Nelson, Mark R^a; Katlama, Christine^b; Montaner, Julio S^c; Cooper, David A^d; Gazzard, Brian^a; Clotet, Bonaventura^e; Lazzarin, Adriano^f; Schewe, Knud^g; Lange, Joep^h; Wyatt, Christinaⁱ; Curtis, Sue^j; Chen, Shan-Shan^j; Smith, Stephen^j; Bischofberger, Norbert^j; Rooney, James F^j; for the Tenofovir DF Expanded Access Team

Introduction

Tenofovir disoproxil fumarate [tenofovir DF (Viread, Gilead Sciences, Inc., Foster City, California, USA)] is a once daily nucleotide reverse transcriptase inhibitor indicated in combination with other antiretroviral agents for the treatment of HIV-1 infection. Tenofovir DF was approved for marketing in the United States in 2001 and in Europe in 2002. Cumulative patient exposure to tenofovir DF since marketing approval is estimated to be 455 392 patient-years. Tenofovir DF is administered as a 300 mg tablet (Viread) or as a combination tablet with 200 mg emtricitabine (Truvada).

The most common adverse reactions in patients receiving tenofovir DF in clinical trials were mild to moderate gastrointestinal events (e.g., diarrhea and nausea). Laboratory abnormalities observed in these studies occurred with similar frequency in the tenofovir DF and placebo or active control groups, with the exception of lipid abnormalities, which were reported at higher frequencies in the control groups ^[1–5]. In one study, reduction in bone mineral density, but not incidence of bone fractures, was greater in the tenofovir DF group compared with the active control ^[3].

The safety of tenofovir DF for the treatment of HIV infection in clinical practice has not been fully characterized. Nephrotoxicity, including renal insufficiency and Fanconi's syndrome, has been reported infrequently but the incidence, risk factors, and time to resolution remain uncertain ^[6–18]. Small reductions from baseline in creatinine clearance, or increases in serum creatinine, have been reported for tenofovir DF-treated patients in some ^[8,9,17] but not all ^[3,16] studies. This study evaluated the safety profile of tenofovir DF for the treatment of HIV infection in adults, as observed in the tenofovir DF expanded access program (EAP) and via spontaneous safety reports submitted to the Gilead Department of Drug Safety and Public Health through the first 4 years following the drug's introduction to clinical practice.

Methods

Tenofovir DF expanded access program

Prior to commercial availability of tenofovir DF, a global EAP was initiated in March 2001 for HIV-1-infected patients who had failed prior HAART and had limited treatment options. Eligibility criteria included age ≥ 18 years, confirmed laboratory diagnosis of HIV-1 infection, unable to construct a viable treatment regimen without tenofovir DF, no hematological (absolute neutrophil count < 750 cells/μl; platelets < 50 000/μl; or hemoglobin < 8.0 g/dl), renal (serum creatinine ≥ 1.5 mg/dl; serum phosphorus < 2.2 mg/dl) or hepatic dysfunction (alanine and aspartate aminotransferases > 5× upper limit of normal), not pregnant or breast feeding and agreeing to use effective barrier contraceptive method while receiving tenofovir DF. Initial study entry criteria included limitations on CD4 cell counts and HIV RNA, but these criteria were removed shortly after the start of enrollment. The type of EAP implemented in each country (e.g., protocol, named patient program, etc.) varied according to local regulations. Standardized case report forms were used for data collection.

A total of 10 343 patients were enrolled in the Tenofovir DF EAP from March 2001 to March 2004. Participating countries (number of patients enrolled) included Australia (654), Belgium (348), Canada (1761), France (1857), Germany (411), Ireland (12), Italy (553), the Netherlands (256), Portugal (58), Spain (576), United Kingdom (819), and the United States (3038).

Serious adverse events (SAEs) are those adverse events meeting the regulatory definition of ‘serious’: result in hospitalization or prolongation of hospitalization, death, disability, or require medical intervention to prevent permanent impairment. Data for SAEs were requested for all patients and the events were tabulated; selected events that were potentially drug or disease related were further characterized. Adverse event terms were determined by the investigator and coded according to standard terminology. Serum creatinine values were evaluated at screening in all countries and were collected at entry and follow up for patients enrolled in the EAP in France, Germany, and Italy. Changes from baseline in serum creatinine, estimated creatinine clearance (CrCl; using the Cockcroft–Gault method ^[19]), and estimated glomerular filtration rate (GFR; using the Modification of Diet in Renal Disease method ^[20]) were calculated. Each of these methods has been used to approximate renal clearance and they provide slightly different estimates ^[21]. Baseline weights were used for calculating the CrCl, as weights were not recorded in follow-up visits. Risk factors for development of graded creatinine abnormalities (single value, not confirmed) were determined using multivariate logistic regression models using backward elimination (stay criterion, 0.05) and forward selection (entry criterion, 0.05). Variables were included if they had a P ≤ 0.10 in a univariate model.

Postmarketing safety data

Adverse drug reactions (ADRs) for Viread (tenofovir DF), both serious and nonserious, received through spontaneous reporting up to 30 April 2005 were collected and analyzed. Reporting rates (number of cases/100 000 person-years of exposure to drug) for individual serious ADR (SADR) were calculated based upon 455 392 person-years of exposure to tenofovir DF (estimated from sales data).

Although reporting rates do not represent the true incidence of SADR (as postmarketing SADR tend to be underreported and are subject to reporting biases), the pattern of SADR reported in the postmarketing database was compared with SAE reported in the EAP. For patients with renal ADR, renal events were grouped into relevant categories (e.g., renal failure, renal tubulopathy, etc.). For other ADRs of interest, such as bone abnormalities or those associated with antiretroviral therapies (e.g., lactic acidosis or neuropathy), individual adverse events of a similar nature were grouped into a clinically relevant single category of interest (e.g., the single adverse events of ‘bone pain’ or ‘traumatic fracture’ were combined into the single category of ‘bone abnormalities’) according to the number of patients who experienced that type of event. This approach is conservative and provides a clinically relevant combined incidence that is equal to or greater than that of single adverse events.

For patients with renal ADR where serum creatinine values were provided, time to onset, median maximal serum creatinine, and time to resolution of serum creatinine (Kaplan–Meier estimate) were determined. Potential confounding causes of renal disease were assessed for a serial subset of spontaneous and serious, drug-related clinical trial renal/urinary adverse event reports received in a 6 month period between 1 November 2004 and 30 April 2005. Of the 107 cases received, 86 that contained information on the patient's medical history were included in the analysis.

Results

Expanded access program

Baseline demographics and disease characteristics of enrolled patients are displayed in Table 1. Enrolled patients were predominantly male with a median age of 42 years. Patients had moderately advanced HIV infection and were heavily treatment experienced (72% of patients had failed three classes of HIV drugs). The median number of drugs coadministered with tenofovir DF at baseline was three, and the most common antiretroviral agents coadministered with tenofovir DF were lopinavir/ritonavir (51%), lamivudine (48%), didanosine (32%), and abacavir (30%). Mean time on tenofovir DF ranged from 12.8 weeks in the United States to 28.7 weeks in Canada.

Table 1

SAEs were reported for 631 patients (6%), and for 211 of these (2%) the SAE was considered related to tenofovir DF. SAEs were typical of those reported in patients with advanced HIV infection, including pneumonia, fever, bacterial infection, and lymphoma (Table 2). No single SAE was reported in > 1% of patients. SAE designated by investigators to be possibly associated with tenofovir DF included pancreatitis (0.3%), fever (0.1%), pneumonia (0.1%), and acute renal failure (0.1%). SAEs sometimes associated with antiretroviral therapy, and of specific interest with respect to tenofovir DF, are listed in Table 2, where specific adverse events of a similar nature are combined into a single clinically relevant category. Pancreatitis (0.5%) was reported more commonly in patients receiving didanosine (0.7%) or stavudine (1.0%) than in patients not on didanosine (0.3%; P < 0.001) or not on stavudine (0.3%; P < 0.001). The percentage of patients experiencing any type of renal SAE was 0.5%, with renal failure (acute and chronic) reported in 0.3% and Fanconi syndrome in < 0.1%. Other renal events, such as nephrogenic diabetes insipidus, nephritis, and proteinuria, occurred in ≤ 0.1% of patients. Other SAEs of interest, including bone abnormalities, were reported in ≤ 0.1% of patients. Rates of lactic acidosis were significantly higher in patients taking stavudine (0.2%) than in those not taking stavudine (0.1%) (P = 0.04). Mean average increases in CD4 cell counts (Davg) through month 6 were less in patients taking didanosine (17 cells/μl) than in those not taking didanosine (32 cells/μl) (P < 0.001).

Table 2

Values for serum creatinine were collected at entry and follow up for patients enrolled in France, Germany, and Italy. Data are available for 2790 patients at baseline and 1704 patients on study (Table 3). Elevations in serum creatinine at or above grade 1 or grade 2 were observed in 2.2% and 0.6% of patients on study, respectively. Median changes from baseline in serum creatinine, CrCl, and GFR through month 13 were small and nonprogressive. Median change from baseline in GFR at months 3, 6, and 9, where the most data were available, were −1.1, −2.7, and 0 ml/min per 1.73 m², respectively. In a multivariate logistic regression analysis (Table 4), baseline characteristics associated with an increase in serum creatinine of ≥ 0.5 mg/dl (44 μmol/l) from baseline were age [odds ratio (OR), 1.05], CD4 cell count (OR, 0.46), weight (OR, 0.94), and concomitant nephrotoxic medications (OR, 2.40). Baseline characteristics associated with developing a serum creatinine of ≥ 2.0 mg/dl (≥ 177 μmol/l) were serum creatinine (OR, 17.4) and weight (OR, 0.88). The percentage of patients taking a concomitant nephrotoxic medication at baseline (other than an antibiotic) was 16%, and the most commonly used were pentamidine (4%), nonsteroidal anti-inflammatory drugs (3%), acyclovir (3%), valaciclovir (2%), and ganciclovir (1%). Other baseline factors in the model not reaching significance in either analysis are listed in Table 4.

Table 3

Table 4

Postmarketing safety surveillance

Reporting rates for all SADRs received up to 30 April 2005 are given in Table 2.

Postmarketing SADR data are used to detect signals of new safety events not observed in clinical trials and to explore changes in patterns of known safety events. The most common SADRs reported for tenofovir DF were renal, including renal failure, Fanconi's syndrome, and serum creatinine increase. The pattern of renal SADRs reported in the postmarketing safety database is similar to that observed in the EAP, where these three renal events were also the most often reported. Other serious renal events (e.g., nephrogenic diabetes insipidus, nephritis, and proteinuria) were reported infrequently in the postmarketing safety database and the EAP. Other SADRs of interest included pancreatitis and lactic acidosis, both of which have been reported to occur with HIV infection and antiretroviral therapy, and the pattern of these SADRs reflected the pattern seen in the EAP. Bone abnormalities, mitochondrial toxicity, and neuropathy were infrequently reported in both the postmarketing safety database and the EAP.

Additional analyses were performed on postmarketing renal events that also had available creatinine data to allow determine of the time to onset, maximum severity, and time to resolution of serum creatinine abnormalities (Table 5). Concomitant risk factors for postmarketing renal events were also assessed. Of all renal and urinary ADRs and clinical trial events reported between 1 November 2004 and 30 April 2005 (107), information on medical history was available for 86 patients. Of these, possible concomitant risk factors for renal adverse events were reported for 72 patients (84%). The most common risk factors reported were sepsis/serious infection (24%), history of renal disease/baseline renal impairment (24%), late stage HIV (22%), concurrent nephrotoxic drugs (19%), and hypertension (16%).

Table 5

Discussion

The data presented in the current report represents the largest dataset yet analyzed to evaluate the safety of tenofovir DF for the treatment of HIV infection in adults, and includes all safety events reported to the product's manufacturer, Gilead Sciences, Inc., to April 2005, 4 years since the initiation of the tenofovir DF EAP.

The findings from the tenofovir DF EAP and postmarketing safety experience presented here confirm and extend the findings from controlled clinical trials and indicate a favorable safety profile for tenofovir DF in the treatment of adults with HIV infection. The most common SAEs reported in the EAP were those commonly associated with advanced HIV infection and included pneumonia, pancreatitis, fever, and bacterial infections. Serious renal events were reported in 0.5% of patients, and graded increases in serum creatinine occurred in 2.2% of those evaluated. Other SAEs of interest, such as bone abnormalities, were reported infrequently in the EAP and during postmarketing safety surveillance.

Abnormalities in renal function have been reported in up to 30% of patients infected with HIV ^[21–26]. The causes of renal disease in HIV-positive patients include HIV-associated nephropathy, comorbid conditions such as diabetes and hypertension, and side effects of medications used to treat HIV infection or its complications ^[21,22]. Case reports have described nephrotoxicity in HIV-positive patients receiving tenofovir DF, including renal failure and Fanconi syndrome ^[6–18]. Preclinical animal data identified the kidney and bone as organs exhibiting toxicity when animals were administered high doses of tenofovir DF ^[5]. Data from controlled clinical trials with tenofovir DF, which generally excluded patients with a past history of renal disease, have indicated a low incidence of renal adverse events in tenofovir DF treated patients comparable to rates in placebo or active treated control groups ^[1–5]. Small reductions from baseline in estimated CrCl or GFR, or increases in serum creatinine, were noted in tenofovir DF-treated patients in some studies, but not in others ^{[3,8,9,16,17]}. In the present report, the incidence of renal SAEs of any kind in the tenofovir EAP was 0.5%, similar to the 0.3% rate reported for tenofovir DF-treated patients in the smaller cohort of Winston et al. ^[9]. The patients enrolled in the EAP were HIV-positive patients with advanced disease and limited treatment options, the group of patients who are at greatest risk for complications of HIV and its management. The most common serious renal abnormalities observed in this population, and in the postmarketing safety database, were renal failure (0.3%), Fanconi syndrome (< 0.1%), and increased serum creatinine (< 0.1%).

The percentage of patients experiencing elevations in serum creatinine ≥ 2.0 mg/dl (177 μmol/l) on treatment was 0.6% (11/1704), and the percentage experiencing any graded serum creatinine abnormality was 2.2% (37/1704). As noted in some previous reports, we observed small median changes from baseline in serum creatinine, CrCl, and GFR that did not progress through 13 months of follow up.

In a multivariate analysis, baseline risk factors leading to development of increased serum creatinine (at or above grade 1 or grade 2) on study were identified. Significant baseline risk factors were elevated serum creatinine, concomitant use of nephrotoxic medications, increased age, lower weight, and lower CD4 cell count. Preexisting kidney disease, nephrotoxic medications, and older age are known risk factors for renal disease ^[21,22]. These results are consistent with the results from smaller previous cohorts. Gallant et al. ^[8], Antoniou et al. ^[18], and Winston et al. ^[9] all identified baseline serum creatinine, baseline CrCl, or decreased renal function at baseline as risk factors for the development of increased serum creatinine or decreased CrCl in tenofovir DF-treated patients, while Gallant et al. ^[8] identified lower CD4 cell count as a significant risk factor.

Coadministration of lopinavir/ritonavir with tenofovir DF has been suggested as a risk factor for nephrotoxicity in selected case reports ^[6–13], but it was not a risk factor for the development of increased serum creatinine in this study, which enrolled over 5000 patients treated with tenofovir DF and lopinavir/ritonavir. Interestingly, we also observed a somewhat higher rate of renal SAE in patients taking lopinavir/ritonavir (1.1%) compared with those not taking lopinavir/ritonavir (0.5%). However, patients treated with lopinavir/ritonavir also had significantly more advanced HIV disease, with lower median CD4 cell counts (190 cells/μl on lopinavir/ritonavir versus 270 cells/μl not on lopinavir/ritonavir; P < 0.001) and higher median viral loads (4.6 log₁₀ copies/ml on lopinavir/ritonavir versus 4.0 log₁₀ copies/ml not on lopinavir/ritonavir; P < 0.001). When evaluated in a multivariate analysis, lower CD4 cell count was significantly associated with increased serum creatinine, but usage of lopinavir/ritonavir was not. Advanced HIV infection may have been a confounding factor for the development of renal disease in some of the previously published case reports.

The time to onset, maximum severity, and time to resolution of nephrotoxicity in tenofovir DF-treated patients has not been fully described. In the postmarketing safety dataset, the median time to onset for renal SAE was 282 days. For patients with renal events and available serum creatinine data, the median maximum serum creatinine was 2.3 mg/dl (202 μmol/l), and the median times after stopping tenofovir DF to resolution of increased serum creatinine to at or below grades 2 and 1 were 29 and 52 days, respectively. These results compare favorably with those reported by Izzedine et al. ^[7], who described 19 tenofovir DF-treated patients with an onset of renal events an average of 6.9 months (∼ 207 days) after starting tenofovir DF, with a mean maximum serum creatinine of 2.6 mg/dl, and a mean time to complete resolution of 4.7 weeks (∼ 33 days) after stopping tenofovir DF. Information regarding time to onset and resolution can provide important guidance to physicians in the recognition and management of nephrotoxicity in patients taking tenofovir DF.

One limitation of this analysis is that all patients were treated with tenofovir DF. Therefore, the specific contribution of tenofovir DF to the development of nephrotoxicity cannot be precisely determined in each case. Potential concomitant risk factors for nephrotoxicity, other than tenofovir DF, were found in 84% of those with renal SAEs reported to Gilead Sciences from October 2004 to April 2005, including sepsis, history of renal disease, late-stage HIV, and concurrent nephrotoxic medications. Jones et al. ^[16] have reported 90% of the tenofovir DF-treated patients who developed renal dysfunction in their clinic had an alternative cause identified ^[16]. Further epidemiological studies, including a control group not taking tenofovir, are needed to define precisely the contribution of tenofovir DF to renal dysfunction. Nonetheless, identification of patient groups at risk may assist physicians in selection, monitoring, and management of patients to minimize the risk of renal dysfunction in those taking tenofovir DF. Determination of CrCl at baseline, to inform patient selection and dosing, and at regular intervals while on therapy, is important for optimal patient management. Patients with risk factors for nephrotoxicity should have more frequent monitoring.

Other limitations of this analysis include the limited duration of treatment in the EAP, which was a mean of 13 weeks in the United States, 24 weeks in the European Union/Australia, and 29 weeks in Canada. It is possible that events which might occur with longer duration of dosing might not be detected in this cohort. However, it is reassuring that the pattern of renal adverse events was similar in the postmarketing database, which includes patients receiving dosing over a much longer term, and that no new major toxicities were identified in the postmarketing safety analysis. Long-term safety is also being evaluated in Study 903, with favorable safety outcomes reported through 4 years of treatment, and with plans to continue follow up through 7 years ^[27]. Another limitation of this analysis is that the EAP recommended against inclusion of patients with abnormal renal function at baseline. Only 22 patients were included with known baseline CrCl < 50 ml/min. Follow-up data were available for 14 of these patients. None of these patients experienced a renal AE, but further studies in this select population will be needed to adequately address safety and dosing issues. One additional and important limitation of this analysis is that adverse event reporting was voluntary in some of the country-specific EAP and for all of the postmarketing safety reporting, so underreporting is expected. Event rates reported in this paper should be viewed as minimum estimates of the actual event rates.

Other adverse events of interest were also evaluated in this study. Pancreatitis was reported for 0.5% of patients in the EAP, an incidence similar to that reported in patients with advanced HIV disease and in EAP for other antiretroviral agents ^[28,29]. Patients receiving didanosine or stavudine along with tenofovir DF had a higher incidence of pancreatitis than patients not receiving those drugs. Pancreatitis is a known adverse event associated with therapy with didanosine or stavudine ^[28,29]. Patients receiving didanosine in addition to tenofovir DF also had a smaller average mean increase in CD4 cell counts through 6 months on study compared with patients not receiving didanosine. Similar reports regarding reduced CD4 cell response in patients treated with tenofovir DF and didanosine have been published from other studies ^[30,31] and may represent toxicity owing to high levels of didanosine when coadministered with tenofovir DF. Tenofovir DF is known to increase serum levels of didanosine ^[5], and most patients in this study who took both drugs did not use a reduced dose of didanosine, as the interaction between the two drugs was not characterized when the EAP was initiated.

Other adverse events of interest, such as bone fractures and neuropathy, were reported infrequently in both the EAP and the postmarketing safety database. Lactic acidosis was reported in 0.1% of patients in the EAP and was significantly more frequent in patients also taking stavudine. Importantly, no new major unexpected toxicities were observed in the postmarketing safety surveillance program through 4 years of observation.

Tenofovir DF is one of the most widely prescribed antiretroviral agents, and is one of the frequently used antiretroviral drugs in ongoing clinical trials evaluating the treatment and prevention of HIV infection in the developed and developing world. Potent simplified therapies, including those containing tenofovir DF, will be a key factor in assuring successful therapeutic outcomes. A single tablet regimen, containing tenofovir DF, emtricitabine, and efavirenz, has recently been approved and could further facilitate treatment ^[32]. The safety of antiretroviral therapy will be very important in developing world settings, where access to medical care and laboratory monitoring is less available than in the developed world. Recent data on the safety of tenofovir DF from studies in Africa and southeast Asia have been encouraging, reporting similar safety outcomes to studies in the developed world ^[33,34] (S. Walker on behalf of the DART Study Team, personal communication, 26 May 2006).

Further studies are planned to characterize the long-term safety profile of tenofovir DF in a clinical trial through 7 years ^[27], safety in combination with new antiretroviral agents, and safety in developing world populations with differences in demographics, genetics, and access to medical care.

Acknowledgements

The Tenofovir DF EAP Team would like to thank the large number of patients, physicians, and research personnel who have participated in the Tenofovir DF EAP and the clinical research organizations that assisted in the conduct of these studies.

M. Nelson, S. Smith, N. Bischofberger, and J. Rooney participated in study design, data collection, data analysis, data interpretation, and writing of the report. C. Katlama, J. Montaner, D. Cooper, B. Gazzard, B. Clotet, A. Lazzarin, K. Schewe, and J. Lange participated in study design, data collection, data interpretation, and writing of the report. S. Curtis participated in data collection, data analysis, data interpretation, and writing of the report. S. Chen participated in data analysis, data interpretation, and writing of the report. C. Wyatt participated in data interpretation and writing of the report.

Conflicts of interest: B. Clotet is a consultant on advisory boards, speakers bureaus, and in the conduct of clinical trials with Boehringer Ingelheim, Roche, Abbott, Bristol-Myers Squibb, GlaxoSmithKline, Gilead, Tibotec, Merck, and Pfizer. M. Nelson and B. Gazzard have received consultant fees, lecture fees, or grant support from Gilead. C. Katlama has participated in drug advisory boards for Bristol-Myers Squibb, BI, Roche, Gilead, and GlaxoSmithKline. J. Lange has received consulting or lecture fees from Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, Gilead, and GlaxoSmithKline. D.A. Cooper is an advisory board member for Gilead and has received research support from Gilead. A. Lazzarin has received consulting fees, lecture fees, or grant support from Boehringer Ingelheim, Bristol-Myers Squibb, Gilead, Roche, Abbott, and GlaxoSmithKline. J. Montaner has received consultant fees, lecture fees, or grant support from Abbott, Boehringer Ingelheim, Bristol Myers Squibb, Gilead, GlaxoSmithKline, Merck, Pfizer, Roche, Tibotec, and Trimeris. K. Schewe is a consultant, participated on speaker bureaus, and has received research grants from Gilead, Bristol Myers Squibb, and Abbott. S. Curtis, S. Chen, S. Smith, N. Bischofberger, and J. Rooney are employees of Gilead Sciences. C Wyatt declares she has no conflict of interest.

Sponsorship: This study was supported by Parexel International, Ingenix Pharmaceutical Services Inc., and Kendle. Gilead Sciences provided financial support for the tenofovir DF EAP and collected the data used in postmarketing safety analyses.

Additional members of the Tenofovir DF Expanded Access Team: H. Gallais (France), A. Adam (Germany), R. Colebunders (Belgium), M.J. Aguas (Portugal), F. Mulcahy (Ireland), S. Follansbee (USA), S. Walmsley (Canada), H. Tilson (USA), J. Elder (Kendle), M. Wulfsohn, L. Metzler, Li Hsu, A. Cheng and J. Toole (Gilead Sciences).

References