Therapeutic Decision Making in 1997: Roundtable Discussion of Five Cases

Therapeutic Decision Making in 1997: Roundtable Discussion of Five Cases

In today's extremely dynamic therapeutic environment, care of the individual patient is more complex than ever. Practitioners feel compelled to use very new drugs in barely tested combinations. Promising early reports from clinical trials are increasingly tempered by concerns about side effects, drug interactions, compliance, and cost. For this issue, ACC asked three expert AIDS physicians -- John Bartlett of Johns Hopkins, Robert Schooley of the University of Colorado, and Roy Gulick of New York University -- to independently recommend approaches to antiviral therapy in five clinical vignettes based on actual cases. Although our panel members' therapeutic philosophies differ, as do their specific approaches, I think you will find their advice thoughtful and pragmatic. Readers are cautioned that some of these drugs and/or approaches have not received FDA approval.

We at ACC are committed to keeping you well informed and confident in these difficult times. To this end, the next two issues will carry detailed reports from the 4th Conference on Retroviruses and Opportunistic Infections, held in January in Washington, D.C. -- Deborah Cotton

Case One: M.K., A Possible Pregnancy in Her Future

M.K. is a 35-year-old African-American female who was found to be HIV seropositive in 1992. She was treated briefly with AZT in 1994, at which time her CD4 count was 420 cells/mm³. This medication was stopped after several months due to mild GI intolerance. She now has a CD4 cell count of 300 and a viral load of about 20,000 copies/ml by HIV-1 PCR. She has heard that AZT can reduce the likelihood of vertical transmission of HIV, and may want to consider pregnancy in the future, although not for at least one year.

Given the possibility of future pregnancy, which antiretroviral regimen would you choose for this patient now? Should AZT be reserved, given its proven efficacy in reducing vertical transmission? If she does become pregnant, what antiviral regimen would you use? Would you stop some or all antiretroviral agents during the first trimester to prevent teratogenicity? How frequently would you monitor viral load during pregnancy and would you adjust therapy if virus remains detectable?

Dr. Bartlett: My initial recommendation is a double nucleoside regimen, probably AZT plus either ddI, ddC, or 3TC. The addition of indinavir or ritonavir to "two nucs" is more aggressive and may be a standard in the near future. The consideration of pregnancy would not alter this decision. However, we need to treat this patient based on the current numbers and clinical features, which show a depleted immune reserve and relatively high viral burden. It is important to provide counseling about the consequences of pregnancy in terms of both vertical transmission and possible need for foster parents. If the patient does become pregnant, we need to distinguish between drugs given to prevent vertical transmission and those given to treat the mother for HIV infection. For pregnancy, AZT is the only drug with established merit and recent reports indicate that even those who show no response in terms of viral load derive substantial benefit in terms of the rate of vertical transmission (see ACC p. 16). Therefore, the viral load would be used primarily to monitor the treatment of the mother. I am not sure we know how to use this information to guide decisions on modifying therapy to reduce vertical transmission. If my initial regimen (such as AZT plus 3TC) worked well, I would continue its use through the first trimester and thereafter. The patient should be informed about the lack of information on AZT during the first trimester and the lack of information about the consequences of 3TC for the infant. The extensive and favorable experience with AZT has made me quite comfortable using this drug, and preliminary data on 3TC in pregnancy indicates it is safe. Because of the AZT experience many feel the entire class of nucleosides is probably safe, but we can't be sure.

Dr. Schooley: M.K. would be best managed by initiating therapy with a combination of two nucleoside analogs and a protease inhibitor in an effort to drive levels of HIV-1 RNA to below 200 to 500 copies. Possible regimens include AZT plus ddI plus either indinavir or ritonavir. If she does not tolerate AZT, alternative nucleoside analog components of the regimen include d4T plus 3TC or ddI plus d4T.

There is no evidence that there is any specificity to the observation that AZT decreases transmission of HIV-1 in the perinatal setting. Recent virologic data suggest that a large part of the effectiveness of AZT in preventing vertical transmission lies with preemptive treatment of primary infection of HIV-1 in the neonate. Thus, the major consideration should be to try to use antiretroviral drugs in such as way as to reserve agents to which the virus will be susceptible at the time of delivery. Since M.K. has clear indications for treatment, it would be appropriate to use the drugs in such a way as to defer the induction of resistance. This is most likely to be accomplished if she is treated with a potent enough combination regimen to drive plasma RNA levels below the limit of detection.

If M.K. becomes pregnant and is on a regimen that has suppressed HIV-1 to levels below detectable, it would be appropriate to leave her on that regimen. The data on teratogenicity of antiretroviral chemotherapeutic agents are not complete, but animal studies and what human data there are do not suggest a major problem with induction of birth defects. Although most transmission of HIV-1 in the perinatal setting appears to occur near the time of delivery, allowing the virus to rebound during the first trimester would not seem to be a good decision. Thus, I would be willing to leave her on antiretroviral chemotherapy during the first trimester.

Her plasma RNA levels should be monitored during pregnancy much as before pregnancy. If plasma viral RNA levels are not detectable, following the levels every three months is reasonable assuming morning sickness or other problems don't compromise compliance. If drugs must be changed, at least two should be changed simultaneously.

Case Two: R.C., Alcoholism and Recurring Bouts of Pancreatitis

R.C. is a 43-year-old white male who suffered penetrating trauma to the abdomen in 1993 and underwent a Whipple procedure. Subsequently he had multiple bouts of pancreatitis attributed to alcohol use. In 1996 he was found to be HIV seropositive with a CD4 count of 205 and a viral load of 230,000 copies. He was started on AZT plus 3TC and TMP-SMX for PCP prophylaxis. He had another bout of pancreatitis, not associated with alcohol use, several months after starting treatment. His most recent CD4 count was 326, with a viral load of 2400 copies/ml by HIV-1 PCR. TMP-SMX was discontinued. He is now hospitalized for yet another bout of pancreatitis.

Would you continue AZT plus 3TC? If not, what regimen would you choose, given his risk for pancreatitis? Should TMP-SMX have been discontinued?

Dr. Bartlett: The major question with R.C. seems to be the cause of pancreatitis, since in the past this has been associated with alcohol use but now he has had relapses of pancreatitis while taking two drugs that have been implicated in this complication. TMP-SMX is probably the easiest drug to change because he now has a CD4 cell count of 326, making PCP less likely. His numbers show a good response to AZT plus 3TC, so I would prefer not to change that part. The last bout of pancreatitis was apparently independent of TMP-SMX, raising a question about the role of the nucleoside analogs. 3TC studies have shown relatively high rates of pancreatitis in pediatric patients and it may cause this complication in adults, but I am not aware of evidence to that effect beyond anecdotal reports for which cause and effect were not clearly defined. Options for new antiretroviral therapy for R.C. are somewhat limited. 3TC is thought to be the offending agent. My choice would be AZT plus a protease inhibitor if compliance can be assured. At our clinic the major cause of missed doses is alcoholic bouts, and this may be a problem here. Another option is ddI alone, but ddI has been implicated as a cause of pancreatitis in 0.5% to 1% of recipients; besides, ddI may be viewed as a BB gun in a setting that requires a cannon.

Dr. Schooley: It is difficult to determine whether R.C.'s pancreatitis is drug-related or, possibly, a late sequelae of anatomic defects associated with prior bouts of pancreatitis. If the pancreatitis is not severe and if he is able to take his drugs by mouth, it would be reasonable to see if the pancreatitis resolves on his antiretroviral drugs and TMP-SMX sulfa. If the pancreatitis is moderately severe or severe (which it probably is, given the need to hospitalize him), he will likely be made NPO and stopping all drugs will be required. After resolution of the pancreatitis, it would be reasonable to restart his antiretroviral drugs, since he would not be a candidate for ddI therapy in the future. Because the antiviral activity of AZT and 3TC in vivo is time-limited and because a loss of control of viral replication on this regimen will require alteration of his nucleosides, he should be started on a protease inhibitor as soon as the pancreatitis resolves. This will prevent having to devise a potent non-ddI-containing regimen when a protease inhibitor is ultimately started. Given the association of pancreatitis with pentamidine, I'd try to resume the TMP-SMX before moving to less efficacious regimens for PCP prophylaxis, if his CD4 cell count falls below 200.

Dr. Gulick: I would continue AZT plus 3TC. Based on his history, R.C. may have chronic pancreatitis as a complication of his previous episodes of alcoholic pancreatitis. AZT has not been associated with pancreatitis. Although 3TC has been associated with pancreatitis in pediatric clinical trials, it is a rare cause of pancreatitis in adults (<0.5%). Given his encouraging responses in both viral load and CD4 cells, I would continue R.C.'s AZT plus 3TC, and consider further modifications in his antiretroviral regimen when his pancreatitis improves. His recurrent episodes of pancreatitis serve as a relative contraindication to ddI-, ddC-, and, to a lesser extent, d4T-containing regimens. Pancreatitis does not contraindicate the use of protease inhibitors. Alcoholism, of course, may complicate compliance, particularly with complex regimens containing protease inhibitors.

TMP-SMX has been associated with pancreatitis, though more often at higher doses than at a typical PCP prophylaxis dose. Drug-induced pancreatitis may be more likely to occur in a patient with preexisting pancreatic disease. Given that this patient never had a documented CD4 count less than 200, and has now experienced a CD4 cell rise on antiretroviral therapy, it was reasonable to discontinue TMP-SMX to avoid risking TMP-SMX-induced pancreatitis. If his CD4 count were to fall below 200, the risks and benefits of prophylaxis would need to be reassessed.

Case Three: D.S., An Antiretroviral-Naive Breast Cancer Patient

D.S. is a 54-year-old white female who has been HIV positive for at least eight years. She has never received any antiretroviral. Over the last two years her CD4 cell count has fluctuated between 525 and 650. Her viral load measurement is now 80,000, which has been confirmed on a second test. She was recently diagnosed with an invasive ductal carcinoma of the breast, without evidence of spread to lymph nodes or bone. Tamoxifen has been started for treatment of her breast cancer. She also takes oral acyclovir for chronic suppression of genital herpes. In addition, she is reluctant to start antiretrovirals, particularly AZT.

Should antiretrovirals be recommended? Should protease inhibitors be included in the initial regimen? Should her cancer diagnosis and treatment alter her management in any way?

Dr. Bartlett: Most would recommend initiation of antiretroviral therapy due to D.S.'s high viral load. The patient ultimately makes this decision, and many patients never fill prescriptions or, if they do, comply poorly. It just doesn't make sense to try to force therapy on a patient who does not want it. Thus, the initial effort should be to simply inform the patient about the options with both doctor and patient understanding that D.S. will ultimately make the decision. Therapy without AZT could be initiated but she should be cautioned about fallacious information from "the underground." Options include monotherapy with ddI or, if she wants something really simple and easy, monotherapy with d4T. This was acceptable after ACTG 175 results were first announced, but monotherapy with any drug has become less justifiable since that time. Alternatives are ddI plus d4T or d4T plus 3TC. I don't think her cancer chemotherapy is an obstacle since the regimens suggested do not suppress the marrow. I would suggest a protease inhibitor with two nucleosides, but am concerned about side effects, given a reluctant patient and cancer chemotherapy. Protease inhibitors should clearly be used if the above regimens fail and CD4 count drops or viral load rises.

Dr. Schooley: Given D.S.'s relatively high level of plasma HIV-1 RNA, she is an excellent candidate for antiretroviral chemotherapy and it should be recommended. The inability of AZT and 3TC to maintain durable control of viral replication, particularly when plasma HIV-1 RNA levels are in this range, make the addition of a protease inhibitor to the regimen advisable. A pharmacologist could be consulted about whether tamoxifen metabolism is affected by inhibition or induction of cytochrome P450 enzymes.

Dr. Gulick: Antiretroviral therapy should be recommended, given the patient's confirmed elevated viral load.

Protease inhibitors should probably not be added. The benefits of protease inhibitor therapy in patients with CD4 counts above 500 (except in acute HIV disease) remain to be established. Given the patient's reluctance to take antiretrovirals, which may affect her compliance, it is probably reasonable to recommend a simple initial regimen: ddI monotherapy (with the new formulation), or perhaps d4T plus 3TC, which are easy-to-take initial regimens. After the patient has some experience and success with antiretrovirals, more complicated regimens could be considered.

Tamoxifen may increase her chances of cancer-free survival and antiretroviral therapy may slow HIV disease progression and improve survival, though this remains to be formally demonstrated in the CD4 >500 group. Optimally treating both her breast cancer and HIV disease now will also give her the best chance to benefit from future advancements in the treatment of both diseases. However, tamoxifen has been associated with nausea, vomiting, and liver function test abnormalities, which may complicate the management of new antiretroviral agents.

Case Four: A.K., Multiple OIs and Neuropathy

A.K. is a 30-year-old white male with multiple opportunistic infections. He was diagnosed with cerebral toxoplasmosis in 1994. AZT monotherapy was instituted but stopped because of GI symptoms and severe neutropenia. He was then enrolled in a trial of ddI (open label) with or without delavirdine (blinded), and over the next six months his CD4 cell count fell from 60 to 30. He then developed cryptococcal meningitis with a protracted hospitalization complicated by hydrocephalus, extrapulmonary PCP, HIV-associated gastroparesis, recurrent pulmonary emboli, and wasting.

He slowly recovered and in the spring of 1996 was placed on d4T plus 3TC plus indinavir. His HIV-1 viral load became "undetectable" (<300 copies/ml) and his CD4 cell count rose to 100. However, he developed CMV retinitis, which was treated with IV foscarnet. During the summer, a painful, debilitating lower extremity neuropathy resulted in the discontinuation of d4T, 3TC, and then foscarnet. He was switched to AZT plus indinavir and cidofovir. This time he was able to tolerate full-dose AZT (200 mg three times daily) without GI or hematologic toxicity. His neuropathy improved but quickly recurred when he was rechallenged with 3TC. On AZT and indinavir his CD4 cell count has continued to rise (to 150), but his viral load has increased to 300,000. He is also receiving TMP-SMX for PCP prophylaxis and clarithromycin for MAC prophylaxis.

What does the discrepancy between CD4 cell count and viral load suggest? Given his neuropathy, what changes in the antiretroviral regimen could be made? If his CD4 count remains above 100, can his MAC prophylaxis safely be stopped, given the many other medications he must take?

Dr. Bartlett: We have seen this discrepancy between CD4 cell count and viral load in our clinic, although it does beg explanation. A possible explanation is the SI phenotype. It is important to make sure both readings are correct. If they are, my preference for monitoring antiretroviral therapy is with the viral burden, but we need to watch CD4 as well. The problem for A.K. is he has limited options if the current antiretroviral regimen is delared a therapeutic failure. The neuropathy will require exclusion of ddI, ddC, and d4T, and, since the "indinavir card" has already been played, options for protease inhibitors are limited. Saquinavir plus ritonavir with or without AZT could be tried. Another option for consideration is the nelfinavir open-label protocol with AZT as the companion drug. The decision about MAC prophylaxis for A.K. would be controversial because it is unclear that CD4 cells are as competent "on their way up" as "on the way down." I would continue clarithromycin if it is well tolerated, although the dose may need to be adjusted if ritonavir is added. I would feel more strongly about continuation of MAC prophylaxis if it were secondary rather than primary.

Dr. Schooley: It is difficult to weigh 30- to 50-cell fluctuations in the CD4 count. Given the debilitating nature of A.K.'s peripheral neuropathy, he is not a candidate for regimens containing ddC, ddI, or d4T. Although delavirdine or nevirapine could be added, neither alone is likely to have a sustained effect in view of the 300,000 copies/ml of HIV-1 RNA in his plasma. It would be appropriate to determine whether he received delavirdine during the study. If delavirdine is added to the current regimen his indinavir dose should be decreased. Nevirapine will have the opposite effect on indinavir levels. All things considered, it might be best to stay with the current regimen and simultaneously add GlaxoWellcome's 1592, a nucleoside analog with a 2 log₁₀ effect on indinavir levels, plus delavirdine when the two new agents become available in 1997. This choice would be more attractive if he is found not to have received delavirdine when the study in which he participated is unblinded.

We as yet lack robust data indicating whether prophylaxis for opportunistic pathogens can be discontinued after CD4 cells rise on potent antiretroviral regimens. If a patient has already had a specific infection and is on secondary prophylaxis, it is generally not prudent to stop prophylaxis. For primary prophylaxis, the situation is less clear. Studies of cessation of MAC prophylaxis are currently under development in the ACTG and the CPCRA. For A.K., if the clarithromycin significantly complicates the ability to use other needed drugs, it would not be unreasonable to stop MAC prophylaxis if his CD4 cell count remains above 100.

Dr. Gulick: In general, although not always, an antiretroviral-induced fall in viral load precedes a corresponding rise in CD4 cell count. The goal of antiretroviral therapy is both to control viral infection and to improve immune function. A rise in CD4 number does not tell us about the actual function of these cells. Recent data suggest that the "new" CD4 cells result from peripheral expansion of memory T cells rather than naive T cells. Some have suggested that a patient whose CD4 count falls below a critical number (50) may lose certain CD4 clones, resulting in permanent immune system deficits. CMV retinitis is unusual with a CD4 cell count of 100 and its occurrence in A.K. raises concern about whether his new antiretroviral regimen has restored immune function.

3TC is an uncommon cause of peripheral neuropathy, but rechallenge of this patient with this agent seemed to provoke a recurrence of symptoms. The severe peripheral neuropathy also likely contraindicates ddI, ddC, and d4T. Because A.K.'s current antiretroviral regimen of AZT plus indinavir appears to be failing (pending confirmation of the viral load of 300,000), other options should be considered. Since he needs maintenance therapy for CMV disease, reinstitution of foscarnet, which has activity against both CMV and HIV, should be considered. The best antiretroviral options when indinavir has failed are unknown. Recent data have shown the short-term safety and antiretroviral activity of saquinavir plus ritonavir (at reduced doses of 400 mg twice daily each) in a group of protease inhibitor-naive patients, many of whom had taken all five approved nucleoside analogs. The risks and benefits of a protease inhibitor combination in a protease-inhibitor-experienced patient are unknown at present. The experimental protease inhibitor nelfinavir is available for compassionate use, but its benefit after failure of indinavir is also unknown. Clinical trials may provide access to other experimental agents, e.g., Burroughs Wellcome 1592, a new nucleoside analog or 141W94, a new protease inhibitor. Nevirapine should not be used as a single agent and its use with protease inhibitors is not recommended since it lowers the blood levels of these drugs.

It is unknown whether OI prophylaxis should be stopped when CD4 cell counts rise on antiretroviral therapy. Since the function of this patient's CD4 cells is in question, particularly given his development of CMV retinitis, I would not stop his MAC prophylaxis. The clarithromycin may incidentally be helping his HIV gastroparesis as well.

Case Five: G.J: Paradoxical CD4 and Viral Load Values

G.J. is a 31-year-old white male, HIV positive since 1985. For the first 10 years of infection, he received no antiretroviral therapy and was free of opportunistic infection. In August 1995, he developed PCP and esophageal candidiasis. In December 1995, he initiated therapy with AZT plus 3TC. Base-line pre-therapy CD4 count was 8 and his viral load was 597,000 by bDNA assay. The patient tolerated the antiretroviral regimen well; three months into therapy his viral load had dropped below the limits of detection of this assay (<10,000 copies/ml). At that time, indinavir became available and was added to his regimen. His viral load was assayed again after he had been on AZT plus 3TC for four months and indinavir for one month; it had risen to 177,000 copies by bDNA. One month later, the patient developed herpes zoster and had a CD4 count of 108. During month three of indinavir therapy and month six of AZT plus 3TC, G.J.'s CD4 count had risen to 148 and his viral load was 208,000 copies (by now HIV-1 RNA PCR was being done). On month six of indinavir therapy and month nine of total therapy, his CD4 count had risen to 199 and his viral load to 358,000.

Can this parallel rise in both CD4 count and viral load be explained? What would now be the most appropriate antiretroviral strategy?

Dr. Bartlett: The paradoxical increase in both CD4 cell count and viral load escapes logic. Assuming these numbers are correct (and they should be checked on repeat assays), one has to decide which number to go with as a guide for antiretroviral therapy. I would weight the viral burden more heavily, and here the count is simply too high. But options are limited when AZT, 3TC, and indinavir have all failed. I would recommend ritonavir plus saquinavir plus either d4T or ddI. Another option is d4T plus ddI.

Dr. Schooley: There are clearly individual patients for whom CD4 and plasma HIV-1 RNA level changes will appear to be inconsistent. The CD4 cell count changes more slowly than viral load and a vaccination or intercurrent infection may transiently increase HIV-1 RNA levels. In such a case it might appear that both CD4 cells (in response to recent control of viral replication) and plasma HIV-1 RNA levels (in response to a more recent intercurrent T-cell activating event) are rising. In cases in which the CD4 cell count and plasma HIV-1 RNA levels appear to be inconsistent, prophylaxis for OIs should be managed according to the CD4 cell count, and antiretroviral therapy decisions should be driven by the plasma HIV-1 RNA level.

G.J.'s predicament illustrates the problems caused by incremental initiation of antiretroviral chemotherapy. The initiation of AZT plus 3TC without a protease inhibitor suboptimally suppressed HIV-1 replication, and this went undetected by the insensitive first-generation bDNA assay. Even with a second-generation assay, an initial regimen of AZT plus 3TC is associated with durable suppression of HIV-1 RNA in only a minority of patients with advanced disease. The subsequent addition of indinavir in such patients results in only a temporary improvement in antiretroviral activity, since this approach adds a single drug to a failing regimen. When this happens, the patient's options are limited. One approach might be to change the regimen to d4T plus ddI plus ritonavir plus saquinavir. An alternate approach might consist of d4T plus ddI plus indinavir (in a reduced dose) plus delavirdine. Neither approach, however, can be expected to result in durable improvement in control of the patient's viral replication.

Dr. Gulick: Several clinical trials have documented a dissociation between responses in viral load and CD4 count. Diverse regimens (AZT plus 3TC; indinavir monotherapy) in several studies have resulted in sustained CD4 count rises in the face of viral load climbs toward baseline over 6 to 12 months. In general, a rising viral load will in time be accompanied by a declining CD4 cell count. In experiencing in nine months a CD4 rise from 8 to 199 while viral load was increasing from undetectable to 358,000, G.J. is an extreme example.

G.J.'s therapy is failing and should be changed. He's had two viral loads over 200,000 within three months and without obvious confounders such as vaccination, herpes, or other opportunistic infection. The best option for a patient whose AZT plus 3TC plus indinavir regimen fails is unknown. A new antiretroviral regimen would ideally consist of two new drugs, perhaps d4T plus ddI or AZT plus ddI plus nevirapine. It would be reasonable to start one of these regimens and recheck the viral load in four to six weeks for an initial response.

— Deborah Cotton, MD, Moderator

John Bartlett is Chief, Division of Infectious Diseases, Johns Hopkins University School of Medicine. Robert Schooley is Tim Gill Professor of AIDS Research and Head, Division of Infectious Diseases, at the University of Colorado Health Sciences Center. Roy Gulick is Instructor in Medicine, New York University School of Medicine and Director, HIV Research Clinic, at Bellevue Hospital.

Published in AIDS Clinical Care February 1, 1997