6 0. IDS0E, T. GUTHE, & R. R. WILLCOX

THE NATURE OF , ITS MODE OF ACTION, AND PRINCIPLES OF PENICILLIN THERAPY

THE NATURE OF PENICILLIN CO2H Penicillin was first produced on a large scale by O H CH2R2 fermenting various strains of Penicillium notatum, BX AJ from which several compounds and derivatives were R'-CO-NH extracted, i.e. penicillin F, G, X, and K (Bacharach & Hems, 1946). In the effort to develop pure Fig. 2. Nucleus of the (7-aminoce- phalosporanic acid). A: dihydrothiazine ring; B: preparations, commercial penicillin came to contain lactam ring; RI, R2: side chains. more N-heptylpenicillin (K fraction) and less benzyl- penicillin (G fraction). Consequently, the results obtained in treating syphilis became temporarily less by adding side chain precursor acids to the fermenta- good (Moore, 1946) since the other fractions proved tion culture. However, only derivatives of acetic less antitreponemal than G (Car- acids were incorporated by the mould. Subsequently penter et al., 1946; Arnold et al., 1947; Eagle & several acid stable (e.g., penicillin V) Fleischman, 1948). In syphilis therapy, therefore, emerged in 1952 as the first oral penicillins to have benzylpenicillin G became the preparation of choice. therapeutic effect that could also be produced easily It was soon to become commercially available in a (Holt & Stewart, 1964). Next it was found that nearly pure form (Willcox, 1947). In 1948, the penicillin nucleus material accumulated in fermenting penicillin molecule was shown to consist of a culture media without the presence of a side chain nucleus, 6-aminopenicillanic acid (Brotzu, 1948), and precursor (Sheehan & Logan, 1957). It therefore a side chain linked to the amino group of the became possible to produce a great number of new nucleus. The effect was found to depend penicillins by adding any acid to nucleus material on the nature of the side chain acid (Chain, 1964). (extracted from the fermentation medium) as a side The addition of phenylacetic acid resulted in benzyl- chain substance, and not only derivatives of acetic penicillin G. The penicillin nucleus contains a fused acid (Chain, 1964). This biosynthetic method was thiazolidine ,B-lactam ring, on the integrity of which costly and toilsome. In 1957 penicillin V was depends the acid stability ofpenicillin. The penicillin synthesized by closing the ,-lactam ring of 6-amino- nucleus is capable of forming salts. Among these the penicillanic acid by means of appropriate reagents procaine penicillin salt of penicillin G (Buckwalter (Sheehan & Logan, 1957). However, this method & Dickison, 1948) and its N,N'-dibenzylethylene- was more expensive and complicated than the bio- diamine salt (DBED or benzathine penicillin G) logical method of penicillin V production. In 1958 (Szabo et al., 1951) have particular therapeutic microbiological enzymes of amidase nature were value-e.g., in the treponematoses (syphilis, yaws, shown to be capable of hydrolysing the side chain and pinta). The characteristics of the penicillin acid of commercial penicillin, releasing the nucleus molecule are shown in Fig. 1, while Fig. 2 illustrates (6-aminopenicillanic acid). Amidase fromEscherichia the structure of the related cephalosporins. coli proved notably efficient in hydrolysing the In the early years of penicillin production, at- phenylacetic acid side chain of benzylpenicillin; tempts were made to manufacture different penicillins inexpensive commercial production of the penicillin nucleus became a reality and a large number of new penicillins were manufactured semi-synthetically. ON CO2H In several fields of medicine, the need for alterna- R-CO-NH tive drugs to benzylpenicillin G had gradually \S/ become apparent. First, allergic reactions were provoked because of sensitization from its use and Fig. 1. Nucleus of the penicillins (6-aminopenicillanic acid). A: thiazolidine ring; B: ,B-lactam ring; R: side abuse. Secondly, microbial resistance developed chain. from the wider dissemination of organisms producing PENICILLIN IN THE TREATMENT OF SYPHILIS 7

penicillinase. This enzyme (also called penicillin constituents (Chain, 1964). Evidence indicates that ,B-lactamase) opens the ,B-lactam ring of the penicillin this action takes place only during phases of growth nucleus (Abraham & Chain, 1940; Stewart, 1966), of bacterial cells. In this context two related aspects which by deacylation removes the side chain of the are of practical importance. First, some inhibited penicillin molecule (Stewart, 1966). In both instances growing microorganisms will continue to produce the effect of penicillin on the particular micro- transpeptidase; thus they will be able to synthesize organism is dissipated. Thirdly, the occurrence was glycopeptide within 20 minutes or so and may noted of superinfection by penicillin-resistant organ- recover unless penicillin continues to be present. isms-e.g., staphylococci, coliform bacteria, Gram- This suggests that penicillin-free intervals or inade- negative bacilli, Candida-already present in the host quate penicillin serum concentrations during treat- or originating from another person (cross-infection ment of penicillin-sensitive infections should be or " hospital " infections). avoided, and that dose/time relationships are Acid-stable penicillin preparations were needed fundamental in this type of therapy. Secondly, for oral administration. Benzylpenicillin is suscep- during penicillin treatment partial or complete tible to hydrolysis by gastric acid and the oral dose damage of the may result in the transforma- must be at least 5 times the parenteral dose in order tion of sensitive microorganisms into insensitive to produce effective serum concentrations (American spheroblasts or atypical forms (Park, 1964). The Medical Association, 1966). Among the new spheroblasts may be lysed or survive depending on penicillins that became available and have found environmental circumstances. They may also revert clinical use (apart from the acid-stable penicillin V) to sensitive forms when no longer under the influence are: (penicillinase-resistant), of penicillin. There is no evidence that this occurs in (penicillinase-resistant and acid-stable), Treponema pallidum, although the findings of Collart (penicillinase-resistant and acid-stable), and co-workers (Collart et al., 1962, 1967; Collart & (penicillinase-resistant and acid-stable), and ampicil- Durel, 1964; Collart, 1965) and others (Yobs et al., lin (acid-stable). , owing to its effect on 1968) of apparently persistent " dormant " trepo- Gram-negative bacilli, may be considered as a nemal forms, particularly in human lymph nodes in broad-spectrum antibiotic, and can be applied orally previously treated late syphilis, may be connected and parenterally (Bull. Johns Hopk. Hosp., 1965). with this aspect. In regard to sudden lysis of an Table 1 shows the chemical structure, names, and invading organism by penicillin, a subsequent release characteristics of various penicillins. ofendotoxins from T. pallidum is commonly believed to be the cause of the Jarisch-Herxheimer reaction. serum albumin THE OF ACTON OF PENICILLIN It was shown as early as 1947 that MODE interferes with the antibacterial action of penicillin The mode of action of penicillin has been particu- (Tompsett et al., 1947). A clearer understanding of larly thoroughly studied in recent years (Tipper & the nature and significance of the protein-binding of Strominger, 1965, 1968). The antibiotic appears to penicillin emerged later. There is now evidence that become irreversibly bound to microorganisms and the inhibition of glycopeptides in the microbial cell interferes with the synthesis of the glycopeptides that wall is caused by free penicillin molecules only. The are essential constituents of the microbial cell wall, actual effect of penicillin on microorganisms in vitro thus leading to the death of the microorganism. and in vivo therefore depends on the availability of There is some evidence that there is a strong affinity free penicillin. Since part of the penicillin is reversi- between the antibiotic and peptidases produced by bly bound to protein (albumin) in serum and tissue, growing microorganisms, particularly transpeptidase. the total concentration of penicillin is the sum of The lack of transpeptidase may cause inhibition of the free and the protein-bound penicillin (Knudsen, the cell wall synthesis (Lancet, 1967). The micro- 1964; Quinn, 1964). organisms are converted into osmotically fragile There are variations in the degree ofserum binding entities that will die, presumably as a result of between different penicillins in the same medium. osmotic lysis (Park, 1964; Montgomerie et al., 1965). For example, penicillin K binds extensively to serum This shows, on the one hand, the specific action of albumin; hence it is less effective in the treatment of penicillin on sensitive microorganisms; and explains, syphilis than penicillin G at a similar dosage (Merrell on the other hand, its extremely low toxicity for & Rider, 1949). Also, individual variations occur animal tissue cells of which glycopeptides are not in the binding capacity of the same penicillin. Thus 8 0. IDS0E, T. GUTHE, & R. R. WILLCOX

Table 1. Chemical structures, names and characteristics of various penicillins

Side chain a Nonproprietary name Characteristics

OCH2- penicillin G

b acid-stable; used orally

°C2H4- pheneticillin b acid-stable; used orally

methicillin penicillinase-resistant c; used parenterally

OCH3

oxacillin b relatively acid-stable; used orally and parenterally; penicillinase-

N c resistant

\O CH, I I cloxacillin relatively acid-stable; used orally; penicillinase-resistant O/ CH,

H-CC- 11 I similar to cloxacillin ai \0/ \",M

nafcillin 12 acid-stable; used orally; penicillinase-resistant

ampicillin b " broad spectrum " antibiotic; used orally and parenterally NH,

b HN N- betacillin similar to ampicillin; not generally available H CH,

b under clinical investigation; spectrum includes and Proteus; not generally available COOH

a The side chains correspond to R in Fig. 1. b International nonproprietary name. c Penicillinase-resistant" means resistant to cleavage by penicillinase. PENICILLIN IN THE TREATMENT OF SYPHILIS 9 in benzylpenicillin G individual variations in binding and tissues. This has been illustrated by Rolinson of up to 30% have been found (American Medical (1964). Such exposure will be prolonged if the Association, 1966). It has further been shown that feeding of penicillin into this system is continued there is a correlation between the loss of activity over a period of time from intramuscular depots, in vitro of a penicillin in the presence of serum and notably when using repository long-acting penicillin the degree of protein-binding of that penicillin in the preparations. An interesting approach to assessing particular serum. This would indicate that, also serum penicillin, i.e., the active free penicillin frac- in vitro, the penicillin concerned will lose some tion, has been made by direct determination of the activity. In the case of benzylpenicillin G this could rate of treponeme immobilization in normal human represent a loss of 3-30%. Estimates of the inhibi- sera nonreactive in the T. pallidum immobilization tory concentration of a penicillin in vivo for a given (TPI) test after intramuscular injection of a fixed microorganism take this loss into consideration, dose of PAM (Durel & Hardy, 1957; Prieto & since it is only the bacteriocidal, i.e. free, penicillin Jimon, 1965). that is measured. In syphilis, this factor is con- sidered to have been taken adequately into account with a theoretical therapeutic penicillinaemia of THE PRINCIPLES OF PENICILLIN THERAPY IN 0.03 international units (IU) per ml of serum. This SYPHILIS is more than 10 times the inhibitory concentration of bznzylpenicillin G found experimentally (Magnuson, The principles of penicillin treatment of early 1952; Turner & Hollander, 1957). It is believed that syphilis were established by experimental and clinical this safety margin will also cover other possible studies (Magnuson et al., 1948; Cumberland & individual variations in serum concentrations, e.g., Turner, 1949; Eagle et al., 1950). In view of their the rate and degree of metabolism, excretion, and importance for the scientific rationale in the practical tissue diffusion (Bond et al., 1965). medical management of syphilis, they are briefly Only the free penicillin fraction is capable of outlined below. diffusion into the tissues from the serum. Once in the tissues a portion of the free penicillin will bind Dose/time relationship to the available protein (albumin). The protein con- centration will vary in different types of tissue; the T. pallidum is one of the most penicillin-sensitive proportions of bound and free penicillin in a given microorganisms known. However, for penicillin to tissue may therefore be different from those in be effective in the therapy of syphilis, two require- other tissues as well as from those in the serum ments are essential: a minimal concentration of (Rolinson, 1964). The equilibrium between free 0.03 IU of penicillin per ml of serum-which gives penicillin, protein (albumin), and protein-bound several times the serum and tissue levels needed to penicillin, however, is characteristic for each medium. kill T. pallidum-should be maintained for at least This equilibrium may be disturbed either by varia- 7-10 days in early syphilis; and penicillin-free or tions of the concentration of free penicillin or by subtreponemicidal intervals during treatment should variations in the amount of protein. If, for example, not exceed 24-30 h in order that treponemes still free penicillin is removed for antimicrobial activity, surviving can be prevented from remultiplying. The it is replaced either through a breakdown of the multiplication time of T. pallidum has been found protein-bound complex or by an influx of free to be 30-33 h (Magnuson et al., 1948; Cumberland penicillin from the serum to restore the equilibrium. & Turner, 1949), at least during the first days of the Experimental investigations show that binding of experimental infection (Collart, 1970). A continuous penicillin to protein as well as breakdown of the penicillin blood-level just above 0.03 IU per ml of protein-bound penicillin occur very rapidly (Rolin- serum is therapeutically more effective and economic son, 1964). In the latter case the antibacterial than high peak serum and tissue concentrations, activity ofthe penicillin is reestablished (Bond, 1964). interrupted by subtreponemicidal levels. It is note- The protein-bound penicillin in serum and tissues worthy that following single injections of the same thus represents a floating penicillin depot in a dose (in international units) of different penicillin sensitive but balanced blood/tissue system, function- preparations, widely varying durations of penicillin- ing by liberating penicillin and exposing micro- aemia may result, ranging from a few hours (aqueous organisms to its antibiotic action in the body fluids benzylpenicillin G) to 2-3 weeks (benzathine 10 0. IDS0E, T. GUTHE, & R. R. WILLCOX penicillin G) (Ids0e et al., 1954). The potential expected duration of the penicillinaemia obtainable effectiveness of penicillin doses (or therapy schedules) with a given dose. Although variations in resorption cannot, therefore, be compared without further and excretion in individual patients occur (Ids0e et qualification. al., 1954), the penicillin blood-level duration depends Any preparation of penicillin G will be effective in to a great extent on the characteristics of the peni- the treatment of syphilis provided the dose and cillin salt and the preparations themselves, including intervals of the injections are adapted to the peni- the size of the penicillin particles (too large particles cillin blood-level duration obtainable with the pre- give an irregular repository effect, and too small paration used to secure "active" treponemicidal particles cause inconvenient viscosity), the type of levels in the serum and tissues. Long-acting peni- aluminium stearate used in PAM, and technical cillin preparations, notably PAM and benzathine aspects of production (Pellerat et al., 1961). Pro- penicillin G in aqueous suspension (DBED) or in oil visional minimum requirements to ensure uniform (BOM), are preferable from a practical point of PAM preparations were set up by WHO in 1952 view. With these preparations, a single or a few (WHO Expert Committee on Venereal Infections injections in one or more sessions suffice to obtain and Treponematoses, 1953), supplemented by further a continuous and long-lasting treponemicidal peni- criteria in 1957 and 1964 (WHO Expert Committee cillinaemia acting on T. pallidum in the infected on Biological Standardization, 1958, 1964). They patient (Ids0e et al., 1954). Short-acting aqueous include: definition of physical and other criteria; penicillin G or similar preparations administered in establishment of an international reference prepara- multiple injections are obviously also effective but tion of PAM; a blood duration test ofpenicillinaemia show rapid resorption and excretion, resulting in in rabbits; and an assay method for penicillin with intermittent peaks of high concentration. Much defined limits ofconfidence (WHO Expert Committee higher total dosages and many more injections are on Venereal Infections and Treponematoses, 1953; then necessary than with long-acting penicillin prep- Lightbown & Sulitzeanu, 1957; WHO Expert arations. Moreover, the management of the patient Committee on Biological Standardization, 1964). generally becomes more complex. PAM is not an ideal repository preparation. By The basic dose/time relationship requirements are progressive doubling of the dose, the initial peni- not adequately identified in many published reports cillin concentration in the serum will-up to a on the results of penicillin treatment of syphilis and point-be approximately doubled, as the follow- the presumed unsatisfactory outcome of therapy may ing tabulation (data of Ids0e et al., 1954) shows: well be associated with the failure to determine the adequacy of the total dose in relation to the time Approximate Duration (days) Single Injection of PAM penicillin concen- of effective factor (Willcox, 1964a; Kern & Elste, 1966). tration In serum penicillinaemia ml IU after 24 hours (0.03 IU/ml In view of the importance of the dose/time (IU/ml) serum) relationship in the treatment of syphilis, the rationale 1 300 000 0.1 3-4 for the use of long-acting repository penicillin pre- 2 600 000 0.2 5-6 parations is evident. The introduction of such pre- 4 1 200 000 0.5 6-8 parations was therefore a significant advance from 8 2 400 000 1.0 7-10 both a clinical and a public health viewpoint. Their use dominates modern syphilotherapy (Danbolt, However, the duration of the penicillinaemia will 1960; Degos & Ebrard, 1960; Perdrup, 1960; be progressively shorter in relation to the dose, in Wiedman, 1964; Willcox, 1964a). Crystalline peni- keeping with the law of diminishing returns (Rey- cillin G in arachis oil and beeswax (POB), introduced nolds, 1952). In practice a treatment schedule of in 1945 (Romansky & Rittman, 1945), was soon 600 000 IU or 1.2 million LU of PAM each day or followed by procaine penicillin G in oil with 2% every second day until the total dosage has been aluminium monostearate (PAM) (Buckwalter & reached appears to be commonly used in syphilis Dickison, 1948). Because of the variations noted in (Willcox, 1964a; US Department of Health, Educa- different preparations of PAM (Guthe et al., 1953), tion, and Welfare, 1968). Under certain circum- their minimum repository characteristics need to be stances, for example when patients cannot be relied defined so that a slow resorption from the muscle on to return for further treatment or when a rapid depot into blood and tissues occurs as constantly as possible and so that reliance can be placed on the 1 Sarcina lutea method. PENICILLIN IN THE TREATMENT OF SYPHILIS 11 spread of infection is likely, it is rational to give of ampoules or bottles may result in heterogeneous initially a large " epidemiological " or " insurance " suspensions and cause remarkable deviations from dose (2.4 million IU) supplemented by further injec- the intended dosage (Luger, 1963). There are also tions until desired blood duration limits are achieved individual variations of blood and tissue penicillin (ibid., 1968). As a whole, the selection of a treatment levels and their duration arising from abnormal schedule should depend on the practical circum- penicillin metabolization in some persons. Repeated stances of the patient as well as on medical con- injections of the same dose and preparation may siderations, provided the dose/time principle is show wide variations of serum level (Willcox, 1970). known and followed. The rate of renal excretion of penicillin also varies A significant simplification of therapy was (Weinstein et al., 1970). In persons where the blood achieved by the introduction of benzathine penicillin and tissue levels fall too rapidly below the treponemi- G (DBED) (Szabo et al., 1951). Its use in syphilis cidal limit, a relapse ofthe infection must beexpected. has been reported by several investigators (O'Brien " Unsatisfactory " serum penicillin levels, below & Smith, 1952; Guthe & Danbolt, 1954; Shafer & measurable values, were thus found (Colli et al., Smith, 1954; Guthe, 1955; Batunin et al., 1959; 1964) after injection of 1.2 million IU of long-acting Delson, 1961; Willcox, 1962). One injection of penicillin in 0.47 % of the persons examined after one 300 000 IU of DBED gives a serum concentration week, 4.95% after 2 weeks, 15.75% after 3 weeks somewhat above 0.03 IU/ml lasting for approxi- and 46.25 % after 4 weeks. mately 7 days, while the same dose of PAM gives an Repeated injections of PAM over long periods, effective penicillinaemia for about 3-4 days. Follow- e.g., for prophylactic purposes in prostitutes, may ing a single injection of, for example, 2.4 million IU cause inconvenient oleostereomas. DBED may ofDBED, the treponemicidal penicillinaemia persists sometimes cause irritation in the subcutaneous and for up to 3-4 weeks (Ids0e et al., 1954; Putnam & muscular tissues at the site of application. This Roberts, 1954). preparation is less thermostable than PAM (Moore Since 1954, reports have appeared from several & Woody, 1960; Maha, 1961). By incorporating countries on the results of syphilis treatment with benzathine penicillin G in arachis oil with aluminium aqueous benzathine penicillin. In early syphilis the monostearate, a preparation (BOM) of satisfactory results are at least equal to those achieved with PAM thermostability that causes insignificant irritation (Shafer & Smith, 1954; Brown, 1956; Brown et al., has emerged. This type of preparation has been 1958; Brukson, 1960; de Graciansky & Grupper, shown to have a therapeutic effect in syphilis equal 1960; Bowszyc, 1961; Durel, 1963; Nicholas, 1964). to that of DBED (Brown et al., 1958). In late latent syphilis and in late complications of syphilis DBED has also given clinical effects equal to Duration ofinfection those of PAM (Fiumara, 1964). Following intra- muscular injection, it is found in the brain of rabbits Eagle (1949) showed experimentally that the in treponemicidal concentrations (Ovcinnikov et al., treponemicidal effect of a given amount of penicillin 1957) and its effectiveness in symptomatic and is directly related to the number of treponemes asymptomatic neurosyphilis in humans has been present. The earlier the course of syphilis treatment demonstrated (Delson, 1961; Kahn, 1961; US is started, the less important is the dose/time rela- Department of Health, Education, and Welfare, tionship, since in incipient disease relatively few 1965a). The prolonged treponemicidal penicillin- treponemes are present. Higher dosages are needed aemia secured by benzathine penicillin may therefore -within limits-as the duration of the infection be important in the late stages of syphilis. The increases. However, from a practical viewpoint, to preparation is also useful in the treatment of pre- ensure sufficient safety margins, all stages of early natal and congenital syphilis because of its ease of syphilis are usually treated as required for secondary application in a few injections only (Jackson et al., syphilis. This is the basis for the recommendations 1962; Harris & Cave, 1965). of the WHO Expert Committee on Venereal Infec- When using long-acting penicillin, the rate of tions and Treponematoses (1953) for a minimum absorption and thereby the repository effect show total dosage of 4.8 million IU of PAM and at least individual variations and depend inter alia on the half this dosage of benzathine penicillin. preparation being correctly injected and deposited In late latent and late symptomatic syphilis-that in the deep layer of the muscles. Inadequate shaking is, syphilis of more than 4 years' duration-there 12 0. IDS0E, T. GUTHE, & R. R. WILLCOX are other objectives of treatment with penicillin than infrequently after 2 years (Stokes et al., 1944; to obtain an immediate treponemicidal effect. Thomas, 1949). The remaining living treponemes Dose/time requirements need to be considered in the would usually have multiplied within 2 years and light ofindividual factors concerning the pathological induced a subsequent " cycle" of the disease. It is process of the diseased organ. The age, general therefore important to observe a patient with early health, and clinical state of the patient affect the disease clinically and serologically for 2 years after prognosis after treatment. treatment, allowing for a needed safety margin. It has been postulated that most patients requiring Immunological aspects re-treatment after adequate penicillin therapy of early syphilis have been reinfected (Jefferiss & The effect of penicillin therapy in syphilis should Willcox, 1963). Relapse and reinfection are some- be appraised in relation to the natural course of times indistinguishable. If syphilis is highly prevalent untreated syphilis. A great part ofpresent knowledge in a community, reinfections are more likely to on the natural course of the infection comes from occur than when the disease is less common. The the Boeck-Bruusgaard-Gjestland patient material in epidemiological situation in an area must therefore Norway where some 1 000 untreated patients were be considered when evaluating the frequency of observed 30-50 years after the infection (Bruusgaard, reinfection. 1929; Danbolt et al., 1954; Gjestland, 1955; Guthe, In late latent and late symptomatic syphilis, the 1960). The findings can be summarized as follows: immunological pattern has been firmly established (a) about 25 % of the patients developed secondary and only to a limited extent can it be influenced by relapses, the majority during the first year; (b) about treatment. At this stage the main aim of therapy is 13% developed benign late syphilis; (c) about 10% clinical arrest of the infection, the serological developed cardiovascular syphilis; (d) about 6% response being of secondary importance, as discussed developed neurosyphilis; and (e) syphilis was the elsewhere in this review. primary cause of death in some 10%. From these findings we may assume that about 60-70% of untreated syphilis patients go through life without, Modification oflesions or with minor, physical and mental consequences of Many individual as well as environmental factors the infection, indicating that important protective influence treponeme/host relationships and differen- immunity processes are involved in the host/trepo- ces occur in the response to infection with T. palli- neme relationship (Chesney, 1927; Cannefax, 1965). dum. There are several influencing factors, e.g., It is not known, however, to what extent humoral the virulence of the strain of T. pallidum, the site of and/or cell-mediated immune mechanisms are in- inoculation and size of inoculum, the mode of volved, nor has the role of reagin or autoinmmune infection, the condition of skin and mucous mem- phenomena in the pathogenesis of syphilis been branes, and the nutritional, genetic and, notably, the clarified (WHO Scientific Group on Treponematoses immunological status of the patient. These aspects Research, 1970). may have a bearing on the incubation and morpho- Intensive penicillin treatment given shortly after logy of early lesions, as pointed out inter alia by infection interferes abruptly with the immunological Chesney (1926) and Almkvist (cited by Stokes et al., mechanism of the host, and the influence of treat- 1944). Considerable variation in the symptomato- ment depends on the immunological state when logy of syphilis during the natural course of the treatment begins. The earlier in the infection this untreated infection was described in the classical occurs, the more likely it is that the host will be able literature, e.g., by Boeck (cited by Bruusgaard, 1929). to regain the original preinfection immunological In the metal therapy era variations in symptomato- state by elimination of the antibodies formed logy were also emphasized (Stokes et al., 1944). (Thomas, 1949). It follows (i) that when adequate There is thus little reason why such variations should treatment has been given early in the infection, not also be observed after penicillin came into use. reinfection of the patient may occur on re-exposure Concerning the suggested modified symptoms of to contagion; (ii) that when treatment has been early syphilis since the introduction of penicillin inadequate relapse may take place. Relapses in early (Durel & Hardy, 1957; Midana, 1960; Perdrup, syphilis usually occur within the first 3-9 months 1960; Malhurst, 1963; Manson & Price, 1963; Degos, after treatment, rarely after one year and most 1963, 1966; Longhin et al., 1966; Olansky, 1966), PENICILLIN IN THE TREATMENT OF SYPHILIS 13 atypical chancres, syphilis d'emble, abnormal clini- members of the WHO Expert Advisory Panel on cal forms, polymorphism, milder signs and symp- Venereal Infections and Treponematoses in 1965. toms, and prolonged incubations have all been Fifty-five per cent of the panel experts did not find described as being more frequent than before. These that the initial manifestations had changed, although aspects were studied in a survey undertaken among occasional atypical manifestations occurred.

EARLY SYPHILIS

TREATMENT RESULTS with 2.4-24.0 million IU of penicillin and followed for 2-10 years, clinical (and serological) normality During the first decade following the introduc- was the outcome. In one group of 211 patients tion of penicillin, its efficacy in early syphilis was ap- (Jefferiss & Willcox, 1963) retreatment was required praised in relatively short-term observation studies in in 3.2% of the 186 cases followed, this being attri- a number of countries (Willcox et al., 1954), notably buted in most cases to reinfection. Among 1 030 the USA (Moore, 1948; Bauer & Price, 1949; and patients treated for seroreactive primary syphilis, others), Austria (Gumpesberger, 1953), France 93-100% obtained serological nonreactivity, as (Jame et al., 1952; Gougerot, 1953), Switzerland determined by lipoidal antigen tests. Some reinfec- (Mayer & Herkenhoff, 1951; Miescher & Brenn, tions were described in these patients also (Jefferiss 1952), Yugoslavia (Grin, 1953), Norway (Danbolt, & Willcox, 1963; Tramier & Oddoze, 1965). 1954), India (Rangiah, 1950), and Turkey (Ceriboglu, 1950). In these early studies and-surprisingly Secondary syphilis enough-also in later investigations, very few Table 3 includes 783 patients with secondary reports ofthe results of penicillin treatment complete- syphilis who were treated on the basis of the same ly satisfy the criteria and methods of analysis well dose/time relationship schedules as those for primary outlined by Iskrant et al. (1951); an important syphilis, or in some cases (Capinski et al., 1969) with exception is the Blue Star study in the USA (1945) somewhat higher total doses. Among these patients (Bauer, 1951; Shafer et al., 1954). Moreover, there 98-100% serological nonreactivity (using lipoidal are differences in definition concerning stages of antigen tests) was achieved. No relapses were disease in relation to characteristics at the start of observed, but in another group of 196 patients there therapy, the nature ofpenicillin preparations, dosages was a 3.2% rate of recurrence among the 158 cases and treatment schedules in relation to time/dose followed, the majority thought to be reinfections considerations, follow-up observations after treat- (Jefferiss & Willcox, 1963). ment, criteria of cure, etc. We have analysed the voluminous medical litera- "Early syphilis " and " early latents" ture on the effect of the use of penicillin in the Table 4 includes 2 458 patients with " early syphi- treatment of early syphilis since 1943. Only 5 640 lis not defined " or " early latent syphilis " treated patients have been retained as meeting generally the on the basis of similar dose/time schedules in appraisal requirements (Tables 2, 3 and 4). Even in Denmark, the Federal Republic of Germany, France, these patients the stages of the disease are not Poland, Sweden, the United Kingdom, the USA, and always fully defined. In the present study a number the USSR and followed for 2-11 years. In a group of patients have therefore been grouped as " early of 50 patients described as early latents, the retreat- syphilis not defined ", and the category " early ment rate was 8.7 % of the 46 cases followed latents " has been retained where the penicillin effect (Jefferiss & Willcox, 1963), the great majority being in relation to the immunological state at the onset of attributed to reinfections. In a group of 146 patients treatment could not be assessed. with " early syphilis " there was one clinical relapse and 12 reinfections (Degos & Ebrard, 1957) and Primary syphilis serological nonreactivity was achieved in 78 %. In a Table 2 shows the treatment results as observed in parallel survey of 103 early syphilis patients treated Austria, France, Poland, Singapore, Switzerland, the with the same penicillin schedules plus bismuth in United Kingdom, and Yugoslavia. In 1 381 cases of courses lasting 2-4 years, the investigators found serologically nonreactive primary syphilis treated 95 % serological nonreactivity and one relapse but