HIV Articles  
Raltegravir central nervous system tolerability in clinical practice: results from a multicenter observational study
  Download the PDF here

28 November 2012
Research Letters

Madeddu, Giordanoa; Menzaghi, Barbarab; Ricci, Elenac; Carenzi, Laurac; Martinelli, Caniod; di Biagio, Antonioe; Parruti, Giustinof; Orofino, Giancarlog; Mura, Maria S.a; Bonfanti, Paoloh; for the C.I.S.A.I GroupaDepartment of Clinical and Experimental Medicine, University of Sassari, SassaribDepartment of Infectious Diseases, Busto Arsizio Hospital, Busto ArsiziocDepartment of Infectious Diseases, Luigi Sacco Hospital, MilandDepartment of Infectious Diseases, Azienda Ospedaliera Universitaria Careggi, FlorenceeDepartment of Infectious Diseases, San Martino Hospital and University of Genoa, GenoafDepartment of Internal Medicine, Unit of Infectious Diseases, Pescara General Hospital, PescaragDepartment of Infectious Diseases, Amedeo di Savoia Hospital, Turin hUnit of Infectious Disesases, A Manzoni Hospital, Lecco, Italy.


Central nervous system (CNS) symptoms have been reported in clinical trials and case reports in patients receiving raltegravir. We investigated CNS symptoms in 453 HIV-infected patients. Of these 47 (10.4%) developed at least one drug-related CNS symptom. Predictors of CNS symptoms were concomitant therapy with tenofovir or with proton pump inhibitors that can increase raltegravir concentration. Thus, our data suggest a possible correlation between high raltegravir plasma concentrations and CNS symptoms, and therefore their monitoring in clinical practice.

Raltegravir is the first HIV integrase inhibitor available in clinical practice for the treatment of HIV infection in both naive and experienced patients [1-4]. Raltegravir inhibits the strand-transfer step of integration by blocking the enzyme's active site, and thus the preintegration complex is unable to bind to host DNA [5,6]. The nonintegrated proviral HIV DNA is repaired via normal cellular DNA repair mechanisms and is rendered inactive [7]. In contrast to most other antiretroviral drugs, raltegravir is metabolized by glucuronidation via UGT1A1 [8,9]. Excretion in feces (51%) and in urine (31%) accounts for most of the elimination. No dose adjustment is required for sex, age, hepatic or renal function, or BMI [10]. Raltegravir has been shown to pass the blood-brain barrier in the majority of patients, even if central nervous system (CNS) concentrations exceed the drug concentration needed to inhibit 95% of viral replication for HIV-1 strains without resistance to integrase inhibitors in only 50% of cases [11].

Randomized clinical trials have shown a good safety profile. However, some CNS symptoms have been reported in clinical trials, and case reports of worsening depression and acutely onset insomnia after starting raltegravir have also been described [12,13].

The aim of our study was to further investigate CNS safety of raltegravir in a multicenter observational study. The Surveillance Cohort Long-Term Toxicity of Antiretrovirals (SCOLTA) Project is an online pharmacovigilance program involving 18 Italian infectious disease departments. The Project has an internet site ( in which grade III and IV adverse events, according to Division of AIDS table, are recorded ( The SCOLTA Project currently includes two cohorts: raltegravir and darunavir. Patients undergo follow-up at 6-month intervals, and adverse events are notified when they are clinically observed. Complete data collection and follow-up procedures for the cohorts are described elsewhere [14]. Patients were asked about the onset of CNS symptoms including headache, dizziness, altered dreams, nightmares, insomnia, anxiety, and depression that were prospectively evaluated and recorded in a standardized form.

A total of 453 HIV-infected patients with a mean age of 45.8 ± 9.2 years were enrolled, of these, 302 (66.7%) were male. Mean CD4 cell count was 378 ± 263 cells/μl, and HIV RNA was 3.01 ± 1.57 log10 copies/ml. A total of 176 (38.8%) were in Centre for Disease Control stage C, and 176 (38.8%) had hepatitis C virus coinfection. In 181 (40.0%), a clinical diagnosis of lipodystrophy was also present. At the time of the analysis, the median follow-up was 23 months (interquartile range 13-30), and 371 (81.9%) were still receiving raltegravir. Complete demographic and therapeutic characteristics of the patients are summarized in Table 1.

Therapy interruptions were caused by patient's choice/low adherence in 15 (3.3%) patients, regimen simplification in four (0.9%), virological failure in 13 (2.9%), death in nine (2.0%), and other reasons in nine (2.1%). Adverse event-related interruptions were recorded in 15 (3.3%) patients, and 17 (3.8%) were lost to follow-up.

During follow-up, 47 (10.4%) patients developed at least one drug-related CNS symptom. Among these, 17 (3.8%) referred headache, 15 (3.3%) depression, eight (1.8%) anxiety, seven (1.5%) dizziness, six (1.3%) insomnia, and one (0.2%) altered dreams.

Among interruptions due to adverse events, four were caused by CNS symptoms. These included headache in two cases, psychomotor agitation and suicide attempt in one case.

At univariate analysis, patients with CNS symptoms were receiving tenofovir more frequently (14.2%) in respect of those without (7.8%, P = 0.03). Patients receiving proton pump inhibitors (PPIs) also had CNS symptoms more frequently when compared with those without PPI (25.9 versus 9.4%, P = 0.006).

At multivariable analysis, after adjustment for age, sex and tenofovir and PPI in turn, the only significant predictors of CNS symptoms were concomitant therapy with tenofovir [odds ratio (OR) 1.9; 95% confidence interval (CI) 1.0-3.5, P = 0.04] or with PPI (OR 3.4; 95% CI 1.3-8.8, P = 0.01).

To our knowledge, this study includes the largest observational cohort of raltegravir-treated patients. Our results confirm raltegravir safety as evidenced by the low proportion of drug discontinuation due to adverse events. Unexpectedly, CNS symptoms, found in more than 10% of patients, represented the second cause of drug discontinuation following muscle adverse events. Of note, concomitant administration of tenofovir has been associated in both univariate and multivariate analysis with CNS symptoms. In pharmacokinetic studies, tenofovir has been shown to increase raltegravir maximum concentration of drug (Cmax) by 64% and area under the concentration-time curve (AUC) by 49% [15] with an unexplained mechanism. Furthermore, CNS symptoms have also been associated with coadministration of PPI, which have been shown to increase raltegravir Cmax by 415% and AUC by 312%, in the case of omeprazole, in healthy volunteers and in HIV-infected patients, but to a lesser extent [16]. This interaction has been explained by increased intestinal absorption [16].

Efavirenz, a nonnucleoside reverse transcriptase inhibitor, has been associated with the onset of CNS side effect in 20-40% of patients. Efavirenz plasma levels seem to predict persistent CNS symptoms and a dose reduction guided by therapeutic drug monitoring (TDM) has been shown to reduce symptoms without loss of virologic response [17,18]. Studies on raltegravir pharmacokinetic and pharmacodynamic are still ongoing, and no conclusive results are available to date. A recent study has evidenced a very high interpatient and intrapatient variability of raltegravir pharmacokinetics in HIV-infected patients on stable HAART, eventually exposing the patients to drug under exposure and increased risk of virological failure [19]. A population pharmacokinetic analysis has further confirmed a very high interpatient pharmacokinetic variability of raltegravir, suggesting a possible relevant role of TDM in some situations [20]. No genetic polymorphism was found to explain the large raltegravir pharmacokinetic variability, except possibly for UGT1A9*3, which needs further confirmation [20].

A correlation between high raltegravir concentrations and the onset of severe insomnia has been found in three patients [19,21]. Furthermore, recent data have shown that raltegravir is present in cerebrospinal fluid (CSF), and that a significant correlation between CSF and plasma concentration exists [22,23].

Thus, our data suggest a possible correlation between high raltegravir plasma concentrations and CNS symptoms. However, a possible limitation of our study is the lack of drug plasma level quantification, given its observational nature.

In conclusion, our results suggest a careful evaluation of patients with psychiatric diseases prior to starting raltegravir and a continuous monitoring of CNS symptoms in clinical practice in those starting the drug. Attention should also be paid to concomitant drugs that can increase raltegravir concentrations. Although TDM is not currently recommended in clinical practice, it could be useful in the management of patients receiving raltegravir with CNS symptoms, especially in those with limited drug options. Further prospective studies are needed to better clarify risk factors, the role of drug interactions, and the clinical significance of CNS symptoms in patients receiving raltegravir.

  iconpaperstack view older Articles   Back to Top