and eventually liver biopsy studies, type and duration of eventual prior antiretroviral therapy and HAART, and laboratory abnormalities related to possible untoward effects of previously administered antiretroviral compounds (with particular attention focused on abnormal body habitus, serum lipid levels, and markers of hepatic, skeletal muscle, and pancreatic toxicity, as well as hyperlactatemia), were recorded (Table 1). With more detail, elevated serum triglyceride levels were considered when their value glutamic-oxaloacetic acid transaminase tested >172 mg/dL; serum cholesterol levels >230 mg/dL; serum and glutamic pyruvic transaminase levels >37 and 40 U/L (respectively); serum amylase, pancreatic isoamylase, and lipase >220, 115, and 270 U/L (respectively); serum creatinphosphokinase (CPK) >40 U/L; and serum lactic acid levels >18 mg/dL.



Back to Top | Article Outline


An NNRTI-based regimen was started by administering efavirenz (600 mg once daily, at bed time), or nevirapine (200 mg/d for the first 2 weeks, followed by 200 mg twice daily), in association with at least 2 NAs.

Adherence levels were estimated by patients’ self-reports, specifically addressed questionnaires, and direct drug distribution and accountability performed at our outpatient center at least every month. Patients with an estimated compliance <90% were ruled out from the present study. Also patients who received a sequential treatment with efavirenz and nevirapine could not be assessed.

The evaluation of efficacy was carried out by an at least quarterly assessment of virologic and immunologic status compared with baseline values (those found at the time of NNRTI introduction), and the eventual development of HIV-associated disorders, or AIDS. The rate of complete viral suppression (viremia <50 copies/mL) was considered for antiretroviral-naive patients.

The assessment of drug safety included a careful active surveillance (telephone interview, eventual clinical examination) during the first 4 weeks of NNRTI administration, followed by systematic clinical examination performed at least monthly, and including body weight and arterial blood pressure measurement. Eventual subjective disturbances were carefully considered, and all patients given an NNRTI for the first time were invited to contact our outpatient service either over the telephone or through a first-aid ambulatory service at our center, which offers immediate and continuous assistance to HIV-infected patients.

The abnormal body habitus was assessed in detail using repeat physical examination, patient’s perception, specifically addressed questionnaires, and evaluation of waist-to-hip ratio. In selected cases, abdominal ultrasonographic examination and dual-energy x-ray absorptiometry were also performed, to provide an assessment of body fat distribution, hepatic steatosis, and bone mineralization rate. A mild amelioration of the lipodystrophy syndrome was considered when patient’s and physician’s assessments were favorable, but they did not match completely, and the waist-to-hip ratio was modified by 10% only; a moderate improvement was defined by a concurrent patient’s and physician’s judgment, associated with a waist-to-hip ratio change >10%.

Laboratory evaluation was carried out at least every 3 months and included hematologic and blood clotting profile, complete urinalysis, and an extensive biochemical workout including complete liver and kidney function tests, and complete serum pancreatic enzymes, serum lipids, glucose, lactic acid, and muscle enzyme levels (at least CPK levels). Further blood and instrumental assessments were performed when deemed necessary, i.e., abdominal ultrasonography or CT, and enlarged liver function tests in the event of liver toxicity, neurologic or psychiatric assessment with eventual electroencephalography, when central nervous system (CNS) disturbances were of concern.

NNRTI withdrawal due to insufficient efficacy or toxicity was decided by physicians in charge based on international recommendations for antiretroviral therapy, 1 and subsequent regimens were discussed with each individual patient. Alternative therapeutic regimens were offered to all patients failing (or not tolerating) prior NNRTI-based regimens.

Efficacy and safety data available at July 15, 2003 are analyzed and discussed in the present report.

Back to Top | Article Outline

Statistical Studies

For statistical analysis, we used Student t test, Mantel-Haenszel χ2 test, or Fisher exact test where appropriate, with significance levels set at P < 0.05.

Back to Top | Article Outline


Of 662 patients, 117 (17.7%) were not evaluable, because they had estimated adherence levels <90% (76 cases: 39 on efavirenz, and 37 on nevirapine) or received sequentially efavirenz and nevirapine (or vice versa), during the study period (41 patients).

Back to Top | Article Outline

Baseline Features

Demographic, epidemiologic, and clinical variables, HBV or HCV chronic hepatitis, and metabolic or toxic disturbances potentially depending on HAART (in pretreated patients) were statistically comparable in the 2 study groups (Table 1).

On the whole, IV drug use was the predominant type of exposure to HIV infection (53.7% of cases), followed by heterosexual risk (25%), and homo/bisexual exposure (21.3%), while men predominated over women (62.5 vs. 37.5%). The mean patient age increased among subjects naive to antiretrovirals (35.0 ± 6.8 years), to experienced persons (38.7 ± 7.9 years; P < 0.0001), to those on salvage regimens (39.5 ± 6.8 years; P < 0.0001), but no difference was seen according to the selected NNRTI in each study group. Prior antiretroviral therapy was longer among patients on rescue regimens (54.6 ± 14.9 vs. 46.6 ± 18.7 months of experienced subjects; P < 0.0001), as well as the mean number of administered antiretroviral lines (P < 0.0001), and the duration of PI-based HAART (39.1 ± 1.9 vs. 25.6 ± 10.1 months; P < 0.0001) (Table 1), but again no difference was detected between efavirenz- and nevirapine-treated subjects. When considering HIV disease stage, patients in CDC stage 3 and those with AIDS were 8.4% and 4.5%, respectively, among the naive patient group, compared with 23.2% and 16.4%, respectively, among antiretroviral-experienced individuals (P < 0.0001), and 46.6% and 25.2%, respectively, among those on salvage therapy (P < 0.0001), in absence of differences between the efavirenz and nevirapine groups (Table 1).

Co-infection with HBV or HCV viruses affected 9 (5.8%) and 36 patients of 154 (23.4%) among naive subjects, compared with 14 (4.7%) and 65 (21.8%) among experienced ones, and 3 (2.9%) and 36 (34.9%) among subjects on salvage anti-HIV therapy, with a greater frequency of HCV chronic hepatitis among patients on rescue therapy, compared with other groups (P < 0.02), but without any difference between efavirenz and nevirapine groups (Table 1). Also, the number of ex-IV drug users on methadone therapy did not differ among patient groups (Table 1).

When focusing on laboratory markers of HIV disease progression, compared with the other 2 patient groups, the 103 patients undergoing a salvage HAART had a greater mean viremia (4.9 ± 1.4 log10 HIV RNA copies/mL vs. 4.5 ± 1.1; P < 0.003) and a lower mean CD4+ count (181.8 ± 116.4 vs. 256.6 ± 89.4 cells/μL; P < 0.0001). Once again, no significant difference in terms of virologic and immunologic status was noticed at baseline (Table 1).

Finally, laboratory abnormalities potentially related to antiretroviral therapy were reported at baseline (i.e., before initiating or adding an NNRTI). When excluding patients naive to antiretrovirals, the lipodystrophy syndrome was more frequent among individuals on salvage therapy (62.1%), compared with experienced ones (30.2%; P < 0.0001). Hypertriglyceridemia and hypercholesterolemia had a greater frequency among patients undergoing rescue HAART (68 of 103 overall cases: 66%) vs. previously treated subjects (45.6%; P < 0.0001). Alterations of liver and pancreatic enzymes, although noticed also in naive patients, were strikingly more frequent among patients on salvage regimens (64 patients of 103: 62.1%) vs. those with multiple prior therapeutic changes (39.9%; P < 0.0001). Finally, abnormal serum skeletal muscle enzymes (CPK) and altered lactacidemia were found with increased incidence in the rescue group (29 of 103: 28.2%) vs. experienced patients (43 of 298 patients: 14.4%; P < 0.0001), with sparse cases among naive subjects. Furthermore, no difference was found as to all the aforementioned laboratory parameters, when evaluating separately subjects who received efavirenz and those treated with nevirapine (Table 1).

Back to Top | Article Outline

Follow-Up of Efficacy Parameters

Virologic and immunologic features of efavirenz- and nevirapine-treated patients are depicted in Figures 1–6: all laboratory determinations carried out at least quarterly are reported from baseline until 18 months of follow-up, with number of patients quoted at each evaluation, in our on-treatment analysis.













The most prominent difference between efavirenz- and nevirapine-treated subjects was detected in the group of patients completely naive to all antiretrovirals: a virologic gain was seen from the 3rd to the 12th month of observation, compared with baseline plasma HIV RNA levels (P < 0.0001) (Fig. 1), while a significant immunologic difference was limited to a shorter period, ranging from 6 (P < 0.02) to 9 months (P < 0.0001) (Fig. 2). Both antiretroviral-experienced patients, and those on rescue HAART, who added an NNRTI for the first time as a part of their regimen did not show any significant difference between the efavirenz and nevirapine groups, as to virologic and immunologic response, from baseline up to the 18th month of treatment, as represented in Figures 3–6. When examining the virologic response of antiretroviral-naive patients, the mean peak HIV RNA drop was −2.4 log10 HIV RNA copies/mL achieved after 9 months and again after 12 and 18 months for efavirenz, while a more limited but progressive gain occurred for nevirapine (highest drop of viremia reached after 18 months, with a mean −2.2 log10 HIV RNA copies/mL) (Fig. 1). Undetectable viremia (<50 copies/mL) was attained between the 6th and the 12th month of follow-up in 51 overall patients belonging to the evafirenz group, and in 40 subjects treated with nevirapine: the difference was statistically significant at the 6th month (P < 0.006) and the 9th month (P < 0.0001). In the same patient group, the greatest mean CD4+ lymphocyte increase was reached after 9 months in the efavirenz group (+41.1% vs. baseline) and after 12 months for nevirapine (+25%) (Fig. 2). Among pretreated patients, the greatest virologic advantage was attained in both groups after 9 months, with a mean −1.6 log10 drop for efavirenz and a mean −1. log10 decrease of HIV RNA levels for nevirapine (Fig. 3), while the more relevant increase of mean CD4 count was reached again after 9 months of therapy for both NNRTIs, with a +47.3% vs. baseline levels for efavirenz, and a +29.3% for nevirapine (Fig. 4). Finally, patients on salvage therapy had an expected, more limited benefit from an NNRTI adjunct, as the highest virologic response was achieved after 6 months, with a − 1.3 drop of log10 HIV RNA levels for efavirenz, and a −0.9 log10 decrease obtained with nevirapine, while the subsequent months of follow-up were characterized by a progressive increase of viral load in both patient groups (Fig. 5). Also, immunologic advantage was blunted among subjects experiencing their rescue therapy, with a maximum increase of mean CD4 lymphocyte count observed at the 6th month for both drugs (+23.7% for efavirenz and +9% for nevirapine), followed by a relevant loss of CD4 cells occurring during the remaining follow-up (Fig. 6). Complete viral suppression was comparable and not sustained in both study groups, when considering experienced patients, and especially those on salvage therapy (data not shown).

When evaluating the duration of a sustained virologic and immunologic response, it included the entire 18-month reported follow-up for antiretroviral-naive patients introducing an NNRTI in their first antiretroviral combination, although virologic response proved greater in the efavirenz group between 3 and 12 months, and immunologic gain was higher between 6 and 9 months (Figs. 1 and 2). For antiretroviral-experienced patients, and those undergoing a salvage therapy including for the first time an NNRTI, a reversal of virologic trend was observed since 12 months in the first group (Fig. 3) and since 9 months among patients on rescue therapy (Fig. 5), regardless of the NNRTI used (efavirenz or nevirapine). The immunologic profile strictly followed the virologic one, showing a decrease of CD4 cell count since 12 and 9 months, again with negligible differences between efavirenz- and nevirapine- treated subjects (Figs. 4 and 6). In patients on a rescue regimen, a relationship between duration of efficacy of NNRTI regimen and length of prior antiretroviral therapy (P < 0.0001), number of experienced antiretroviral lines (P < 0.0001), and duration of PI including HAART (P < 0.0001) was evident, probably reflecting an increased pattern of antiviral resistance over time (Table 1), although we are not able to present a complete evaluation of genotypic or phenotypic resistance patterns for all treated patients.

Moreover, a more advanced (stage C) HIV disease or a concurrent diagnosis of AIDS paralleled a greater mean viral load and a lower CD4 lymphocyte at baseline, comparing antiretroviral-naive patients with experienced ones, and with subjects on salvage therapy (P < 0.0001) (Table 1), thus explaining a blunted and less sustained laboratory response paralleling the duration and extent of prior antiretroviral treatment. When comparing baseline and final (18-month) laboratory profile, we can easily notice that a virologic and immunologic gain was maintained throughout the observation period in naive patients (Figs. 1 and 2) and to a lesser extent in pretreated patients (Figs. 3 and 4), whereas the 18-month study period allowed us to observe a return to baseline levels of both mean virologic-immunologic figures (representing a laboratory failure) after the entire follow-up period (Figs. 5 and 6). A progression to stage C disease and the occurrence of full-blown AIDS were reported in 3 and 2 patients, respectively, in the efavirenz group, and in 2 and 2 subjects, respectively, in the nevirapine group, all belonging to the patient subgroup on salvage therapy.

Back to Top | Article Outline

Tolerability Evaluation

Most drug safety issues are summarized in Table 2, which compares overall efavirenz- with nevirapine-treated patients, regardless of their status as naive, experienced, or salvage subjects.



The early (first month) discontinuation rate proved similar for efavirenz (4.2%) and nevirapine (4.3%) (Table 2), but it was prompted by the typical CNS disturbances in 11 cases of 12 treated with efavirenz and by an allergic rash in all patients experiencing early nevirapine withdrawal.

From a metabolic point of view, the lipodystrophy (lipoaccumulation or lipoatrophy) syndrome had a mild-to-moderate improvement in a greater percentage of nevirapine-vs. efavirenz-treated patients throughout the entire 18-month study period (24% vs. 11.8%; P < 0.05), whereas gynecomastia appeared in 7 male patients belonging to the efavirenz group and without any other risk factor for this event, compared with no cases among nevirapine-treated patients (P < 0.02), regardless of the antiretroviral treatment stage (3 antiretroviral-naive vs. 2 experienced, vs. 2 subjects on rescue therapy; all experienced patients switched from a PI-based association). This last untoward effect, confirmed by ultrasonography, was responsible for some discomfort but never led to drug discontinuation or surgery; no association with prior or underlying NAs as part of the administered antiretroviral regimen was found. Hypertriglyceridemia was already present in 34 patients starting efavirenz and in 30 subjects initiating nevirapine, whereas high serum cholesterol levels affected 21 subjects in the efavirenz group and 15 in the nevirapine group. Throughout the study period, we observed an amelioration of these parameters among patients on nevirapine, compared with the entire efavirenz group (P < 0.0001 for hypertriglyceridemia, P < 0.04 for hypercholesterolemia) (Table 2), with an unexpected overall increase of frequency of dyslipidemia among efavirenz-treated patients, compared with nevirapine-treated ones, particularly for hypertriglyceridemia, whose frequency increased from 34 to 55 overall cases (P < 0.0001 vs. the nevirapine group) (Table 2). Serial body weight and arterial blood pressure measurements did not show relevant modifications among efavirenz- and nevirapine-treated patients (data now shown).

A hepatotoxicity characterized by an at least 2-fold increase of serum transaminases occurred significantly in the nevirapine vs. the efavirenz group throughout the 18-month study period: the number of patients with increased serum liver enzymes dropped in the efavirenz group from 93 cases at baseline to 51 overall patients throughout the study period, while it concurrently increased in nevirapine-treated patients from 80 cases observed at baseline, up to 134 patients during the 18-month observation time (P < 0.0001) (Table 2), regardless of a concurrent HBV/HCV coinfection and concomitant methadone use (Table 1). Time to peak altered liver enzymes tested significantly shorter among nevirapine- vs. efavirenz-treated patients (P < 0.0001) (Table 2), and hepatotoxicity prompted nevirapine withdrawal in 9 patients vs. none for efavirenz (P < 0.002) (Table 2).

Laboratory abnormalities of pancreatic function were still present at the time of start of NNRTI-containing regimens (30 cases in the efavirenz group and 25 in the nevirapine group), but during the study period an increased frequency of serum amylase or pancreatic amylase or lipase clearly increased in the efavirenz group (55 cases on the whole) vs. the nevirapine group (22 patients) (P < 0.0001) (Table 2), probably as a consequence of persisting dyslipidemia among the efavirenz-treated patients, and regardless of the greater nevirapine-related liver toxicity.

Conversely, no significant difference occurred when examining the rate of serum CPK and lactic acid levels, comparing the frequency of baseline abnormalities with those occurring throughout the 18-month study period (Table 2); probably these toxicities rely mostly on associated NA, as an effect of mitochondrial abnormalities prompted by these last drugs.

The overall drug discontinuation rates attributable to untoward events were comparable: 11.5% in the efavirenz group vs. 12% in the nevirapine group (Table 2), and the time to drug withdrawal tested similar, although great variations occurred in both patient groups, as expressed by the elevated SD figure (Table 2), so that interruptions occurred throughout the entire 18-month study period for both NNRTIs.

Back to Top | Article Outline


When assessing efficacy parameters, several comparative studies are available for efavirenz vs. nevirapine, sometimes pointing out a possible greater effect for efavirenz only. However, data are still limited, patients samples and follow-up time are sometimes insufficient, and the study design is often retrospective or nonrandomized, so that conclusive statements regarding the first-choice NNRTI cannot be drawn. 4,5,9–11,14,15

The Euro-SIDA study group performed a large retrospective analysis including 2203 mainly antiretroviral-experienced patients who started an NNRTI-containing regimen: virologic failure was considered as the main endpoint. When comparing 1325 patients on nevirapine with 878 on efavirenz, a significantly lower (P < 0.0001) failure rate was recognized for efavirenz, but many confounding variables were recognized by the authors, including type and duration of prior antiretroviral therapy and clinical and laboratory markers at baseline. 4 A randomized pilot study compared 31 naive patients treated with efavirenz with 26 who received nevirapine together with didanosine and stavudine and did not disclose significant differences as to antiviral efficacy and safety after a 48-week assessment 10: in this study, the patient sample was too small to allow significant statements. Nonrandomized data from an Italian cohort of antiretroviral-naive HIV-infected patients compared 460 patients who started nevirapine with 234 who initiated efavirenz: the adjusted relative hazard of virologic failure was higher for nevirapine-treated subjects (P = 0.0006), whereas immunologic response was poorly different between the 2 groups. 11 A cohort study performed on 1078 antiretroviral-naive subjects established a more elevated and sustained potency of efavirenz- vs. nevirapine-based triple-drug HAART, when comparing time to failure and virologic gain (P < 0.001) in a multivariate analysis. 9 A recent English report of 171 patients who switched from PI-based regimens because of virologic failure demonstrated that efavirenz and abacavir adjunct acted favorably during a 1-year follow-up; in particular, a significant rate (P < 0.003) of virologic success of efavirenz vs. nevirapine was demonstrated. 14 An attempt to perform a meta-analysis with other HAART regimens did not disclose relevant differences between efavirenz- and nevira-pine-based antiretroviral combinations, which even demonstrated a slightly more favorable outcome of NNRTI-based regimens compared with PI-based associations. 5 Finally, a recent multicenter study presented at the Tenth Conference on Retroviruses and Opportunistic Infections in February 2003 demonstrated a comparable efficacy of nevirapine and efavirenz in an extensive, prospective, randomized study involving 1216 antiretroviral-naive patients. 15

Conversely, more extensive data are available on patients abandoning PI-based regimens toward a simplified HAART often based on an NNRTI, aimed at demonstrating an increased adherence, a sustained activity, and a reduced toxicity (especially when metabolic abnormalities and lipodystrophy were of concern during PI administration). 2,3,8,12–14 In a Spanish study, 77 patients with long-lasting viral suppression were switched from a PI-based therapy to nevirapine (26 patients), efavirenz (25 patients), or continued PI treatment (26 patients): after 12 months, viral suppression was slighlty more elevated in the efavirenz group (96%), compared with nevirapine and PI ones (92% each). The improvement of quality of life demonstrated in both groups switching to an NNRTI-based therapy was associated with reduced serum lipid levels in the nevirapine group, but increased CNS complaints among efavirenz-treated patients. 12 In an Italian study performed on patients on a long-term effective PI-based HAART, 85 patients simplified their regimen starting efavirenz, and 54 took nevirapine: this last NNRTI was associated with a worse (P < 0.02) virologic outcome, limited to patients who started their long-term antiretroviral therapy with NAs only, and progressed to AIDS. 13

Unfortunately, extensive cross-resistance between nevirapine and efavirenz strongly limits their sequential use due to low genetic barrier to mutations, 16–19 so that patients who had been treated with both drugs were ruled out from our assessment. Heavily pretreated patients and those undergoing salvage regimens are at risk for a limited benefit from an NNRTI adjunct even when naive to this last drug class, 20 because of the extensive spectrum of transcriptase inhibitor mutations developed over time, including those at codons 103 and 181, conferring cross-resistance. 6,7,20–24 However, both laboratory 25,26 and anecdotal in vivo experiences 15,27 showed a possible synergistic effect of dual NNRTIs against HIV. Moreover, a sensitization to NNRTIs may emerge from altered viral fitness descending from key mutations conferring NA phenotypic resistance, 19,28 and a slow decline of mutations encoding NNRTI resistance is observed after drug suspension. 18 Other NNRTI compounds with a different pattern of viral resistance are under development, and some will enter clinical use soon. 19,25,29

In our experience, although limited by the absence of a systematic evaluation of viral resistance and the observational, prospective open-label nature of our field study, patients introducing either efavirenz or nevirapine as their first NNRTI were prospectively compared for several efficacy and tolerability features, and very limited statistical differences were found between the 2 groups, according to the virologic and immunologic features, for a prolonged (18-month) observation period.

When considering the effects on laboratory and clinical markers of HIV disease progression, the only statistically significant advantage was found in efavirenz-treated patients who were naive to all antiretrovirals, but this difference vs. nevirapine-treated subjects was limited to a 9-month period for mean viral load and attained complete viral suppression, and to 3 months only for mean CD4 lymphocyte count, while 12- to 18-month follow-up did not disclose any relevant difference between the 2 study drugs. Comparable features were found in all other patient groups, including patients who were pretreated with several antiretroviral lines, and those introducing an NNRTI as a part of a late-rescue anti-HIV treatment. As expected, both virologic and immunologic profile showed advantages until 9 months among experienced patients and was limited to 6 months in those on salvage therapy, as a consequence of baseline higher viremia, advanced underlying disease, and multiple experienced therapeutic lines (Table 1). Virologic effects were stronger among naive patients (as expressed by a mean peak −2.4 drop of log10 HIV RNA copies/mL for efavirenz and a mean drop of − 2.2 log10 HIV RNA copies/mL for nevirapine) but were also found among experienced patients (mean peak of −1.6 and −1.2 for efavirenz and nevirapine, respectively), and among those on rescue regimens (mean peaks of −1.3 and −0.9 for efavirenz and nevirapine, respectively). Immunologic gain was less evident, also in antiretroviral-naive patients (in whom a significant difference in favor of efavirenz was limited to 3 months only, when the mean increase of CD4 lymphocyte count vs. baseline levels reached >40% for efavirenz and 25% for nevirapine). Interestingly, pretreated patients who started from a significantly lower mean CD4 cell count showed a greater increase when an NNRTI was added (mean peak rise of CD4 count of >47% for efavirenz and nearly 30% for nevirapine). Unfortunately, there was limited success for patients on rescue treatment, who started from a more compromised clinical-laboratory situation and experienced a transient (6-month) virologic and immunologic success, but after 18 months returned to baseline laboratory levels, as a probable effect of an extensive resistance profile, following a more elevated number and duration of prior antiretroviral treatments. The differences between the 2 studied NNRTIs did not reach statistical significance in all experienced patients, regardless of baseline situation and therapeutic background, which tested significantly different at baseline. Laboratory changes did not parallel significant clinical outcomes, because disease progression occurred in a very limited number of patients, all belonging to the salvage group.

In our series, considerations about safety have to take into careful account the different baseline features of patients undergoing for the first time an NNRTI-based therapy, and eventual co-morbid conditions (such as chronic hepatitis). A comparable first-month discontinuation rate was found for both drugs (mean 4.2% for efavirenz, 4.2% for nevirapine), although CNS disorders and allergic rash prompted almost all cases for efavirenz and nevirapine, respectively. Overall, crude drug discontinuation rates prompted by toxicity or poor tolerability reached 11.5% in efavirenz-treated patients and 12% in the nevirapine group, confirming the absence of significant differences between the 2 NNRTIs, also in terms of time to withdrawal (<6 months for both compounds). When considering the lipodystrophy syndrome and serum lipid abnormalities, only experienced patients could be evaluated, due to the predominant causative role prompted by PIs and NAs in previous therapeutic lines 30,31: in this circumstance, nevirapine had a more favorable effect on the perceived and measured abnormal body habitus and on both serum triglyceride and cholesterol levels. In particular, efavirenz-treated patients experienced an appearance or a worsening of hypertriglyceridemia (and poor modifications of hypercholesterolemia rates) throughout the entire study period, and an unexpected occurrence of gynecomastia (a recently recorded possible untoward effect of HAART), in 7 cases (although concurrent NA administration seems to play a role in this disturbance). 32 However, prior and concurrent antiretroviral therapy can play a significant and confounding role in many toxic events, so that NNRTI introduction cannot be considered a remarkable, independent variable in this field; also, multivariate analysis becomes very hard in pretreated patients. Conversely, liver toxicity had a more significant increase in the nevirapine than in the efavirenz group compared with baseline, and regardless of an eventual, concurrent HBV/HCV chronic hepatitis. A mean 6.5-month period was sufficient to reach peak hepatotoxicity for nevirapine-treated patients, who had to withdraw their NNRTI for liver toxicity in 9 cases, vs. none for efavirenz. The untoward effects on serum pancreatic enzymes proved more frequent in the efavirenz group but were probably influenced by a higher frequency of hyperlipidemia in this group, while concurrent damage prompted by concurrently administered NAs and HIV itself cannot be excluded. 33 Markers of skeletal muscle damage and hyperlactatemia did not show relevant differences between the 2 study drugs, as an effect of a probable mitochondrial toxicity promoted by drugs other than NNRTIs. 34–36

When considering adverse events and toxicity of NNRTIs, in a multicenter study including >1000 patients treated with efavirenz (>62% IV drug users and 20% antiretroviral-naive), an overall rate of adverse events of 29.3% was found, and treatment withdrawal became necessary in 8.2% of cases. CNS disturbances affected 24.1% of subjects who started efavirenz, followed by cutaneous rash (5.9%), gastroenteric complaints (1.45%), and liver toxicity (0.68%). 37 Patients who concurrently took methadone (6.6% of the population) had a more elevated incidence of adverse events (39.7%), 37 but this last feature was not evident in our study population.

Liver enzyme abnormalities emerge during HAART as a potential consequence of a broad spectrum of variables, most of them associated with factors other than NNRTI use (direct drug damage, mitochondrial toxicity, multiple metabolic abnormalities, concurrent viral hepatitis, substance and alcohol abuse, and opportunistic infections). 2,3,35,36,38,39 However, NNRTIs (especially nevirapine) have been implicated in hepatic toxicity, although time, mode, severity, and frequency are still extensively discussed. 2,3,8,35,39–43 In fact, isolated dual-NA therapy seems associated with a nonnegligible hepatotoxicity. 38 In many reports in the literature, a chronic viral hepatitis appeared to significantly increase the risk of liver toxicity, 3,35,41–43 but it was not the case in our experience. In a retrospective study comparing 162 nevirapine-treated and 136 efavirenz-treated patients, with a 45% overall rate of HCV co-infection, Martin-Carbonero et al. 42 showed an increased liver toxicity in the nevirapine group (12%) compared with the efavirenz group (4%) (P < 0.02), after a median time of 5.5 months, which was shorter vs. that of our patients who developed (or had a worsened) liver injury, where a different time to elevated liver enzymes was recognized between nevirapine and efavirenz. (mean 6.5 ± 2.1 vs. 9.6 ± 3.2 months, respectively). Female gender and alcohol abuse also acted as independent risk factors for hepatic toxicity, according to this recent Spanish experience. 42 In a large prospective study conducted by Sulkowski et al. 43 on 568 patients suffering from an underlying HCV infection in 43%, and HBV infection in 7.7% of patients, and undergoing an NNRTI-based HAART, severe (grades 3 and 4) hepatotoxicity was assessed among 312 and 256 patients to whom efavirenz and nevirapine were prescribed, respectively. Severe liver toxicity occurred in 15.6% of patients given nevirapine and 8% of those treated with efavirenz, but only 32 and 50% of episodes (for nevirapine and efavirenz, respectively) were detected in the first year of follow-up, thus confirming our observation of long-term development of eventual liver injury. The risk proved greater when chronic hepatitis was present, and PIs were concurrently administered (reaching 82% among NNRTI-PI cotreated patients), whereas even 84% of subjects with chronic HCV/HBV coinfection did not experience relevant hepatotoxicity, 43 as also demonstrated by the 18-month follow-up of our cohort. On the same line, Palmon et al. 40 did not disclose liver toxicity among 272 NNRTI-treated patients, 52% of whom were treated with nevirapine, although coinfection with HBV or HCV affected 9 and 12% of cases, respectively. This last feature confirms the unpredictable role of chronic hepatitis coinfection in prompting liver toxicity, as demonstrated in our experience.

Effects on serum lipid abnormalities prompted by prior PI-based HAART were improved after switching to nevirapine compared with efavirenz, 3 as suggested by our data. Preliminary data from a large randomized trial conducted in naive patients 15 disclosed that nevirapine use prompted an elevation of serum high-density lipoprotein cholesterol levels and led to a 8% reduction of the total cholesterol/high-density lipoprotein cholesterol rate, compared with efavirenz.

In conclusion, our experience points out that NNRTIs have a comparable activity on virologic, immunologic, and clinical parameters of HIV infection, save a more evident (but not sustained) efficacy on mean viral load and (at a lesser extent) on mean CD4+ lymphocyte count levels, among antiretroviral-naive patients. Including antiretroviral-experienced patients and those on salvage therapy, we noticed a progressively decreasing intensity and duration of NNRTI activity on laboratory markers of HIV disease, without any significant difference between efavirenz and nevirapine. NNRTI tolerability must be taken into careful account, since short- and long-term toxicity has a broad spectrum and proves extremely variable between efavirenz and nevirapine, in terms of frequency, features, and timing. Although the overall frequency of drug withdrawal was comparable between the 2 study groups, it proved very different as to causative events and related time of occurrence. A prolonged follow-up (at least 18 months in our study) shows a tendency to cumulative hepatotoxicity for nevirapine, while efavirenz may not resolve (or might even prompt) metabolic abnormalities, with particular emphasis on serum lipid levels, abnormal body habitus, and gynecomastia. The pathogenesis of this surprisingly different spectrum of toxicity between efavirenz and nevirapine, despite their belonging to the same class of NNRTIs, warrants urgent and extensive, controlled epidemiologic, clinical, and laboratory investigation. In the meantime, a close monitoring of allergy, hepatic function, serum lipid metabolism, and other metabolic abnormalities is strongly suggested in all patients undergoing an NNRTI-based HAART, especially those with preexisting hepatic and laboratory disorders.

Back to Top | Article Outline


1. Yeni PG, Hammer SM, Carpenter CCJ, et al. Antiretroviral treatment for adult HIV infection in 2002. Updated recommendations of the International AIDS Society-USA Panel. JAMA. 2002;288:161–178.
2. Moyle GJ. Considerations in the choice of protease inhibitor-sparing regimens in initial therapy for HIV-1 infection. Curr Opin Infect Dis. 2000;13:19–25.
3. Barreiro P, Garcia-Banayas T, Soriano V, et al. Simplification of antiretroviral treatment: how to sustain success, reduce toxicity and ensure adherence avoiding PI use. AIDS Rev. 2002;4:233–241.
4. Phillips AN, Pradier C, Lazzarin A, et al. Viral load outcome of non-nucleoside reverse transcriptase inhibitor regimens for 2203 mainly antiretroviral experienced patients. AIDS. 2001;15:2385–2395.
5. Torre D, Tambini R, Speranza F. Nevirapine or efavirenz combined with two nucleoside reverse transcriptase inhibitors compared to HAART: a meta-analysis of randomized clinical trials. HIV Clin Trials. 2001;2:113–121.
6. Manfredi R, Calza L, Chiodo F. A prospective comparison of the two main indications of efavirenz in 2001 highly active antiretroviral therapy (HAART) regimens: first-line versus salvage use. J Antimicrob Chemother. 2002;49:723–729.
7. Manfredi R, Rizzo E, Calza L, et al. The use of efavirenz as a part of a late rescue antiretroviral treatment. HIV Clin Trials. 2001;2:413–420.
8. Bell C, Matthews GV, Nelson MR. Non-nucleoside reverse transcriptase inhibitors: an overview. Int J STD AIDS. 2003;14:71–77.
9. Keiser P, Nassar N, White C, et al. Comparison of nevirapine- and efavirenz-containing antiretroviral regimens in antiretroviral-naive patients: a cohort study. HIV Clin Trials. 2002;3:296–303.
10. Nuñez M, Soriano V, Martin-Carbonero L, et al. SENC (Spanish Efavirenz vs. Nevirapine Comparison) trial: a randomized, open-label study in HIV-infected naive individuals. HIV Clin Trials. 2002;3:186–194.
11. Cozzi-Lepri A, Phillips AN, D’Arminio Monforte A, et al. et al. Virologic and immunologic response to regimens containing nevirapine or efavirenz in combination with 2 nucleoside analogues in the Italian Cohort Naive Antiretrovirals (I.Co.N.A.) study. J Infect Dis. 2002;185:1062–1069.
12. Negredo E, Cruz L, Paredes R, et al. Virological, immunological, and clinical impact of switching from protease inhibitors to nevirapine or to efavirenz in patients with human immunodeficiency virus infection and long-lasting viral suppression. Clin Infect Dis. 2002;34:504–510.
13. Chiesa E, Bini T, Adorni F, et al. Simplification of protease inhibitor-containing regimens with efavirenz, nevirapine or abacavir: safety and efficacy outcomes. Antivir Ther. 2003;8:27–35.
14. Asboe D, Mandalia S, Gazzard BG. Sequencing to NRTI plus NNRTI-only combinations after virological failure of protease inhibitor-based combination HIV-1 therapy. HIV Clin Trials. 2003;4:1–10.
15. Van Leth F, Hassink E, Phanuphak P, et al. A randomised comparative open label trial of first line antiretroviral therapy with regimens containing either nevirapine, efavirenz, or both drugs combined, in addition to stavudine and lamivudine. Paper presented at: Tenth Conference on Retroviruses and Opportunistic Infections (CROI); February 10–14, 2003; Boston, MA. Abstract 176.
16. Antinori A, Zaccarelli M, Cingolani A, et al. Cross-resistance among non-nucleoside reverse transcriptase inhibitors limits recycling efavirenz after nevirapine failure. AIDS Res Hum Retroviruses. 2002;18:835–838.
17. Casado JL, Moreno A, Hertogs K, et al. Extent and importance of cross-resistance to efavirenz after nevirapine failure. AIDS Res Hum Retroviruses. 2002;18:771–775.
18. Quiros-Roldan E, Airoldi M, Moretti F, et al. Genotype resistance profiles in patients failing an NNRTI-containing regimen, and modifications after stopping NNRTI therapy. J Clin Lab Anal. 2002;16:76–78.
19. Campiani G, Ramunno A, Maga G, et al. Non-nucleoside reverse transcriptase (RT) inhibitors: past, present, and future perspectives. Curr Pharm Des. 2002;8:615–657.42.
20. Manfredi R. A cross-sectional survey of HIV genotype mutations conferring resistance to all nonnucleoside reverse transcriptase inhibitors and related features, at the time of failure of antiretroviral therapy in the real world. J Acquir Immune Defic Syndr. 2001;28:97–99.
21. Manfredi R, Chiodo F. Substitution of protease inhibitors during treatment of patients with human immunodeficiency virus infection: frequency, mode, reasons and mid-term outcome. J Antimicrob Chemother. 2000;45:261–263.
22. Manfredi R, Chiodo F. Limits of deep salvage antiretroviral therapy with nelfinavir plus either efavirenz or nevirapine, in highly pretreated patients with HIV disease. Int J Antimicrob Agents. 2001;17:511–516.
23. Manfredi R, Chiodo F. Switch of protease inhibitor-containing HAART in routine clinical practice: a four-year prospective observational study. Int J STD AIDS. 2001;12:84–88.
24. Manfredi R, Chiodo F. First-line nevirapine in combination with nucleoside analogues compared with nevirapine added to a salvage HAART. Infection. 2002;28:167–170.
25. King RW, Klabe RM, Reid CD, et al. Potency of nonnucleoside inhibitors (NNRTIs) used in combination with other human immunodeficiency virus NNRTIs, NRTIs, or protease inhibitors. Antimicrob Agents Chemother. 2002;46:1640–1646.
26. Kollmann C, Tremblay C, Giguel F, et al. In vitro anti-HIV-1 synergy between non-nucleoside reverse transcriptase inhibitors nevirapine and efavirenz. Antivir Ther. 2001;6:143–144.
27. Olivieri J. Nevirapine + efavirenz based salvage therapy in heavily pretreated HIV infected patients. Sex Transm Infect. 2002;78:72–73.
28. Whitcomb JM, Huang W, Limoli K, et al. Hypersusceptibility to non-nucleoside reverse transcriptase inhibitors in HIV-1: clinical, phenotypic and genotypic correlates. AIDS. 2002;16:F41–F47.
29. Rizzo RC, Udier-Blagovic M, Wang DP, et al. Prediction of activity for nonnucleoside inhibitors with HIV-1 reverse transcriptase based on Monte Carlo simulations. J Med Chem. 2002;45:2970–2987.
30. Manfredi R, Chiodo F. Disorders of lipid metabolism in patients with HIV disease treated with antiretroviral agents: frequency, relationship with administered drugs, and role of hypolipidaemic therapy with bezafibrate. J Infect. 2001;42:181–188.
31. Calza L, Manfredi R, Farneti B, et al. Incidence of hyperlipidaemia in a cohort of 212 HIV-infected patients receiving a protease inhibitor-based antiretroviral therapy. Int J Antimicrob Agents. 2003;22:54–59.
32. Manfredi R, Calza L, Chiodo F. Gynecomastia associated with highly active antiretroviral therapy. Ann Pharmacother. 2001;35:438–439.
33. Manfredi R, Calza L, Chiodo F. Gabexate mesilate and/or octreotide in human immunodeficiency virus-associated pancreatic abnormalities. Aliment Pharmacol Ther. 2002;16:1791–1794.
34. Manfredi R, Motta R, Patrono D, et al. A prospective case-control survey of laboratory markers of skeletal muscle damage during HIV disease and antiretroviral therapy. AIDS. 2002;16:1969–1971.
35. Manfredi R. HIV infection, antiretroviral therapy, and hepatic function: emerging epidemiological, pathogenetic, and clinical issues, and their consequences on disease management. AIDS. 2003;17:2253–2256.
36. Manfredi R, Motta R, Patrono D, et al. Frequency, risk factors and features of hyperlactatemia in a large number of patients undergoing antiretroviral therapy. AIDS. 2003;17:2131–2133.
37. Perez-Molina JA. Safety and tolerance of efavirenz in different antiretroviral regimens: results from a national multicenter prospective study in 1,033 HIV-infected patients. HIV Clin Trials. 2002;3:279–286.
38. Verucchi G, Calza L, Manfredi R, et al. Incidence of liver toxicity in HIV-infected patients receiving isolated dual nucleoside analogue antiretroviral therapy. J Acquir Immune Defic Syndr. 2003;33:546–548.
39. Manfredi R. HIV disease, antiretroviral treatment and the liver: emerging problems. Recenti Prog Med. 2003;94:149–153.
40. Palmon R, Koo BC, Shoultz DA, et al. Lack of hepatotoxicity associated with nonnucleoside reverse transcriptase inhibitors. J Acquir Immune Defic Syndr. 2002;29:340–345.
41. Kontorinis N, Dieterich DT. Toxicity of non-nucleoside analogue reverse transcriptase inhibitors. Semin Liver Dis. 2003;23:173–182.
42. Martin-Carbonero L, Nuñez M, Gonzalez-Lahoz J, et al. Incidence of liver injury after beginning antiretroviral therapy with efavirenz or nevirapine. HIV Clin Trials. 2003;4:115–120.
43. Sulkowski MS, Thomas DL, Mehta SH, et al. Hepatotoxicity associated with nevirapine- or efavirenz-containing antiretroviral therapy: role of hepatitis C and B infections. Hepatology. 2002;35:182–189.

antiretroviral therapy; efavirenz; nevirapine; comparison; efficacy; tolerability; observational prospective study; current clinical practice

© 2004 Lippincott Williams & Wilkins, Inc.