AEDs. AED serum concentrations should be monitored

after delivery, particularly when dosage escalations have

been made during pregnancy.

Frequent, “routine” determinations of serum AED concentrations are costly and not warranted for patients whose clinical

status is stable. Clinicians may tend to focus attention on normal

variability in serum concentrations rather than on the patient’s

clinical status; as a result, unnecessary dosage adjustments may

be made to make serum concentrations fit the “normal range.” A

plan of action for what the clinician is going to do with the information once it is obtained should be in place before obtaining the

sample. Therefore, the results of individual serum concentration

determinations must be evaluated carefully to decide whether a

significant, clinically meaningful change has occurred.26

Interpretation of Serum Concentrations

Several factors can alter the relationship between AED serum

concentration and the patient’s response to the drug. Whenever

a change in serum concentration is apparent, pharmacokinetic

factors (Table 58-4) should be considered (along with the patient’s

clinical status) before a decision is made to adjust the AED dosage.

Laboratory variability can cause minor fluctuations in reported

AED serum concentrations. Under the best conditions, reported

values for serum concentrations may be within plus or minus

10% of “true” values.27,28 Therefore, the magnitude of any apparent change must be considered. Therapeutic ranges are not well

established for some drugs (e.g., valproate and the newer AEDs

such as lamotrigine, topiramate, and tiagabine). Published therapeutic ranges may have been determined in small numbers of

patients or may more accurately represent average serum concentrations at usual doses. As an example, many clinicians agree

that “pushing” valproate serum concentrations up to 150 to 200

mcg/mL may be beneficial for some patients; these higher concentrations are not consistently associated with specific toxicity

symptoms.29–32 Nevertheless, most laboratories report 50 to 100

mcg/mL as a therapeutic range for valproate. Inappropriate sample timing can result in inconsistent and clinically meaningless

changes in AED serum concentrations.22 Generally, serum concentrations of AED should not be measured until a minimum

of four to five half-lives have elapsed since initiation of therapy

or a dosage change. Blood samples should be obtained in the

morning, before any doses of the AED have been taken; this

practice provides reproducible, postabsorptive (i.e., “trough”)

serum concentrations. Interindividual variability in response to

a given serum concentration of medication is common. Excellent therapeutic response or even symptoms of intoxication may

be associated with AED serum concentrations that are classified

as “subtherapeutic.”33 Active metabolites of AEDs usually are

not measured when serum concentrations are determined.28,34

Alterations in the relative proportion of parent drug and active

metabolite may result in an apparent alteration in the relationship between the serum concentration of the parent drug and

the patient’s response. Binding to serum proteins is significant

for some AEDs (e.g., phenytoin, valproate, tiagabine). Changes

in protein binding can result from drug interaction, renal failure, pregnancy, or changes in nutritional status. These changes

can alter the usual relationship between the measured total drug

concentration (bound and unbound to plasma proteins) and the

unbound (pharmacologically active) drug concentration. This

change may not be apparent when only total serum concentrations are measured. Determination of serum concentrations of

free (i.e., unbound) AEDs is available from many commercial

laboratories; these determinations are expensive and results may

not be available for several days. If significant changes in protein

binding are suspected, measurement of free concentrations of

AED may provide additional information useful for adjustment

of doses or interpretation of the patient’s symptoms.27,28,34

MONOTHERAPY VERSUS POLYTHERAPY

Decades ago, epilepsy was often treated initially with multiple

AEDs (polytherapy). A second, third, or even fourth drug was

added when seizures were incompletely controlled with a single

fulness in patients with persistent low-level viremia.

4. To prevent the development of resistance, one new drug

should never be added to a failing regimen. An exception

to this rule is if the initial response to a first regimen has been

TABLE 73-7 Drugs That Should not be Coadministered With Antiretrovirals (at any dose unless otherwise indicated)5 Cardiac Agents Lipid-Lowering Agents Anti-Infectants Gastrointestinal Drugs Psychotropic Agents Anticonvulsants Pulmonary Agents Herbs Others ARV Agents ATV ± RTV Bosentana Lovastatin Pitavastatin Simvastatin Rifampin Rifapentine Voriconazoleb Cisapride PPIsc Midazolam Triazolam Salmeterol Sildenafil (for PAH) Fluticasoned St. John’s wort Alfuzosin Irinotecan Buprenorphinee ETR NVP DRV/r Lovastatin Pitavastatin Simvastatin Rifampin Rifapentine Voriconazoleb Cisapride Midazolam Triazolam Salmeterol Sildenafil (for PAH) Fluticasonec St. John’s wort Alfuzosin FPV ± RTV Flecainide Propafenone Lovastatin Pitavastatin Simvastatin Rifampin Rifapentine Voriconazoleb Cisapride Midazolam Triazolam Salmeterol Sildenafil (for PAH) Fluticasonec St. John’s wort Alfuzosin ETR LPV/r Lovastatin Pitavastatin Simvastatin Rifampin Rifapentine Voriconazoleb Cisapride Midazolam Triazolam Carbamazepine f Phenobarbital f Phenytoin f Salmeterol Sildenafil (for PAH) Fluticasonec St. John’s wort Alfuzosin TPV/r Amiodarone Flecainide Propafenone Quinidine Lovastatin Pitavastatin Simvastatin Rifampin Rifapentine Voriconazoleb Cisapride Midazolam Triazolam Salmeterol Sildenafil (for PAH) Fluticasonec St. John’s wort Alfuzosin ETR EFV Rifapentine Voriconazoleb Cisapride Midazolam Triazolam St. John’s wort Other NNRTIs ETR Rifampin Rifapentine Carbamazepine Phenobarbital Phenytoin St. John’s wort Clopidogrel Unboosted PIs ATV/r, FPV/r, or TPV/r Other NNRTIs NVP Rifapentine Ketoconazole St. John’s wort ATV ± RTV Other NNRTIs MVC Rifapentine St. John’s wort a Bosentan not recommended if ATV is unboosted. b Do not coadminister with boosted PIs unless benefit outweighs risk and do not coadminister with EFV at standard doses (voriconazole 400 mg BID, EFV 300 mg daily). c PPIs are not recommended if ATV is unboosted or in treatment experienced patients. d Do not coadminister inhaled or intranasal fluticasone with boosted PIs unless benefit outweighs risk of systemic corticosteroid adverse effects. e Do not coadminister with unboosted ATV. f Not recommended if LPV/r is given once daily. ATV, atazanavir; DRV/r, darunavir/ritonavir; EFV, efavirenz; ETR, etravirine; FPV, fosamprenavir; LPV/r, lopinavir/ritonavir; MVC, maraviroc; NVP, nevirapine; RTV, ritonavir; TPV/r, tipranavir/ritonavir.

1708

1709Pharmacotherapy of Human Immunodeficiency Virus Infection Chapter 73

inadequate (e.g., undetectable viral load at 16–20 weeks). In

this situation, some clinicians may intensify therapy with an

additional agent provided that the viral load measurements

were trending downward since initiation of therapy.

5. If possible, a regimen that has failed in the past should not

be reinitiated, as an isolate resistant to the failed regimen

could continue to reside within various compartments of the

body. If a regimen to which the patient had previously failed

were restarted, unimpeded viral replication of the resistant

strain would occur, repopulate the host, and eventually result

in treatment failure. In some situations (e.g., patients with

advanced disease, limited treatment options, and prior exposure to most antiretroviral agents), it may be necessary to

reinitiate agents or regimens in combination with additional

new agents with the goal of suppressing viral replication.

6. When treatment failure is a direct result of drug toxicity

(rather than poor drug efficacy), the offending agent should

be replaced with an alternative drug from a similar class, provided that the potential for cross-resistance is minimal.

7. If an agent in a given regimen must be stopped, it is recommended that all agents in the regimen be stopped and restarted

simultaneously to prevent the development of resistance. An

exception to this rule is when components of a regimen have

differing half-lives, such as NNRTIs and NRTIs. In this situation, if all drugs are discontinued simultaneously, continued

monotherapy exposure with the NNRTI is likely to result,

due to the much longer NNRTI half-life. Consequently, many

experts recommend continuing the NRTIs for 1 to 2 weeks

past NNRTI discontinuation to provide combination therapy

while the NNRTI is eliminated from the body (covering the

“tail” of the pharmacokinetic profile).

It should be recognized that many alternative regimens are

based on theoretical benefits or limited data. In addition, many

potential options could be limited in some patients based on

prior antiretroviral use, toxicity, or past intolerances. Therefore,

the clinician should carefully discuss these issues with the patient

before changing therapy.

Because E.J. is failing to respond to his current regimen, a

new antiretroviral regimen must be chosen. In addition to selecting susceptible agents from resistance testing, the new regimen

should take into consideration quality-of-life issues. In E.J.’s situation, it is reasonable to switch to a ritonavir-boosted PI regimen,

with two or more nucleoside agents as dictated by the viral resistance profile.

Considerations in AntiretroviralExperienced Patients

CASE 73-2

QUESTION 1: H.G. is a 46-year-old, HIV-positive man with

an extensive history of treatment with a variety of antiretroviral agents. He took ZDV monotherapy in the late 1980s

and early 1990s. When 3TC became available, he took the

combination of ZDV and 3TC until he failed therapy about

15 years ago. At that time, he began experiencing ZDVinduced myopathies. Since that time, H.G. has been “on

and off” various regimens without sustained clinical benefit. He is currently taking TDF, FTC, and lopinavir/ritonavir

with a CD4 count and viral load measurement of 55 cells/μL

and 48,000 copies/mL, respectively. These laboratory values

have been stable for the last 9 months. How do patients with

extensive antiretroviral histories differ from antiretroviralna¨ıve patients? Are there any special considerations when

selecting therapeutic regimens for patients such as H.G.?

Patients who have been infected for 15 or more years may have

been treated with many different regimens, both experimental

and FDA-approved. As a result, many potential regimens have

already been exhausted; thus, there are not many viable choices

left. Many patients previously treated with multiple antiretroviral regimens will exhibit a decreased response and decreased

durability to older protease inhibitors such as indinavir and

nelfinavir.5,6 Several newer agents, such as tipranavir, darunavir,

etravirine, and enfuvirtide, have been developed specifically for

highly treatment-experienced patients: These may achieve viral

suppression in these patients. Maraviroc, the first CCR5 receptor

antagonist approved by the FDA, is also an option for treatmentexperienced patients infected with HIV-1 utilizing this coreceptor

and failing current treatment. Maraviroc is not recommended for

use in patients with dual-trophic viral populations (i.e., able to use

CXCR-4 or CCR5 as coreceptors) or CXCR-4-trophic virus and,

thus, patients with extensive treatment histories should undergo

tropism testing before maraviroc is initiated. Additionally, raltegravir has a novel mechanism of action that allow for its use in

patients with extensive reverse transcriptase or protease mutations. Some highly treatment-experienced individuals will have

extensive resistance patterns that limit their therapeutic options.

In such patients, full suppression of viral load and immune reconstitution may not be possible.5,6

Patients who have experienced several antiretroviral regimens

present other unique challenges for clinicians and require consideration of the following factors.

1. Regimen tolerability: Patients with advanced HIV disease display decreased tolerability to many medications, including

antiretroviral agents. Although this is not fully understood, it

is probably a result of HIV-induced immune alterations and

cytokine dysregulations. Subsequently, clinicians evaluating

patients with advanced disease should be alert for possible

drug-induced adverse events.

2. Drug interactions: Many patients with advanced disease take

numerous medications for primary or secondary prophylaxis

of various opportunistic infections, as well as other medications for comorbid disease states. Subsequently, the risk for a

drug–drug interaction is increased. The addition of any new

medication, either prescription or over-the-counter, should be

carefully evaluated for potential interactions with the patient’s

current antiretroviral regimen (see Table 73-7). In addition,

any change to the current antiretroviral regimen should also

be checked against the patient’s current medication list.

3. Altered bioavailability: Patients with advanced HIV infection

may have unreliable absorption of many medications due to

severe diarrhea, anorexia, weight loss, wasting, and gastric

achlorhydria. As a result, the bioavailability of some agents,

especially certain PIs that require specific dietary requirements, may be affected (Table 73-3). Any changes in dietary

habits or bowel function should be carefully assessed in light

of the potential impact on the antiretroviral regimen.

4. Antiretroviral drug histories and resistance testing: The most useful information to guide the choice of alternative regimens is

a detailed drug history, in conjunction with appropriate resistance testing. Among patients with extensive prior antiretroviral use, it is critical to identify previously failed regimens and

determine the precise cause of the failures. In an experienced

patient who has taken many different regimens over a lifetime,

the number of remaining viable agents and regimens may be

limited. Therefore, it is important to determine whether prior

regimens truly failed for virologic reasons or some other cause

1710 Section 14 Infectious Disease

(e.g., regimen intolerability or an inadequate trial period). In

addition, detailed knowledge of regimen intolerabilities and

which agent(s) caused the adverse event will help in the selection of a new appropriate regimen. In some situations, the

offending agent may be reinitiated if the adverse event was

minimal or can be appropriately managed.

RESISTANCE, VIRAL GENOTYPING,

PHENOTYPING, AND VIRAL FITNESS

CASE 73-2, QUESTION 2: Will viral genotyping and phenotyping assist in selecting an appropriate therapeutic regimen for H.G.? What are these tests? What are their limitations and when should they be used? What is viral fitness,

and does it have a role in clinical decision making?

Viral genotyping and phenotyping reveal resistance patterns

to antiretroviral agents. Genotyping evaluates mutations in the

virus’s genetic material, whereas phenotyping assesses the ability

of the virus to grow in the presence of increasing concentrations

of antiretroviral agents. The three potential causes for the development of resistance are as follows.

1. Initial infection with a resistant isolate69–71,129

2. Natural selection of a resistant isolate as a consequence of

inefficient, error-susceptible viral replication36,130,131

3. Generation of resistant isolates via selective pressures from

antiretroviral therapies that do not fully suppress viral

replication6,132–141

Mutations are generated when naturally occurring amino

acids in the HIV genome are replaced with alternative amino

acids. For example, resistance to 3TC occurs when the amino acid

methionine (M) is replaced by valine (V) at the 184th amino acid in

the protein chain.142,143 This mutation is subsequently referred to

as an M184V mutation. These amino acid substitutions change

the proteins that are produced and may alter the shape, size,

or charge of the viral enzyme’s substrate or primer.144–146 Subsequently, antiretroviral drug binding to the active site is decreased,

affinity for natural substrates is increased, or there is an increased

removal of the antiretroviral agent from the enzyme by the virus

(known as pyrophosphorylation).147 Whether a mutation results

in a clinically resistant, less viable, or indifferent isolate depends

on which amino acid(s) is replaced. In addition, certain mutations

or combination of mutations have been shown to produce viral

isolates that display increased sensitivity to various antiretroviral

agents (known as hypersusceptibility). Alterations to certain key

amino acids can also result in cross-resistance between various

antiretrovirals.6

Two key enzymes have been extensively studied with regard

to their potential for development of resistance: reverse transcriptase (RT) and protease. Replication by the RT enzyme is highly

susceptible to errors. Given that the HIV genome is approximately 10,000 nucleotides in length and that mutations via the RT

enzyme occur approximately once in every 10,000 nucleotides

copied, it has been estimated that a mutation occurs with every

viral replication cycle. With up to 10 billion particles of virus being

produced per day, the potential exists for 1,000 to 10,000 mutations occurring at each site in the HIV genome every day.35 Integrase resistance, including resistance mutations to raltegravir,

exists, but has not yet been fully described, because use is not

widespread.51 Key mutations to the antiretroviral agents are identified in Figure 73-5 and are updated periodically by expert panels

of clinicians.6

Over time, countless viral subpopulations known as “quasispecies” develop. In any given host, at any given time, many

different quasi-species can exist. In addition, within any compartment of the body (e.g., CNS, testes, lymph nodes), many different

quasi-species can also exist. Because these mutant strains represent only a small number of isolates in the total viral population,

they must have some replicative disadvantage when compared

with the “wild-type” virus.138,139 Under selective pressures from

antiretroviral therapies, however, these mutant isolates can replicate. For example, if wild-type viral replication is inhibited by an

antiretroviral regimen, and if any one viral strain of the quasispecies is more fit for growth in the presence of that regimen,

then the viral mutant will have a competitive advantage.148 It

should be recognized that for resistance to develop, viral replication must occur. When viral replication is completely inhibited,

the development of resistant isolates is uncommon.

Genotypic analysis involves sequencing the viral genetic material via PCR amplification. Mutations are identified by analysis

of key sequences of the RT or protease enzymes. These tests can

be rapidly processed; however, they detect mutations present

only in more than 25% of all HIV isolates in the body, and can

only reliably detect resistance patterns in samples with HIV RNA

greater than 1,000 copies/mL. Because pressures from antiretroviral therapies select resistant isolates, these tests may not provide

information regarding rare, yet potentially clinically significant,

isolates.5,6

Phenotypic analysis involves growing virus in the presence of

various concentrations of drug and then determining viral susceptibilities (e.g., half-maximal inhibitory concentration [IC50]).

Phenotyping is limited because it evaluates only one viral isolate at a time and could fail to identify other clinically relevant

isolates.5 An additional methodology, known as the virtual phenotype, is also currently available, and compares the genotype

information from a sample of interest to a large database of

viral isolates where both genotype and phenotype have been

performed.5 This method is only as reliable as the databases from

which the virtual phenotype is generated, and may be limited

for newer agents with less available genotype-phenotype data.

Genotyping and phenotyping have the potential to provide

useful information to clinicians who treat patients with HIV

infection. Evidence to date suggests that these tests are effective at predicting which agents will not work, but they are less

useful in predicting those that will work.5,6,149 For example, if

testing identifies resistance to an agent, it is highly unlikely this

agent will be effective. If testing identifies an agent as susceptible, however, it does not necessarily predict that this agent will be

effective. Clinical trials evaluating the use of genotyping and phenotyping for selection of alternative antiretroviral regimens have

shown positive treatment responses.5,6 As with all clinical decisions made for those who are HIV infected, careful assessment of

the results in combination with the treatment history are essential

for proper clinical decision making. Consultation with an expert

in antiretroviral drug resistance patterns is highly recommended.

One other measure available to clinicians is “viral fitness” or

“replication capacity.” The genetic changes that occur in a virus

to become resistant to antiretroviral therapy often impair the

virus’ ability to replicate.41,150–152 This measure is called “fitness”

and is quantified as replication capacity during phenotypic evaluation. During amplification of the virus before the phenotype

is measured, the replication capacity of the virus is evaluated and

compared to a reference wild-type, drug-sensitive virus. A virus

with normal fitness has a replication capacity between 70% and

120% of the reference viral strain; isolates with values less than

70% are considered to be less fit than wild-type virus. In general,

the more mutations that occur to the virus, the more compromised the virus becomes, and the lower the fitness (although, in

1711Pharmacotherapy of Human Immunodeficiency Virus Infection Chapter 73

FIGURE 73-5 Mutations associated with resistance to the various classes of antiretroviral agents. Amino acid abbreviation: A,

alanine; C, cysteine; D, aspartate; E, glutamate; F, phenyalanine; G, glycine; H, histidine; I, isoleucine; K, lysine; L, leucine; M,

methionine; N, asparagine; P, proline; Q, glutamine; R, arginine; S, serine; T, threonine; V, valine; W, tryptophan; Y, tyrosine.

(Reprinted with permission from the IAS–USA. Johnson VA, Brun-Vezinet F, Clotet B, G ´ unthard HF, Kuritzkes DR, Pillay D, Schapiro ¨

JM, and Richman DD. Update of the Drug Resistance Mutations in HIV-1: December 2010. Topics in HIV Medicine.

2010;18(5):156–163. Updated information and User Notes are available at www.iasusa.org.) (Continued )

some situations, the interplay between mutations can result in

a viral isolate that is relatively fit). Recent data have shown that,

despite persistent viral replication, unfit viruses may not cause

the same degree of immune destruction as fit viruses.151 This

finding is important for patients with limited treatment options

because it may allow for continuation of a HAART regimen that

is not completely suppressive, but that produces an unfit virus,

with less T-cell depletion. In these situations, it may be best to

keep patients on their current therapy despite measurable viral

load measurements (provided their CD4 cell counts are stable)

until newer treatment options become available.

Given H.G.’s extensive antiretroviral drug history and his current failure, a genotype, phenotype, or virtual phenotype will

likely provide some insights into potential therapeutic options.

SPECIAL CIRCUMSTANCES

Discordant Surrogate Markers

CASE 73-2, QUESTION 3: H.G. was started on a regimen

of tenofovir, 3TC, etravirine, and darunavir–ritonavir. During the following 6 months, H.G.’s T-cell counts increased

to 325 cells/μL and his viral load value declined to

5,000 copies/mL. At the last two clinic visits, H.G.’s viral load

(copies/mL) and CD4 counts (cells/μL) were 2,000/275 and

less than 50/225, respectively. H.G. reports no new clinical

complaints or adverse drug events. In addition, H.G. states

1712 Section 14 Infectious Disease

FIGURE 73-5 (Continued )

that he has been compliant with therapy. Are changes to

therapy necessary?

In most cases, declines in viral load result in increases in CD4

cell counts, and increasing viral loads result in declining CD4 cell

counts. In approximately 20% of cases, however, T cells decline

along with viral load decline or T cells increase along with viral

load increases.5 These situations are referred to as discordant

surrogate marker data.

In this situation, it may be wise to continue current treatment

and monitor the patient closely, because changing or adding

an additional drug may not increase CD4 cell counts.5 When

T-cell counts increase despite increasing viral load measurements, decisions regarding therapeutic options are unknown.

Although these patients may derive clinical benefit from the

1713Pharmacotherapy of Human Immunodeficiency Virus Infection Chapter 73

current antiretroviral regimen for a short period of time, data

suggest that long-term stability of the CD4 count is unlikely.5

Because this may be an early warning sign of impending CD4 cell

destruction, many clinicians would change antiretroviral therapies. Other clinicians, however, might follow patients carefully

and change therapies at the first sign of T-cell declines or new

clinical signs and symptoms. In these situations, other potential

causes of viral load increases (e.g., recent infection, vaccination)

should be carefully evaluated to prevent inappropriately changing a nonfailing regimen.

Therapeutic Drug Monitoring

CASE 73-3

QUESTION 1: P.P. is a 30-year-old HIV-positive man who

has failed prior antiretroviral regimens. Based on phenotypic resistance testing, it appears that a regimen including

lopinavir–ritonavir may be able to overcome the viral isolate’s resistance pattern. Will plasma measurements of P.P.’s

antiretroviral drug concentrations be useful? What measurement should be used, which agent(s) should be measured,

and how should the samples be collected?

Pharmacokinetic evaluations of various antiretrovirals,

including the NNRTIs, PIs, and raltegravir have shown wide

interpatient variability in drug exposures among cohorts of

patients taking the same dose of drug under the same

conditions.47,153–157 Many factors contribute to interpatient variability in drug exposure, such as pharmacogenetics, environment, different physiologic conditions, regimen adherence, and

drug interactions. For most antiretrovirals, a drug exposure–

response relationship exists (e.g., the higher the exposure, the

faster and more prolonged the viral suppression). In addition,

most antiretrovirals also have well defined exposure–toxicity relationships.

Therapeutic drug monitoring (TDM) is currently recommended for selected clinical situations.156 In patients failing a first

regimen, or a regimen that was initially fully suppressive, TDM

may identify suboptimal drug concentrations. For patients with

uncharacterized drug interactions and those with impairments in

gastrointestinal, hepatic, or renal function, drug concentrations

may help identify low or high exposure that can be corrected with

a dosage adjustment. TDM may also be useful for assuring that a

novel antiretroviral combination does not have any unpredictable

adverse drug interactions. In treatment-experienced patients,

knowledge of drug exposure with viral susceptibility may assist in

designing an optimal dosage regimen. Conversely, patients experiencing toxicities thought to be concentration-dependent (e.g.,

neuropsychiatric effects of efavirenz) may benefit from TDM

as well. Monitoring adherence and evaluating pharmacokinetics in special populations, such as pregnant women or pediatric

patients, are additional indications for TDM.

Pharmacology experts currently recommend obtaining

trough concentrations immediately before the next dose of a

PI or an NNRTI. For efavirenz (which is usually taken in the

evening), samples obtained 12 hours postdose will closely reflect

the concentration at 24 hours postdose due to its long half-life.

Many factors can affect the trough values of various antiretroviral

agents. For proper interpretation of the concentrations, patients

should provide a dosing history from the last several days, a list

of concomitantly administered drugs to screen for interactions,

and the exact time the last dose was taken. The exact time the

TDM sample was collected should also be recorded. Another

factor that can affect interpretation of drug concentrations over

time is intrapatient variability. Although not fully evaluated, it

appears that under stringent pharmacokinetic study conditions

(e.g., study conditions in which dosing, concomitant food administration, drug interactions, and adherence all are controlled) the

day-to-day variability of drug concentrations over time is minimal. Outside of a clinical study, however, where patients take

their medications under conditions that vary from day to day, the

concentrations could be highly variable at clinic visits over time,

and several samples might be required to determine trends in

the drug exposure before making a dose adjustment.158 Finally,

to minimize laboratory variability and error in measuring these

concentrations, it is also recommended that a laboratory that routinely measures antiretrovirals and participates in both internal

and external quality control programs be used. A list of such laboratories can be obtained from the Clinical Pharmacology Quality

Assurance and Quality Control Program (https://www.fstrf.

org/apps/cfmx/apps/cpqa/cpqaDocs/public/index.html).

In general, total drug concentration ranges for defining efficacy and toxicity have been developed for PIs and NNRTIs.156 A

limitation of obtaining total drug concentrations is that unbound,

active drug fraction is not quantitated. These agents are generally

highly protein-bound, and alterations in the bound or unbound

fraction can contribute to both reduced efficacy and increased

toxicity within accepted total drug concentration ranges. Despite

this fact, total concentrations are generally used owing to technical challenges inherent in determining unbound drug concentrations, and because a reasonable relationship exists between these

concentrations and clinical outcomes. The plasma or serum concentrations of NRTIs, however, may or may not be reflective of

the active intracellular triphosphate moiety, but may be useful to

measure in certain situations. Limitations to determining intracellular concentrations include the sophisticated processing of

the cellular components before storage, the technical difficulty

of the assay itself (generally only found in pharmaceutical companies and academic centers), and the limited availability of instrumentation to measure the low concentrations of these moieties

(e.g., fmol/106 cells).

The interpretation of the drug concentration itself will vary

according on the clinical situation. Among treatment-na¨ıve

patients with wild-type viral isolates, an assessment should be

made whether the concentration lies within a range of the population concentrations—how that range is defined tends to be

clinician-specific, but can be interquartile range, or a 90% confidence interval, among others. If the concentration is less than

the desired cut-off, and the patient is not responding well to the

therapy, the clinician should consider increasing the dose or using

a pharmacokinetic-enhancing technique (e.g., adding ritonavir).

If the concentration is higher than the cut-off, and the patient

is experiencing drug toxicity, a dose reduction should be considered. Clinical trials evaluating TDM among antiretroviral-na¨ıve

patients initiating PI-based HAART have shown improved clinical responses compared with patients given standard of care

without TDM. However, many of these improved responses

were seen with older protease inhibitors such as indinavir and

nelfinavir that had narrower therapeutic ranges than the newer

antiretroviral drugs recommended as first line agents. A study

performed assessing TDM of more recent PI-based HAART regimens in treatment experienced patients failed to show an overall

benefit.159

Among patients with antiretroviral resistance, the interpretation of the concentration is much more complex and may

require the additional assessment of phenotypic resistance data.

Higher drug exposure may be needed for patients with drugresistant viruses to achieve efficacy. Preliminary investigations

have evaluated the ratio between drug exposure and viral susceptibility (in a similar fashion to the antimicrobial efficacy ratios of

Cmin/minimal inhibitory concentration [MIC] or area under the

1714 Section 14 Infectious Disease

curve [AUC]/MIC), also called an inhibitory quotient. Interpretation of the current inhibitory quotient data is difficult, however,

owing to the various methodologies used for its calculation.156

Ongoing research is focused on standardizing calculation methods to improve the prediction of appropriate antiretroviral concentrations for maximal efficacy.

In the current case, assuring that P.P.’s lopinavir trough concentration is at least greater than the IC50 value from the phenotype may help improve his chances of optimal antiviral response.

If P.P.’s response to treatment is inadequate (e.g., unsatisfactory

decline in HIV RNA), then a concentration should be obtained

from a trough sample, with a careful concomitant drug and dosing history, and sent to a reliable laboratory for analysis.

Drug Interactions

CASE 73-4

QUESTION 1: J.F. is a 37-year-old HIV positive man who

has been taking emtricitabine, tenofovir, and atazanavir/

ritonavir for the past 4 years. Lately, he has been experiencing heartburn and was diagnosed with gastroesophageal

reflux disease (GERD). What needs to be considered when

starting a new medication for J.F.’s acid reflux?

Drug interactions are very common with antiretrovirals and

need to be considered when changing HIV-related or non-related

drug regimens. In this case, stomach acid suppressants will significantly decrease the concentration of atazanavir as it requires

an acid environment for complete dissolution.51 If J.F.’s GERD

needs treatment, H2 receptor antagonists should be given either

simultaneously with ritonavir-boosted atazanavir or at least

10 hours before administration of the antiretrovirals. A dose of

famotidine or its equivalent should not exceed 40 mg twice a

day (BID) for treatment-na¨ıve patients, and famotidine 20 mg

BID or its equivalent should not be exceeded in treatmentexperienced patients. Additionally, atazanavir AUC exposures are

decreased by 25% with concomitant tenofovir and Cmin can be

decreased 23% to 40%. Therefore, in treatment-experienced

patients taking tenofovir, atazanavir should be given at an

increased dose of 400 mg daily in combination with 100 mg of

ritonavir. Proton pump inhibitors (PPIs) are not recommended

for use in protease-inhibitor–experienced patients. In treatmentna¨ıve patients who require a PPI, only atazanavir–ritonavir 300

mg/100 mg should be used, and separated from the PPI by 12

hours or more.5

Pregnancy and Breast-Feeding

CASE 73-5

QUESTION 1: T.D. is a 32-year-old woman infected with HIV

whose antiretroviral therapy includes efavirenz, tenofovir,

and emtricitabine. She has had a positive home pregnancy

test result. Is her current antiretroviral regimen appropriate

for use during pregnancy and for prevention of mother-tochild transmission?

The current perinatal HIV guidelines75 recommend that

women receiving and tolerating a currently suppressive regimen

when they become pregnant continue on that regimen unless

it contains efavirenz. Efavirenz is the only antiretroviral that is

currently pregnancy category D and is not recommended in pregnancy or breast-feeding. It has been shown to be teratogenic in

animal studies, particularly in the first trimester of pregnancy,

and retrospective case reports of neural tube defects in humans

have been documented.160–163 Perinatal transmission prevention

guidelines recommend lamivudine and zidovudine as first-line

NRTIs and abacavir, didanosine, emtricitabine, and stavudine

as alternative agents. Nevirapine is the recommended NNRTI

if the CD4 count is less than 250/μL and lopinavir–ritonavir is

the recommended protease inhibitor during pregnancy. Many of

the pregnancy recommendations are based on those drugs which

have the most safety and efficacy data available during pregnancy

(both animal and human).75

In women who have a viral load persistently greater than

1,000 copies/mL, it is recommended that a planned 38-week

cesarean section be performed to reduce the risk of motherto-child transmission. In women with a viral load less than

1,000 copies/mL, there is little evidence to show that cesarean

section decreases transmission rates compared to vaginal delivery. In this case, the decision is made at the discretion of the

physician in consultation with the mother. Additionally, it is recommended that intravenous zidovudine be administered to the

mother at 2 mg/kg intravenously (IV) for 1 hour at the onset of

labor, then given as a continuous infusion of 1 mg/kg/hour during the intrapartum period (during labor and postpartum). Once

delivered, it is recommended that oral zidovudine at 2 mg/kg

every 6 hours be started in the infant