Adverse reactions to penicillin with particular reference to treatment of syphilis

40 0. IDS0E, T. GUTHE, & R. R. WILLCOX penicillin treatment in neurosyphilis could be syste- matically measured by laboratory examinations of the spinal fluid. Thus, in asymptomatic neuro- syphilis the follow-up control should comprise quantitative serology and spinal fluid examinations at 3-monthly intervals in the first year, and at 6-monthly intervals in the second year. The cell count and total serum protein in the spinal fluid should return to normal within the observation period, as a measure of the arrest of the syphilitic process. Following treatment, the spinal fluid reagin (and treponemal antibody) tests may remain reactive for several years at a low titre without indication of active syphilis (Perdrup, 1968). There is evidence that the IgM levels in the spinal fluid decrease more rapidly than the IgG levels and that therefore quantitative flocculation tests, which are more reac- tive to IgM immunoglobulins than the complement fixation test, would be of particular value in assessing the results of therapy (Oxelius et al., 1969). In cardiovascular syphilis and other late compli- cations, the serological,follow-up care after treat- ment will be similar to that in late latent syphilis. The clinical results and the functioning of the cardio- vascular system will depend on the damage already present when treatment begins (Perdrup, 1964). Follow-up control for many years will obviously be necessary. Supporting therapy, including surgery and medical care, may be required. After treatment of early syphilis in pregnancy, quantitative serological tests should be carried out monthly until delivery. Newborn babies of ade- quately treated mothers should be followed with quantitative serological tests monthly for 3 months and a final test after 6 months. Particular attention should be paid to the possibility of passive trans- placental transfer of reagins and treponemal anti- bodies, should the initial tests prove to be positive without clinical or radiological signs of the disease. In such circumstances the reagin titre should be checked at more frequent intervals. If it increases, syphilis is likely; if it decreases, passive transfer is the cause (special consideration being given to the results of the quantitative flocculation tests because of their particular affinity to the IgM immuno- globulins). Alternatively the modified FTA test may be used if available to demonstrate IgM immuno- globulins in the infant (see page 32). In late con- genital syphilis the spinal fluid should be examined before treatment. Otherwise, the follow-up prin- ciples are the same as for late latent and late syphilis in adults. ADVERSE REACTIONS TO PENICILLIN WITH PARTICULAR REFERENCE TO TREATMENT OF SYPHILIS TOXIC REACTIONS Among the antitreponemal antibiotics penicillin stands out because of its very low toxicity (Stewart, 1964). Toxic reactions occur almost exclusively when penicillin is given in excessive doses to patients with reduced renal function hampering its excretion (McGovern et al., 1970; Kurtzman et al., 1970). The toxic effect in such cases is associated with the pharmacological properties of the penicillin prepara- tion used, and almost entirely with the cations of the potassium or sodium salts (Stewart, 1965; McGovern et al., 1970). In particular, neurotoxic reactions have been observed (Smith et al., 1967; Cohill et al., 1967; Kurtzman et al., 1970), especially if brain tissue has been weakened by cerebral disease (Deisenhammer, 1969). When administered intrathecally, a procedure now rarely used, even very low doses ofpenicillin may cause neurotoxicphenomena (Kurtzman et al., 1970). Other rare complications may result from the physical nature of the penicillin preparation used, e.g., pain and other local reactions at the injection site, and Hoigne's syndrome (sensory phenomena, elevated blood pressure) (Hoigne & Schoch, 1959; Hoign6, 1962), which may end fatally (Kieswetter & Ernst, 1968). There is evidence that reactions of this type are due to emboli caused by large penicillin crystals (20-100 IU or more) in the pre- paration used (Batchelor et al., 1951), and that therefore only microcrystalline or micro-particle size penicillin preparations should be employed (WHO Scientific Group on Treponematoses Re- search, 1970). Embolic toxic reactions arising from accidental intravenous injection of procaine peni- cillin, suggesting the procaine element as the major factor (Reuter, 1970), are also rare (Popper, 1964; Freedman, 1965). PENICILLIN IN THE TREATMENT OF SYPHILIS ALLERGIC REACTIONS On the other hand, allergic reactions due to immunological responses to penicillin have become a major medical problem. Penicillin is now the most common cause of drug allergy (Weinstein, 1970). It has been estimated that in the USA some 2.2 million of the population may be allergic to penicillin (Smith et al., 1966). Surveys and investigations in several countries indicate that the frequency of total allergic reactions ranges from 0.7% to 10% in patients treated with penicillin, depending on methods of observation, criteria of recording, and nature and composition of the treated groups, including the number of injections given (Willcox, 1964b; Goodman & Gilman, 1970). There is no clear indication that severe allergic reactions, e.g., anaphylactic shock, have increased in propor- tion to the increasing consumption of penicillin in the last decennium (Ids0e et al., 1968); this is largely because more care has been taken to prevent them. The frequency of allergic reactions among patients treated with penicillin for syphilis and gonorrhoea has been found to be relatively low. Thus surveys in Norway (Oslo) showed 1 % total reactions and no anaphylaxis (Gjessing, 1960), in France 0.3% total reactions and 0.04% anaphylaxis (Thiers et al., 1964), and in the USA 0.67% total reactions and 0.015% anaphylaxis (US Department of Health, Education and Welfare, 1965b). No deaths were observed in these studies. Among venereal disease patients in several other countries (Willcox, 1962, 1964b), 1 % reactions of all types and 0.014% fatalities were noted. The allergic reactions encountered in penicillin treatment for syphilis and gonorrhoea are of the same types and follow the same immunochemical pattern as generally seen. They are outlined below. Allergic penicillin reactions may be conveniently classified according to the interval between the administration of the drug eliciting them and the onset of symptoms. Each group has its different immunological pattern and symptomatology, al- though urticaria and some other skin eruptions are frequently seen in all groups, accounting for more than half of the total allergic penicillin manifestations (Calnan, 1964). Immediate reactions usually occur from seconds to half an hour after the administration of penicillin. There may be initial weak and ill-defined malaise and skin symptoms, which may either subside or develop into anaphylactic shock, or shock may occur without prodromes. Onset, recovery or death may take place dramatically within a few minutes (Vidal et al., 1960). Accelerated reactions occur from 1-2 to 48 hours after penicillin administration (McGovern et al., 1970) and are dominated by urticaria; occasionally laryngeal oedema may occur (Westerman et al., 1966). The character of the immediate and acceler- ated reactions indicates that antibodies may already be present in the patient as a result of previous exposure to penicillin or related substances (e.g., cephalosporins) (Estrada-Parra & Salazar Mallen, 1965; Bendixen, 1966; Grieco, 1967) In late allergic reactions there may have been no previous exposure to penicillin and the reactions occur only some 3 days or more after the introduc- tion of the drug, during which time antipenicillin antibodies have been formed. They include serum- sickness-like symptoms, cutaneous manifestations (usually with urticaria), stiffness and pain in the joints, fever, haemolytic anaemia, thrombocytopenia, and other blood dyscrasias (McGovern et al., 1970). Hepatitis, myocarditis, nephritis and acute psychotic syndromes (Bjornberg & Selstam, 1960), pachy- meningitis (Willcox, 1964b) and some other mani- festations can occur on an allergic as well as a pharmacological basis (McGovern et al., 1970). Contact dermatitis results from local application of, or exposure to, penicillin (Voronina, 1965; Epstein, 1966), and thus has its own particular immunopathological mechanism. Considerable knowledge of the immunochemical mechanisms of penicillin allergy has been acquired during the last decade, particularly in regard to the antigenic determinants involved (Levine, 1960; de Weck & Eisen, 1960; Parker et al., 1962). Benzyl- penicillin has a low molecular weight. To induce immune responses it must first combine with larger molecules, usually proteins, to form hapten-protein conjugates (Landsteiner, 1945; Levine, 1966; Mc- Govern et al., 1970). Some of the antigenic determin- ants from penicillin and its derivates were illustrated by de Weck (1967). The major antigenic determinant has been shown to be the penicilloyl protein con- jugate (de Weck & Eisen, 1960; Levine & Ovary, 1961; Parker, 1964; Maciejowska, 1969). As this is a potent immunogen (i.e., inducer of antibody formation) it cannot be used for skin-testing in non- 41 0. IDS0E, T. GUTHE, & R. R. WILLCOX allergic humans. On the other hand, the conjugate of penicilloyl with a polymerized lysine forms penicilloyl-polylysine (PPL), which appears to be nonimmunogenic although antigenic and therefore can be used with fair safety for skin-testing (de Weck & Eisen, 1960; Parker et al., 1965; Pedersen- Bjergaard, 1969; Lentz & Nicholas, 1970). Minor antigenic determinants include hapten conjugates of benzylpenicillin itself and its degradation products (Levine & Price, 1964; de Weck, 1967). Protein impurities preformed during the process of fermentation of penicillin or during storage may also form immunogenic conjugates with penicillin or its derivatives. Also, immunogenic macromolecules resulting from polymerization of penicillin and its derivatives have been found in semisynthetic peni- cillins (Stewart, 1968; Pedersen-Bjergaard, 1969; Schneider, 1970; Feinberg, 1970). These findings have led to the search for " purified " penicillins of less antigenicity than the commercial penicillins now available (Knudsen et al., 1967). Antibodies and immunological testing Several antibodies are involved in penicillin allergy. Skin-sensitizing antibodies (immunoglobulin E- IgE), are mostly specific to the minor and to a lesser extent to the major antigenic determinants. They are associated with immediate penicillin reactions, including anaphylaxis, while skin-sensitizing anti- bodies specific to the major antigenic determinants may predominate in accelerated urticarial reactions (Levine & Zolov, 1969). For testing of potential immediate and accelerated reactors, therefore, both PPL and benzylpenicillin antigens (or preferably a mixture of minor antigenic determinants) should be employed. This procedure has been shown to have prospective value in patients prior to penicillin treat- ment (Levine & Zolov, 1969; Lentz & Nicholas, 1970). Positive reactions to such skin tests indicate a high probability of an immediate or accelerated allergic reaction to penicillin (Levine & Zolov, 1969). Skin-sensitizing antibodies may also mediate some late allergic penicillin reactions, particularly of the urticarial and arthralgic types, while in other late reactions IgG and IgM antibodies are involved (Fellner et al., 1970). However, skin testing would have no value in predicting late reactions, since antibody production may begin after the administra- tion of penicillin (Levine & Zolov, 1969). Haemagglutinating antibodies have been found to be IgG and IgM immunoglobulins. They occur frequently, together with skin-sensitizing antibodies, in different types of allergic penicillin reactions (de Weck & Blum, 1966; Voss et al., 1966). Identifica- tion of haemagglutinating antibodies by haemag- glutination testing appears to have no value in predicting immediate or accelerated reactions to penicillin in patients with negative skin tests, since these antibodies do not mediate such reactions (Levine & Zolov, 1969). However, as IgG antibodies may act as " blocking" antibodies, i.e., compete with IgE for antigen, IgG titres may be useful in judging the possibility of immediate reactions in patients with positive skin tests (Levine & Zolov, 1969). Other immunological tests, e.g., the direct and indirect basophil degranulation tests (Shelley, 1963; Resnik & Shelley, 1965), the measurement of hist- amine release from rabbit blood incubated with penicillin and patients' serum (Shelley & Comaish, 1965), the lymphoblast transformation test (Lazzaro & Buzzoni, 1968; Ky et al., 1970), and the passive cutaneous anaphylaxis (PCA) test (Ovary & Biozzi, 1954), have so far not found practical application. It should be noted that sensitivity can be induced and allergic reactions (including anaphylaxis) pro- voked by any amount of penicillin, by any type of penicillin preparation administered in any commonly used vehicle, and by any method of application (Ids0e et al., 1968). Sources of small amounts of penicillin are numerous (McGovern et al., 1970). Thus hidden penicillin contacts may have immuno- genic and antigenic effects, e.g., penicillin in milk and milk products (Wone, 1966), penicillin-con- taminated air (Gould, 1958), penicillin-contaminated sterilizer water (through boiling syringes used for penicillin injections) (Coleman & Siegel, 1955), penicillin-added vaccines (WHO Expert Committee on Venereal Diseases and Treponematoses, 1960), mycotic infections, and airborne moulds (Stewart et al., 1970b). There is evidence that patients with past or present allergic diathesis, particularly bronchial asthma, are more disposed to penicillin hypersensitivity than are normal individuals (Siegal et al., 1953; J. Amer. med. Ass., 1955; Ids0e et al., 1968). Penicillin reactors have been found among all age groups, although most commonly in adults between 20 and 49 years old (Feinberg & Feinberg, 1956). They are rare among children under 12 years old, although severe anaphylactic reactions have been observed in infancy (Matheson & Elegant, 1955), probably due to skin- sensitizing antibodies transplacentally transferred from a mother treated with penicillin (Niebauer & Kraft, 1969). 42 PENICILLIN IN THE TREATMENT OF SYPHILIS Prevention and treatment ofpenicillin reactions This subject has been discussed elsewhere in some detail (Ids0e et al., 1968). The main features are outlined below. The prevention of penicillin reactions should be undertaken at the public health level as well as at the patient level. At the public health level, the primary objective should be to reduce the possibility of sensitizing the population at large. This implies particularly that the use of penicillin for nonmedical purposes (e.g., in agriculture and food preservation) should be regulated or prohibited; that distribution of penicillin should be regulated; that possibilities for hidden contacts should be reduced to the lowest possible level; and that the medical profession should be advised to limit the use of penicillin to clinical and public health indications, and to avoid unwarranted " prophylactic " treatment with penicillin (and other antibiotics). The public should be informed of the danger resulting from the misuse of antibiotics. Penicillin should not be employed for external treat- ment or on mucous membranes. At the patient level, no penicillin should be given to patients with a history of previous penicillin reactions, except in serious infections where peni- cillins are the drug of choice. In such cases, carefully applied and interpreted skin tests will help in assessing the risk of serious allergy (Green, 1970). Desensitization (or hyposensitization) should not be attempted unless penicillin treatment is regarded as life-saving (Pedersen-Bjergaard, 1969; Fellner et al., 1970; Green, 1970), In patients with allergic dia- thesis (asthma bronchiale), indications for penicillin treatment should be severely restricted. An emergency kit for treatment of allergic reac- tions should be readily available whenever penicillin is administered. The treatment of anaphylactic reactions should be initiated with immediate injec- tions of 0.1 % adrenaline hydrochloride (or epi- nephrine) followed by appropriate shock therapy. Otherwise corticosteroids, antihistamines and peni- cillinase may hasten the recovery, although these remedies are chiefly recommended in accelerated urticarial and late reactions. MICROBIOGENIC EFFECTS OF PENICILLIN Microbial resistance So far, there is no experimental evidence that T. pallidum is less susceptible to penicillin today than when it was first used in syphilis therapy 3 decades ago. However, clear scientific proof on this point is lacking because it is not possible to cultivate the organism in vitro. The finding, despite penicillin treatment for syphilis, of treponemal structures in lymph nodes, spinal fluid, and the aqueous humour has been discussed in some detail on pages 30-31. Symptoms related to the oral cavity, respiratory and intestinal tracts, genitalia and skin due to over- growth of resistant coliforms and monilia (Guthe et al., 1958) are more common after oral treatment with tetracyclines than after penicillin, which in syphilis therapy is most frequently given parenterally. Furthermore, as benzylpenicillin G is ineffective against Escherichia coli, overgrowth of monilia is not common with this preparation. But with the wider spectrum of ampicillin, which is active against E. coli, intestinal complications need to be considered (Daikos, 1964). The Jarisch-Herxheimer reaction This reaction is believed to be caused by endo- toxins released from killed treponemes following initial administration of the treponemicidal drug, resulting in both systemic disturbances and local reactions of the syphilitic lesions (Stokes et al., 1944; Willcox, 1964a). Some investigators have indicated that the Jarisch-Herxheimer reaction might be an allergic phenomenon induced by released substances (Moore et al., 1948; Sheldon & Heyman, 1949). However, evidence of causative hypersensitivity has not been substantiated (Skog & Gudjonsson, 1966). In primary seronegative and seropositive syphilis the Jarisch-Herxheimer reaction occurs in more than half the cases treated either with penicillin or with arseno-heavy metals (Hochleitner, 1965). In primary seropositive cases the incidence can rise to 95% (Putkonen et al., 1966). Treatment in early syphilis can continue without other complications occurring (Sablan & Best, 1964). There is no need for preven- tive pretreatment, either with small doses of peni- cillin or with bismuth or mercury (Willcox, 1964a). During the 3 decades of experience with penicillin therapy in syphilis, only a few cases has a probable causative association been reported between peni- cillin treatment of late syphilis and a Jarisch- Herxheimer reaction (Willcox, 1964a). Furthermore, there is no important difference in the course of the reaction whatever antitreponemal agent may have been used (Knudsen & Aastrup, 1965). Experimental investigations also indicate that the reaction follows an " all or nothing " law (Willcox, 1961); it will occur whether treatment is initiated with bismuth or 43 44 0. IDS0E, T. GUTHE, & R. R. WILLCOX mercury, or with small penicillin doses (Heyman et al., 1952). The majority of clinicians therefore also initiate treatment of late syphilis with normal peni- cillin doses. Only in cases in which the potential risk of increased local damage is more than normally serious (primary optic atrophy or nerve deafness) is initial treatment with very small doses of penicillin, or with steroid cover, believed justified (Huriez & Agache, 1958; Huriez & Vanoverschelde, 1965). Corticosteroids have been reported to exert some effect on the Jarisch-Herxheimer reaction in general (Gudjonsson & Skog, 1968; Arfouilloux, 1969; Viegas et al., 1969). However, while corticosteroids may reduce fever, they have been considered not capable of preventing the Jarisch-Herxheimer reac- tion (Bien & Suchanek, 1969) and the benefits they provide may be outweighed by the risks involved (Viegas et al., 1969). Few instances of therapeutic paradox-i.e., clinical progression in spite of " biological " cure due to rapid replacement of organic tissue by fibrotic scar tissue-have been reported following penicillin treatment of late syphilis (Reynolds, 1948; Mohr & Hahn, 1952), and in some of them the worsened condition may have been due to inadequate treat- ment (Mohr & Hahn, 1952). THERAPY OF SYPHILIS WITH ANTIBIOTICS OTHER THAN PENICILLIN GENERAL Penicillin should not be used for treatment of syphilis in persons known to be actual or potential penicillin reactors. Cross-sensitivity occurs in per- sons sensitized to this antibiotic regardless of the type of penicillin, since the basic immunological mechanism in penicillin allergy relates to the anti- genicity of protein-hapten combinations with derivatives of the penicillin G nucleus, 6-aminopeni- cillanic acid. Although therapy of early syphilis with a-aminobenzylpenicillin (ampicillin) has shown clini- cal and serological results almost comparable to those obtained with benzylpenicillin (Cannata, 1965; Argenziano & Montagnani, 1967), this drug offers no alternative for the treatment of benzylpenicillin- sensitized patients. Also it appears that skin rashes occur more frequently after treatment with ampi- cillin than with other penicillins (Kronig & Dennig, 1970). This is possibly due to an ampicillin-specific allergenic effect of the side-chain component of ampicillin (Lancet, 1969). Clemizole penicillin (Muckter et al., 1954, 1957) requires daily injections of 800 000-1 000 000 IU for the maintenance of effective penicillin concentrations in serum and tissue (Walter & Heilmeyer, 1965), but has insignificant effect in preventing immediate penicillin reactions in sensitized persons (Rosenthal, 1958; Sciple et al., 1959; Calnan, 1964; Green, 1970). Experimental and clinical evidence shows cross- sensitivity between conventional penicillins and the cephalosporins (Brandriss et al., 1964; Batchelor et al., 1966; Mashimo et al., 1967). This is mainly due to the fact that the ,B-lactam ring, through which invisible binding to the proteins takes place, is a constituent of the penicillin nucleus as well as of the cephalosporin nucleus, 7-aminocephalosporanic acid (see Fig. 1 and 2) (Feinberg, 1970). There are reports of anaphylactic reactions to the initial injection of cephaloridine in patients previously allergic to peni- cillin (Thobum et al., 1966; Rothschild & Doty, 1966; Pedersen-Bjergaard, 1967). Also, skin-sensitiz- ing antibodies to cephalothin, cephaloridine, and 7-aminocephalosporanic acid have been demonstrat- ed in a patient 5 months after anaphylactic reaction to orally administered penicillin (Grieco, 1967). Furthermore, sensitivity to cephaloridine has been acquired by handling cephaloridine preparations, with anaphylactic shock after one injection of the drug (Kaplan & Weinstein, 1967). Apparently, 7-aminocephalosporanic acid derivatives, e.g., ce- phaloridine, cephalosporin C, cephalothin, cephalex- in, and cephaloglycin, are not completely safe substitutes for penicillin in allergic patients (J. Amer. med. Ass., 1967). Above all, they should not be used in patients who have previously had reactions of an anaphylactic nature to penicillin. It is probable that similar immunologic mechanisms are also associated with the use of synnematin B, the structure of which closely resembles that of penicillin. Metals are not attractive substitutes for penicillin in view of their toxicity, the chronic suppressive nature of their action, the inconvenient mode of administration, and the need for application over a long period of time. Effective antibiotics other than penicillin must be considered a more desirable alter- native. Many antibiotics have been shown to have anti-