THE RELATION OF UPPERRESPIRATORYINFECTIONS TORHEUMATICFEVER IN CHILDREN

I. THE SIGNIFICANCE OF HEMOLYTIC STREPTOCOCCIIN THE PHARYNGEALFLORA DURING RESPIRATORYINFECTION 1

By MAYG. WILSON, EUGENIA INGERMAN, ROBERT0. DuBOIS,AND BENJAMIN McL SPOCK

(From the New York Hospital, and the Department of Pediatrics, Cornell UniversityMedical College, New York)

(Received for publication December 1, 1934)

The view that the rheumatic process is initiatedor activated by streptococcal respiratory infectionshas met with wide acceptance. The evidencewhich has accumulated in support of this concep-tion originated in the long recognized associationof tonsillitis and rheumatic fever, as well as theso-called rheumatic sequelae of scarlet fever.

The more recent observations of Glover (1),Schlesinger (2), Coburn (3), Collis (4) andothers have stressed a broader view of strepto-coccad respiratory infections and have includedsuch infections as pharyngitis, common cold,sinusitis, otitis, bronchitis and cervical adenitis.It is important to note that the presence ofhemolytic streptococci in the flora of the throatduring these infections has been considered bymany observers as diagnostic of a streptococcalrespiratory infection, the criteria for diagnosisbeing bacteriologic rather than clinical. Coburnand Pauli (5) from extensive epidemiological,bacteriological and immunological studies postu-late that the hemolytic streptococcus initiates therheumatic process.

The majority of investigators emphasize thequiescent period (of apparent health), a period ofapproximately seven to twenty-one days betweenthe respiratory infection and the subsequent rheu-matic recurrence, lending support to the hypothe-sis that rheumatic fever is allergic in nature.Paul and Salinger (6) from historical familystudies have described simultaneous waves of res-piratory infections and rheumatic recrudescencesin several members of the household. Similarobservations have been reported by various in-vestigators, occurring in schools, camps, hospitalwards and convalescent homes. In some instances

1 This study was conducted under a special grant fromthe Commonwealth Fund.

these episodes have assumed epidemic propor-tions.

It must be conceded that these various observa-tions appear quite convincing, but their possiblespecific etiological significance may be questionedin view of the known frequency of respiratoryinfections in children in the seasons when rheu-matic activity is likely to occur. This is illus-trated by the following observations conductedfor a two year period, 1930 to 1932, in which 222ambulatory rheumatic subjects five to fifteenyears of age were under close observation as partof another investigation. These children wereunder the medical supervision of the Heart Clinicof the New York Nursery and Child's Hospitaland were seen at least monthly. The majority ofthe subjects were under our care from the onsetof the first rheumatic symptom. Every effort wasmade to record accurately any illness suffered bythese children.

The seasonal incidence of respiratory infectionsand rheumatic recurrences experienced by thesesubjects is graphically presented in Chart 1.

Here may be noted the three seasons of in-creased incidence of respiratory infections-fall,winter and spring-common in this section of thecountry. The trend of the curves representingthe seasonal incidence of rheumatic recurrences,while somewhat similar to the respiratory curve,is significantly lower, showing the greatest inci-dence in the spring season. During this two yearperiod 222 subjects experienced 783 respiratoryinfections and 401 rheumatic recurrences. It isof some significance that less than 10 per cent ofthese rheumatic episodes were preceded within aperiod of three weeks by a respiratory infection.Bacteriological studies of the pharyngeal florawere not made in those years.

325

M. G. WILSON, E. INGERMAN, R. 0. DUBOIS AND B. MCL. SPOCK

1930 1931

~Wintenr Sprinit Summer FalIl - rnter Sprinr Summer Fallpce.Jan.FOc hMnApr'.M aZYunaJultiA¶ Sept.Oct% gec.Jn.Feb flr.Apr.Im Junwe u Septct.Nowt

Total No. Z18 ZZZ 219 2tO 193 204 200 198Resp.lnq. 140 130 49 193 I P9 9 8 3 5 89Rh.Rec. 65 93 41 39 43 93 4C

7.70

60~~~~~~~~~~~~~~~

50

30~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0

2 0

10 ~ Rispwratornj lniections Follows bi4 Rheumati~Rcurwences

-Totck No. II 8Ilii i i Z3iZ iii 3o- -Respir'ator4 ln4ections *-RheumoeAc Recurrences

CHART 1. SEASONALINCIDENCE OF RHEUMATICRECURRENCESAND RESPIRATORY INFECTIONS 1930-1931This chart illustrates the similar seasonal incidence of respiratory and rheumatic infections for 222 ambu-

latory rheumatic subjects.

In view of the etiological significance attributedto the presence of hemolytic streptococci in theflora of the throat during respiratory infections,it was considered worth while to include bacteri-ological and immunological observations in thefollowing investigations.

To evaluate the possible relation of respiratoryinfections to rheumatic fever, the family wastaken as a unit for study. Ninety-two familiesincluding 123 rheumatic children, members of theCardiac Clinic, were selected for this study.These families have been under our observationfor a period of two to fourteen years. Since Sep-tember 1932 all the children attended the NewYork Hospital. About one-half of the rheumaticsubjects are included in the 1930-1931 study.

Information as to the incidence and time rela-tionship of respiratory and rheumatic attacks wasobtained as accurately as possible. Trained in-vestigators visited the homes weekly and recordedupon a family chart the symptoms of the illness ofevery member of the household. When neces-sary, physicians visited the homes. Weekly throatswabs for cultures were taken routinely from allrheumatic children, from a series of controls and

from every available member of the householdduring illness. Children visited the clinic for ex-amination at least monthly. The majority of themembers of the household were examined duringthe year. Observations were continued on chil-dren admitted to the pavilions and convalescentcottages of the New York Hospital during theperiod of study.2

A more detailed discussion of the bacteriologicaltechnique and criteria followed in these studies isincluded in the third paper of this series (7).Throat culture specimens were taken at regularweekly intervals, in some instances more fre-quently. After inspection of the pharynx, thesurface was rubbed with a sterile dry cottonswab. Blood agar plates were inoculated andincubated for twenty-four hours. The relativenumber of each type of colony was estimated andrecorded numerically from 0 to 4. The relativenumber of colonies of hemolytic streptococciisolated from the pharyngeal flora of each subject

2We wish to acknowledge our indebtedness to theresident staff of the New York Hospital and to MissMarie L. Troup, Superintendent of the New York Hos-pital Convalescent Cottages.

326

HEMOLYTIC STREPTOCOCCUS

was recorded and correlated with the clinicalsigns and symptoms.

The following rheumatic manifestations wererecognized as evidence of activity: active carditis,chorea, arthritis, growing pains, pains in thejoints, eruptions of the skin and rheumaticnodules.

Respiratory infections included the commoncold, nasopharyngitis, tonsillitis, otitis, cervical

throat and the high incidence of respiratory in-fections in contrast to the insignificant rise ofrheumatic recurrences during the months ofMarch, April and May.

During the year, 98 per cent of the subjects ex-perienced a total of 649 respiratory infections.About one-third of the respiratory infections wereassociated with either marked local or constitu-tional symptoms, the remaining were common

CHART2. MONTHLYINCIDENCE OF RHEUMATICRECURRENCESAND RESPIRATORYINFECTIONS EXPERIENCEDBY 123 RHEUMATICSUBJECTS 1933-1934

adenitis and bronchitis. When such infectionswere associated with the presence of hemolyticstreptococci in the pharyngeal flora, they were

designated as "streptococcal" respiratory infec-tions for brevity. This, we believe, is in accordwith the criteria apparently accepted in compar-

able investigations, although contrary to our own

conception.A comparison of the monthly incidence of res-

piratory infections and rheumatic recurrences ex-

perienced by the 123 rheumatic subjects over a

period of twelve months, from September 1933 toSeptember 1934, is graphically presented in Chart2. Of particular interest is the increased inci-dence of hemolytic streptococci in the flora of the

colds. Many children suffered from persistentethmoiditis and sinusitis during the period of ob-servation. Three hundred and fifty-three (54 per

cent) of the total respiratory infections were as-

sociated with hemolytic streptococci in the pha-ryngeal flora. In about one-third hemolyticstreptococci predominated in the pharyngeal flora.

The designation, streptococcal respiratory in-fection, judging from the recent literature, issolely a bacteriological distinction and does notappear to be based on any diagnostic local signsor symptoms in the majority of instances. Thatsuch a bacteriological differentiation of respira-tory infections would not appear to be a valid one,

is indicated by the following. During the months

ISpt. Oc t. Nov. Dec. Jao Fab Malr Apr. tiay June Jury Aug.uVror s.ll1l jig IA.l 122 III2 1 19 1 17 1 llt |113 11% I II |L2

i+ H.S;. ut Throat MmsSL

0- - RorpJ nf-. Total/- H.-St Resp Int.

Rh. Aorivit; /

1r r \.>i, *1

,Al(

p..' .*--o*/A~~~~ \\i~~

vc a/(,

327

M. G. WILSON, E. INGERMAN, R. 0. DUBOIS AND B. MCL. SPOCK

Ttat No.o. R e.,r.lbe170 fections

700..

64911

600.

.500_

400_

300.

aoo-

100.

n

-5

*m"i Ief..t*s.R", toit* &sres

Rh. o"ti be.rr1

10+ 167.

Tetel Nt4RhowmaticRoewrttoe*"

Rh~eumoic Adlivi

Nece4d.d b1

P*sprfot.rl lntection

22 -10 16. 11%

I ce / A

CHART3. RELATIVE INCIDENCE OF RESPIRATORYINFECTIONS AND RHEUMATICRECURRENCES1933-1934 (123 RHEUMATICSUBJECTS)

of March, April and May, hemolytic streptococciappeared in the pharyngeal flora in the majorityof the subjects. It is significant that during thisperiod hemolytic streptococci predominated in thepharyngeal flora of 50 per cent of the subjectswho were well, 40 per cent during respiratory in-fections and 10 per cent during rheumatic ac-

tivity. It is of interest that the majority of theserespiratory infections were the common cold.

In marked contrast to the prevalence of res-

piratory infections during the year was the rela-tively low incidence of rheumatic recrudescences.

Of the total 649 respiratory attacks only 16 percent were associated with rheumatic activity.There did not seem to be any direct relation be-tween the pharyngeal flora, or the type or severityof the respiratory infection and the rheumaticactivity that happened to follow such infection.It is perhaps of some significance that duringthese observations the majority (84 per cent) ofthe respiratory infections experienced by theserheumatic subjects were not associated with rheu-matic activity.

Sixty-one subjects (49 per cent) experienced

TABLE I

Relation of rheumatic activity to respiratory infection

Preceded by During respiratory Not associated withinfection respiratory infection

Total

Rhumti ctviy number -ClassificationRheumatic activity numrof "Strepto- Non- "Strepto- Non- Preceded doubtfulcases coccal " streptococcal coccal " streptococcal by Negative

respiratory respiratory respiratory respiratory streptococciinfection infection infection infection in throat

cases cases cases cases cases cases casesCarditis, arthritis, nodules 8 0 0 4 2 1 1Arthritis 9 2 1 3 1 0 2Chorea 16 1 2 5 3 2 3Skin (purpura, urticaria) 6 2 0 2 0 1 0 1Joint pains 100 11 19 27 19 13 11

Total cases 139 16 22 41 25 17 17 1Per cent 11.5 16 29 18 12 12

328

Rh c v m Q tic. Activity Htmo%ytic. 5tr. Resp.Inf. P71 Rtsp. lf%t.L::fA 0

HEMOLYTIC STREPTOCOCCUS

139 rheumatic symptoms; about three-fourths,100, were recurring joint pains, 6 purpura, 16chorea, 9 arthritis and 8 carditis, arthritis andnodules. Analysis of the time relationship be-tween these rheumatic recurrences and respira-tory infections, summarized in Table I, is of in-terest.

month before the onset of rheumatic activity.The intervening period of apparent health wasdesignated as the quiescent interval. Sixteen percent of the rheumatic episodes were preceded bya respiratory infection (unassociated with hemo-lytic streptococci in the flora of the throat) andonly 11.5 per cent of the 139 rheumatic recur-

1933-1934

| Sept. Oct. Nov. Dec. Jan. Feb.SuvecfCs 1 I 14 17 16 : So

9o t̂1oathW1lvt.Idence of4-- .S. iThfO.Cr frIo.s0- - Resp In f.To to /

-_-H .5%t. Rec p 1^e410Rh. Acfrivit~

70- /0\ t/V

i'- \, ,

rij JuIt AuX.to to

fl- -

20_ / I k

10-

C-.

CHART4. MONTHLY INCIDENCE OF RHEUMATIC RECURRENCESAND RES-PIRATORY INFECTIONS EXPERIENCED BY 62 RHEUMATIC SUBJECTS AT THE NEWYORKHOSPITAL CONVALESCENTCOTTAGES.

One-fourth (24 per cent) of the 139 rheumaticepisodes were unrelated to respiratory infections.About one-half (47 per cent) of the rheumaticrecurrences occurred either simultaneously with arespiratory infection or more frequently during achronic respiratory infection, such as persistentethmoiditis so commonly present in children. Insome instances the respiratory infection occurredduring the course of the rheumatic activity.

Of particular interest is a consideration of theincidence of rheumatic recurrences which werepreceded (within a month) by a respiratory in-fection. Respiratory infections were consideredto precede rheumatic activity if the termination ofthe respiratory infection was within a period of a

rences were preceded by a " streptococcal " res-piratory infection. The quiescent interval subse-quent to respiratory infections was infrequentlyobserved to precede rheumatic activity. This isnot in accord with the observations of other in-vestigators.

In children it is particularly difficult to deter-mine the onset of rheumatic disease when res-~piratory symptoms are present. Fever is fre-quently the only early objective symptom of rheu-matic activity. It is suggested that in such casesrheumatic fever may be present at the onset ofwhat appears to be a respiratory infection, al-though not easily detected clinically, and that thequiescent period may then represent the apparent

329

M. G. WILSON, E. INGERMAN, R. 0. DUBOIS AND B. MCL. SPOCK

symptomless period, so commonly observed dur-ing the course of rheumatic activity (see ProtocolJ. V., Paper II of this series).

The occurrence of a crop of nodules, a sharprise of temperature or a pericardial rub, follow-ing an apparently symptomless period of a weekor ten days, is a characteristic train of symptomsduring the course of rheumatic activity in chil-dren, unassociated with respiratory infections.Such episodes frequently lead one to attribute er-roneously etiological significance to events, or totherapeutic or diagnostic measures, just preced-ing the appearance of these objective symptoms.

A consideration of the relation between res-piratory infections in the household and theoccurrence of rheumatic fever is of interest.Complete data for analysis were obtained for 79rheumatic families comprising 117 rheumatic in-dividuals under sixteen years of age, 164 non-rheumatic children and 183 adults. The monthlyincidence of household infections for ten months(September to July) was as follows: two or moremembers per month experienced respiratory in-fections 193 times, respiratory infections withstreptococci in the pharyngeal flora 102 times, andcarried hemolytic streptococci in the throat 106times (in apparent good health). During thisperiod 50 children experienced 106 rheumatic re-currences. These rheumatic episodes did not ap-pear directly related to the presence of householdinfections, occurring as frequently in the absenceof such respiratory infections. It is perhaps ofsome significance that rheumatic subjects experi-enced an average of five respiratory attacks com-pared with three in the control group for the tenmonth (September to July) period.

During this investigation a series of 62 rheu-matic children who were under daily observationat the New York Hospital Convalescent Cottagesfor a period of from two months to one year gavean opportunity for a more accurate study of therelation of respiratory infections to rheumaticfever in children living in close contact. The ma-jority of the subjects were convalescing fromsevere rheumatic recrudescenses, many were com-pletely at rest in bed, others were ambulatory.Non-rheumatic convalescent children were alsounder the same supervision, and no attempt wasmade to isolate the rheumatic child. During theperiod of observations 105 respiratory infections,

of which 53 were accompanied by hemolyticstreptococci in the pharyngeal flora, were experi-enced by the rheumatic children. Six subjectssuffered nine rheumatic recurrences; three of therheumatic episodes were preceded by "strepto-coccal " respiratory infections. Although threesevere epidemics of respiratory infections asso-ciated with a predominance of hemolytic strepto-cocci in the pharyngeal flora (October, March,May) were observed among these susceptiblerheumatic subjects, there was no appreciable in-crease in the incidence of rheumatic activity dur-ing this period such as might have been expectedfrom the reported epidemics in institutions undersimilar circumstances.

It is significant that during these various ob-servations, although the majority (98 per cent)of the rheumatic children experienced one ormore respiratory infections, only 16 per cent ofthese infections were associated with rheumaticactivity. About one-half (49 per cent) of thechildren suffered one or more rheumatic recur-rences, but only 11.5 per cent of the rheumaticepisodes were preceded by "streptococcal " res-piratory infection.

COMMENT

This investigation was conducted over a periodof several years, comprising studies of a largegroup of rheumatic children observed in thehomes, hospital wards and convalescent cottages.These studies are comparable to those reported bythe majority of investigators with the exceptionthat our investigations were limited to children.

It is evident from the data that our observationsdo not give support to the conception of a specificetiological relationship between respiratory infec-tions and rheumatic fever. Their associationwould seem almost inevitable because of the ob-served frequency of respiratory infection in chil-dren. It is also possible that the rheumatic childis constitutionally vulnerable to both infections.This possibility is indicated by the frequency of asimultaneous onset of rheumatic activity and res-piratory infection (following a common excitingfactor). It is also suggested by the greater sus-ceptibility of the rheumatic child to respiratoryinfection compared with that of a non-rheumaticbrother or sister.

330

HEMOLYTIC STREPTOCOCCUS

Although it is generally well recognized thatthe organisms isolated from throat cultures do notrepresent an accurate picture of the pharyngealflora, merely indicating the fluctuation in the rela-tive numbers of the various organisms rather thantheir presence or absence, recent investigationstend to ascribe etiological significance to the isola-tion of hemolytic streptococci from the flora dur-ing respiratory infections. Emphasis seems to beplaced on the relative numbers of colonies ofhemolytic streptococci present rather than on thelocal signs or symptoms.

Our observations would tend to minimize thediagnostic significance of the presence or absenceof hemolytic streptococci in the pharyngeal flora.It was found that the relative numbers of theseorganisms in repeated consecutive cultures variedfrom predominance to few or none. The speci-ficity of " streptococcal " respiratory infectionsbased solely upon bacteriological criteria does notappear convincing in view of the observed com-parable carrier rate of hemolytic streptococci inthe pharyngeal flora of sick and well children.The immunological studies reported in the secondpaper of this series, showing a rise in the titre ofantistreptolysin following respiratory infectionsunassociated with hemolytic streptococci in thepharyngeal flora is pertinent.

It is of interest to note in passing that sincetonsillectomy has become an early routine pro-cedure in children, tonsillitis is observed infre-quently in the course of rheumatic disease. Re-cent investigators tend to substitute in its placenasopharyngitis.

The etiology of the rheumatic sequelae follow-ing scarlet fever is still controversial. Of inter-est is the evidence of Paul and Salinger (8) intheir studies of the relation of rheumatic sequelaeto scarlet fever that "like respiratory infections,scarlet fever activates the rheumatic process inthe rheumatic child."

Wilson, Lingg and Croxford (9) suggestedthat the common age and seasonal incidence ofrheumatic fever and scarlet fever may be factorsinfluencing their association. In their observa-tions, scarlet fever was more frequently followedby rheumatic recrudescences within the age periodof five to nine years. Seventy-five per cent ofthe rheumatic subjects developing scarlet fever atother age periods did not have rheumatic sequelae.

During these observations (as can be seen inProtocol 4 (10)) the occurrence of scarlet feverduring a severe carditis did not appear to influ-ence the course of the latter.

While it is appreciated that the occurrence ofrespiratory infections in a rheumatic child may notbe a fortuitous event, it would seem from our ob-servations to bear no more specific etiological re-lationship to rheumatic disease than would be at-tributed to similar episodes occurring in the tu-berculous child.

SUMMARY

1. There are presented investigations conductedover a period of several years comprising epi-demiological, bacteriological and clinical studies ofa large group of rheumatic children, observed inthe homes, hospital wards and convalescent cot-tages.

2. Two hundred and twenty-two ambulatoryrheumatic subjects five to fifteen years of age ex-perienced 783 respiratory infections and 401 rheu-matic recurrences (for a two-year period of ob-servation 1930 to 1932). Less than 10 per centof the rheumatic attacks were preceded withinthree weeks by a respiratory infection.

3. Of a total of 123 rheumatic subjects underclose observation for twelve months, September1933 to September 1934, 98 per cent suffered 649respiratory attacks; of which 353 were associatedwith the presence of hemolytic streptococci inthe throat flora. Eighty-four per cent of the res-piratory infections were not associated with rheu-matic activity.

4. Forty-nine per cent of the subjects experi-enced 139 rheumatic episodes; a quiescent in-terval, subsequent to " streptococcal " respiratoryinfection, preceded 11.5 per cent of the rheumaticepisodes.

5. Sixty-two rheumatic subjects were underdaily observation at the Convalescent Cottages fora two to twelve month period. During three epi-demics of respiratory infection associated with apredominance of hemolytic streptococci in thepharyngeal flora, there was no appreciable increaseof rheumatic activity.

6. Rheumatic subjects experienced an averageof five respiratory attacks compared with an aver-age of three for the non-rheumatic children oftheir respective households.

331

M. G. WILSON, E. INGERMAN, R. 0. DUBOIS AND B. MCL. SPOCK

7. During the spring of 1934 hemolytic strepto-cocci appeared in the pharyngeal flora of the ma-jority of the rheumatic and non-rheumatic sub-jects.

8. During this season of its highest incidence,hemolytic streptococci predominated in the pha-ryngeal flora of 50 per cent of the rheumatic sub-jects during health, 40 per cent during respira-tory infections, and 10 per cent during rheumaticactivity.

CONCLUSIONS

The evidence presented does not support theconception of a specific etiological relationship be-tween respiratory infections and rheumatic feverin children.

Our observations would tend to minimize thediagnostic significance of the presence or absenceof hemolytic streptococci in the pharyngeal floraduring respiratory infections. The designationstreptococcal respiratory infection, based solelyon bacteriological findings, would not appear to bejustified.

BIBLIOGRAPHY1. Glover, J. A., Milroy Lectures on the incidence of

rheumatic diseases. Lancet, 1930, 1, 499.2. Schlesinger, B., The relationship of throat infection

to acute rheumatism in childhood. Arch. Dis.Childhood, 1930, 5, 411.

3. Coburn, A. F., The Factor of Infection in the Rheu-

matic State. Williams and Wilkins Co., Balti-more, 1931.

4. Collis, W. R. F., The contagious factor in the etiologyof rheumatic fever. Am. J. Dis. Child., 1932, 44,485.

5. Coburn, A. F., and Pauli, R. H., Studies on the re-lationship of streptococcus hemolyticus to therheumatic process. III. Observations on the im-munological responses of rheumatic subjects tohemolytic streptococcus. J. Exper. Med., 1932,56, 651.

6. Paul, J. R., and Salinger, R., The spread of rheu-matic fever through families. J. Clin. Invest.,1931, 10, 33.

7. Wheeler, G. W., Wilson, M. G., and Leask, Mar-guerite M., The relation of upper respiratory in-fections to rheumatic fever in children. III. Theseasonal bacterial flora of the throat in rheumaticand non-rheumatic children. J. Clin Invest., 1935,14, 345.

8. Paul, J. R., Salinger, R., and Zuger, B., The relationof rheumatic fever to postscarlatinal arthritis andpostscarlatinal heart disease. A familial study.J. Clin. Invest., 1934, 13, 506.

9. Wilson, M. G., Lingg, C., and Croxford, G., Sta-tistical studies bearing on problems in classifica-tion of heart disease. III. Heart disease in chil-dren. Am. Heart J., 1928, 4, 164.

10. Wilson, M. G., Wheeler, G. W., and Leask, M. M.,The relation of upper respiratory infections torheumatic fever in children. II. Antihemolysintitres in respiratory infections and their significancein rheumatic fever in children. J. Clin. Invest.,1935, 14, 333.

332