Download - Streptococcal skin infection acute glomerulonephritis erisipelas and wound infections were all incriminated as antecedent infections of AGNin early studies (Dillon, 1967). Furthermore,

Postgraduate Medical Journal (November 1970) 46, 641-652.

Streptococcal skin infection and acute glomerulonephritis

HUGH C. DILLON, JRM.D.

Professor of Pediatrics and Associate Professor of Microbiology,University of Alabama in Birmingham, Birmingham, Alabama 35233

IntroductionThe role of streptococcal skin and soft tissue in-

fections in the development of acute glomerulo-nephritis (AGN) is now clearly established. Amongthe early studies comparing the epidemiology ofAGN and rheumatic fever it was noted that theformer complication often followed deep or sup-purative streptococcal infection, including woundinfection, whereas rheumatic fever was commonlyassociated only with superficial pharyngeal infec-tions. Impetigo, erisipelas and wound infectionswere all incriminated as antecedent infections ofAGN in early studies (Dillon, 1967). Furthermore,as recently postulated (Wannamaker, 1970), the so-called 'trench nephritis', common during WorldWar I, probably represented post-streptococcalAGN, with streptococcal skin or wound infectionbeing of primary aetiologic importance.

Interest in the role of skin infection in AGNappears to have been relatively dormant during theperiod of World War II and the immediate post-warera. In contrast, intensive interest in streptococcalpharyngeal infection was evident and it was duringthis period that the relation between streptococcalpharyngitis and acute rheumatic fever became un-equivocally established (Rammelkamp, Denny &Wannamaker, 1952). Shortly thereafter the nowclassic studies of epidemic AGN following throatinfection with streptococcal type M-12 led to theestablishment of the concept of there being 'nephri-togenic' strains of streptococci (Rammelkamp &Weaver, 1953). It soon became apparent fromstudies in widely different areas that limited serotypesof group A streptococci, notably M-types 1 and 12,accounted for most cases ofAGN following strepto-coccal sore throat (Wannamaker, 1967).

Following the observations that led to the hypo-thesis of there being nephritogenic streptococci,epidemiologic studies of impetigo in Great Britain(Parker, Tomlinson & Williams, 1955) suggestedthat limited serotypes of streptococci, most of whichwere typable at that time only by slide agglutinationserotyping (which identifies T-antigens, see below)

accounted for most cases of impetigo. These investi-gators coined the term 'impetigo streptococci'.AGN was not reported in that particular investiga-tion but it has become subsequently apparent thatthese studies represented a fundamental contributionto the understanding of differences in the epidemio-logy of streptococcal skin and pharyngeal infections.Sporadic reports of AGN following impetigo, mostof which were from the south eastern area of theUnited States where impetigo has long been recog-nized to be a perennial problem during the warmand humid months of summer and early autumn,were notable in that the streptococci isolated fromthese patients were not further identified other thanbeing beta-haemolytic organisms or group A strepto-cocci. Limitations in ability to serotype the skinstreptococci were clearly a problem in these investi-gations (Blumberg & Feldman, 1962; McCulloughet al., 1951).

In 1954, there occurred a significant epidemic ofAGN at the Red Lake Indian Reservation in northernMinnesota. Two-thirds of the sixty-three patientsreported to develop AGN were definitely consideredto have pyoderma as the antecedent of that complica-tion, with pharyngitis playing a minimal, if any, rolein the outbreak (Kleinman, 1954). The epidemicpeak ofAGN at Red Lake occurred in late summerwhen weather in that particular area is hot andhumid, not unlike that experienced in the southeastern area of the United States. The first outbreakof AGN at Red Lake is perhaps best rememberedbecause the epidemic strain of streptococcus provedto be a new serotype-M-49 (Updyke, Moore &Conroy, 1955) and was the first serotype of strepto-coccus found to be clearly related to AGN followingskin infection. This strain, subsequently known asthe Red Lake strain, has now been incriminated as acause of skin infection and AGN in many parts ofthe world.During the mid 1960s renewed interest in the role

of skin infection and AGN developed largely as aresult of extensive investigations in three particularareas: Alabama, Red Lake and Trinidad. It is

copyright. on 29 M

arch 2019 by guest. Protected by

http://pmj.bm

j.com/

Postgrad M

ed J: first published as 10.1136/pgmj.46.541.641 on 1 N

ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Hugh C. Dillon

apparent that certain findings have been common toeach area. Each study has contributed distinct new

information and collectively they have significantlyfurthered our understanding of the epidemiology ofpyoderma and AGN. The more exciting observa-tions include: the recognition of certain serotypes,including new or provisional M-types, to be ofmajor importance in AGN; further documentationof 'impetigo streptococci' as a distinct group ofstrains in which serotype and site of infectionappear clearly related; the demonstration thatstreptococcal antibody responses in skin andrespiratory infection clearly differ and some interest-ing studies of subclinical nephritis that have implica-tion with regard to the pathogenesis of AGN. Interms of numbers of cases of AGN, the Alabamastudies have revealed a consistent number of cases

to occur on an endemic basis within a given popula-tion each year; limited cases of AGN occurred atRed Lake, with a second epidemic occurring in 1966and interestingly it, too, was associated with type 49streptococci (Anthony et al., 1967). In Trinidad therehave been massive epidemic waves of AGN. Type49 streptococci and at least two new serotypes havebeen incriminated in these epidemics (Parker, 1969).Some of the observations and results of studies at

Alabama are being reported in detail here. Theseresults and certain of those from Red Lake andTrinidad are compared. Individual or unique con-

tributions of studies from the three particular areas

will be emphasized.

Characteristics of streptococcal impetigoStreptococcal impetigo lesions characteristically

begin as small vesicles, becoming purulent andfinally developing thickened honey-coloured crusts.A typical example of the latter type lesions is illus-trated (plate 2, between pp. 679 and 680). It is notunusual to see lesions in varying stages in a givenpatient, or among siblings within a given family. Thestreptococcal skin lesions seen among Indian childrenat Red Lake and among natives of Trinidad areremarkably similar to those illustrated here andlesions of this type clearly predominate in each area.

Chronic indolent lesions ('ecthyma'), are also ob-served and collectively the various skin lesions are

simply referred to as pyoderma. The clinical appear-

ance of streptococcal impetigo is similar in patientswith and without the complication of AGN.The similarity of bacteriologic findings at Ala-

bama, Red Lake and Trinidad, have also been re-markable (Dillon, 1968; Anthony, Perlman &Wannamaker, 1967; Potter et al., 1968; Parker et al.,1968). Table 1 illustrates the frequency of recoveryof group A streptococci, with or without staphylo-cocci, from children seen during recent years inAlabama. It is to be noted that presence of staphylo-cocci in such skin lesions does not alter theirappearance. Indeed, staphylococci appear for themost part to play a secondary role. In our experiencethe presence of penicillin-resistant staphylococci inskin lesions has in no way interfered with effectivetreatment of streptococcal impetigo with penicillin(Dillon, 1970).

Limited numbers of patients with bullous impetigoare also seen in Alabama and it is in lesions of thistype that the staphylococcus is of primary aetiologicsignificance (Dillon, 1968). Most of the bullousstaphylococcal isolates are members of phage groupII, usually type 71, and may or may not be sensitiveto penicillin. When such lesions become secondarilyinfected with streptococci, as has been observed tobe the case among patients studied here, they assumethe typical purulent appearance of streptococcallesions. Infection that remains limited to staphylo-cocci leads to formation of a very thin and varnish-like crust with a bit of serous exudate at the marginof the lesion; staphylococci may be isolated fromthis margin until the lesion is healed. There are otherclinical differences observed among patients withstreptococcal impetigo as opposed to bullous im-petigo which the author has described in more detailelsewhere (Dillon, 1968). One of the principal dif-ferences, however, is the more pronounced associa-tion of lymphadenopathy with streptococcal im-petigo.

Epidemiologic features of streptococcal impetigo andacute glomerulonephritisThe peak seasonal incidence of impetigo in the

south-eastern area of the United States occurs duringthe hot and humid months of late summer and earlyautumn, and the seasonal pattern ofAGN associatedwith impetigo parallels this pattern. Fig. 2 depicts the

TABLE 1. The recovery rate for group A streptococci, with or without concomitant isolation of staphylo-cocci are shown for patients with nion-bullous impetigo seen over a period of 3 years. Patiernts with bullous

impetigo are shown for comparison

Number Total Streptococciof Group A Pure + Pure No

Clinical type patients streptococci streptococci staphylococci staphylococci growth

Non-bullous 878 92% 340 66%0 400 400Bullous 49 - - - 100%0

642

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Streptococcal skin infection and AGN

35I-

30-

cna)20

0

Q)

D51Ez

EM= Skin infection=Skint resp.= Resp.

emEM

10o

5

.......

::4;

JAN FEB MAR APRIL MAY JUN JULY AUG SEPT OCT NOV DEC

FIG. 2 Epidemiologic features of 120 cases ofAGN, seen in Alabamaover a period of 3 years are illustrated. The relation between seasonand primary antecedent site of infection reveals that skin infection ispredominant,and peak months are late summer and early autumn.

experience in Alabama over a 3-year period (1964-66) when altogether 120 children with AGN werestudied. It is apparent from this figure that the skinis the principal site of antecedent infection of AGNseen in this particular population. August andSeptember are annually peak months forAGNfollow-ing skin infection. The seasonal pattern of epidemicAGN at Red Lake is noteworthy in that both theoriginal and the second outbreak of AGN followingpyoderma occurred during summer and early autumnmonths, following an almost identical pattern to theAlabama experience illustrated in this figure.

In Trinidad, where massive waves of AGNfollowing skin infection have been reported, strepto-coccal impetigo appears to be common at all seasonsof the year. The outbreaks of AGN have been non-seasonal, occurring in an irregular fashion (Parker,1969). The lack of a definite seasonal pattern forskin infection and AGN in Trinidad is most likelyexplained by that tropical island's climate, whichvaries little year round. Non-seasonal, endemic orepidemic skin infection and AGN also occurs inother areas in the Caribbean, as well as in other partsof the world including Ceylon (personal communica-tion from Dr Barry M. Panter). The role of humidityand/or a rainy season in contributing to upswings ofskin infection and AGN has not been completelyclarified but in general such conditions seem tofavour their increase. Such climatic factors maycreate conditions favourable for breeding of bitinginsects, such as the mosquito, which contribute toan increased incidence of minor skin trauma that inturn predisposes to streptococcal skin infection.We know less about the mode of transmission of

streptococcal impetigo than that of streptococcal

respiratory infections. In the latter, direct spreadoccurs via the respiratory tract; factors of impor-tance include active infection, quantity of organismsshed and duration of the carrier state (Wannamaker,1954). Among the factors currently considered tocontribute to the development of or spread of im-petigo are: overcrowding, poor skin hygiene, anincreased incidence of minor skin trauma and in-adequate health-care facilities resulting in neglect ofskin infection, permitting persistence of organismswhich can then infect other contacts. The indolentnature of impetigo may further contribute to suchneglect. Direct contact with lesions, hand con-tamination, and insects that serve as mechanicalvectors in transporting viable organisms may allplay a role in the spread of impetigo. In support ofthe latter, are interesting observations from Trinidadthat the Hippelates fly, a small gnat-like insect, iscapable of carrying viable organisms from patient topatient. These insects feed upon skin sores and arethought to play a role in the passive transfer ofstreptococci from patient to patient (Bassett, 1967;Parker, 1969).

Recent interesting studies from Red Lake(Dudding et al., 1970) suggest that streptococciassociated with impetigo may reside on the intactskin of patients prior to the development of skinlesions. It appears that patients with impetigo mighttransmit organisms to siblings or other intimatecontacts and that streptococci peculiar to skin infec-tion have the capability of survival on normal skinfor days or weeks prior to the actual development oflesions. Colonization of 'healthy' skin also precededthe recovery of streptococci from the respiratorytract when the latter was found to be colonized.

- o E f o q s e o v=. - ^ - ::w B - - o . - o Xd-jefZIA-k

643

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Hugh C. Dillon

Such observations offer intriguing evidence of pos-sible important differences in streptococci causingskin and respiratory infection.

Patients with streptococcal impetigo not in-frequently harbour organisms in the upper respira-tory tract. In our own experience these organisms arecommonly of the same serotype as that found in theskin lesions, being serotypes related to impetigorather than pharyngitis (Dillon et al., 1967). Theconcomitant isolation of streptococci from skin andrespiratory sites in patients with impetigo (includingthose with AGN) studied in Alabama has averagedapproximately 30%. during the different years ofstudy. Typically, small numbers of organisms arefound in the respiratory tract. Acute infection of therespiratory tract is uncommon, and it seems unlikelythat colonization of the respiratory tract plays amajor role in the spread of impetigo (Dillon, 1968).However, among individual patients or families, therespiratory tract may serve as a transient reservoirfor organisms capable of infecting the skin. TheTrinidad and Red Lake experiences with regard tothe simultaneous recovery of streptococci from theskin and respiratory tract have been similar to thatin Alabama, and there is general agreement that therespiratory tract is likely colonized secondarily fromskin infection rather than vice versa. This concept isstrongly supported by the detailed observationsprior to and during the second outbreak of nephritisat Red Lake (Anthony et al., 1969).

Returning to Fig. 2, certain patients included inthis figure are shown as having both skin andrespiratory infection; skin infection was the pre-dominant or major site of antecedent infection inthis group. In most of these patients, respiratorysymptoms were minor and could not be correlatedwith the recovery of streptococci from the respiratorytract. It is further apparent from this figure that whencases of AGN have followed throat infection only,the typical seasonal pattern of respiratory infection,namely, autumn, winter or early spring months hasbeen observed. In contrast to the group of patientsjust mentioned with skin infection and minorrespiratory symptoms, the latter group of childrenhave uniformly had a primary complaint of sorethroat, and occasionally scarlet fever, in the absenceof any superficial skin infection.Although the majority of patients in these recent

studies are members of darker skin races, it is notpossible at present to state clearly that there is aracial predeliction for skin infection. Socio-economicfactors might account for this aspect of the problem.Age is of interest in that there is a definite predelic-tion for younger children to have streptococcal skininfection in contrast to pharyngitis, which is morecommon in early school years. It is not unusual toencounter streptococcal impetigo in infants and

young children. A high proportion of those childrenstudied in Alabama include pre-school children, themean age being just under 6 years. It has not beenunusual to see children under 3 years of age withAGN following impetigo.

Characteristics of streptococci associated with skininfection and acute glomerulonephritisSerological typing of skin streptococciA brief description of the methods of serotyping

of streptococci is necessary in order to understandthe rather difficult problems that have been en-countered in the proper identification of thoseorganisms isolated from skin lesions. Current know-ledge of the epidemiology of acute glomerulo-nephritis associated with impetigo has resulted fromextensive efforts of several laboratories to identifystreptococci causing skin infections.Group A streptococci are serotyped by two

methods, the Griffith slide agglutination method(Griffith, 1934) and the precipitin method of Swift,Wilson & Lancefield (1943), which identifies theM-antigen of the streptococcus. The T-antigens ofgroup A streptococci are serotyped by the agglutina-tion method, employing trypsinized suspensions ofwhole streptococcal cells. T-antigens are cell-wallproteins but their role in streptococcal infectionremains uncertain. Their principal usefulness is inepidemiologic studies. The Griffith method has beenused for many years in major reference laboratories.It is possible to identify over 90V. of strains of groupA streptococci regardless of their source of isolationby this method. The principal disadvantage is arelative lack of specificity, particularly for certainstrains common in skin infections, since thesestrains often react with more than one T-antiserumresulting in agglutination patterns or complexes.

M-antigens of group A streptococci are respon-sible for virulence and since they are type-specific (itis rare for a strain to possess more than one M-antigen), the identification of the M-antigen is ahighly useful serologic-epidemiologic too]. Protec-tive antibody to M-antigen develops in the infectedhost, and type-specific immunity to a particularserotype of streptococcus can be determined. Theprimary disadvantage of the serotyping of strepto-cocci by the M-precipitin method alone is that it isnot uncommon to encounter upwards of 50%. of thestrains in any large collection of streptococci whichcannot be identified with available M-antisera.Indeed, this has been a chronic problem in studies ofskin infection. It is presently recognized that the dualsystem ofM and T-typing provides clear advantagesto the investigator in that strains which might nototherwise be identified can be initially characterizedby T-typing; M-typing often becomes a simpler anda less time consuming procedure, since T-antigens

644

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/


or complexes can generally be related to certain M-serotypes, limiting the number of M-antisera re-quired for initial testing of the strain in question.

Impetigo streptococciThe observations of Parker and colleagues that

limited strains of streptococci were commonly iso-lated from skin lesions but seldom from other formsof infection first drew attention to the possibility thatthe biologic characteristics of skin streptococcimight somehow differ from those strains isolatedfrom other sources (Parker et al., 1955). A relation-ship between the serotype of the streptococcus andthe site of infection was thus postulated and theseinvestigators coined the term 'impetigo strepto-cocci'. In the original studies, few of the impetigostreptococci could be identified with available M-antisera and it was suggested that these skin strainseither lacked M-antigen, or were representative ofstrains for which antisera was difficult to prepare, orwere perhaps representative of new serotypes. Wenow know that all three of these postulates weretrue. There are among skin streptococci strainswhich are representative of serotypes for which M-antisera are difficult to produce; there are others thatseem to be weak in their production of M-antigen orappear to have lost M-antigen; and finally, there arenow several examples of distinct new M-serotypeswhich have been described as a result of epidemio-logic studies of skin infection and AGN. It hasbecome increasingly apparent that streptococci mostoften associated with skin infection are a rich sourceof new M-types.Table 2 contains a list of streptococci now

recognized to be clearly associated with skin infec-tion. T-agglutination patterns are shown and M-serotypes which share these particular T-antigens oragglutination complexes are indicated as being

found either predominantly or infrequently amongskin strains. The first three agglutination complexesnoted in this particular table, 3/13/B3264,8/25/Imp.19, and 5/27/44, were those originallyreported to be 'impetigo streptococci'. It has becomeapparent (from studies in Alabama, at Red Lake,Trinidad and elsewhere) that the first two agglutina-tion complexes illustrated are particularly importantin skin infection wherever such infections have beenstudied. In addition, it is now recognized that certainother group A streptococcal serotypes appear to becommonly involved in skin infections but are lessimportant or infrequently found in other forms ofinfection. This includes certain strains which can beidentified by agglutination as T-type 11 or 12, oftennon-M-typable with available reference sera, butsome of which are representative of provisional newtypes of streptococci (see below); strains agglutinatedas T-14/49 which are M-type 49; strains which areagglutinated as T-4 and may be members of pro-visional M-type 60; and strains of the agglutinationcomplex 5/17/19/23/47 among which the new M-serotype 54 was found. The relationship of certainof these serotypes of streptococci to AGN followingpyoderma is described in detail shortly. It is perti-nent to point out here that of the last ten M-sero-types of streptococci to be recognized as new orprovisional new serotypes, all have been found as aresult of studies of skin infection. These include M-types 52, 53, and 54, from Red Lake (Top et al.,1967); type 55 (Potter et al., 1968), type 57 (Parkeret al., 1968), provisional type 58 (personal communi-cation from Dr W. R. Maxted) from Trinidad;provisional types 59, 60 and 61 identified in ourpopulation in Alabama (Dillon & Avery-Reeves,manuscript in preparation). Type 56 was isolatedfrom a patient with pyoderma in Tennessee (Johnisonet al., 1968).

TABLE 2. Serotypes of streptococci predominantly associated with skin infection are shown.M-types found among the prevalent T-agglutination complexes common to impetigo strepto-cocci are listed. M-types with similar T-types, but infrequently associated with skin infection,

are shown for comparison

M-serotypes

T-agglutination Predominantly skin Rare!y skin

3/13/B3264** 33, 39, 41, 43, 52*, 53* 3, 138/25/lmp.19** 2, 55*, 57*, Prov. 58* 25

(occ. 8 and 56)5/27/44 (Originally 5/11/12/27/44)** Usually non-typable 5, 27, 4411, 12, or 11/12, or 'W' Pool Usually non-typable 11, 12, 22

Occasionally Prov. 59*, 61*, M-I114/49 49 144 Prov. 60*,

Occasionally M-45/17/19/23/47 54*

** Original agglutination pattern considered by Parker et al. to be 'impetigo streptococci'.* New or provisional new M-serotypes.

645

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Hugh C. Dillon

Streptococci found in the agglutination complex3/13/B3264, the complex most commonly identifiedwhere impetigo lesions are observed, include severaldifferent M-serotypes, not all of which are prevalentin a given area. Thus, in Alabama M-types 33, 39and 52 have more often been observed whereas inMinnesota types 41, 52 and 53 and more recently 43have been found. Type 52, also found in Trinidad, iscommon in all three areas. Many strains in thiscomplex remain unidentified as to M-type. Amongthe 8/25/Imp.19 strains, those incriminated in en-demic or epidemic AGN, most of which can now beidentified by M-type, are discussed subsequently.The 8/25/Imp.19 strains from patients with uncom-plicated impetigo seen in Alabama and at Red Lakeare frequently non-M-typable. (M-types 8, 25 and56 are occasionally found). Most of the strains withinthe agglutination complex 5/27/44 have been non-M-typable; they are found less frequently in im-petigo than the above two complexes.

In Alabama, strains identified only as T-12 arecommonly associated with impetigo but are rarelyM-typable. (One of the provisional new serotypes,provisional type 59, has been observed amongchildren with uncomplicated impetigo and thisstrain may show an agglutination reaction with eitherT-type 11, 12, or both). M-type 54, on the basis ofcurrent epidemiologic information, is not widelydistributed.

The nephritogenic skin streptococciThe serotypes of streptococci now recognized of

importance in causing AGN following skin infectiondiffer from those commonly associated with pharyn-gitis and AGN as clearly as pharyngitis itself differsfrom impetigo. This is truly one of the fascinatingfeatures of the epidemiology of AGN.

Table 3 summarizes the relation between serotypesof streptococci and the common site of infection ofthe currently recognized nephritogenic streptococciincluding certain provisional new types which appearof definite, though as yet limited importance.Nephritogenic strains associated with throat infec-tion include types 1, 3, 4, 12 and 25, types 1 and 12having been most often associated with epidemicAGN. In spite of vigorous efforts to identify strainsof these particular types where nephritis has followedskin infection, they have been rarely found.

Current knowledge of the nephritogenic serotypesof streptococci associated with skin infection islargely based on the recent Alabama, Red Lake andTrinidad studies being discussed herein. The strainsof principal importance are: M-2, M-55 and M-57,all of which share a common agglutination complex,8/25/Imp.19 and M-49. The provisional new types,59, 60 and 61, are further discussed below. Neither

TABLE 3. Serotypes of streptococci associatedwith AGN are listed, according to T- and M-antigens and the common site from which suchtypes are isolated. The limited distribution of

certain types is indicated

Serotype Infection site

T-antigen M-antigen Skin Throat

I 1 +8/25/Imp.19 2 +3/13/B3264 3 +4 4 +4 Prov. 60* 46 6 +12 12 +8/25/Imp.19 25 +14/49 49 +3/13/B3264 52**8/25/Imp.19 55 +8/25/Imp.19 56*** +8/25/lmp.19 57 +'W' Pool Prov. 59* +11 Prov. 61* +

* Alabama Studies: Provisional 60 stronglyimplicated; prov. 59 and 61 limited cases.

** Limited cases, Red Lake; Alabama;Trinidad.

*** Limited cases, Tennessee.

,I

It'.

- {) c,CM - 2 . .,/. M- 49 Prov Prov Others

(8/ 25 /Imp 19) 60 :M59

43% 30% ....0 IO%10%

%/O 20 30 40 50 60 70 80 90 1OO

FIG. 3. Major nephritogenic serotypes of streptococciisolated from skin lesions of patients with AGN (149 =49%) and their siblings (155 = 51 %) over a 5-year period(1964-68) of study in Alabama are shown. M-type 2 hasbeen most common. The total number of strains was 304.

type 52 nor 56 appear of major importance in AGNat this time.Type 49 streptococci, from an historical point of

view, are of special interest since this was the firstserotype to be clearly related to skin infection andAGN and remains the one type which has been clear-ly implicated as a significant cause of skin infectionand AGN in many parts of the world. Its distribu-tion includes widely different areas in the UnitedStates, Europe and Trinidad where it has been amajor epidemic strain. The author has received type49 strains isolated from children in Ceylon reportedto have AGN. Unlike the other strains presentlyassociated with skin infection and nephritis, type 49has also been implicated in AGN following phiaryn-gitis, but is most often associated with pyoderma.

M-type 2 streptococci have been the major causeof AGN in patients studied in Alabama (Dillon,Maxted & Reeves, 1968). As shown in Fig. 3, thisserotype accounted for nearly half the cases ofAGN

,, . I.

6546

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/


from which the skin isolates were identified duringthe first 5 years of investigation. It is annually themost prevalent type among the AGN patients andtheir siblings; of thirty patients with skin infectionandAGN seen in 1969, two thirds had M-2 infection.

These type M-2 strains causing skin infection andAGN differ from 'classical' M-2 streptococci foundin throat infection in their agglutination reactions.The latter possess the T-type 2 antigen only; thetype 2 skin strains agglutinate only with T-antisera8/25/lmp. 19. In this regard, they resemble M-types55 and 57, two of the major nephritogenic strainsisolated in Trinidad.M-2 streptococci with the agglutination complex

8/25/Imp.19 have been isolated from the pharynx ofsome patients with type 2 skin infection, but in nocase have we identified them in patients with pharyn-gitis and AGN. Even more impressive, however, isthe infrequency with which they have been isolatedfrom patients other than in families where one ormore cases of AGN have been identified.

Other types of importance in AGN in Alabamaare M-49, the second most common type we find,and the provisional types 60 and 59. Provisionaltype 60, one of three new provisional types identifiedin the Alabama studies (Dillon & Reeves, in prepara-tion) have been isolated annually from patients withAGN in this population. This strain is agglutinatedwith T-4 antisera, as is M-type 4. The M-antigens (4and prov. 60) do not cross-react, however, and wehave not found M-4 strains to be related to pyo-derma and AGN. Provisional type 60, like M-2, hasbeen infrequently found other than in families inwhich AGN has occurred. Provisional type 59 hascaused sporadic cases of AGN, and during 1 yearwas fairly commonly isolated from patients with'uncomplicated impetigo'. Provisional type 61 is theleast common of the three new types; the future roleof these strains as a major cause of AGN remains tobe determined. Their identification, however, hasserved to further clarify the epidemiology of skininfection and AGN.

It has been previously mentioned that many ofthe impetigo streptococci have been difficult toidentify other than by T-agglutination. By employ-ing currently available antisera, plus that for pro-visional new types, we have been able to identifyover 9500 of the streptococcal isolates from skinlesions of patients with AGN. The majority ofnephritogenic strains at Red Lake have been M-type49. In the Trinidad studies the preparation of anti-sera for the new types 55 and 57 made their identi-fication possible. There is no longer doubt that thenephritogenic streptococci from skin lesions possessM-antigens.There remain among the impetigo streptococci

many strains which as yet are classified only by

T-agglutination typing. Experience has shown, how-ever, that new serotypes are to be found among suchstrains. It is increasingly apparent that the skinstreptococci will require M-antisera prepared againstappropriately selected wild strains causing skininfection, if maximum success in identifying M-antigen is to be achieved.

Characteristics of acute glomerulonephritis associatedwith skin infectionThe clinical picture of AGN associated with

pyoderma is identical to that seen with streptococcalpharyngitis. The classical manifestations of hyper-tension, oedema, and typical urine findings prevailregardless of the antecedent site of infection. Alldegrees of severity of AGN have been observed inchildren with pyoderma including some who havepresented with acute encephalopathy, and otherspresenting with acute congestive heart failure. Themajority have presented with an acute illness charac-terized by a variable degree of hypertension andoedema (Dillon et al., 1967). The observed sex ratioin patients with AGN has been approximately 60%males to 40%° females until 1969, during which timethere was a 2 to 1 male to female ratio. In contrast,there is no difference in sex incidence of childrenwith 'uncomplicated impetigo'. This prevalence ofmales in the AGN population remains unexplained.The average hospital stay for children with overtevidence of AGN has been 10 days. Extended post-hospital follow-up studies have revealed that micro-scopic haematuria persists in some instances formany months. In the majority of patients however,clinical and iaboratory parameters have returned tonormal within a matter of weeks following the onsetof the disease.Although we have not performed kidney biopsies.

they were obtained from children studied during the1966 Red Lake epidemic and in Trinidad (personalcommunication from Dr B. F. Anthony and DrElizabeth Potter). The histopathological studies wereessentially the same as those observed previously inchildren whose AGN followed pharyngitis. It thusappears from both the clinical and pathologicalstandpoint that there is no difference in the diseasewith regard to the antecedent site of infection.

Complement in clinical and subclinical nephritisThe determination of beta 1-A/beta 1-C-globulin

(C13 component of complement) has proven ofparticular value in our studies of AGN followingskin infection. Immunoelectrophoretic and radialimmunodiffusion techniques have been used, thelatter being preferable because of s-mplicity and theease with which results can be quantitated (Derrick,Reeves & Dillon, 1970). Complement is low duringthe acute phase of pcst-streptccoccal nephritis, but

647

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Hugh C. Dillon

returns to normal within a few weeks of onset of thedisease.Many children with clinically overt AGN admitted

to our studies have had both clinical and laboratoryevidence, including low complement, of AGN whenfirst examined. Additional cases of overt and sub-clinical AGN have been identified among siblingcontacts of such index cases. Observations in thelatter group have been of particular interest. Quanti-tative results of beta 1-C/beta 1-A-globulin levels inchildren with overt AGN, their siblings and a selectedgroup of patients with 'uncomplicated impetigo'seen during one particular year are shown in Fig. 4.

0 Overt AGNs

0

A

of Normal0

~:c 00 0oOoO

Without haematuriaor I BP 0 Uncomplicatedimpetigo with

With haematuria haematqria6 and/or t BP and/orl BP

ll

A'00 0

.' ' l..*- * i^ 0o ° °

A °

FIG. 4. This figure illustrates beta 1-C/beta 1-A-globulindeterminations found in the three major clinical cate-gories of patients studied with skin infection: AGN;siblings of AGN and 'uncomplicated impetigo'. Onlyten of 255 patients in the latter group are illustrated.None of the remaining 245 children had levels lower thanthose shown here, and only 1/255 had a value less thannornmal.

Serial determinations (three or more) were done inall patients admitted to that study. Initial or lowestlevels are shown in the figure. The lower limit ofnormal is 100 mg/100 ml, and all but one patientwith clinically overt evidence ofAGN fell well belowthis level. The latter patient was seen relatively lateduring the course of AGN. As just described, sib-lings of patients with overt AGN are commonlyinfected with streptococcal serotypes identical tothat isolated from the index case of AGN in the

family. Siblings shown here most often had infectionwith either M-2 or M-49 streptococci, both prevalentnephritogenic strains that year. Among the seventy-two AGN siblings nine had a low beta 1-C/beta1-A-globulin, including two in whom this was theonly evidence of subclinical AGN. In several othersiblings, the drop in complement preceded theoccurrence of haematuria. This relation of haema-turia and low complement is further discussed below.A total of 255 patients with 'uncomplicated

impetigo' had serial complement determinations butfor clarity only those ten individuals found to haveeither microscopic haematuria or a transient eleva-tion of blood pressure are depicted in this particularfigure, since none of the remaining 245 patients hadcomplement levels less than these ten. Children with'uncomplicated impetigo' are typically infected withstrains infrequently associated with nephritis; iden-tical serotypes are usually to be found amongsiblings within a given family. The one patient inthis group of 255 children with 'uncomplicatedimpetigo' found to have a low beta 1-C/beta 1-A-globulin, however, was an only child who was in-fected with a nephritogenic streptococcus. He wasalso the only patient in the 'uncomplicated impetigo'group found to have both microscopic haematuriaand a transient elevation of blood pressure.

It is difficult to be certain that transient micro-scopic haematuria, or a single elevation of bloodpressure is of significance in evaluating patients withstreptococcal skin infection. As a result of observa-tions now extending over a 3-year period, we havefound that beta 1-C/beta 1-A determinations areparticularly sensitive in detecting laboratory evidenceof subclinical AGN. A total of 198 siblings of patientswith overt AGN have been screened. A summary ofabnormalities in the eighteen children we consideredto have evidence of subclinical AGN is shown inTable 4. The nephritogenic serotype causing infectionin both index case and affected sibling is included.Seventeen of the eighteen had a low beta 1-C/beta1-A-globulin demonstrated during the course ofinfection.The most striking observation however, was the

demonstration of a low beta 1-C/beta 1-A-globulinlevel whether or not haematuria occurred. Low

TABLE 4. The finding of low beta 1-C/beta 1-A-globulin, with or withoutother abnormalities, in siblings of patients with clinically overt AGN isshown. Nephritogenic serotypes are included, and reveal the predominance

of M-2

Streptococcal Number of Elevatedserotype patients Low C13 Haematuria blood pressure

Type 2 14 13 7 4Type 49 2 2 0 1Prov. 60 2 2 1 0Totals 18 17/18 8/18 5/18

2502302101901701501301109070503010

0

E*)

648

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/


complement occurred in the absence of haematuriain ten of seventeen patients, and preceded haema-turia in four of the remaining patients. Complementremained normal in two of the eight siblings withhaematuria and one of the five with transient hyper-tension. The combination of low complement,microscopic haematuria, and transient hypertension,occurred together in only one of the eighteensiblings. None of this particular group progressedto develop clinically overt AGN. The fact thatchanges in complement occurred prior to the de-velopment of microscopic haematuria, or in itsabsence, fits the current concept that an immuno-logic mechanism presumably induced by infectionwith a nephritogenic streptococcus, and involvingutilization of complement, is responsible for therenal injury observed in AGN.

In contrast to these exciting observations in thesiblings of children with overt AGN, among thatlarge group of children with 'uncomplicated im-petigo' studied during the same period of years(nearly 500 children) only fourteen had any of theabove mentioned abnormalities. Only 3/14 had adrop in complement, and in each case they were foundto be infected with a recognized nephritogenicstreptococcus prevalent in this population. Two ofthe three children with low complement also hadmicroscopic haematuria. Only one of the three hadall three of the abnormalities (the patient illustratedpreviously in Fig. 4). It is thus apparent that in amajority of children with 'uncomplicated impetigo'microscopic haematuria is infrequently observedand not often associated with or preceded by changesin complement. This observation is in keeping withearlier studies (Stetson et al., 1955) which demon-strated a higher incidence of haematuria in militaryrecruits infected with a recognized nephritogenicstrain of streptococcus than in those infected withserotypes not commonly associated with AGN.We have concluded that beta 1-C/beta 1-A-

globulin determinations are particularly useful inconfirming cases of clinically overt post-strepto-coccal AGN, in detecting subclinical cases of AGNin children at risk for this complication (notablythose infected with a recognized nephritogenicstrain) and in excluding the likelihood of post-streptococcal AGN in children with uncomplicatedstreptococcal skin infections caused by a variety ofstreptococci which infrequently caused AGN.

It has long been recognized that multiple cases ofAGN occur within families within a short period oftime of one another. We have documented four casesof post-streptococcal AGN following type 2 strepto-coccal skin infection in one family and three cases ofovert AGN in two other families and several ex-amples of two or more cases of clinical and/or sub-clinical AGN occurring within a given family. These

observations further emphasize the necessity forprompt and early studies of siblings of children withovert AGN for either detection of or possible pre-vention of further cases.

Streptococcal antibody responses-impetigo andnephritisThe relation between streptococcal immune

response and site of infection is also of particularinterest in considering differences in skin andrespiratory infection. Limited observations fromearlier studies (Burnett, 1962; Markowitz et al., 1965)suggested a somewhat irregular or erratic ASOresponse in children with streptococcal impetigo. Inreviewing these data, it is difficult to clearly relatebacteriologic to serologic findings; however, thepattern of poor ASO response in proven strepto-coccal impetigo was clearly suggested. We nowknow from the various epidemiologic studies thathave been mentioned thus far that the nature of thestreptococcal antibody response is clearly affectedby the site of antecedent infection and in limitedinstances by the serotype of the infecting strain.Anthony and co-workers (1967) found the anti-

DNAse-B test to be definitely superior to the ASOtest in demonstrating serologic evidence of strepto-coccal infection in patients with pyoderma at RedLake. In Alabama (Dillon & Reeves, 1969) we havefound that in patients with streptococcal impetigo,whether or not AGN complicates the infection,anti-DNAse-B titres are approximately twice asoften elevated as ASO titres. Anti-DNAse-B titreshave been found to be strikingly elevated in patientswith skin infection and AGN in this population(Derrick et al., 1970); thus both the frequency of andmagnitude of immune response is greater to DNAse-B antigen than streptolysin 0. An extensive studycomparing differences in antistreptolysin 0 andanti-DNAse-B titres in patients with impetigo, AGNand siblings of AGN, has recently been completedand will be reported in detail (Dillon & Reeves, inpreparation). Certain findings of particular interestwith regard to nephritis are shown in Fig. 5. (Theseantibody determinations were done in that samegroup of patients illustrated earlier in Fig. 2 inwhich the seasonal pattern of AGN in this popula-tion was demonstrated.) These patients, seen over a3-year period, represented the first group with AGNin whom both anti-streptolysin 0 and anti-DNAse-Bdeterminations were performed on one or more seraduring the course of AGN. Findings in succeedingyears have been quite similar to those illustratedhere. Approximately 85% of the patients includedin Fig. 5 had AGN following impetigo. Consideringfirst anti-DNAse-B titres, the upper limit of normalin our population for children over 2 years of age isapproximately 500, but under 2 nearer 90. Among

649

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Hugh C. Dillon

12500800050003170

X 20001250

0 833500

o 3175L 200

125

er 83

50

25

ANTI-DNAse B ANTI-STREPTOLYSIN 0

geo. mean titre 1250 geo. meantitre 200

*: 00*-::::*e-~~~~~000*

*-: 0--00*00-

*-o

5 10 15 20Number of patients

(Total 117)

A-A-&&

A-----

*--&&&*----&& --*-------

5 10 15 20Number of patients

(Total 110)

FIG. 5. A comparison of two streptococcal antibodytests-anti-DNAse-B and ASO-in children with AGNis shown. (This group of patients same as illustrated inFig. 2.) Anti-DNAse-B titres are both higher and morefrequently elevated than ASO titres in patients with skininfection.

those children illustrated here with titres less than500, were several under 2 years of age; of the totalnumber of children with skin infection and AGN90°4 had an elevated anti-DNAse-B titre. The meananti-DNAse-B titre of 1250 is quite impressive andreflects the vigorous response seen particularly inpatients with skin infection and AGN. Included inthis figure are a limited number of patients whoseprimary antecedent site of infection was the throat.Although anti-DNAse-B titres were elevated in 80%.of that group, the titres were less often as strikinglyelevated as in patients with skin infection. Consider-ing the ASO titres (seven fewer patients had ASOtitres done) the upper limit of normal for children of2 years of age and over was 125; that for childrenunder 2 approximately 80. It is clear that titres ofmany with AGN were not elevated. Among thechildren with AGN following skin infection, only50%. had an elevated ASO titre. Highest ASO titreswere found among the small group of childrenstudied with respiratory infection and AGN. Thetitres were also more frequently elevated in thatgroup (80%4) than in those with skin infection (50%4).The relatively low mean ASO titre (200) in AGNpatients reflects the dominance of skin infection asthe antecedent site in this population.We have concluded that the anti-DNAse-B is a

superior streptococcal test in the study of childrenwith skin infection and AGN, and indeed, considerit the primary streptococcal antibody test for thispopulation or others where streptococcal skininfection is predominant. We consider the ASO testless reliable as a means of serologic documentationin children with skin infection and AGN, since thefinding of a normal ASO titre in such patients couldclearly be misleading.

In studies of AGN and skin infection in Trinidadthe antihyaluronidase has been used as a second testalong with the ASO test for serologic studies. It hasbeen demonstrated there that the antihyaluronidaseis a more efficient streptococcal antibody test than isthe ASO test (Potter et al., 1968). It is thus apparentthat in the three widely different areas where exten-sive studies of skin infection and AGN have recentlybeen done, anti-streptolysin 0 patterns are similar ineach area and a second antibody test has been foundnecessary to afford maximum serologic definition ofstreptococcal infection.The anti-NADase (formerly called anti-DPNase)

streptococcal antibody test was previously found tobe of value in studies of AGN following respiratoryinfection. It is now apparent that this test is un-reliable in studies of streptococcal skin infection andAGN (Wannamaker, 1970).The relatively feeble anti-streptolysin 0 responses

in streptococcal skin infections are not completelyexplained. Several explanations present themselvesfor consideration. Streptolysin 0 production or itsantigenicity is conceivably inhibited to some degreewhen streptococci are growing in cutaneous sites.Serotypes commonly producing impetigo may bepoor producers of streptolysin 0. In any event,present evidence suggests that the marked differencesobserved in ASO and anti-DNAse-B responses inpatients with streptococcal skin infection as opposedto respiratory infection are a function of the site ofinfection.

Treatment of streptococcal infectionsIt is not intended here to discuss at length the

appropriate therapy of streptococcal pyoderma, butthe question commonly arises as to whether or notthe treatment of streptococcal skin infection canadequately prevent the development of AGN. Itseems pertinent to emphasize several key points.First, there is the common question regarding therelative merits of topical therapy versus systemictherapy for streptococcal impetigo. In our recentstudies (Dillon, 1970) we have been able to clearlydemonstrate the marked superiority of parenteralversus topical therapy in eradicating streptococci,rapidly clearing lesions clinically, eliminating orga-nisms also present in the upper respiratory tract.More recently (Derrick & Dillon, 1970), wehave shown that either penicillin or erythromycinadministered orally in appropriate doses is quiteeffective in prompt bateriologic and clinical clearingof skin infection. The advantages of such therapyover topical therapy have been thoroughly discussed.The occasional patient with one or two discretelesions (particularly children whose infection isbrought to the attention of a physician quite early)can possibly be treated safely with topical therapy.

650

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Streptococcal skin infection and AGN 651

Our own experience, and that of others studyingpopulations where the economicaliy disadvantagedgroup of patients is seen, suggests that in caseswhere impetigo is fairly extensive and where lesionsare commonly present for days or weeks prior toseeking attention, topical therapy is impracticable,and indeed, an unwise choice.

In spite of vigorous efforts to do so, we have notbeen able to prove unequivocally that we can preventcases of AGN by prompt treatment of streptococcalskin infection. We do feel, however, that prompttreatment of patients with skin infections with arecognized nephritogenic strain is a most importantpublic health measure in preventing further spreadof these organisms. This is yet another reason topursue family studies of patients with skin infectionwhere an index case of AGN has occurred. Prompttherapy in infected sibling contacts, if infection hasnot been present too long, might prevent AGN inthis 'high risk' group.

AcknowledgmentsThis work was done under the sponsorship of the Com-

mission on Streptococcal and Staphylococcal Diseases ofthe Armed Forces Epidemiological Board, and was sup-ported in part by the U.S. Army Medical Research &Development Command, Department of the Army, underresearch contract No. DA-49-193-MD-2635. Additionalsupport was granted by Ross and Wyeth Laboratories.

ReferencesANTHONY, B.F., KAPLAN, E.L., CHAPMAN, S.S., QUIE, P.G.& WANNAMAKER, L.W. (1967) Epidemic acute nephritiswith reappearance of type 49 streptococcus. Lancet, ii, 787.

ANTHONY, B.F., PERLMAN, L.V. & WANNAMAKER, L.W.(1967) Skin infections and acute nephritis in AmericanIndian children. Pediatrics, 39, 263.

ANTHONY, B.F., KAPLAN, E.L., WANNAMAKER,L.W., BRIESE,F.W. & CHAPMAN, S.S. (1969) Attack rates of acutenephritis following type 49 streptococcal infection of theskin and of the respiratory tract. Journal of Clinical In-vestigation, 48, 1697.

BASSETT, B.C.G. (1967) Hippelates flies in acute nephritis.Lancet, i, 503.

BLUMBERG, R.W. & FELDMAN, D.B. (1962) Observations onacute glomerulonephritis associated with impetigo. Jouirnalof Pediatrics, 60, 677.

BURNETT, J.W. (1962) Management of pyogenic cutaneousinfections. Newv Enigland Journal of Medicine, 266, 164.

DERRICK, C. WARREN, REEVES, MARY S. & DILLON, HUGH C.(1970) Complement in overt and asymptomatic nephritisafter skin infection. Jolurnal of Clinical Investigation, 49,1178.

DERRICK, C. WARREN & DILLON, H.C., JR (1970) Furtherstudies on the treatment of streptococcal skin infection.Journal of Pediatrics. (In press).

DILLON, H.C. (1967) Pyoderma and nephritis. Annual Reviewof Medicine, 18, 207.

DILLON, H.C. (1968) Impetigo contagiosa: Suppurative andnon-suppurative complications. American Journal ofDiseases of C/lildren, 115, 530,

DILLON, H.C. (1970) The treatment of streptococcal skininfections. Journal of Pediatrics, 76, 676.

DILLON, H.C., MOODY, M.D., MAXTED, W.R. & PARKER,M.T. (1967) The epidemiology of impetigo and acute

glomeruloniephritis-results of serological typing of groupA streptococci. Amnerican Journal of Epidemiology, 86, 710.

DILLON, H.C., MAXTED, W.R. & REEVES, M.S. (1968) Acuteglomerulonephritis following skin infection due to strepto-cocci of M-type 2. Lancet, i, 543.

DILLON, H.C. & REEVES, M.S. (1969) Streptococcal antibodytiters in skin infection with AGN. Pediatric Research,(Abstract), 3, 362.

DUDDING, B.A., BURNETT, J.W., CHAPMAN, S.S. & WANNA-MAKER, L.W. (1970) The role of normal skin in the spreadof streptococcal pyoderma. Journal of Hygiene, 68, 19.

GRIFFITH, F. (1934) Serologic classification of Str. Pyogenes'Journal of Hygiene, 34, 542.

JOHNSON, J.C., BASKIN, REED C., BEACHEY, EDWIN H. &STOLLERMAN, G.H. (1968) Virulence of skin strains ofnephritogenic group A streptococci: New M protein sero-types. Journal of Immunologi', 101, 187.

KLEINMAN, H. (1954) Epidemic acute glomerulonephritis atRed Lake. Minnesota Medicine, 37, 479.

MARKOWITZ, M., BRUTON, H.D., CUTTNER, A.G. & CLUFF,L.E. (1965) The bacteriologic findings of streptococcalimmLine response and renal complications in children withimpetigo. Pediatrics, 35, 393.

MCCULLOUGH, G.C., COFFEE, J.Y., TRICE, P.A., STONE, J.J.& CRANDALL, H.L. (1951) Acute glomerulonephritis:Impetigo as an etiological factor. A review of 124 cases atLloyd Noland Hospital, Fairfield, Alabama. Journal ofPediatrics, 38, 346.

PARKER, M.T., TOMLINSON, A.H.J. & WILLIAMS, R.E.O.(1955) Impetigo contagiosa: The association of certaintypes of Staphylococcus aureus and Streptococcus pyo-genes with superficial skin infections. Journal of Hygiene,53, 458.

PARKER, M.T. (1969) Streptococcal skin infection and acuteglomerulonephritis. British Journal of Derniatology, 81,Suppl. 1, 37.

PARKER, M.T., BASSETT, D.C.J., MAXTED, W.R. & ARNEAUD,J.D. (1968) Acute glomerulonephritis in Trinidad: Sero-logical typing of group A streptococci. Journal of Hygiene,66, 657.

POTTER, E. V., SIEGEL, A.C., SIMON, N.M., MCANINCH, J.,EARLE, D.P., POON-KING, T., MOHAMMED, 1. & ABIDH, S.(1968) Streptococcal infections and epidemic acuteglomerulonephritis in South Trinidad. Jouirnal ofPediatrics,72, 871.

RAMMELKAMP, C.H., JR, DENNY, F.W., JR & WANNAMAKER,L.W. (1952) Studies on the epidemiology of rheumaticfever in the Armed Services. In Rheumatic Fever (Ed. byL. Thomas), p. 72. University of Minnesota Press, Minnea-polis.

RAMMELKAMP, C.H., JR & WEAVER, R.S. (1953) Acuteglomerulonephritis. The significance of the variations inthe incidence of the disease. Journal of Clinical Investiga-tion, 32, 345.

STETSON, C.A., RAMMELKAMP, C H, JR, KRAUSE, R.M.,KOHEN, R.J. & PERRY, W.D. (1955) Epidemic acutenephritis: Studies on etiology, natural history and pre-vention. Medicine (Baltimore), 34, 431.

SWIFT, H.F., WILSON, A.T. & LANCEFIELD, R.C. (1943)Typing group A hemolytic streptococci by M precipitinreactions in capillary pipettes. Journal of ExperimnentalMedlicine, 78, 127.

Top, F.H., JR, WANNAMAKER, L.W., MAXTED, W.R. &ANTHONY, B.F. (1967) M-antigens among group A strepto-cocci isolated from skin lesions. Journal of ExperimlentalMelicine, 126, 667.

UPDYKE, E.L., MOORE, M.S. & CONROY, E. (1955) Provisionalnew type of group A streptococci associated with nephritis.Science, 121, 171.

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

652 Hugh C. Dillon

WANNAMAKER, L.W. (1970) Differences between strepto-coccal infections of the throat and of the skin. New EnglandJournal of Medicine, 282, (Part I), 23, (Part II), 78.

WANNAMAKER, L.W. (1967) Epidemiology ofacute glomerulo-nephritis. In Acute Glomerulonephritis, 17th Annual Sym-

posium on the Kidney, National Kidney Foundation (Ed.by J. Metcoff), p. 39. Little, Brown, Boston.

WANNAMAKER, L.W. (1954) The epidemiology of strepto-coccal infections. In Streptococcal Infections (Ed. by M.McCarty), p. 157. University Press, Columbia.

copyright. on 29 M


http://pmj.bm

j.com/

Postgrad M


ovember 1970. D

ownloaded from

http://pmj.bmj.com/

Download - Streptococcal skin infection acute glomerulonephritis erisipelas and wound infections were all incriminated as antecedent infections of AGNin early studies (Dillon, 1967). Furthermore,

Top Related