Management of the Adverse Effects of Antiretroviral Therapy and Medication Adherence
- 1CORE Center for Prevention, Care, and Research of Infectious Disease (formerly the Cook County HIV Primary Care Center), Division of Infectious Disease, Department of Medicine, Cook County Hospital, and Rush Medical College, Chicago, Illinois
- 2University of Illinois College of Pharmacy, Chicago, Illinois
- Reprints or correspondence: Blake Max, Pharm.D., CORE Center for Prevention, Care, and Research of Infectious Disease, 2020 West Harrison St., Chicago, IL 60612 (bmax{at}corecenter.org)
Abstract
A commonly cited cause of poor adherence to highly active antiretroviral therapy (HAART) is adverse drug reactions. Short-term adverse effects are potential threats to successful introduction and maintenance of HAART. The long-term toxicities of HAART are still emerging and being defined, as evidenced by the recently described metabolic disorders (i.e., the syndrome of maldistribution, hyperlipemia, glucose intolerance and insulin resistance). With 14 licensed agents in 2000, other agents in common use, and numerous combinations of ⩾3 drugs, awareness and recognition of adverse effects are increasingly important for clinicians and patients. The common adverse drug reactions encountered with HAART, including new agents and their impact on patient adherence, are reviewed. Current strategies to anticipate and mitigate adverse effects are summarized.
New evidence of reduced mortality and morbidity due to HIV disease was reported in 1998. Whitman and Murphy [1] reported a 65% decline in deaths due to AIDS in Chicago from 1995 to 1997, which included a 35% decline among women, a 57% decline among Hispanics, a 47% decline among African-Americans, and a 47% decline among persons with current or past injection drug use. These data show broader populations with significant mortality reductions than do national data [2] and suggest that a system of care and support that gives access to potent antiretroviral therapy can benefit diverse populations of people with HIV disease.
Nonetheless, there are numerous reasons for caution regarding the current state of treatment (as of 2000). In the best of circumstances, the goal of maximal virus suppression is achieved in only 50% of patients [3]. Highly active antiretroviral therapy (HAART) regimens are often complex. Regimens can include numerous pills with frequent dosing and various, sometimes conflicting, food requirements. Adverse events are common and may lead to discontinuation of therapy, dose interruption, and significant reductions in quality of life. Adherence may be compromised because of adverse events, and adherence is increasingly recognized as an important determinant of successful antiretroviral therapy [4].
This article will review the adverse effects of the currently approved agents for the treatment of HIV disease and 2 unapproved agents in clinical use and management strategies to prevent or reduce adverse effects. Consideration will be given to new dosing strategies for existing drugs, new agents, newly observed adverse events associated with HAART (including the so-called metabolic disorders), and the implications of these data for current treatment strategies.
General Principles of HAART, Adverse Effects, and Adherence
Poor adherence has been strongly associated with virological failure of potent antiretroviral therapy by several investigators. In a study of 143 patients in an urban HIV clinic, in which viral load reductions of <500 copies/mL at 6 months were achieved only for 47% of patients, Deeks [5] reported a 15-fold increase in the likelihood of virological treatment failure in the presence of poor adherence (as measured by patient self-reports). Surveys of people receiving HAART have shown that 30% of patients missed doses within the previous 3 days, and adverse effects account for 10%–15% or more of those discontinuations of treatment [6]. Because adverse events are common with all available antiretroviral agents, it is critical to anticipate, recognize, and manage them when providing primary care for HIV-infected patients. Because several recent studies and initiatives address adherence [7–11], only the association between adherence and adverse events will be briefly considered here.
Adherence requires a reliable, trusting patient-physician relationship, and it results from a negotiation in which the patient is an active participant [12, 13]. With regard to adverse events, patients should be informed of potential side effects during consideration of the first regimen, options for subsequent regimens (i.e., sequencing), and possible management strategies in case of adverse events. Adverse events may influence the initial treatment decision. For example, medications associated with a high risk of diarrhea may be less desirable for patients who are experiencing diarrhea or HIV-related wasting. In view of the unforgiving nature of HIV infection, it is essential to establish the patient's readiness for HAART before the first prescription, including knowledge of and treatment for possible adverse events [14]. Similarly, adherence is not a “one-shot” problem; repeated monitoring, education, and intervention are needed to ensure durable adherence, including close monitoring for adverse events. Adherence requires an individualized approach, since no single strategy will work for all patients. Adherence improves with recruitment of family, friends, or peer support staff [15].
Potentially beneficial interventions to improve adherence are summarized in table 1. In a recent study, an off-site clinical pharmacist referral clinic was associated with improved virus suppression [17]. A similar strategy using an on-site clinical pharmacist has been in place at the Cook County HIV Primary Care Center (now the CORE Center for Prevention, Care, and Research of Infectious Disease, Chicago) for the past 2 years. The clinical pharmacist is available for patient education, support for readiness before HAART, adherence follow-up and monitoring, discussion of adverse events, pill and food strategies, and education of family and friends. In addition, the clinical pharmacist provides real-time consultation services to the clinicians in the clinics. As noted in table 1, other members of the health care team can provide additional support for this consultative service, such as nurses, case managers, peer educators, and community volunteers [18].
Table 1
Possible strategies to improve adherence to antiretroviral therapy.
Antiretroviral Therapy and Common Adverse Effects
Guidelines for antiretroviral therapy that contain current options for antiviral therapy and their common side effects have recently been updated [3]. Although some adverse effects are common to the entire class, such as rash for the nonnucleoside reverse transcriptase inhibitors (NNRTIs), there are important distinctions for each drug within each class that require the physician and health care team to review and monitor every drug with the patient. Below, possible strategies for management and reduction of adverse effects are reviewed. Note that in some cases, optimal management requires that treatment be discontinued to eliminate or reduce the effect. In the text and tables in this article, data on the incidence of adverse effects are taken from previously reported data from the pivotal approval trials of each agent, as indicated in the package insert.
The patient's medical diagnoses and treatments may influence the management of adverse effects. In some cases, toxicity may be intolerable in the presence of another drug with similar toxicity. For example, both zidovudine and ganciclovir may cause bone marrow suppression, and their concurrent use requires careful monitoring. This example illustrates that it is essential to take an individualized, case-by-case approach to drug regimen selection and assessment of adverse effects.
Nucleoside Reverse Transcriptase Inhibitors (NRTIs)
NRTIs have been a cornerstone of treatment for HIV infection since zidovudine became available in 1986. With the recent approval of abacavir in 1998 by the US Food and Drug Administration (FDA), there are now 6 NRTIs for the treatment of HIV infection. NRTIs are nucleoside analogues (zidovudine and stavudine are thymidine analogues, lamivudine and zalcitabine are cytosine analogues, didanosine is an adenosine analogue, and abacavir is a guanosine analogue), and although they all have the same mechanism of action, many adverse effects of each agent are unique. NRTIs inhibit the HIV reverse transcriptase enzyme and terminate proviral DNA chain elongation. NRTIs must be phosphorylated intracellularly by specific host cell enzymes to carry out their critical function. Unfortunately, NRTIs inhibit not only viral reverse transcriptase but also cellular DNA polymerases (particularly mitochondrial polymerase γ); mitochondrial toxicity may explain some of their long-term toxicity [19–21].
Results of a recent clinical study of the combination of 3 NRTIs (zidovudine, lamivudine, and abacavir) compared favorably with the treatment standard of 2 NRTIs plus a protease inhibitor (PI) [22]. Although these data are promising, more data are necessary on the durability of triple NRTI treatment and the potency in patients with high viral loads (>100,000 copies/mL). NRTI-based regimens also have a lower pill burden than PI-based regimens. The choice of NRTIs is based largely on drug resistance, tolerability, and the individual patient. Although patients seldom inquire about drug resistance, drug adverse effects are often a focal point of patient concern. Clinicians should be knowledgeable about the adverse effects of antiretrovirals, so they can both keep the patient well informed and appropriately manage and monitor potential adverse effects. With the selection of NRTIs available, if a patient is suffering from adverse effects of one NRTI, then it would be prudent to change it to another that would conceivably be better tolerated [23].
There are few adverse effects that characterize NRTIs as a class. However, lactic acidosis with severe hepatomegaly with steatosis, although uncommon, has caused deaths in patients receiving treatment with nucleoside analogues [24–26]. The mechanism, which has yet to be completely elucidated, appears to be mitochondrial toxicity [27]. Clinicians also need to be aware that many adverse effects of NRTIs are dose-related, and dose adjustments are often necessary for patients with organ dysfunction. For example, the only NRTI that has data to suggest a dose adjustment in hepatic failure is zidovudine (recommended dose reduction of 50%) [28]. For patients with renal dysfunction, particularly those undergoing hemodialysis, dose reduction is required for all NRTIs except abacavir [29]. Common adverse effects of NRTIs are noted in table 2, and dosage alterations of NRTIs for patients with renal insufficiency are noted in table 3.
Table 2
Common adverse effects of nucleoside reverse transcriptase inhibitors (NRTIs).
Table 3
Dosage modifications of nucleoside reverse transcriptase inhibitors (NRTIs) for patients with renal insufficiency.
Zidovudine. Over the years, there has been a decrease in the recommended dose of zidovudine from 1500 mg/day to 600 mg/day, which has improved tolerability. The most severe adverse effect is bone marrow suppression, which causes anemia and/or neutropenia, but the most common adverse effects are nausea, malaise, myalgias, insomnia, and headache [30–32]. Bone marrow suppression appears to be more common in those patients with advanced disease and is related to the dose and the duration of treatment. In patients with CD4+ cell counts >100 cells/mL, hematologic effects occur in 2%–14% of patients; however, the incidence is much greater among patients with CD4+ cell counts <100 cells/mL [31]. Data from the AIDS Clinical Trials Group showed that the incidence of anemia and neutropenia associated with zidovudine ranged from 1% to 31%, depending on the stage of disease and dose (in most of these early studies, the dosage was 1200–1500 mg/d) [32, 33]. Nowadays the dose of zidovudine is much lower, thus zidovudine-induced bone marrow suppression has become less frequent. Anemia can occur as soon as 1–2 months after zidovudine is started but is more likely to develop after 2–4 months of therapy [30]. Zidovudine-induced anemia appears to result from thymidine triphosphate deficiency leading to inhibition of stem cell maturation [34]. Macrocytosis is also frequently observed in patients taking zidovudine, so much that elevated mean corpuscular volume (MCV) increased 20% or more from pretreatment values in 96% of patients in one study [30]. This phenomenon of zidovudine-induced macrocytosis is not necessarily associated with anemia and not due to vitamin B12 or folate deficiencies [35]. Some patients have developed thrombocytosis while taking zidovudine, including patients who previously were thrombocytopenic [31].
Although zidovudine has a multitude of CNS side effects, the 2 most common are headache and nausea (occurring in 42% and 46%, respectively, of patients); however, these side effects have occurred in patients receiving 1500 mg/d [31], and the incidence is lower when the dosage is 600 mg/d. Nausea is also the most common gastrointestinal side effect and is more likely to occur in patients with advanced disease. Myalgia has been described in ∼8% of patients receiving zidovudine treatment [36]. Myopathy has also been reported and can be severe, comparable with a polymyositis syndrome [37]. Zidovudine-induced myopathy usually occurs after 6–12 months of therapy and is characterized by myalgias, muscle tenderness, weakness, and elevated serum concentrations of muscle enzymes (creatine kinase and lactate dehydrogenase). Symptoms usually resolve upon discontinuation of zidovudine treatment; however, resolution may take >2 months. Most cases are related to skeletal muscle, but 1 study did suggest that zidovudine and other NRTIs might cause cardiomyopathy [38]. Hyperpigmentation of fingernails and toenails is also associated with zidovudine treatment [36, 39].
It is recommended that the daily dose of zidovudine be reduced by ∼50% for patients with creatinine clearances <25 mL/min [29]. Removal of zidovudine by dialysis is minimal; however, as for other NRTIs for which doses must be adjusted for hemodialysis, the dose should be administered after completion of dialysis. Although zidovudine's primary metabolite (GZDV) accumulates in the presence of renal dysfunction, this is not of clinical significance since GZDV is inactive and does not add to zidovudine's toxicity.
The hemoglobin level and hematocrit should be monitored closely when zidovudine treatment is initiated, particularly in those patients with advanced disease, those who have a history of anemia, and those taking other myelosuppressive drugs (e.g., ganciclovir, hydroxyurea, and pyrimethamine). Although zidovudine's adverse effects are dose-related, patients who have developed toxicity should not have their dose reduced to improve tolerability. This action could lead to subtherapeutic levels of zidovudine and potentially to HIV drug resistance. Granulocyte colony-stimulating factor and erythropoietin have been used to correct bone marrow suppression. Erythropoietin is recommended for use if the serum erythropoietin level is <500 IU/mL [40]. Granulocyte colony-stimulating factor has been shown in a number of studies to resolve neutropenia caused by zidovudine [41]. If the patient is unable to tolerate adverse effects of zidovudine, it may be more appropriate to switch treatment to another NRTI, such as stavudine or abacavir. Headache and myalgias can be treated symptomatically with analgesics, such as acetaminophen or nonsteroidal anti-inflammatory drugs. Most of the adverse effects of zidovudine (e.g., headache, malaise, nausea, and myalgia) are usually transient and diminish after the first few weeks of therapy [42]. It is important to educate patients about these adverse effects and their appropriate management.
Didanosine. Didanosine is the only NRTI for which food has a significant effect upon absorption. Because didanosine is extremely acid labile, each tablet is buffered with calcium carbonate and magnesium hydroxide. Didanosine plasma concentration (area under the curve, AUC) decreased 55% when administered up to 2 hours after a meal [43]. Therefore, didanosine should be administered before meals and on an empty stomach. The magnesium content may account for the gastrointestinal adverse effects, such as diarrhea, which occurs in 15%–20% of patients.
Didanosine is partially metabolized in the liver, and one of the metabolites is uric acid. Approximately 60% of didanosine is eliminated via the kidneys as unchanged drug; therefore, it is recommended to administer one-quarter of the standard dose every 24 h if patients are undergoing hemodialysis [29]. Because hemodialysis removes ∼20% of total body stores of didanosine, the dose should be administered after dialysis, and no supplemental doses are necessary. Dose reduction is necessary for patients with renal impairment to prevent magnesium accumulation. The dose of didanosine is also based on body weight, and clinicians should ensure that the dose is adjusted accordingly.
Each dose of didanosine should contain at least 2 tablets to provide the necessary buffering capacity. Didanosine is also available as a buffered powder for solution; however, a larger dose is required because the bioavailability of powder is 20%–25% less than that of the tablets. The buffered powder has fallen out of favor because it is associated with more gastrointestinal adverse effects than are the tablets. Didanosine has an intracellular half-life of 8–43 h; therefore, once-daily dosing is frequently used in clinical practice and recently received FDA approval [44].
The most serious dose-limiting adverse effects associated with didanosine are peripheral neuropathy and pancreatitis, which have occurred in 10%–20% and 5%–9%, respectively, of patients in clinical trials [43]. Data from recent clinical trials have shown a much lower incidence of peripheral neuropathy (2%–3%) [45].
Like peripheral neuropathy, pancreatitis occurs more frequently in patients receiving treatment with higher doses of didanosine. Didanosine-associated pancreatitis may be mild, presenting as mild abdominal pain, but several cases of fatal pancreatitis have been reported [46, 64]. Patients at risk for pancreatitis are those with advanced disease, who have a history of pancreatitis and/or alcohol abuse, who have hypertriglyceridemia, who are being treated with high doses of didanosine and/or other pancreatotoxic medications, and/or who have renal impairment [47]. These patients should have pancreatic function monitored carefully. Many clinicians avoid treating former alcoholics with didanosine, although the available data suggest that careful use with close monitoring allows didanosine to be administered in this setting. Pretreatment testing of amylase and lipase levels is of unproven value, but it is a reasonable precaution to screen for subclinical chronic pancreatitis. Didanosine treatment should be discontinued if signs and symptoms of acute pancreatitis are evident. This may be particularly important with combination therapy that includes a PI, which has been associated with metabolic abnormalities, including elevated triglyceride levels. Symptoms usually resolve, and the serum amylase level returns to normal within 2 weeks after didanosine treatment is discontinued [48].
Diabetes mellitus occurred in patients receiving didanosine before the introduction of HAART [49]. Some patients who developed hyperglycemia also had elevated serum amylase and lipase levels; therefore, patients who develop hyperglycemia during didanosine therapy should also be monitored closely for signs of pancreatitis.
The most common adverse effects associated with didanosine are gastrointestinal; however, these symptoms should be treated with caution because some could be signs of evolving pancreatitis. Diarrhea can be ameliorated with antidiarrheals, such as loperamide. Pancreatitis and peripheral neuropathy usually require discontinuation of therapy, and patients with pancreatitis should not be rechallenged with didanosine. Symptoms of peripheral neuropathy could be treated with appropriate agents (see stavudine section). Of note, results from clinical trials on the combination of stavudine and didanosine did not show that the incidence of peripheral neuropathy was higher in association than with either drug alone [50]. Once-daily dosing of didanosine also has not been associated with any increase of adverse effects.
Zalcitabine. Zalcitabine is used less frequently because the dosing schedule is unfavorable, it has overlapping resistance with other nucleotides, and it has significant toxicity: phase I/II studies showed that peripheral neuropathy was the most significant dose-limiting toxicity that required discontinuation of treatment [51]. Studies have shown that neuropathy was dose-dependent [52]; however, 23% of patients developed peripheral neuropathy at the approved dosage of 2.25 mg/d.
Although no clear predictors of zalcitabine-induced peripheral neuropathy have been identified, some studies have shown that low CD4+ cell counts (<50 cells/mL), pre-existing neuro-pathy, nutritional deficits, and high alcohol consumption increase the risk [52, 53, 63]. Clinicians must monitor patients closely to detect symptoms early and to modify treatment in order to minimize its negative impact on quality of life. Other adverse effects include ulcerative stomatitis, which was described in 3%–17% of patients from various studies [54, 55] and unlike other NRTIs, zalcitabine is associated with a maclopapular rash involving the trunk and extremities [56].
Zalcitabine is primarily eliminated via the kidneys, and dose adjustment is required for patients with diminished renal function (see table 3).
Stavudine. Stavudine is structurally similar to zidovudine but has a different adverse effect profile. The primary dose-limiting toxicity is sensory peripheral neuropathy with symptoms similar to the neuropathy associated with didanosine and zalcitabine. The incidence of sensory peripheral neuropathy is dose-related; the highest frequency is associated with dosages of 4–8 mg/kg/d, which are much higher than the recommended dosage of ∼1 mg/kg/d [57]. One-year rates of peripheral neuropathy associated with dosages of 0.1, 0.5, and 2 mg/kg/d were 6%, 17%, and 37%, respectively, whereas the cumulative dose was not associated with development of peripheral neuropathy [58]; in 1 study, the incidence of stavudine-induced neuropathy was 55% [59]. Symptomatic patients typically have tingling, burning, and pain in the lower extremities, especially at night. Symptoms usually begin to diminish and/or resolve within 1–9 weeks after stavudine therapy is discontinued. However, some patients who developed peripheral neuropathy have persistent symptoms, despite discontinuation of treatment. It is not clear whether these symptoms are due to permanent toxicity from stavudine or an entirely different etiology. It can be difficult to distinguish between HIV-induced peripheral neuropathy and drug-induced peripheral neuropathy. The diagnosis is often made on clinical grounds, and, if it is drug-induced peripheral neuropathy, the symptoms usually resolve when therapy is discontinued.
In a pivotal study on stavudine, 63% of patients had grade 1–2 peripheral neuropathy that resolved after discontinuation within a median time of 17 days [60]. Patients and clinicians should be informed that symptoms may intensify for several weeks before improving. Some patients may tolerate reintro-duction of stavudine therapy at a lower dose, but approximately one-half will have a recurrence of peripheral neuropathy. In a phase I study of patients who resumed stavudine treatment at one-half the dose, 60% were able to tolerate therapy for an average of 9 months [62]. However, what is not clear is if dose reduction still provides appropriate antiviral activity. Like zalcitabine, risk factors for peripheral neuropathy are a CD4+ cell count <100 cells/mL, history of peripheral neuropathy, use of other neurotoxic agents, excessive alcohol intake, and vitamin B12 deficiency [53, 57]. Although these factors are not a contraindication for prescribing stavudine, close monitoring is essential to prevent symptomatic progression.
Stavudine has little hematologic toxicity [57, 65]. However, macrocytosis has been observed in 68% of patients in phase I and II trials with changes in MCV typically less than those observed in zidovudine treated patients [57]. Other adverse effects associated with stavudine that may or may not be dose-related are asthenia, headache, malaise, insomnia, abdominal pain, and modest increases in liver transaminase levels [61].
To reduce the incidence of adverse effects of stavudine, the dose requires adjustment based on weight and renal function (see table 3).
Treatment of stavudine-induced peripheral neuropathy is often frustrating for both the patient and the clinician. Agents with partial efficacy include analgesics and topical capsaicin, which may be helpful in a limited number of patients [66]. Tricyclic antidepressants such as amitriptyline, desipramine, or nortriptyline are frequently used to treat peripheral neuropathy and in controlled studies have shown some efficacy for diabetic neuropathy [67]. Amitriptyline is commonly used at a starting dosage of 25 mg once a day at bedtime (maximum dose, 150 mg). Tricyclic antidepressants have a number of unwanted anticholinergic side effects, such as sedation, dry mouth, and orthostatic hypotension, which can be exacerbated if patients are taking PIs due to inhibition of the metabolism of tricyclic antidepressants [68]. Anticonvulsants, such as carbamazepine and phenytoin, may also relieve symptomatic neuropathic pain. However, both of these agents must be used cautiously, if at all, in patients with HIV infection. Because of their enzyme- inducing properties, they are not recommended to be prescribed for patients taking PIs or NNRTIs [3].
Gabapentin, another anticonvulsant, has received recent attention as therapy to relieve the symptoms of peripheral neuropathy [69]. If peripheral neuropathy is disabling, narcotics may be required to control pain. Other agents that are receiving attention include recombinant nerve growth factor, the anticonvulsant lamotrigine, and alternative therapies like acupuncture [70]. The antiarrhythmic mexiletine has been shown to be no better than placebo for relieving symptoms of peripheral neuropathy [71]. The clinician should inform patients in advance of the possibility of neuropathy and its symptoms and closely monitor the patient for those symptoms.
Lamivudine. Lamivudine has relatively few adverse effects and is well tolerated. Unlike other nucleoside analogues, lamivudine is not likely to be incorporated into mitochondrial DNA [72].
The most common adverse effects of lamivudine that were reported in dose-ranging clinical studies were diarrhea, malaise, fatigue, headache, and sleep disturbances [73, 74]. However, there are anecdotal reports of a variety of nonspecific CNS adverse effects that recur on rechallenge [75]. Lack of placebo-controlled safety data in most studies of lamivudine makes determination of the true incidence of adverse effects difficult. Hematologic toxicity and peripheral neuropathy are rarely associated with lamivudine [76]. Studies evaluating lamivudine for the treatment of hepatitis B have found that tolerability was identical between treated and placebo groups (albeit at a lower dose than that used for treatment of HIV infection) [77]. In phase III clinical trials of lamivudine in combination with zidovudine, the incidence of adverse effects was no higher than that seen in the zidovudine monotherapy arm [78].
Lamivudine is primarily excreted via the kidneys; therefore, dose modification is required for patients with renal dysfunction. It is not known whether lamivudine is removed by hemodialysis or by peritoneal dialysis, and data are lacking to make dosing recommendations. In the absence of such data, the dosage should be based on the creatinine clearance with administration of daily doses after dialysis (see table 3). Lamivudine (150 mg) has been combined with zidovudine (300 mg) as a single tablet (Combivir; Glaxo Wellcome, Research Triangle Park, NC), which is a potent double nucleoside combination with convenient administration (1 pill twice daily). Lamivudine also has a long intracellular half-life (∼15 h), which permits once-daily investigational dosing strategies.
Abacavir. Abacavir is a novel guanine-based NRTI that was approved for use in October 1998. Abacavir has excellent potency when part of a triple NRTI regimen, (compared with a regimen of 2 NRTIs and 1 PI) [22], is administered twice daily, and has low selectivity for mammalian mitochondrial DNA synthesis. Abacavir is not metabolized by cytochrome P-450 isoenzymes; therefore, drug interactions are not expected between abacavir and other agents metabolized by these enzymes. Abacavir is not eliminated via the kidneys, and dose reduction is probably unnecessary for patients with renal dysfunction. Abacavir is primarily metabolized by alcohol dehydrogenase and glucuronyl transferase. Preliminary studies have not shown a significant drug interaction between abacavir and alcohol [79].
Clinical trials have shown that the most commonly reported adverse effects of abacavir are fatigue, asthenia, rash, headache, nausea, vomiting, and diarrhea [79–81]. Because abacavir was used in combination with other antiretrovirals in clinical trials, it is not clear whether these adverse effects are strictly due to abacavir or partially or wholly attributable to other antiretrovirals given concomitantly.
The most clinically significant adverse effect of abacavir is a hypersensitivity reaction characterized by “flulike” symptoms, which is a systemic reaction that typically involves multiple organ systems. The most common symptoms are fever, rash, and gastrointestinal complaints and 85% of cases occur in the first 6 weeks [79]. Diagnosis requires >2 of the following symptoms: fever, rash, malaise, nausea, vomiting, diarrhea, arthralgias, abdominal pain, dyspnea, and/or paresthesias. Uncommonly, respiratory complaints, such as cough and pharyngitis have been observed. Abacavir treatment is often started with other antiretroviral therapy that can also cause similar or identical symptoms, which can complicate the diagnosis. If the patient is believed to have developed the hypersensitivity reaction, abacavir treatment should be discontinued, and patients should not be rechallenged. If abacavir is the cause of the symptoms, they usually resolve within 24–48 h after discontinuation of therapy. If a patient is rechallenged, severe symptoms can recur within hours, which may include life-threatening hypotension, anaphylaxis, and death [79, 82].
Approximately 5% of patients involved in clinical studies of abacavir have developed the hypersensitivity reaction. The symptoms generally occur within the first 6 weeks after the start of abacavir treatment, although symptoms may occur at any time during therapy. Physical findings often include lymphadenopathy, mucus membrane lesions, and rash. It should be noted that hypersensitivity reactions have occurred without rash, and rash has occurred without other symptoms of the reaction. It is extremely important for medical providers to be aware of this reaction, to diagnose it properly, and to inform and/or educate patients about this reaction. Patient education is critical but so is education of the medical community: for example, emergency physicians will undoubtedly see patients with the hypersensitivity reaction.
The hypersensitivity reaction requires discontinuation of abacavir treatment. After discontinuation of abacavir therapy, patients can be treated with analgesics, antihistamines, anti-diarrheals, and/or fluids to relieve symptoms. Abnormal results of laboratory studies have included elevated transaminase levels, increased creatine phosphokinase or serum creatinine level, and lymphopenia; these are usually mild and self-limited. Nonetheless, careful monitoring is necessary for patients with underlying hepatic or renal dysfunction.
Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTIs)
With approval of nevirapine in 1996 and delavirdine in 1998 by the FDA and impressive results from clinical trials of efavirenz [83], NNRTIs have gained a definitive place in the treatment of HIV infection. NNRTIs possess the same mechanism of action as NRTIs, although they are quite different in molecular structure. NNRTIs inhibit HIV replication by binding to a specific nonsubstrate hydrophobic pocket of HIV type 1 (HIV-1) reverse transcriptase [84]. This binding site is distinct from the NRTI binding site but also inhibits viral replication. The NNRTI binding site is located close to the reverse transcriptase catalytic site; allosteric binding inactivates HIV-1 reverse transcriptase by altering its conformation.
The NNRTIs are characterized by similar pharmacokinetic parameters, including excellent oral absorption, a long half-life, and metabolism by the cytochrome P-450 enzyme system. NNRTIs have the potential for serious drug interactions due to cytochrome P-450 enzyme induction (nevirapine), inhibition (delavirdine), or both (efavirenz). NNRTIs are also associated with the rapid development of viral resistance, especially if they are not part of a maximally suppressive regimen; resistance requires only a single amino-acid codon substitution. For example, the K103N mutation is associated with resistance to nevirapine, delavirdine, and efavirenz. Because of the high degree of cross-resistance between the currently available NNRTIs, it is doubtful that a second NNRTI will provide any virological benefit after resistance has developed. However, HIV-1 strains resistant to NRTIs often remain susceptible to NNRTIs. The NNRTIs are also characterized by their adverse effect profiles; all 3 can cause rash and can increase transaminase levels. The adverse effects of NNRTIs are noted in table 4.
Table 4
Common adverse effects of nonnucleoside reverse transcriptase inhibitors (NNRTIs).
Nevirapine. Nevirapine has an advantageous pharmacological profile that includes convenient administration (1 tablet twice daily), excellent bioavailability, good CNS penetration, and a prolonged half-life (25–30 h), which makes once-daily dosing possible [85, 86]. Nevirapine also induces its own metabolism; maximal induction occurs within 2–4 weeks after initiation of treatment, which is why the recommendation is to give 50% of the dose for the first 2 weeks of therapy [87]. This method of dose escalation has improved tolerability of nevirapine by decreasing the incidence of rash. As with other antiretrovirals, nevirapine is associated with a number of drug interactions. Nevirapine is considered a moderate inducer of the cytochrome P-450 3A4 isoenzyme; however, few drug interaction studies have been conducted. Nevirapine has been found to decrease indinavir AUC; however, there was large between-patient variability. Because of this finding, it is recommended to increase indinavir to 1000 mg q8h if it is given concomitantly with nevirapine [3]. Nevirapine has also been shown to induce symptoms of opiate withdrawal in patients receiving methadone treatment [88]. Careful consideration of drug interactions is needed, particularly for those drugs that are metabolized by the cytochrome P-450 3A4 isoenzyme.
Nevirapine is well tolerated; rash is the most common adverse effect that requires discontinuation of treatment. Other common adverse effects include increased transaminase levels, headache, diarrhea, and nausea [87, 89, 90]. These effects usually diminish in a few weeks and are more prevalent at initiation of treatment. Nevirapine can be taken with food, which may alleviate nausea. Mild adverse effects can be treated symptomatically; however, if adverse effects persist, particularly rash, it may be necessary to change to another NNRTI or another antiretroviral regimen. Data is lacking regarding cross-sensitivity between NNRTIs. One report had 9 of 19 patients who discontinued nevirapine due to rash and switched to efavirenz also developed a mild-moderate rash, however, only two patients discontinued therapy because of rash [98].
The rash associated with nevirapine occurred in up to 32% of patients in 1 clinical study [89]; however, the overall incidence of rash is ∼17% [91]. The rash tends to occur early, usually in the first 6 weeks of therapy. Patients should be instructed that if any rash develops during the 2-week induction period, the dose should not be escalated until the rash resolves. There are no known risk factors for the development of the nevirapine-associated rash. In general, most nevirapine-associated rashes are mild to moderate in severity; they are maculopapular, erythematous cutaneous eruptions, with or without pruritus [87]. The rash typically appears on the trunk, face, and extremities. Severe rashes have occurred that require discontinuation of treatment (overall incidence, 7%) [87]. Life-threatening and rare fatal skin reactions, such as Stevens-Johnson syndrome and toxic epidermal necrolysis, have occurred in patients treated with nevirapine [92]. The incidence of Stevens-Johnson syndrome across all trials was 0.5% [91].
Elevation of transaminase levels has been observed, but increases occurred at rates similar to those observed in non-nevirapine-containing treatment arms [91]. However, cases of severe acute hepatitis associated with nevirapine have been reported in the literature [93]. For patients with pre-existing liver disease, close monitoring for the first 6 months is recommended.
Management of the nevirapine-associated rash should be based on the type and severity of symptoms. Grades 1–2 (mild to moderate) rashes are usually self-limited, and the rash usually resolves without discontinuing nevirapine treatment. Symptoms may be relieved with oral antihistamines or topical steroid cream. Some investigators have postulated that prophylaxis with antihistamines or corticosteroids during the induction phase may reduce the incidence of rash, but this hypothesis has not been formally evaluated.
Rash accompanied with fever, severe pruritus, vomiting, severe gastrointestinal manifestations, oral lesions and/or ulceration, blistering, or muscle or joint pain requires discontinuation of nevirapine therapy. Rechallenge with nevirapine in these cases should not be performed. It is important for the clinician to instruct patients to call immediately if rash or other constitutional symptoms occur so that they can be properly evaluated and treatment can be discontinued, if necessary, to prevent possible escalation to Stevens-Johnson syndrome.
Delavirdine. Delavirdine, like nevirapine, has excellent bio-availability and can be taken with or without food. However, delavirdine is different from nevirapine because it has poor penetration into the CSF, inhibits the cytochrome P-450 3A4 isoenzyme, approved for 3-times-a-day dosing, and has a large pill burden (12 tablets per day) [94]. The bioavailability of delavirdine increases by ∼20% when a slurry is prepared, which is accomplished by allowing delavirdine to disintegrate in water [94].
The most common reported adverse effects of delavirdine are rash (18% of cases), nausea (7%), diarrhea (4%), headache (6%), fatigue (4%), and increases in transaminase levels (5%) [94, 95]. Delavirdine is considered to be well tolerated. The rash, which has an incidence that has been observed to be as high as 44%, is described as erythematous, maculopapular, mildly pruritic, and mild to moderate in intensity [96]. It typically develops 7–15 days after the onset of drug therapy and, like the nevirapine-associated rash, often resolves without discontinuing therapy. The rash has occurred in all dosing groups and is neither dose-dependent nor dose-limiting (only 4.3% of patients discontinued treatment because of rash) [94]. In contrast to therapy with nevirapine, dose titration is not necessary, nor does it significantly reduce the incidence of rash. It has been observed that continuing delavirdine treatment without interruption does not lead to a higher incidence of complications or delay resolution, as compared with an alternative approach of discontinuing delavirdine treatment until the rash has resolved, followed by reintroducing delavirdine [96].
Delavirdine is also different from nevirapine because it inhibits the cytochrome P-450 3A4 isoenzyme [97]. Increased plasma concentrations, which are potentially harmful, are expected for a number of drugs that are metabolized by the cytochrome P-450 3A4 isoenzyme (e.g., astemizole, midazolam, triazolam, alprazolam, cisapride, and dihydropyridine calcium channel blockers) [94]. The inhibitory effects of delavirdine on enzymes can also be advantageous when it is used in combination therapy with PIs. Other significant drug interactions include antacids, which decrease delavirdine absorption. Therefore, if delavirdine and didanosine are given concomitantly, administration should be separated by at least 1 h. In addition, both rifabutin and rifampin are not recommended to be given concomitantly with delavirdine because they substantially decrease delavirdine plasma concentrations (by 80% and 96%, respectively) [94].
The management of the delavirdine-associated rash is the same as described for the nevirapine-associated rash. Although the incidence of severe rash is lower than that associated with nevirapine, any patient who develops severe rash with fever, blisters, or other constitutional symptoms should discontinue delavirdine treatment and see their medical provider.
Efavirenz. Efavirenz received FDA approval in September 1998. Forty-eight week data from on-treatment and intent-to-treat analyses were similar for efavirenz in combination with zidovudine and lamivudine and superior compared with indinavir, zidovudine, and lamivudine [83]. Efavirenz is recommended as a preferred agent for the initial treatment of HIV infection [3]. Efavirenz also has a favorable pharmacokinetic profile that includes excellent absorption, no food restrictions, and a long half-life that allows for once-daily dosing [98]. However, efavirenz is metabolized by the cytochrome P-450 3A4 isoenzyme and is a mixed inhibitor and inducer, which makes potential drug interactions difficult to predict. Because malformations were observed in fetuses from efavirenz-treated primates, efavirenz is contraindicated in pregnancy, and women of child-bearing age should use adequate contraception while receiving efavirenz treatment [98]. The adverse effects of efavirenz are similar to those of other NNRTIs, but it also commonly causes unusual CNS side effects.
CNS adverse effects occurred in 52% of patients receiving efavirenz treatment in clinical trials [98]. Symptoms included dizziness, confusion, abnormal thinking, impaired concentration, agitation, hallucinations, nightmares, vivid dreams, and euphoria. Of patients in clinical trials, 2.6% had to discontinue efavirenz treatment because of CNS adverse effects.
Skin rash occurred in 10% of patients in controlled clinical trials [98]. Severe rash that included symptoms of blistering or ulcerations occurred in 1% of patients. The rash is most likely to occur in the first 2 weeks of therapy. Efavirenz also can increase transaminase levels, which require monitoring particularly in patients coinfected with hepatitis B or C virus.
Data supported by pharmacokinetic studies will be important to determine significant drug interactions with efavirenz. Fiske et al [99] have shown that efavirenz decreases the indinavir AUC by 31%; when coadministered, the dosage of indinavir should be increased to 1000 mg q8h. Efavirenz also significantly decreases saquinavir AUC by 62% and therefore should not be coadministered with this agent [99] and amprenavir AUC 36%, requiring dose modification [3]. Efavirenz decreases rifabutin AUC by 38% when coadministered; it is recommended that the dosage of rifabutin be increased to 450 mg once a day. Recent data have also shown that efavirenz decreases methadone plasma levels, which has resulted in symptoms of opiate withdrawal [100].
CNS adverse effects usually appear within the first few days of therapy and generally resolve after 2–4 weeks. The administration of efavirenz is recommended once daily at bedtime to minimize the CNS side effects. Administering efavirenz as a divided dose or initiating a lower lead-in dose followed by dose escalation has not been shown to decrease the incidence or severity of symptoms. It is important for clinicians to educate their patients about the adverse effects of efavirenz and that they are likely to decrease with continued therapy. Low doses of haloperidol, benzodiazepines, tricyclic antidepressants, and antihistamines have all been used anecdotally to ameliorate CNS adverse effects with some benefit.
Protease Inhibitors (PIs)
Inhibitors of HIV aspartyl protease are the most potent inhibitors of HIV replication and have been the subject of recent reviews [101–106]. Like all antiretrovirals, PIs require appropriate dosing to prevent the development of resistance mutations. Because of the highly error-prone action of reverse transcriptase and the high replication rate (∼1010 virions daily), HIV is said to be “unforgiving” [101]. Wild-type HIV in treatment-naive patients contains polymorphisms at all amino-acid base pairs, and typically ⩾4 mutations are required for resistance to PIs [107]. As a consequence, patients and physicians must take every possible step to ensure full adherence to PI treatment by using appropriate doses from the outset of therapy. The risk of producing resistant HIV overshadows the potential danger of adverse events or other reasons for incomplete patient adherence. Therefore, the first principle of the management of adverse effects of PIs is clear: do not decrease the dose in an effort to reduce adverse effects. In all cases, there are only 2 options: (1) continue to administer full doses as determined by appropriate pharmacological study, or (2) switch to therapy with an alternative agent or agents. Increasing the dose of ritonavir to mitigate the adverse effects is an obvious exception to these rules.
Adverse effects of all PIs are common (table 5). In addition, numerous reports have identified a syndrome of metabolic disorders in patients receiving PI treatment, which appears to occur in association with all PIs and has also uncommonly been described in drug-naive patients and in patients receiving treatment with PI-sparing regimens [108] (see below).
Table 5
Common adverse effects of protease inhibitors (PIs).
Other infrequent adverse events associated with PIs are noteworthy. A syndrome of increased and uncontrollable bleeding has been described in hemophiliacs receiving HAART [109]. The mechanism is unclear; patients generally have advanced HIV infection, are receiving multiple therapies, and are often affected by frequent bleeding episodes before HAART. Severe hemarthrosis and intracranial hemorrhage have been reported [109]. Case-by-case analysis of a hemophiliac's response to HAART and frequent consultation with the primary hematologist are clearly required.
At recommended doses, PI serum levels may vary 3-fold to 10-fold. Recent data that show an association between antiviral effect and therapeutic drug level have raised interest in monitoring of drug levels for antiviral effect and have renewed interest in investigating possible associations between drug levels and adverse effects (see below) [110–112]. Drug interactions are common with all PIs, although ritonavir has the greatest potential for interactions with the largest number of drugs. Unlike the NRTIs, PIs are not excreted by the kidney; therefore no dose alterations are required in patients with renal insufficiency.
As noted above, the occurrence of moderate-to-severe side effects of PIs may best be managed in this era of multiple treatment options by switching therapy. On the other hand, because adverse effects occur in association with all agents and tend to decrease during the first 4–6 weeks of therapy, the clinician is justified in treating mild to moderate adverse effects symptomatically when possible and preserving future treatment options.
Saquinavir. Saquinavir was the first PI to be approved by the FDA in November 1995. The original formulation of saquinavir has been replaced in favor of the soft-gel formulation (Fortovase; Roche Laboratories, Nutley, NJ). Although the bioavailability is substantially improved with the soft-gel formulation, the rate of bioavailability remains low (8%–12%), the side effect profile is similar, and the dosing requirements are still difficult, requiring 6 tablets 3 times daily with a large high-fat meal [113, 114]. These difficulties have limited the use of soft-gel formulation of saquinavir as a single PI, although its use in combination therapy remains common (see dual PI section). The most common adverse effects are gastrointestinal: 20%–25% of patients experience nausea, vomiting, abdominal bloating and pain, dyspepsia, and diarrhea. The gastrointestinal effects of the soft-gel formulation are moderate to severe in 10%–20% of patients. Other adverse effects include headache (12% of patients) and elevated transaminase levels (2%–6%). In addition, both forms of saquinavir have been associated with a moderately high incidence of metabolic abnormalities in comparison with other PIs (see below). Whether the high-fat diet required for the soft-gel formulation plays a role is yet to be determined.
Mitigation of gastrointestinal adverse effects is difficult. Some success is reported anecdotally with antidiarrheals, such as loperamide, but there is no clear benefit from histamine receptor blockade or antacids. Headache is usually mild and responds to acetaminophen or nonsteroidal anti-inflammatory drugs. Hepatic enzyme levels should be monitored in all patients receiving treatment and more closely followed in patients with pre-existing liver disease. Management of the metabolic abnormalities is discussed below.
Saquinavir is a weak inhibitor of the cytochrome P-450 system and is associated with relatively mild drug interactions [3].
Indinavir. Indinavir is one of the better tolerated PIs. The adverse effects of indinavir include headache, nausea and gastrointestinal distress, elevations in hepatic transaminase levels, nephrolithiasis, asymptomatic hyperbilirubinemia, dry skin, and taste perversion [115]. Nephrolithiasis, which occurs in 10%–15% of all patients and in 5% when “high-risk” patients are excluded, is the most serious adverse effect. However, in a 3-year follow-up study of 33 patients receiving treatment with indinavir, zidovudine, and lamivudine, the incidence of nephrolithiasis was 39% [116]. Indinavir can crystallize in the urine and form stones or “sludge,” which is caused by precipitation due to the increased concentrations of drug in the renal tubules. Indinavir absorption is impaired when it is taken with food; in particular, foods with a high-fat content. Therefore, indinavir should be taken on an empty stomach or with a fat-free snack.
Indinavir-related nephrolithiasis can be minimized or prevented by instructing patients to drink 40 oz (1.5 L) of fluids daily. Avoidance may be prudent in patients with underlying renal disease or a structural genitourinary deficiency, in whom additional renal impairment may be serious and irreversible. Other adverse effects of indinavir are generally mild and infrequently require intervention. Headache is usually self-limited and responds to analgesics, and taste perversion rarely leads to discontinuation of therapy. Hyperbilirubinemia is generally of no clinical consequence; monitoring of liver function during initiation of treatment is indicated for patients with underlying liver disease. Indinavir has also been associated with the metabolic disorders of HAART (see below).
Indinavir is also a weak inhibitor of the cytochrome P-450 3A4 isoenzyme. Caution is therefore needed when indinavir is administered with other medications that are metabolized by these enzymes. Medications not recommended for concomitant use include rifampin, phenytoin, carbamazepine, midazolam, triazolam, terfenadine, and ergot alkaloids [3].
There is much interest in using indinavir for dual PI combinations, for example indinavir and ritonavir. Among other potential benefits, ritonavir (400 mg twice daily) with indinavir (400 mg twice daily) blunts both the peak and trough concentrations of indinavir, which may allow for twice-daily dosing and may reduce or eliminate the strict food and fluid dosing requirements (see dual PI section); other doses of this combination, such as indinavir 800 mg with ritonavir 200 mg twice daily, are under study [117].
Ritonavir. Numerous adverse effects have limited the usefulness of ritonavir as a single agent at the recommended dosage (i.e., 600 mg b.i.d.). These side effects include asthenia, nausea, vomiting, diarrhea, perioral dysesthesia, headache, dizziness, and taste perversion [118]. Occasionally, patients tolerate full-dose ritonavir quite well, but there is no way to predict reliably at present who will have significant adverse effects, which occasionally are severe and poorly tolerated. Other adverse effects include abnormal liver enzyme levels and metabolic disorders (see below). Most comparisons of the prevalence of the metabolic disorders have found that the frequency of these disorders in association with ritonavir is higher than that in association with other PIs [119]. Many of the adverse effects associated with ritonavir, much like those associated with zidovudine, decrease during the first weeks of administration.
Perhaps the most important initial action to reduce ritonavir-associated side effects is a gradual dose escalation. Clinicians should initiate ritonavir at a dosage of 300 mg twice daily for 4 days, then 400 mg for 4 days, and 500 mg for 5 days followed by 600 mg twice daily thereafter. The most significant side effects occur with doses >400 mg. Again, for this reason, the dosage of ritonavir of 400 mg twice daily is preferred in combination with saquinavir (or possibly other agents) because its adverse effects are relatively well tolerated. Anecdotally, chocolate milk and Nutella are reported to improve the palatability of ritonavir. Ritonavir capsules require refrigeration but may be kept at room temperature for 24 h.
In August 1998, manufacturing problems caused cessation of the production of ritonavir capsules, and ritonavir liquid (400 mg/5 mL), which has an unpleasant aftertaste but requires no refrigeration, was available for substitution. Strategies to minimize the aftertaste of ritonavir liquid are diverse and patient-specific; these strategies may include having the patient eat hard candy, chocolate, or peanut butter candy or drink chocolate milk, or instilling the liquid with a syringe or dropper in the posterior oral cavity behind the taste buds. In June 1999, a revised formulation of the ritonavir capsule received approval for use from the FDA. The pills are formulated as 100 mg and require refrigeration; the adverse effects are comparable with those of the previous capsule formulation. Thus, one of the serious limitations to ritonavir use (i.e., the foul taste of the liquid formulation) has been overcome. Some patients prefer the oral liquid form, and its use in those patients and in children will continue.
Mitigation of the adverse effects of ritonavir with other medications has proven to be difficult. Analgesics will help relieve headache and abdominal pain, and antiemetics and antispasmodics may decrease the gastrointestinal side effects. At present, no treatment has been shown to reduce perioral dysesthesia or taste perversion.
Ritonavir is a potent inhibitor of the several cytochrome P-450 isoenzymes and therefore is associated with numerous drug interactions. The list of contraindicated medications is very long and includes rifampin, triazolam, terfenadine, and ergot alkaloids. Clinicians must tell patients to consult with them before taking any prescription or over-the-counter medications [3].
Nelfinavir. Nelfinavir is a generally well-tolerated PI, with mild diarrhea occurring in 14%–32% of patients [120]. The diarrhea is characterized by loose stools 3–4 times daily. In the pivotal nelfinavir studies, 11% of patients discontinued therapy because of adverse reactions, of which 2% were diarrhea [121]. Other adverse effects of nelfinavir are infrequent, occurring in <3% of patients and with no greater incidence than with trial controls (such as zidovudine and lamivudine).
Nelfinavir-associated diarrhea can be readily controlled with antidiarrheal agents, such as diphenoxylate atropine or loperamide. Patients starting treatment with nelfinavir should be informed of the possibility of diarrhea and be given prescriptions for these agents with the first prescription of nelfinavir. Nelfinavir-associated diarrhea is nonetheless an important consideration in the initiation of HAART, particularly for patients with pre-existing gastrointestinal illness and wasting (in whom a mild to moderate increase in diarrhea can have an adverse impact on the quality of life and on nutritional recovery). For such patients, aggressive antidiarrheal therapy should be undertaken from the start of nelfinavir treatment, and some consideration to alternative agents should be given.
Nelfinavir is also a weak inhibitor of the cytochrome P-450 system and therefore is associated with numerous drug interactions. Contraindicated medications include rifampin, triazolam, terfenadine, and ergot alkaloids [3].
Amprenavir. Amprenavir was approved for use by the FDA in February 1999. Amprenavir in combination with dual NRTI therapy was associated with sustained viral load suppression similar to that associated with the other approved PIs. The usual dose is 1200 mg (8 150-mg tablets twice daily) [122]. Side effects include rash, gastrointestinal problems, fatigue, headache, paresthesias, and mood disorders. In general, these side effects are more pronounced during the first 2 weeks of treatment and decrease somewhat thereafter.
The rash occurs in 28% of patients, of whom 4% have grade 3 or 4 rashes. To date, 3% of patients have discontinued treatment because of rash. Stevens-Johnson syndrome occurs in 1% of patients receiving amprenavir treatment and 4% of patients with an amprenavir-related rash. The rash is usually maculopapular and mild to moderate in severity. The incidence of rash associated with amprenavir is higher than that associated with other PIs because of a sulfa moiety in amprenavir; however, patients with known allergy to sulfa drugs may take amprenavir. Gastrointestinal side effects include nausea (38%–73% of patients), vomiting (20%–29%), and diarrhea (33%–56%). Gastrointestinal symptoms led to discontinuation of treatment in 11% of patients in 2 pivotal trials (PRO3001 and PRO3006) [122]. Headache occurred in 7%–44% of patients and was mild to moderate in severity. Perioral dysesthesias were observed in 26%–30% of patients and rarely led to discontinuation of treatment. Psychiatric side effects, including depression and mood alteration, occurred in 4%–15% of patients and were infrequently associated with discontinuation of treatment.
Amprenavir therapy may safely be continued for patients with mild rash without pruritus or systemic symptoms, such as fever or vomiting. If treatment is interrupted because of mild amprenavir-associated rash, the drug can be readministered without an increased risk of severe rash. Antihistamines may ameliorate the mild discomfort of pruritus. On the other hand, if there is moderate-to-severe rash that involves mucous membranes or is accompanied by fever, chills, vomiting, or intense pruritus, treatment should be discontinued, and rechallenge should be avoided. Stevens-Johnson syndrome should be managed with conventional use of antihistamines and systemic corticosteroids.
The mild gastrointestinal side effects of amprenavir may respond to antiemetics and antidiarrheal therapies. Nausea is difficult to control and may be decreased by dosing immediately after meals. Moderate to severe nausea, vomiting, and diarrhea may require discontinuation of treatment. Headache is usually self-limited and responds to analgesics, such as acetaminophen or nonsteroidal anti-inflammatory drugs. There is too little data to date on the appropriate management of amprenavir-related mood disorders and dysesthesias.
New protease dosing strategies and dual PI combinations. The growing recognition of the importance of adherence to HAART has generated intense interest in simplifying dosing intervals and reducing pill burdens. Peterson et al. [123] recently reported results of long-term follow-up study of twice-daily versus 3-times-daily therapy with nelfinavir in combination with stavudine and lamivudine for 286 patients. Perhaps most importantly, the side effects of both doses were equivalent, with any diarrhea occurring in 45% of patients and serious diarrhea occurring in 11%–12% of patients in both groups. After 48 weeks, there were no significant differences in the percentage of patients with viral loads below the level of detection. The accompanying data on plasma pharmacokinetics were comparable for both regimens. On the basis of these and other supporting data, many clinicians have begun to use nelfinavir at the 1250-mg twice-daily dosage, and twice-daily dosing recently received FDA approval.
Clinicians are well advised to view new dosing strategies with skepticism and follow the original dosing recommendations until alternative strategies have been subjected to thorough study. Twice-daily indinavir dosing appeared promising in preliminary uncontrolled trials of 1000 or 1200 mg twice daily, but associated pharmacological data suggested that the drug levels often fell below the IC90 for indinavir [123]. A controlled clinical trial showed that twice-daily indinavir dosing was significantly inferior to indinavir every 8 h; this trial has been halted, and twice-daily indinavir dosing is not recommended [124].
Some dual PI combinations offer the potential advantages of decreasing dose frequency, reduced adverse effects, and comparable potency. Among dual PI combinations, the most extensive experience is with 400 mg of ritonavir and 400 mg of saquinavir given twice daily. Mellors et al. [125] recently reported 72-week follow-up data for 100 patients receiving treatment with ritonavir and saquinavir, which showed that 90% of patients had viral loads <200 copies/mL at 72 weeks. The mean increase in CD4+ cell count was 185 cells/mL, and adverse events were infrequent; 10%–12% of patients had gastrointestinal events. Nineteen percent of patients had a minor increase in viral load within 72 weeks, after which intensification of treatment was instituted with stavudine and lamivudine; of these patients, 78% had reductions in viral loads below the level of detection.
The combination of ritonavir and saquinavir is also being studied in a first-line capacity. Kirk et al. [126] reported results of a comparison of the combination of ritonavir and saquinavir with ritonavir or indinavir alone with 2 NRTIs; at 24 weeks, ritonavir and saquinavir with 2 NRTIs had reduced viral loads 82% below the level of detection; saquinavir with 2 NRTIs had reduced viral loads 71%; ritonavir with 2 NRTIs, 67%. There were also fewer serious adverse events: ritonavir and saquinavir with 2 NRTIs, 12.9%; saquinavir with 2 NRTIs, 16%; ritonavir with 2 NRTIs, 25%. Twenty-one PI-naive patients treated with ritonavir (400 mg b.i.d.) and saquinavir (400 mg b.i.d.) with 2 nucleoside inhibitors were followed up for 2 years at Cook County Hospital (Chicago); the rate of adverse events was 12%, and 76% of patients had viral loads below the level of detection (500 copies/mL) [127].
Numerous other studies of dual PI combinations are under way. The combination of nelfinavir and saquinavir (soft-gel formulation) appears to be efficacious in clinical studies [128]. Because there is only a modest pharmacological interaction, these combinations are associated with only a modest reduction in pill burden or a change in dietary requirements. Similarly, these preliminary data suggest changes in the adverse effect profiles, with 35%–45% of patients having diarrhea and discontinuation of treatment by 15%–18% of patients (which may be a disincentive for these combinations). Further study is nonetheless warranted.
Another combination under study is ritonavir and indinavir with which peak and trough levels of indinavir that are up to 10 times greater than the indinavir IC90 are achieved with dosages of 400 mg of both drugs b.i.d. [117]. This combination has potent antiviral efficacy. In 1 uncontrolled study, 57 patients were treated for a mean of 42 weeks without the usual food and fluid requirements, and nephrolithiasis did not occur [129, 130]; data on clinical outcomes are pending, and further study is warranted. More recently, alternate dosing strategies for ritonavir and indinavir are under study, including 1–200 mg of ritonavir and 800 mg of indinavir b.i.d. [131].
Finally, not all PI combinations are promising; indinavir and saquinavir are antagonistic in vitro, and this combination should be avoided [3]. Similarly, indinavir and nelfinavir have been studied in a twice-daily combination with promising preliminary pharmacological data [132]; however, the failure of treatment with indinavir twice daily led this combination to be abandoned, since it did not confer other desired benefits: reduction of pill burden, decreased dose frequency, no food requirements, comparable potency, and reduced adverse effects.
New Adverse Events (as of 1999)
New reports of unexpected adverse effects have emerged as HAART has become more widespread. Recently, a number of metabolic abnormalities (i.e., fat maldistribution, hyperlipidemia, and glucose intolerance) have been described in patients receiving PI-containing HAART. These abnormalities have 3 common elements: (1) glucose intolerance and rare diabetes or diabetic ketoacidosis, (2) hyperlipidemia with hypertriglyceridemia and hypercholesterolemia, and (3) fat redistribution with accumulation of abdominal fat and “buffalo humps,” loss of peripheral fat in the face, arms, legs, and buttocks, as well as ectodermal dysplasia characterized by ingrown toenails [119, 133–139]. Breast enlargement has been described in women and rarely in men [140]. Note that each manifestation can occur in the absence of the other 2 and that it remains uncertain whether the syndrome should be so defined.
Carr et al. [133, 141] recently reported that 68% of patients treated with PIs for an average of 11 months had fat redistribution when they were assessed by sensitive dual energy radiographic absorptiometry, and a large cohort study in France showed a similar prevalence (58%) [119]. However, other investigators have reported substantially lower prevalences of these changes (i.e., 17%–32%) [137, 138, 140]. This constellation of symptoms has been primarily described in patients receiving PI treatment (including all known members of the class), although there are case reports of the syndrome in patients receiving PI-sparing regimens (including 16% of patients receiving dual NRTIs in an Australian cohort [142]). In a well-controlled Women's Interagency HIV study, in which extensive pretreatment metabolic data were available, 16% of women reported an abnormal habitus, including breast enlargement of 2 sizes in 71% of those reporting changes [140]; 41% had elevated cholesterol levels, and 38% had triglyceride levels greater than baseline values.
Whether these findings are linked to so-called metabolic disorders, as has been proposed [133], or are disparate manifestations of PI therapy or viral load reduction is as yet unknown and unproven. Nonetheless, they are worthy of close attention by clinicians, particularly because of the real possibility of significant short- and long-term clinical consequences. For example, anecdotal reports of premature cardiovascular disease in patients receiving PI treatment have emerged [143]. Henry [144] described 4 patients with hypertriglyceridemia-associated pancreatitis. Some of these abnormalities in lipid concentration would place patients into distinctly higher risk classifications for cardiovascular disease. When the observed increases in lipid levels were analyzed in a model of cardiovascular disease from the Framingham study [145], the estimated increase due to increased lipid levels alone was only 1.4 cases per 100,000 population every 10 years. To date, diet and exercise alone have not been associated with reversals of hyperlipidemia, although modest reductions in lipid levels of 10%–20% have been observed with atorvastatin, gemfibrozil, and other agents that lower lipid levels [146]. Importantly, one-half of the 14 cases of diabetic ketoacidosis that were reported to the FDA in 1997 occurred in patients with no known risk or predisposition to diabetes.
Management of Metabolic Disorders
Given the current uncertainty regarding the prevalence, incidence, and clinical significance of these events, the following recommendations can be made. First, baseline lipid levels and random glucose levels should be determined before patients start PI treatment, and these patients should be monitored for changes in fat distribution and abnormalities in carbohydrate or lipid levels. Medical management should include examination of the patient's diet and recommendations for a low-fat diet and regular exercise; oral hypoglycemic agents or agents that lower lipid levels may be necessary and should be used according to current medical standards, such as those of the National Cholesterol Education Program [146]. Data do not exist at present to evaluate the relative contribution of each of these interventions. Nonetheless, in view of the potential for simple management measures of some of the observed changes, treatment for patients who are presently virologically and clinically stable should not be changed unless there is a compelling rationale for doing so (e.g., severe hypertriglyceridemia with pancreatitis, recurrent diabetic ketoacidosis or refractory diabetes mellitus, or accelerated ischemic heart or vascular disease).
Currently, it is not known whether discontinuation of PI treatment will reverse the metabolic adverse events. Preliminary data from studies of switches from PIs to NNRTIs or abacavir suggest that virologic control is maintained in >90% of patients, triglyceride and cholesterol decline modestly (10–20 mg/dl), but no clear change in fat maldistribution has been observed [147, 148]. In addition, there are anecdotal reports of improvement with discontinuation of treatment and/or administration of agents that lower lipid levels [149]. For patients at high risk for cardiovascular disease, extra vigilance may be necessary. As with all patients, general measures to reduce the risk of cardiovascular disease should be pursued aggressively, such as stopping smoking and treating hypertension.
Two other antiretroviral agents were available for clinicians in 1999, either by nonapproved use of an existing drug or by release according to compassionate protocol. These agents are hydroxyurea and adefovir dipivoxil (table 6).
Table 6
Common adverse effects of hydroxyurea and adefovir dipivoxil.
Hydroxyurea. Hydroxyurea (Hydrea; Bristol-Myers Squibb, Princeton, NJ) is an antimetabolite used in the treatment of hematoproliferative disorders, such as polycythemia vera and sickle-cell disease. Its use in the treatment of HIV infection has evolved with the recognition of its inhibition of ribonucleotide reductase, the cellular enzyme in the production of nucleotides that also mediates the metabolism of didanosine and other antiviral agents. By inhibiting ribonucleotide reductase, the intracellular pool of nucleic acids (particularly adenosine) decreases. This decrease increases the likelihood that didanosine, an adenosine analogue, will be incorporated into the viral RNA, thus halting viral replication [150]. Recent clinical data have shown that hydroxyurea enhances the antiviral effect of didanosine by increasing suppression of the viral load by 0.5 log, even in patients with prolonged past treatment with didanosine [151]. Importantly, hydroxyurea reduces the overall lymphocyte count; therefore, it is not associated with increases in absolute CD4+ cell counts, as are other antiviral therapies. However, CD4+ cell percentages do increase significantly in patients who are treated with didanosine and hydroxyurea.
Hydroxyurea has a well-known adverse effect profile for patients with myeloproliferative disorders [152], but a complete description of adverse effects in people with HIV disease is still lacking. The most important adverse event is myelosuppression of all lineages, which occurs in 10%–15% of patients and is reversible with discontinuation of treatment. Other reported adverse effects that occur at moderate frequency include diarrhea, stomatitis, and maculopapular pruritic rash. In addition, hydroxyurea is associated with complications of therapy arising from its antimetabolic effect on rapidly reproducing cell lines, such as hyperuricemia, aggravation of gout, and urate-induced renal disease. To date, these complications have not been described in patients with HIV disease.
The absence of several key pieces of data has limited the use of hydroxyurea. The optimal dose has yet to be established, although dosages of >1200 mg/d are associated with more frequent adverse events. The most commonly used dosage at present is 500 mg b.i.d.; dose-ranging studies are under way. Similarly, the full extent of the side effect profile in HIV disease has not been characterized. These data should be available within 1 year.
Hydroxyurea-associated cytopenias have been reported to respond to erythropoietin and granulocyte colony-stimulating factor, although the long-term use of these agents is discouraged. In general, hydroxyurea should be used with caution as treatment for patients with pre-existing cytopenias, whether from infection due to HIV, parvovirus, or other causes; patients with these conditions should be monitored closely for cytopenia, particularly within the first 2 months of hydroxyurea use. Hydroxyurea-associated diarrhea is generally mild and is readily controlled with antidiarrheal agents. Although the rash associated with hydroxyurea is generally well tolerated and responsive to antihistamines and topical steroids, discontinuation of treatment is necessary in severe cases with systemic manifestations or mucosal involvement.
Adefovir dipivoxil. Adefovir dipivoxil is a novel nucleotide reverse transcriptase inhibitor that has modest activity against HIV (viral load reduction, ∼0.5 log) and also has activity against cytomegalovirus and hepatitis B virus [153–155]. Nucleotide analogues differ from the NRTI class, because nucleotides require only 2 phosphorylation steps via ubiquitous cellular enzymes and therefore may have greater activity in a broader range of cell types. Adefovir dipivoxil is a pro-drug that has 2 pivalic acid moieties attached to improve bioavailability. Pivalic acid derivatives have been associated with reduced serum carnitine concentrations; therefore, L-carnitine (500 mg) daily is given concomitantly with adefovir dipivoxil [156]. A novel resistance profile makes it potentially attractive for combination regimens, as does its once-daily dosing profile.
Renal toxicity is the most important adverse effect of adefovir dipivoxil; it has been described in 38% of patients treated for >6 months. It presents as interstitial nephritis that causes a dose- and duration-dependent disorder similar to Fanconi's syndrome (i.e., elevation of creatinine concentration with or without renal tubular dysfunction, proteinuria, glycosuria, and/or wasting of electrolytes, particularly phosphate, bicarbonate, or potassium). Of 83 events in a phase III trial of 120 mg of adefovir dipivoxil, 21 took >2 months to resolve, and some proximal tubular damage may be irreversible [157, 158]. The symptoms were more prevalent after patients had been receiving therapy for >24 weeks. The original dose under study was 120 mg, but for one-half of the patients the original doses had to be reduced to 60 mg. Further studies are required to evaluate the overall antiviral efficacy of adefovir dipivoxil at this dose level.
Other adverse effects of adefovir dipivoxil include gastrointestinal symptoms, such as nausea and diarrhea, headache, abdominal pain, and elevations of creatine kinase and hepatic enzyme levels. The gastrointestinal symptoms occur in 5%–8% of patients and are mild in severity; in the expanded access program, <1% of 2000 patients receiving adefovir dipivoxil treatment had grade 3 or 4 gastrointestinal side effects [156]. Elevations in creatine kinase levels occur in 2% of patients and are transient and asymptomatic. Elevations in liver enzyme levels occur in 4% of treated patients and are mild and well tolerated.
An unexpected interaction between adefovir dipivoxil and delavirdine was observed recently in the AIDS Clinical Trials Group 359 Study that resulted in a 42% lower level of delavirdine than expected [159].
The management of adefovir dipivoxil-related nephrotoxicity is discontinuation of treatment. As noted above, most of the proximal tubular disorder appears to be reversible with discontinuation of treatment. Generally, metabolic acidosis and hypophosphatemia are mild and tolerated; in severe cases, replacement of phosphate and treatment of acidosis acutely are needed. Proximal tubular function returns to normal in 40% and 90% of patients by 4 weeks and 14 weeks after cessation of therapy, respectively. Elevations of creatine kinase levels are generally mild and asymptomatic. Elevations of hepatic enzyme levels require monitoring during the initiation of therapy, and more careful monitoring is needed for patients with pre-existing liver disease.
Therapeutic Drug Monitoring
Several recent studies have renewed interest in the potential for clinical use of therapeutic drug monitoring of antiretroviral therapies. Early studies of PIs demonstrated striking variability in drug plasma levels, from 3-fold to ⩾10-fold. Because the action of these drugs depends on intracellular concentrations, the clinical importance of this variability is unclear, although their role in adverse effects was more clearly associated with a dose response. More recently, 3 studies found an association between PI levels and changes in viral load [110–112], although other investigators have found no such association. In 2 studies, higher nelfinavir levels correlated with greater declines in viral load [110, 112]. In another study [111], inadequate viral load responses to saquinavir were associated with trough levels below the known IC50 of saquinavir. Both of these latter studies were uncontrolled and retrospective, but they raise the possibility that inadequate drug levels may lead to inadequate treatment responses and that clinical monitoring at certain periods may be of value in clinical management, both for the interpretation of treatment failures and for mitigation of adverse effects.
There are several potential limitations to therapeutic drug monitoring for the class of PIs, including the above-noted variability, potential costs, logistical difficulties of linking drug levels to meaningful trough and peak concentrations, and uncertain correlation between serum levels and intracellular activity. Nonetheless, further study of the association between drug levels and virological and clinical outcomes is needed in the future. In particular, the possibility should be explored that excessive adverse effects or suboptimal virological responses could be evaluated by determination of serum levels and compensated by dose increases.
Summary and Implications for Treatment
We have outlined several strategies to reduce or prevent adverse effects of HAART and to improve adherence to HAART. In each case, before adding medications to therapy or prematurely discontinuing treatment, the clinician should re-examine the dose (relative to the patient's weight and renal function, when appropriate), the recommendations for food administration, and the relevant details of the patient's life, such as his or her values and priorities, daily milestones or habits that may be used for pill-taking strategies, past behaviors with medication adherence, and other pertinent personal information.
Although the occurrence of new adverse effects, such as metabolic disorders, is of concern, an overreaction by physicians or patients might make matters worse. At present, careful study of the extent and clinical impact of the problem is needed, whereas the remarkable successes of the last 2 years are recognized and further extended. Patients whose conditions are clinically and virologically stable should continue with their current treatment regimens unless the metabolic disorders are severe or unacceptable to the patient's quality of life, and clinicians should increase their surveillance for underlying metabolic disorders while the natural history and clinical impact of these events are being defined. Because the long-term effects of these adverse events is unknown, a new note of caution may exist for patients with marginal indications for starting HAART, such as patients with CD4+ cell counts >500 cells/mL and viral loads of 1–10,000 copies/mL. Arguments for early therapy and reasonable caution have recently been articulated and should be fully shared with patients [160–162]. Nonetheless, the recommendation for patients with advanced disease to begin potent antiviral therapy immediately is stronger than ever, in view of recent evidence that the immune system can be partly restored (even in patients with advanced disease).
New alternatives for long-term sequencing of antiretroviral therapy exist (as of 1999). These include nucleoside-only regimens, such as the combination of stavudine, didanosine, and hydroxyurea or zidovudine, lamivudine, and abacavir, and other PI-sparing regimens, such as the combination of zidovudine, didanosine, and nevirapine or zidovudine, lamivudine, and efavirenz. New strategies for first-line PI dosing, such as ritonavir combined with saquinavir or twice-daily nelfinavir, are enabling improved patient adherence. Reasonably successful salvage therapy with a second-line PI regimen has also been identified, such as ritonavir combined with saquinavir. Although the rate of cross-resistance among the PIs has not yet been confirmed, it has been as high as 65%–75% in some studies [163–165]. These new options offer numerous opportunities for feasible multiple-drug sequences and expand the potential range of treatment options for newly identified patients.
The use of genotypic and phenotypic resistance testing appears promising and has been the subject of recent reviews [166, 167]. An association between such testing and virological outcomes has recently been shown [168–170], and in at least 2 clinical settings (acute or recent HIV infection and first virological treatment failure), there may be a role for such testing in clinical practice in the future. An additional benefit pertinent to this discussion is that unnecessary adverse effects may be reduced by avoiding regimens and drugs to which the patient's virus exhibits resistance. Two prospective clinical trials have shown improved short-term virological outcomes when clinicians made treatment decisions at the first HAART failure with the benefit of genotypic resistance testing [169, 171], and their use in clinical practice in defined settings has been endorsed by the Health and Human Services/Kaiser Family Foundation guidelines [3]. For all patients for whom therapy is undertaken, careful attention to adherence and the anticipation and reduction of adverse effects before therapy is begun are imperative for successful outcomes.
- © 2000 by the Infectious Diseases Society of America
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
-
-
-
- ↵
- ↵
- ↵
- ↵
- ↵
-
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
-
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
-
-
-
-
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
-
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
