Journal of Clinical Immunology, Vol. 11, No. 2, 1991
Special Article
Allergic Manifestations of Human Immunodeficiency Virus (HIV) Infection ANDREW CARR, 1 DAVID A. COOPER, 1'2 and RONALD PENNY 1
INTRODUCTION
Accepted: November I5, 1990
Drug allergy is the most common and significant allergic manifestation of HIV 3 infection. Initially described in patients treated with SMX-TMP for PCP, allergy is now known to involve a multitude of drugs. The pathogenesis of, and risk factors for, allergy in HIV infection are poorly understood, although there is evidence suggesting that allergy is more common with advancing immunodeficiency. HIV-negative subjects with sulfonamide allergy may have drug-specific antibodies and drug metabotiteinduced lymphocyte cytotoxicity, abnormalities that could partly explain the allergic mechanisms and which may have future diagnostic potential; these abnormalities have not been described in HIV-infected subjects. Therapy includes avoidance, suppressive agents such as corticosteroids, and desensitization, although the appropriate role for each is not entirely clear. Serum IgE levels have been shown to rise with progressive disease; those patients with higher levels may have a worse prognosis. The mechanisms of this rise are multifactorial, probably a combination of altered T-lymphocyte regulation of IgE synthesis and of production of specific IgE directed against microbial antigens.
Study of allergic disease in HIV 3 infection has focused on two areas, namely, drug allergy and regulation of IgE synthesis by T lymphocytes. Drug allergy was the first clinical manifestation of allergy described in HIV infection and remains the most important. HIV-infected patients treated with highdose SMX-TMP for PCP were found to develop cutaneous reactions at a greater frequency than either immunocompetent or HIV-negative immunodeficient subjects (I, 2). These reactions are presumed to have an allergic (i.e., immune-mediated) basis. The clinical features and frequency of drug allergy have been described, and drug allergy has since been identified in HIV-infected patients with the use of multiple drugs. The occurrence of drug hypersensitivity appears paradoxical, considering the profound anergy that develops in these patients to other immune stimuli; this paradox has yet to be explained. An improved understanding of both the underlying risk factors and mechanisms, neither of which have been studied to any significant degree, will allow for improved strategies to circumvent this hypersensitivity. One therapeutic strategy has been desensitization, which has allowed for improved therapy for some allergic patients. Abnormalities of IgE synthesis have been described in immunodeficiency states of T lymphocytes other than HIV infection, both congenital and acquired, such as the Wiskott-Aldrich syndrome and acute graft-versus-host disease. In HIV infection, a progressive rise in serum IgE levels has been found and attributed to alterations in the levels of T
KEY WORDS: Drug hypersensitivity; human immunodeficiency virus (HIV); IgE; desensitization; Pneumocystis carinii pneumonia.
1Centre for Immunology, St. Vincent's Hospital, Darlinghurst, Sydney, Australia 2010. ZNational Centre of HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia. 3Abbreviations used: HtV, human immunodeficiency virus: SMX-TMP, sulfamethoxazole-tlimethoprim: PCP, Pneumocystis carinii pneumonia; IgE, Immunoglobutin E; EBV, EpsteinBarr virus; CMV, cytomegalovirus; IFN, interferon; IL, interleukin.
55 0271-9142/91/0300-0055506.50/0© 1991PlenumPublishingCorporation
56
lymphocyte-derived cytokines and, also, to specific microbial antigens. Specific IgE may have clinical significance; Pneumocystis carinii-specific IgE has been proposed as a possible contributing factor in the acute pneumonitis, seen early in the treatment of PCP, that can be suppressed with corticosteroids. The rise in total serum IgE has also been shown to have prognostic value in some patients and should prove of clinical value if found to be independent of other prognostic markers.
CARR, COOPER, AND PENNY
Table I. Frequency of Drug Allergy • Allergy to sulfonamides complicating treatment and prophylaxis of PCP in 30-65% of patients; prevalence with other drugs unclear • Allergy more common in HIV-infected than in immunocompetent or HIV-negative immunosuppressed patients • Allergic manifestations may resolve without ceasing therapy in 30% of cases • Up to 80% of patients may tolerate sulfonamide for secondary PCP prophylaxis despite allergy during PCP therapy (t4) • Degree of cross-reactivity between different sulfonamides varies, and may be as low as 10% • Multiple drug allergies may occur in the one patient
FREQUENCY OF DRUG ALLERGY
Sulfonamides Before the advent of HIV, SMX-TMP had been shown to be the therapy of choice for PCP, as it caused less toxicity than pentamidine and was more effective (3, 4). However, in HIV-infected patients with PCP, toxicity with the use of SMX-TMP occurred in 65% of patients, compared to just 12% of HIV-negative immunosuppressed patients (5). More specifically, the frequency of SMX-TMP allergy differs markedly among subjects who are immunocompetent, HIV-infected, or immunodeficient but HIV-negative. In the largest study of SMX-TMP allergy in the general population, 38 cutaneous reactions were found in 1121 patients (3.3%) who received SMX-TMP; 19 of these reactions occurred within 3 days, and the remainder within 2 weeks (6). This patient cohort was based upon admissions to several general medical, surgical, and psychiatric wards unlikely to contain immunosuppressed patients and, as such, corresponds best with the general community. Similarly, of 19 H/V-negative immunosuppressed adults with PCP treated with SMX-TMP, only I patient developed a drug rash (7, 8). However, the frequency of SMX-TMP allergy in HIV-positive patients has been estimated at between 29 and 65% (5, 9-11) (Table I). However, only 0-33% of those in the above studies needed to discontinue therapy as a result of hypersensitivity. This wide variation was partially dependent upon the severity both of the reaction and of the PCP itself. In those patients in whom allergy was associated with high fever, StevensJohnson syndrome, or presumed visceral involvement, therapy was usually ceased (9). In contrast, patients with milder PCP may have required a shorter course of therapy and, so, would have been
less likely to have had SMX-TMP continued once an allergic reaction had occurred. Given that 85% of patients will develop PCP at least once (12) and that, futhermore, some 60% of patients will relapse within 1 year of PCP therapy without subsequent prophylaxis (13), both primary and secondary prophylaxis against PCP is now mandatory. Despite the frequency of drug allergy with SMX-TMP, it has the advantages of oral administration, home-based therapy, minimal cost, better protection against extrapulmonary pneumocystosis than aerosolised pentamidine, and perhaps a lower overall relapse rate as well. Hypersensitivity may develop to SMX-TMP administered as primary or secondary prophylaxis against PCP. SMX-TMP hypersensitivity during secondary prophylaxis following uncomplicated therapy of PCP with SMX-TMP has not been reported. Consequently, SMX-TMP may be administered to these patients without the development of drug allergy. Conversely, not all patients will develop hypersensitivity to SMX-TMP given as secondary prophylaxis if they developed hypersensitivity to SMX-TMP during therapy of PCP. Fourteen patients who developed fever and rash during SMX-TMP treatment of PCP received subsequent oral SMX-TMP as prophylaxis; only three of the paients (21%) developed rashes (14). The frequency of SMX-TMP hypersensitivity complicating primary prophylaxis was found to be as high as 50%, with a further 30% developing pruritis of uncertain significance (I5). However, only 14% of these subjects needed to discontinue therapy, all of whom developed fever and mucosal inflammation consistent with Stevens-Johnson syndrome. The different reaction rates in various reports are no doubt a reflection of differing dosage regimes and differing criteria for defining drug Journal of Clinical Immunology, Vol. 11, No. 2, 1991
ALLERGIC MANIFESTATIONS OF HIV INFECTION
hypersensitivity, for timing drug withdrawal, and for patient selection. Therapy of PCP with dapsone caused allergy in 17 to 53% of patients (16-18), although all studies were small. Only 4% of these patients needed to cease therapy because of drug hypersensitivity. However, in the comparative studies of dapsone and SMX-TMP hypersensitivity in PCP patients, there was no difference between the two groups in either the frequency of hypersensitivity or the discontinuation due to hypersensitivity (17, 18). Dapsone has been more widely used as prophylaxis than therapy. The frequency of hypersensitivity appears to be less than with SMX-TMP. There is also limited cross-reactivity with SMX-TMP, and it would appear that the majority of patients who do react to SMX-TMP can safely tolerate dapsone. However, there are no good data available to confirm these suppositions. The chief caution in the use of dapsone would be for those patients in whom SMX-TMP had induced StevensJohnson syndrome or visceral involvement, such as hepatitis or pneumonitis. Fansidar, a combination of pyrimethamine and sulfadoxine, a long-acting sulfonamide, has been widely used for PCP prophylaxis. The long-acting sulfonamides are no longer used as therapy for common bacterial infections in the general community because of the increased risk of Stevens-Johnson syndrome. The other major use of Fansidar has been as prophylaxis against malaria. However, because of multiple deaths associated with its use, it is likewise no longer recommended by the Centers for Disease Control (19). These deaths have been a result of severe cutaneous exfoliation and sepsis. To date, one fatality has been reported in HIV-infected subjects with its use (20); whether severe reactions are more common than with SMX-TMP or dapsone has not been determined. Moreover, given also that the efficacy of Fansidar as prophylaxis appears to be inferior to SMX-TMP or pentamidine, Fansidar is probably best reserved for use as prophylaxis in those in whom they are inappropriate. Two reports of its use in predominantly SMX-TMP allergic subjects found hypersensitivity developed to Fansidar in only I0 and 12% of patients (21, 22), although the nature of their allergic reactions is not always clear from the reports.
Other Drugs Pentamidine has long been assumed to be a nonallergenic alternative therapy of PCP. However, four reports have documented the occurrence of
Journal of Clinical Immunology, Vol. 11, No. 2, 1991
57
morbilliform eruptions presumed secondary to pentamidine therapy, including one prospective study with a prevalence of 15% (9). No patient was required to stop pentamidine therapy. One patient developed transient urticaria, without other features of anaphylaxis, while receiving intravenous pentamidine (23). The urticaria was localized immediately proximal to the cannula site and occurred within 5 min of the infusion commencing. Following a single intravenous injection of diphenhydramine, all subsequent infusions were administered without incident. Whether this was type I (i.e., IgEmediated) hypersensitivity or an anaphylactoid response is unknown. Similarly, the mechanism responsible for bronchospasm that occurs in a large percentage of patients receiving aerolized pentamidine is unknown. Experience with other PCP therapy is less well documented. Rash developed in 9 of 49 patients treated for PCP with trimetrexate, and resolved in all but 1 when concurrent sulfadiazine was ceased. The other patient had ceased all other therapy at least 8 days previously but did not develop a rash with a second course of trimetrexate given several months later (24). The development of drug hypersensitivity is not limited to the treatment of HIV-infected patients with PCP, also occurring with the use of sulfadiazine (SDZ) as therapy of central nervous system toxoplasmosis (25-27). Reaction rates of 10-40% have been observed, usually resulting in cessation of SDZ and substitution with clindamycin. The latter may also be used as initial therapy in those patients previously allergic to SMX, although clindamycin is also known to induce drug rashes. Other antibiotics reported to induce hypersensitivity include ampiciUin (28), amoxicillin-clavulanic acid (29), phenytoin (30), carbamazepine (28, 30), isoniazid and other antimycobacterial therapy (31), clindamycin (31), and flucytosine (32). In addition we have also seen reactions to ciprofloxacin, rifampicin, fusidic acid, and ganciclovir. Anti-HIV therapy has also been associated with allergic reactions, including zidovudine (33, 34). didanosine (ddI) (35), and recombinant soluble CD4 (36). As with sulfonamide hypersensitivity, all reactions were delayed in onset, with the exception of reexposures. All had clinical manifestations of rash and fever, with the exception of zidovudine. Rechallenge of three patients with zidovudine resulted in fever without rash on at least two occasions, although there was some evidence that this was
58
infectious, and not allergic, in nature (33). In another patient, however, rechallenge with zidovudine was associated with the rapid onset of fever and biopsy-proven leukocytoclastic vasculitis (34). Multiple drug allergies may develop in the one patient. A patient developed rash and fever on challenge testing to penicillin, clindamycin, trimethoprim, erythromycin, rifampicin, isoniazid, pyrazinamide, ethambutol, prothionamide, streptomycin, and clofazimine (31). Management problems arise when allergy develops to several drugs used for the treatment of one condition, such as to sulfadiazine and clindamycin for the treatment of toxoplasmosis. In such instances, there may be no effective alternative, and desensitization may be indicated.
Clinical Features Hypersensitivity reactions in HIV-infected patients show a high degree of uniformity, despite the multiplicity of causative drugs. The reaction usually begins gradually after 7-12 days of therapy with fever and a morbilliform eruption. Patients with prior exposure may, however, develop immediate, and often severe, reactions. The rash typically consists of erythematous, maculopapular, pruritic, and confluent lesions, most prominent on the body and upper limbs and rarely involving the head. The fever may be accompanied by rigors and be difficult to distinguish from those due to infections; for example, in patients with PCP, superimposed bacterial bronchopneumonia or mycobacteriosis may also cause fevers. Distinguishing allergy from infection is especially difficult if the fever occurs without an accompanying rash; in particular, mycobacteriosis, lymphoma, Kaposi's sarcoma, and HIV itself may all cause fever without other clinical manifestations. Although the frequency of isolated fever due to drug allergy has not been reported, this appears to be less frequent than combined rash and fever. Other features of hypersensitivity reported in HIV-positive patients include anicteric hepatitis, pneumonitis, and the Stevens-Johnson syndrome (37, 38). There have been no reports, however, of true anaphylaxis (i.e., IgE-mediated allergy) in patients with HIV infection. Whether the hematological cytopenias associated with drug therapy are due in part to hypersensitivity as well as to bone marrow toxicity is not known. The more recent observation that Pneumocystis carinii may infect extrapulmonary sites suggests that some of these
CARR, COOPER, AND PENNY
Table II. Clinical Features of Drug Allergy • Cardinal features Morbiltiform eruption Fever • Other features Stevens-Johnson syndrome Hepatitis Acute interstitial pneumonitis ? Hematological toxicity • Occurs after 7-12 days of therapy and may worsen despite cessation
allergic features may be related to disease rather than to treatment (Table II). The clinical features of hypersensitivity complicating PCP prophylaxis are essentially those described above, namely, progressive onset of rash and fever, although they need not both occur or commence concurrently. In our experience, rash and fever may take as long as 4 weeks to develop. In some patients treated at our institution with intermittent SMX-TMP (cotrimoxazole DS, one tablet bd twice weekly), the rash may wax and wane following each day's doses. In other patients, however, rash may occur immediately following the first dose, implying prior exposure and sensitization and, perhaps, an IgE-mediated mechanism.
Therapy Cessation of the drug responsible for an allergic reaction is usually sufficient to ensure resolution of the reaction (Table III). This is not always necessary as, for reasons that are unclear, the rash may recede in up to one-third of patients despite continued therapy (9). However, the intensity of drug hypersensitivity may increase for up to 7 days after the offending drug has been ceased and persist for several weeks. For persistent or severe reactions, corticosteroids and antihistamines have been used, although neither has been shown actually to accelerate the resolution of drug reactions (15, 23). The use of corticosteroids, in particular, must be weighed against their ability to reactivate latent infections, such as cytomegalovirus (CMV) and mycobacteria. Corticosteroids are now being used Table IlL Therapy of Drug Allergy • Cessation of drug usually adequate • Therapy of hypersensitivity with corticosteroids and/or antihistamines not of proven benefit • Desensitization of benefit to patients requiring therapy for cerebral toxoplasmosis or PCP
Journal of Clinical Immunology, Vol. 11, No. 2, 199t
ALLERGIC MANIFESTATIONSOF HIV INFECTION
therapeutically for PCP, where they are thought to reduce the inflammatory component (38); this use could reduce the future incidence of hypersensitivity reactions to SMX-TMP. Desensitization may be appropriate if there is no suitable alternative therapy to an allergenic drug. Early case reports demonstrated successful desensitization to SMX-TMP (14, 40, 41) and sulfadiazine (SDZ) (42) of patients with PCP and cerebral toxoplasmosis, respectively. This was especially of benefit for those with cerebral toxoplasmosis, as indefinite maintenance therapy is required at relatively high dose, SDZ-pyrimethamine has been regarded by many as the more effective therapy, and patients may also be allergic to the other mainstay of treatment, clindamycin. Since 1987, as has been reported (43), some 27 patients have been desensitized on 33 occasions to SDZ at our institution to enable therapy for presumed cerebral toxoplasmosis. This was especially essential for the five patients who had also developed rashes to clindamycin. This protocol has the advantages of oral administration and a lack of dependence upon corticosteroids (42) or antihistamines to suppress allergy. Of the 27 patients to undergo SDZ desensitisation, a final daily dose of 4 g was achieved in 15 patients and of 2 g in a further 6 patients. Desensitization failed in eight patients, secondary to allergy in five and patient noncompliance in three. All of these noncompliant patients were subsequently successfully desensitized. No conclusive reports of desensitization to other agents have been reported. Other situations in which desensitization may be of use are to B-lactam antibiotics, antituberculous, anticonvulsant, and even anti-HIV therapy. The mechanisms of successful desensitization are unknown; knowledge in this area could help predict those patients in whom desensitization would be successful. Mechanisms
There are minimal data regarding risk factors for drug allergy in HIV disease (Table IV). By understanding such factors, and associated mechanisms, it might be possible to identify those at high risk for developing hypersensitivity and modify their treatment accordingly. The similarity to the ampicitlin-associated rash of acute Epstein-Barr virus (EBV) (44) and acute CMV (45) infections is striking. Subclinical infec-
Journal of Clinical Immu,7ology, Vol. 11, No. 2, 1991
59
Table IV. Mechanisms and Risk Factors in Drug Allergy • Mechanisms unknown No data regarding IgE or immune complex involvement ? Molecular structure of drug and/or metabolite(s) (? via hepatic cytochrome P450) • Risk factors uncertain Degree of immunodeficiency(CD4+ celt count,
Special Article
Allergic Manifestations of Human Immunodeficiency Virus (HIV) Infection ANDREW CARR, 1 DAVID A. COOPER, 1'2 and RONALD PENNY 1
INTRODUCTION
Accepted: November I5, 1990
Drug allergy is the most common and significant allergic manifestation of HIV 3 infection. Initially described in patients treated with SMX-TMP for PCP, allergy is now known to involve a multitude of drugs. The pathogenesis of, and risk factors for, allergy in HIV infection are poorly understood, although there is evidence suggesting that allergy is more common with advancing immunodeficiency. HIV-negative subjects with sulfonamide allergy may have drug-specific antibodies and drug metabotiteinduced lymphocyte cytotoxicity, abnormalities that could partly explain the allergic mechanisms and which may have future diagnostic potential; these abnormalities have not been described in HIV-infected subjects. Therapy includes avoidance, suppressive agents such as corticosteroids, and desensitization, although the appropriate role for each is not entirely clear. Serum IgE levels have been shown to rise with progressive disease; those patients with higher levels may have a worse prognosis. The mechanisms of this rise are multifactorial, probably a combination of altered T-lymphocyte regulation of IgE synthesis and of production of specific IgE directed against microbial antigens.
Study of allergic disease in HIV 3 infection has focused on two areas, namely, drug allergy and regulation of IgE synthesis by T lymphocytes. Drug allergy was the first clinical manifestation of allergy described in HIV infection and remains the most important. HIV-infected patients treated with highdose SMX-TMP for PCP were found to develop cutaneous reactions at a greater frequency than either immunocompetent or HIV-negative immunodeficient subjects (I, 2). These reactions are presumed to have an allergic (i.e., immune-mediated) basis. The clinical features and frequency of drug allergy have been described, and drug allergy has since been identified in HIV-infected patients with the use of multiple drugs. The occurrence of drug hypersensitivity appears paradoxical, considering the profound anergy that develops in these patients to other immune stimuli; this paradox has yet to be explained. An improved understanding of both the underlying risk factors and mechanisms, neither of which have been studied to any significant degree, will allow for improved strategies to circumvent this hypersensitivity. One therapeutic strategy has been desensitization, which has allowed for improved therapy for some allergic patients. Abnormalities of IgE synthesis have been described in immunodeficiency states of T lymphocytes other than HIV infection, both congenital and acquired, such as the Wiskott-Aldrich syndrome and acute graft-versus-host disease. In HIV infection, a progressive rise in serum IgE levels has been found and attributed to alterations in the levels of T
KEY WORDS: Drug hypersensitivity; human immunodeficiency virus (HIV); IgE; desensitization; Pneumocystis carinii pneumonia.
1Centre for Immunology, St. Vincent's Hospital, Darlinghurst, Sydney, Australia 2010. ZNational Centre of HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, Australia. 3Abbreviations used: HtV, human immunodeficiency virus: SMX-TMP, sulfamethoxazole-tlimethoprim: PCP, Pneumocystis carinii pneumonia; IgE, Immunoglobutin E; EBV, EpsteinBarr virus; CMV, cytomegalovirus; IFN, interferon; IL, interleukin.
55 0271-9142/91/0300-0055506.50/0© 1991PlenumPublishingCorporation
56
lymphocyte-derived cytokines and, also, to specific microbial antigens. Specific IgE may have clinical significance; Pneumocystis carinii-specific IgE has been proposed as a possible contributing factor in the acute pneumonitis, seen early in the treatment of PCP, that can be suppressed with corticosteroids. The rise in total serum IgE has also been shown to have prognostic value in some patients and should prove of clinical value if found to be independent of other prognostic markers.
CARR, COOPER, AND PENNY
Table I. Frequency of Drug Allergy • Allergy to sulfonamides complicating treatment and prophylaxis of PCP in 30-65% of patients; prevalence with other drugs unclear • Allergy more common in HIV-infected than in immunocompetent or HIV-negative immunosuppressed patients • Allergic manifestations may resolve without ceasing therapy in 30% of cases • Up to 80% of patients may tolerate sulfonamide for secondary PCP prophylaxis despite allergy during PCP therapy (t4) • Degree of cross-reactivity between different sulfonamides varies, and may be as low as 10% • Multiple drug allergies may occur in the one patient
FREQUENCY OF DRUG ALLERGY
Sulfonamides Before the advent of HIV, SMX-TMP had been shown to be the therapy of choice for PCP, as it caused less toxicity than pentamidine and was more effective (3, 4). However, in HIV-infected patients with PCP, toxicity with the use of SMX-TMP occurred in 65% of patients, compared to just 12% of HIV-negative immunosuppressed patients (5). More specifically, the frequency of SMX-TMP allergy differs markedly among subjects who are immunocompetent, HIV-infected, or immunodeficient but HIV-negative. In the largest study of SMX-TMP allergy in the general population, 38 cutaneous reactions were found in 1121 patients (3.3%) who received SMX-TMP; 19 of these reactions occurred within 3 days, and the remainder within 2 weeks (6). This patient cohort was based upon admissions to several general medical, surgical, and psychiatric wards unlikely to contain immunosuppressed patients and, as such, corresponds best with the general community. Similarly, of 19 H/V-negative immunosuppressed adults with PCP treated with SMX-TMP, only I patient developed a drug rash (7, 8). However, the frequency of SMX-TMP allergy in HIV-positive patients has been estimated at between 29 and 65% (5, 9-11) (Table I). However, only 0-33% of those in the above studies needed to discontinue therapy as a result of hypersensitivity. This wide variation was partially dependent upon the severity both of the reaction and of the PCP itself. In those patients in whom allergy was associated with high fever, StevensJohnson syndrome, or presumed visceral involvement, therapy was usually ceased (9). In contrast, patients with milder PCP may have required a shorter course of therapy and, so, would have been
less likely to have had SMX-TMP continued once an allergic reaction had occurred. Given that 85% of patients will develop PCP at least once (12) and that, futhermore, some 60% of patients will relapse within 1 year of PCP therapy without subsequent prophylaxis (13), both primary and secondary prophylaxis against PCP is now mandatory. Despite the frequency of drug allergy with SMX-TMP, it has the advantages of oral administration, home-based therapy, minimal cost, better protection against extrapulmonary pneumocystosis than aerosolised pentamidine, and perhaps a lower overall relapse rate as well. Hypersensitivity may develop to SMX-TMP administered as primary or secondary prophylaxis against PCP. SMX-TMP hypersensitivity during secondary prophylaxis following uncomplicated therapy of PCP with SMX-TMP has not been reported. Consequently, SMX-TMP may be administered to these patients without the development of drug allergy. Conversely, not all patients will develop hypersensitivity to SMX-TMP given as secondary prophylaxis if they developed hypersensitivity to SMX-TMP during therapy of PCP. Fourteen patients who developed fever and rash during SMX-TMP treatment of PCP received subsequent oral SMX-TMP as prophylaxis; only three of the paients (21%) developed rashes (14). The frequency of SMX-TMP hypersensitivity complicating primary prophylaxis was found to be as high as 50%, with a further 30% developing pruritis of uncertain significance (I5). However, only 14% of these subjects needed to discontinue therapy, all of whom developed fever and mucosal inflammation consistent with Stevens-Johnson syndrome. The different reaction rates in various reports are no doubt a reflection of differing dosage regimes and differing criteria for defining drug Journal of Clinical Immunology, Vol. 11, No. 2, 1991
ALLERGIC MANIFESTATIONS OF HIV INFECTION
hypersensitivity, for timing drug withdrawal, and for patient selection. Therapy of PCP with dapsone caused allergy in 17 to 53% of patients (16-18), although all studies were small. Only 4% of these patients needed to cease therapy because of drug hypersensitivity. However, in the comparative studies of dapsone and SMX-TMP hypersensitivity in PCP patients, there was no difference between the two groups in either the frequency of hypersensitivity or the discontinuation due to hypersensitivity (17, 18). Dapsone has been more widely used as prophylaxis than therapy. The frequency of hypersensitivity appears to be less than with SMX-TMP. There is also limited cross-reactivity with SMX-TMP, and it would appear that the majority of patients who do react to SMX-TMP can safely tolerate dapsone. However, there are no good data available to confirm these suppositions. The chief caution in the use of dapsone would be for those patients in whom SMX-TMP had induced StevensJohnson syndrome or visceral involvement, such as hepatitis or pneumonitis. Fansidar, a combination of pyrimethamine and sulfadoxine, a long-acting sulfonamide, has been widely used for PCP prophylaxis. The long-acting sulfonamides are no longer used as therapy for common bacterial infections in the general community because of the increased risk of Stevens-Johnson syndrome. The other major use of Fansidar has been as prophylaxis against malaria. However, because of multiple deaths associated with its use, it is likewise no longer recommended by the Centers for Disease Control (19). These deaths have been a result of severe cutaneous exfoliation and sepsis. To date, one fatality has been reported in HIV-infected subjects with its use (20); whether severe reactions are more common than with SMX-TMP or dapsone has not been determined. Moreover, given also that the efficacy of Fansidar as prophylaxis appears to be inferior to SMX-TMP or pentamidine, Fansidar is probably best reserved for use as prophylaxis in those in whom they are inappropriate. Two reports of its use in predominantly SMX-TMP allergic subjects found hypersensitivity developed to Fansidar in only I0 and 12% of patients (21, 22), although the nature of their allergic reactions is not always clear from the reports.
Other Drugs Pentamidine has long been assumed to be a nonallergenic alternative therapy of PCP. However, four reports have documented the occurrence of
Journal of Clinical Immunology, Vol. 11, No. 2, 1991
57
morbilliform eruptions presumed secondary to pentamidine therapy, including one prospective study with a prevalence of 15% (9). No patient was required to stop pentamidine therapy. One patient developed transient urticaria, without other features of anaphylaxis, while receiving intravenous pentamidine (23). The urticaria was localized immediately proximal to the cannula site and occurred within 5 min of the infusion commencing. Following a single intravenous injection of diphenhydramine, all subsequent infusions were administered without incident. Whether this was type I (i.e., IgEmediated) hypersensitivity or an anaphylactoid response is unknown. Similarly, the mechanism responsible for bronchospasm that occurs in a large percentage of patients receiving aerolized pentamidine is unknown. Experience with other PCP therapy is less well documented. Rash developed in 9 of 49 patients treated for PCP with trimetrexate, and resolved in all but 1 when concurrent sulfadiazine was ceased. The other patient had ceased all other therapy at least 8 days previously but did not develop a rash with a second course of trimetrexate given several months later (24). The development of drug hypersensitivity is not limited to the treatment of HIV-infected patients with PCP, also occurring with the use of sulfadiazine (SDZ) as therapy of central nervous system toxoplasmosis (25-27). Reaction rates of 10-40% have been observed, usually resulting in cessation of SDZ and substitution with clindamycin. The latter may also be used as initial therapy in those patients previously allergic to SMX, although clindamycin is also known to induce drug rashes. Other antibiotics reported to induce hypersensitivity include ampiciUin (28), amoxicillin-clavulanic acid (29), phenytoin (30), carbamazepine (28, 30), isoniazid and other antimycobacterial therapy (31), clindamycin (31), and flucytosine (32). In addition we have also seen reactions to ciprofloxacin, rifampicin, fusidic acid, and ganciclovir. Anti-HIV therapy has also been associated with allergic reactions, including zidovudine (33, 34). didanosine (ddI) (35), and recombinant soluble CD4 (36). As with sulfonamide hypersensitivity, all reactions were delayed in onset, with the exception of reexposures. All had clinical manifestations of rash and fever, with the exception of zidovudine. Rechallenge of three patients with zidovudine resulted in fever without rash on at least two occasions, although there was some evidence that this was
58
infectious, and not allergic, in nature (33). In another patient, however, rechallenge with zidovudine was associated with the rapid onset of fever and biopsy-proven leukocytoclastic vasculitis (34). Multiple drug allergies may develop in the one patient. A patient developed rash and fever on challenge testing to penicillin, clindamycin, trimethoprim, erythromycin, rifampicin, isoniazid, pyrazinamide, ethambutol, prothionamide, streptomycin, and clofazimine (31). Management problems arise when allergy develops to several drugs used for the treatment of one condition, such as to sulfadiazine and clindamycin for the treatment of toxoplasmosis. In such instances, there may be no effective alternative, and desensitization may be indicated.
Clinical Features Hypersensitivity reactions in HIV-infected patients show a high degree of uniformity, despite the multiplicity of causative drugs. The reaction usually begins gradually after 7-12 days of therapy with fever and a morbilliform eruption. Patients with prior exposure may, however, develop immediate, and often severe, reactions. The rash typically consists of erythematous, maculopapular, pruritic, and confluent lesions, most prominent on the body and upper limbs and rarely involving the head. The fever may be accompanied by rigors and be difficult to distinguish from those due to infections; for example, in patients with PCP, superimposed bacterial bronchopneumonia or mycobacteriosis may also cause fevers. Distinguishing allergy from infection is especially difficult if the fever occurs without an accompanying rash; in particular, mycobacteriosis, lymphoma, Kaposi's sarcoma, and HIV itself may all cause fever without other clinical manifestations. Although the frequency of isolated fever due to drug allergy has not been reported, this appears to be less frequent than combined rash and fever. Other features of hypersensitivity reported in HIV-positive patients include anicteric hepatitis, pneumonitis, and the Stevens-Johnson syndrome (37, 38). There have been no reports, however, of true anaphylaxis (i.e., IgE-mediated allergy) in patients with HIV infection. Whether the hematological cytopenias associated with drug therapy are due in part to hypersensitivity as well as to bone marrow toxicity is not known. The more recent observation that Pneumocystis carinii may infect extrapulmonary sites suggests that some of these
CARR, COOPER, AND PENNY
Table II. Clinical Features of Drug Allergy • Cardinal features Morbiltiform eruption Fever • Other features Stevens-Johnson syndrome Hepatitis Acute interstitial pneumonitis ? Hematological toxicity • Occurs after 7-12 days of therapy and may worsen despite cessation
allergic features may be related to disease rather than to treatment (Table II). The clinical features of hypersensitivity complicating PCP prophylaxis are essentially those described above, namely, progressive onset of rash and fever, although they need not both occur or commence concurrently. In our experience, rash and fever may take as long as 4 weeks to develop. In some patients treated at our institution with intermittent SMX-TMP (cotrimoxazole DS, one tablet bd twice weekly), the rash may wax and wane following each day's doses. In other patients, however, rash may occur immediately following the first dose, implying prior exposure and sensitization and, perhaps, an IgE-mediated mechanism.
Therapy Cessation of the drug responsible for an allergic reaction is usually sufficient to ensure resolution of the reaction (Table III). This is not always necessary as, for reasons that are unclear, the rash may recede in up to one-third of patients despite continued therapy (9). However, the intensity of drug hypersensitivity may increase for up to 7 days after the offending drug has been ceased and persist for several weeks. For persistent or severe reactions, corticosteroids and antihistamines have been used, although neither has been shown actually to accelerate the resolution of drug reactions (15, 23). The use of corticosteroids, in particular, must be weighed against their ability to reactivate latent infections, such as cytomegalovirus (CMV) and mycobacteria. Corticosteroids are now being used Table IlL Therapy of Drug Allergy • Cessation of drug usually adequate • Therapy of hypersensitivity with corticosteroids and/or antihistamines not of proven benefit • Desensitization of benefit to patients requiring therapy for cerebral toxoplasmosis or PCP
Journal of Clinical Immunology, Vol. 11, No. 2, 199t
ALLERGIC MANIFESTATIONSOF HIV INFECTION
therapeutically for PCP, where they are thought to reduce the inflammatory component (38); this use could reduce the future incidence of hypersensitivity reactions to SMX-TMP. Desensitization may be appropriate if there is no suitable alternative therapy to an allergenic drug. Early case reports demonstrated successful desensitization to SMX-TMP (14, 40, 41) and sulfadiazine (SDZ) (42) of patients with PCP and cerebral toxoplasmosis, respectively. This was especially of benefit for those with cerebral toxoplasmosis, as indefinite maintenance therapy is required at relatively high dose, SDZ-pyrimethamine has been regarded by many as the more effective therapy, and patients may also be allergic to the other mainstay of treatment, clindamycin. Since 1987, as has been reported (43), some 27 patients have been desensitized on 33 occasions to SDZ at our institution to enable therapy for presumed cerebral toxoplasmosis. This was especially essential for the five patients who had also developed rashes to clindamycin. This protocol has the advantages of oral administration and a lack of dependence upon corticosteroids (42) or antihistamines to suppress allergy. Of the 27 patients to undergo SDZ desensitisation, a final daily dose of 4 g was achieved in 15 patients and of 2 g in a further 6 patients. Desensitization failed in eight patients, secondary to allergy in five and patient noncompliance in three. All of these noncompliant patients were subsequently successfully desensitized. No conclusive reports of desensitization to other agents have been reported. Other situations in which desensitization may be of use are to B-lactam antibiotics, antituberculous, anticonvulsant, and even anti-HIV therapy. The mechanisms of successful desensitization are unknown; knowledge in this area could help predict those patients in whom desensitization would be successful. Mechanisms
There are minimal data regarding risk factors for drug allergy in HIV disease (Table IV). By understanding such factors, and associated mechanisms, it might be possible to identify those at high risk for developing hypersensitivity and modify their treatment accordingly. The similarity to the ampicitlin-associated rash of acute Epstein-Barr virus (EBV) (44) and acute CMV (45) infections is striking. Subclinical infec-
Journal of Clinical Immu,7ology, Vol. 11, No. 2, 1991
59
Table IV. Mechanisms and Risk Factors in Drug Allergy • Mechanisms unknown No data regarding IgE or immune complex involvement ? Molecular structure of drug and/or metabolite(s) (? via hepatic cytochrome P450) • Risk factors uncertain Degree of immunodeficiency(CD4+ celt count,
